acanthosis nigricans in the head and neck regionaipmed.org/publications/june2008edition/vol. 6 no.1...

INTRODUCTIONAcanthosis nigricans is a skin disorder characterizedby dark, thick, velvety plaque of the skin which usuallyoccurs in body folds but may involve other parts ofthe body 1. It is commonly seen on the neck, in thearmpit and groin. It could also be found on the elbow,knee and knuckles 2,3,4. It is a disorder found commonlyamong the African descents with darker skinpigmentation. It is a disorder that may begin at anyage and has no gender differences 5. It can affectotherwise healthy people or be associated withunderlying medical disease 3. Although, the actual causeis unknown, it is most likely caused by factors thatstimulate epidermal keratinocyte and dermal fibroblastproliferation. There are two broad categories ofacanthosis nigricans namely; benign and malignant. Inthe benign form of acanthosis nigricans, some casesare genetically inherited. The mode of inheritance hasnot been established, however, autosomal recessiveor polygenic inheritance has been proposed 6.Acanthosis nigricans is most commonly associated withbeing overweight (BMI >25 kg/m²) and it is frequentlyfound in people with diabetes 7-10. The stimulatingfactor is probably insulin or an insulin-like growthfactor that incites the epidermal cell propagation.Hyperinsulinemia results in binding of insulin to insulin-like growth factor receptors on keratinocytes andfibroblasts, with resultant hyperplasia of the skin 11.Insulin resistance due to obesity underlies thehyperinsulinemia in obesity-associated acanthosis

nigricans 12, 13. Sometimes acanthosis nigricans is dueto other forms of endocrine (glandular) disorder suchas polycystic ovary syndrome, acromegaly, pituitary oradrenal adenomas, and cushing syndrome 14. It couldalso result from the use of exogenous drugs likehormones (contraceptive pills), cytotoxics, steroid, largedose of niacin etc 1, 15. The malignant form is by farless common and the patients with this type tend tobe thin and older than 40 years of age, and theireruption is usually of recent origin 16,17. The stimulatingfactor in malignant acanthosis nigricans is hypothesizedto be a substance secreted either by the tumor or inresponse to the tumor. Transforming growth factor-alpha is structurally similar to epidermal growth factorand it has been implicated 18-20. Acanthosis nigricans israrely associated with a tumor and if it does, unusuallysevere. Development of acanthosis nigricans may bethe first clue to an underlying neoplastic process in anotherwise healthy adult. Ninety percent of theneoplasms responsible for the development ofacanthosis nigricans originate in the abdomen 21.Nonetheless, there is no information on acanthosisnigricans among the head and neck patients in ourenvironment. Hence, our aim is to describe the patternof presentation of acanthosis nigricans, to determineits prevalence and to identify the underlying associatedmedical and oncologic factors among head and neckpatients managed at department ofOtorhinolaryngology, University College Hospital,Ibadan.

ACANTHOSIS NIGRICANS IN THE HEAD AND NECK REGIONJames A. Fasunla M.B.Ch.B, Gbolagunte T. Ijaduola FWACS

Department of Otorhinolaryngology, University College Hospital, Ibadan.

Corresponding Author :Dr. Ayotunde J. FasunlaSenior RegistrarDepartment of Otorhinolaryngology,University College Hospital,PMB 5116, Queen Elizabeth road,Ibadan, Nigeria.Telephone number: 0803 374 0220Email address: [email protected]

ABSTRACTBackground: The sudden appearance of acanthosis nigricans onthe skin of an individual is highly suggestive of an underlyingclinical disease. However, there is no information on its possibleassociation with head and neck disease in our environment.Objective: To determine the prevalence of acanthosis nigricansand identifying the underlying associated medical and oncologicfactor in an Otorhinolaryngologic, head and neck practice.Method: This was a 12-month prospective study of patients seenat Otorhinolaryngology department of University CollegeHospital, Ibadan. Relevant clinical, anthropometric andbiochemical information were obtained with an interviewer assistedquestionnaire and data analyzed using descriptive statistics.Results: Of 764 patients, 85(11.13%) had head and neckmalignancies, 15(1.96%) had diabetes mellitus in addition to theOtorhinolaryngologic, head and neck diseases. The body massindex values for male and female ranged from 18.26 – 25.68kg/m²and 17.94 – 30.25kg/m² respectively. Only two patients, one withnasopharyngeal cancer and the other with obesity in addition tochronic rhinosinusitis had acanthosis nigricans.Conclusion: Acanthosis nigricans is not common amongOtorhinolaryngologic patients in our clinical setting with aprevalence of 0.003. Nevertheless, its presence should herald thesuspicion of an underlying metabolic or oncologic disease.

Keywords: Acanthosis Nigricans, Otorhinolaryngology, Head, Neck

Annals of Ibadan Postgraduate Medicine. Vol.6 No.1 June, 2008 53

MATERIALS AND METHODA prospective study of all patients that presented tothe ENT clinic and ward of University CollegeHospital, Ibadan from August 2005 to August 2006was carried out. An understood informed consent wasobtained from all the patients that participated in thestudy. A structured interview assisted questionnaire wasused to collect the biodata (Age, Sex, and Nationality),Ear, nose and throat disease, history of presence of athick dark area on the skin, and the duration for whichit had been noticed. History of diabetes mellitus,malignancy, and use of cytotoxics, hormonal drugs(oral contraceptives) and steroids were obtained. Allthe patients had their head and neck region, armpits,breasts, elbow, groin, knees and knuckles examinedfor presence of acanthosis nigricans. The weight (Kg)and height (m) of the patients were measured todetermine their body mass index. All the patients withhistory of diabetes mellitus had fasting plasma glucosedone. Data were analyzed using descriptive statistics.

RESULTSThere were seven hundred and sixty four patients; 347(45.42%) males and 417(54.58%) females with ageranged between 3 and 63 years old and overall meanage of 47.78 and standard deviation of 21.08 (Table1). There were 85(11.13%) patients with head and neckmalignancies, 47(55.29%) males and 38(44.71%)females (Table 2). Forty six (54.12%) patients hadsinonasal cancer, 17(20%) had nasopharyngeal cancer,3(3.53%) had ear cancer, 12(14.12%) had laryngealcancer. Three (3.53%) patients had metastatic neckdisease. One of the female patients withnasopharyngeal cancer had acanthosis nigricans aroundthe neck folds.

Table 1: Age range distribution of participants

Table 2: Distribution of head and neck malignanciesand acanthosis nigricans

There were 15(1.96%) patients, Five (33.33%) malesand 10(66.67%) female with associated diabetesmellitus in addition to the head and neck lesion. Allhad their fasting blood sugar checked and were allwithin normal limit. Two patients with diabetes mellitushad associated head and neck malignancy(nasopharyngeal and laryngeal cancer). The body massindex values for male ranged from 18.26–25.68 kg/m² while that of female ranged from 17.94–30.25kg/m². Sixteen (2.09%) of the patients wereoverweight with body mass index >25 kg/m² (8malesand 8females). One of the female obese patients(BMI=30.25 kg/m²) with chronic rhinosinusitis hadacanthosis nigricans in her armpit and under the breastfolds. Thirteen (6.60%) female patients had history ofusage of contraceptive pills for average period of 4months (range 2 – 9 months). Twenty-one (2.75%) ofthe patients had history of treatment with cytotoxicdrugs. The head and neck disease distribution in ourpatients is shown in Table 3.

DiscussionAcanthosis nigricans is a rare disease. It is not in itselfdangerous but it can be associated with a potentiallydangerous diseases. The actual prevalence of the diseaseis unknown. Various researchers had reported differentprevalence of the disease in different regions of theworld 7, 22. Acanthosis nigricans is much more commonin people with darker skin pigmentation 9. Theprevalence in whites is less than 1%. In Hispanics, the


Age range(YRS) Frequency Percentage (%)

1 – 10 69 9.03

11 – 20 67 8.77

21 – 30 169 22.12

31 – 40 229 29.97

41 – 50 163 21.33

51 – 60 58 7.59

61 – 70 9 1.19

Total 764 100.00

Site of Tumour Frequency Presence/ Absence of AN

Sinonasal 16 (18.82%) -

Larynx 18 (21.18%) -

Nasopharynx 24 (28.23%) 1

Ear 3 (3.53%) -

Metastatic neck disease 4 (4.71%) -

Oropharynx 3(3.53%) -

Mandible 6 (7.06%) -

Oral Cavity 3(3.53%) -

Hypopharynx 5 (5.88%) -

Parotid 3(3.53%) -

Total 85(100.00%) 1


prevalence is 5.5%, and, in African Americans, theprevalence is the highest at 13.3% 7, 9, 22. However, inthis study, the prevalence of acanthosis nigricans amonghead and neck patients was 0.003%. This lowprevalence rate might be due to the fact that most ofour patients were without the previously documentedassociated factor of acanthosis nigricans. In addition,there was non availability of a similar study amonghead and neck patients for comparison. A previousstudy had reported 2 out of 12,000 patients withinternal malignancy having acanthosis nigricans 21. Inour series, only one patient with head and neckmalignancy had acanthosis nigricans. Although, nogender inequality had been reported, the two cases ofacanthosis nigricans seen in this study were female.

Acanthosis nigricans can occur at any age; however,the malignant type is usually seen above the age of 40years 18-20, 23. This is in agreement with what was foundin this study as our patient with nasopharyngeal tumorwho had acanthosis nigricans was 42 years old. Thepresentation of acanthosis nigricans is common at thebody folds especially on the neck, in the armpit andgroin. This is the case in this study as it was foundexclusively around the patient’s neck fold.

The benign type of acanthosis nigricans is commonerthan the malignant type. Normal range of BMIaccording to WHO are 18.5 – 24.9 kg/m².Overweight is e” 25 kg/m². Obese individual has aBMI of >30 kg/m² and this is associated with a greatmorbidity. Sixteen (2.09%) of our patients in this studywere overweight but only one obese patients hadacanthosis nigricans in her armpit and under the breastfolds. The head and neck disease in this patient waschronic rhinosinusitis. In addition, the fifteen patientswith diabetes mellitus were already on medications andtheir blood sugar also controlled. Treatment ofunderlying causes of acanthosis nigricans may preventits development and usually results in its resolution 21.

There were thirteen (5.04%) patients with the historyof having used oral contraceptive pills for average of4 months. None of these patients showed evidenceof acanthosis nigricans on their skin. The reason forthese could actually be as a result of short term usageor because they had stopped the usage of themedication for an average period of 3 years 8 months

before they were seen and recruited into this study.This is a good enough time for the skin lesion to havecleared if it was actually present on their skin.

In conclusion, Acanthosis nigricans is a raredermatologic condition in our Otorhinolaryngologicpractice with a prevalence of 0.003. Its identificationon the skin of patients presenting to theOtorhinolaryngologic clinic must not be casuallydismissed without searching for possible associatedunderlying metabolic or oncologic disorder. The earlydetection of an underlying disease will in no doubtresult in a better treatment outcome.

REFERENCES1 Schwartz R.A. Acanthosis nigricans. J Am Acad

Dermatol 1994; 31:1-19.2 Burke J.P., Hale D.E., Hazuola H.P. and Stern

M.P. A quantitative scale of acanthosis nigricans.Diabetes care 1992, 22: 1655 – 1658.

3 Stuart C.A., Driscoll M.S., Lundquist K.F.,Gilkison C.R., Shaheb S. and Smith M.M.Acanthosis nigricans. J. Basic Clin PhysiolPharmacol 1998; 9: 407 – 418.

4 Stuart C.A., Gilkison C.R. and Smith M.M.Acanthosis nigricans as a risk factor for non-insulindependent diabetes mellitus. Clin Pediatr (Phila)1998; 37: 73 – 79.

5 Alberta S.K., Robert L.W., Melissa S., AndrewL.S. and Robert L.R. Acanthosis Nigricans andDiabetees Risk Factors: Prevalence in YoungPersons Seen in Southwestern US Primary CarePractices. Ann Fam Med 2007; 5: 202 – 208.

6 Burke J.P., Duggirala R. and Hale D.E. Geneticbasis of acanthosis nigricans in Mexican Americansand its association with phenotypes related to type2 diabetes. Hum Genet 2000; 106: 467-472.

7 Hud J.A. Jr, Cohen J.B. and Wagner J.M.Prevalence and significance of acanthosis nigricansin an adult obese population. Arch Dermatol1992; 128: 941-944.

Otorhinolaryngologic, head Frequency No. of patients withand neck disease Acanthosis Nigricans

Otology 354 (46.34%) -

Rhinology 247 (32.33%) 2

Laryngology, Head and Neck 163 (21.33%) -

Total 764 (100%) 2

Table 3: Otorhinolaryngologic, head and neck disease distribution and Acanthosis Nigricans


8 Fagot-Campagna A., Pettitt D.J., EngelgauM.M. Type 2 diabetes among North Americanchildren and adolescents: an epidemiologicalreview and a public health perspective. J Pediatr2000; 136: 64-72.

9 Nguyen T.T., Keil M.F., Russell D.L. Relationof acanthosis nigricans to hyperinsulinemia andinsulin sensitivity in overweight African Americanand white children. J Pediatr 2001; 138: 474-480.

10 Stuart C.A., Smith M.M., Gilkison C.R.Acanthosis Nigricans among Native Americans:an indicator of high diabetes risk. Am J PublicHealth 1994; 84: 1839-1842.

11 Cruz P.D. Jr, Hud J.A. Jr. Excess insulin bindingto insulin-like growth factor receptors: proposedmechanism for acanthosis nigricans. Journal ofInvestigative Dermatology 1992; 98: 82-85.

12 Davidson M.B. Clinical implications of insulinresistance syndromes. Am J Med 1995; 99:420-426.

13 Moller D.E., Flier J.S. Insulin resistancemechanisms, syndromes and implications. NEngl J Med 1991; 325:938-948.

14 Kurzrock R., Cohen P.R. Cutaneousparaneoplastic syndromes in solid tumors. AmJ Med 1995; 99:662-671.

15 Stals H., Vercammen C. and Peeters C.Acanthosis nigricans caused by nicotinic acid:case report and review of the literature.Dermatology 1994; 189:203-206.

16 Ellis D.L., Kafka S.P., Chow J.C., Nanney L.B.,Inman W.H. and McCadden M.E. et al.Melanoma, growth factors, acanthosis nigricans,the sign of Leser – Trelat, and multipleacrochordons. A possible role for alpha –transforming growth factor in cutaneousparaneoplastic syndrome. N Engl J Med 1987;317: 1582 – 1587.

17 Clarke J. Malignant acanthosis nigricans. ClinExp Dermatol 1977; 2: 167 – 170.

18 Sabir S., James W.D. and Schuchter L.M.Cutaneous manifestations of cancer. Curr OpinOncol 1999; 11: 139-144.

19 Yeh J.S., Munn S.E. and Plunkett T.A.:Coexistence of acanthosis nigricans and the signof Leser-Trelat in a patient with gastricadenocarcinoma: a case report and literaturereview. J Am Acad Dermatol 2000; 42: 357-362.

20 Kurzrock R., Cohen P.R. Cutaneousparaneoplastic syndromes in solid tumors. Am JMed 1995; 99:662 - 671.

21 Safai B., Grant J.M., Kurtz R. et al. Cutaneousmanifestation of internal malignancies: Acanthosisnigricans. Int J Dermatol 1978; 17: 312 – 315.

22 Mukhtar Q., Cleverley G. and Voorhees R.E.et al. Prevalence of acanthosis nigricans and itsassociation with hyperinsulinemia in New Mexicoadolescents. J Adolesc Health 2001; 28: 372-376

23 Kihiczak N.I., Leevy C.B. and Krysicki M.M.Cutaneous signs of selected systemic diseases.Am J Med 1999; 30: 3-12.

INTRODUCTIONMatthew Baillie first described the complete mirrorimage reversal of the thoracic and abdominal organsin situs inversus in the 18th century. Earlier in 1643,Marco Severino had described the first case ofdextrocardia. Since then several cases of bothconditions separately or together have been reportedin western literature1-4 but relatively rarely in thedeveloping world5-9. Situs abnormalities are congenitalwithout racial or sexual predilection10.

We present the case of a young lady with situs inversusdiscovered during management for appendicitis. As isoften the case in situs situations, there was initialconfusion and some significant delay in offering neededtreatment due to the unusual disposition of the organs.This is highlighted in our presentation.

Case SummaryA 32 year old Nigerian woman first presented to uswith complaint of right sided abdominal pain of 9days duration described as biting in nature and radiatingto the lower back. She had three similar attacks in thepreceding thirteen years, usually short-lived andassociated with nausea and vomiting. She had a bowelmovement every three 3 or 4 days but the stools werenormal.

She had unproductive cough since childhood, workedas an executive secretary, and had three children atpresentation all from spontaneous vaginal deliveries.She was otherwise in good health.

On abdominal palpation, there was tendernessmaximal at the right iliac fossa but extending to thesuprapubic area. White blood cell count as well as urineanalysis results were unremarkable. A clinical diagnosisof acute on chronic appendicitis was made with somedegree of confidence and surgery was planned.

The chest X-ray (Figure 1) showed a right-sided cardiacapex indicative of dextrocardia. However, due to thepoor visualization of the gastric air bubble, anabdominal ultrasound was done to rule out a situsinversus. The abdominal ultrasound confirmed situsinversus affecting the intraabdominal viscera. No signof gynaecological problem was visualised.

Figure 1. The chest radiograph of the patient showingdextrocardia and gastric air bubble on the right

It became difficult attributing the mainly right-sidedpain to appendicitis. A laparoscope was not availableso a barium enema was ordered to excludesymptomatic diverticular disease and other colonicpathology. Barium enema (Figure 2) showed no largebowel pathology apart from confirming situs inversus.CXR showed no evidence of bronchiectasis or otherlung lesion. A normal left sided electrocardiogramshowed inversion of the waves in leads aVR, I, aVL,V4-6. A reverse (Right sided ECG ) demonstrated anormal pattern (Figures not shown) Echocardiographyindicated atrio-ventricular and great arteries-ventricularconcordance.

APPENDICITIS AND SITUS INVERSUS VISCERUM IN A 32-YEAR-OLD FEMALENIGERIAN: A CASE REPORTA.E.O. Adeniyi1, Cynthia O. Akisanya2, O.S. Ogah3, Akinremi, Titilola O., Charles A. Erinle4

1. Department of Surgery 2. Departments of Radiology 3. Departments of Medicine 4. Department of Family Medicine5. Department of Pathology, (Federal Medical Centre, Abeokuta, Ogun State, Nigeria)

Corresponding AuthorDr O.S. OgahDepartment of Medicine,Federal Medical Centre,Idi-Aba, Abeokuta,PMB 3031 Sapon PO, Abeokuta,Ogun State, NigeriaE-mail: [email protected]

Situs inversus is a relatively rare congenital abnormality inwhich the internal organs are disposed in a mirror imageof the normal. It is said to be an autosomal recessive ge-netic condition. Its prevalence varies with population. Inthe United States, Situs inversus is found in 0.01% of thepopulation. The incidence in Nigeria is unknown.We present the case of a young lady with situs inversusdiscovered during management for appendicitis. As is of-ten the case in situs situations, there was initial confusionand some significant delay in offering needed treatmentdue to the unusual disposition of the organs.

Key Words: Situs Inversus, Appendicitis, Dextrocardia

ABSTRACT


CASE REPORT

Figure 2. The Barium enema of the patient showingsplenic flexure on the right and hepatic flexure on theleft.

The pain gradually subsided on the interim medicaltreatment (antibiotics –Ceftriazone and Metronidazole)being given and the patient, who had been apparentlytrying to come to terms with the notion that she was“abnormal”, opted to go home.

About 18 months later she re-presented with featurestypical of appendicitis, complete with reboundphenomenon, but now sited over the left iliac fossa.

She admitted that she had a number of similar attacksof the pain in the interval between the last admissionand the current presentation. Also the location of thepain varied from left to right and centrally in the lowerabdomen.

She had an uneventful appendicectomy (via a midlineinfra umbilical laparotomy incision so as to be able toexplore the abdomen more satisfactorily).

Situs inversus was confirmed at operation with thecaecum and appendix in the left iliac fossa.

The appendix was 11cm long and was buried inadhesions involving almost its entire length binding itto both the caecum and the terminal ileum and wasdisposed in the preileal position. Subsequent histologyrevealed section of appendix showing mucosal andsub-mucosal lymphoid hyperplasia, ulceration andfibrosis as well as inflammatory cellular infiltration withlymphocytes, eosinophils and neutrophils. The muscleand serosal layers were oedematous with lymphocyticand eosinophilic cellular infiltrate in keeping with acuteon chronic appendicitis (figure 3).

She made an uneventful recovery and was dischargedhome obviously greatly relieved.

DISCUSSIONSitus inversus is a relatively rare congenital abnormalityin which the internal organs are disposed in a mirrorimage of the normal. It is said to be an autosomal

recessive genetic condition. Its prevalence varies withpopulation. In the United States, Situs inversus is foundin 0.01% of the population10 .The incidence in Nigeriais unknown. There is no racial predilection and sexincidence is 1:1

Situs describes the position of the cardiac atria andviscera. Situs solitus is the normal position while situsinversus is the mirror image of solitus. In situs inversus,the morphologic right atrium is on the left while themorphologic left atrium is on the right. The left lunghas 3 lobes while the right lung has 2 as opposed tothe normal pulmonary anatomy. The liver andgallbladder are situated on the left while the stomachand spleen are on the right.

Individuals with situs inversus are predisposed tocommon medical as well as surgical conditions.

The recognition of situs inversus is of surgicalimportance, as it will prevent surgical mishaps resultingfrom an atypical history and failure to recognisereversed anatomy.

Acute appendicitis is the commonest cause of surgicalabdominal pain and it is not surprising that appendicitisis widely reported in situs inversus despite the fact thatthat condition is by itself relatively rare.

Figure 3. Photograph of the histology of the removedappendix showing features of acute on chronicappendicitis.

Not helping matters is the report that abdominal painin situs inversus can be poorly localized and that in upto 50% of cases the pain is referred to the right iliacfossa[2]. This situation seems to have been highlightedby our patient whose pain seemed to migrate acrossher lower abdomen. She was saved from a potentiallynon-effective standard right iliac fossa appendicectomyincision by the chest x-ray ordered for other reasons,which exposed a dextrocardia, which subsequentlytriggered a search for the related condition of situsinversus.

The situation underscores the importance oflaparoscopy in the management of acute abdominal


pain of uncertain origin in general and in situs inversusin particular. Contini et al. advised the use oflaparoscopy when appendicitis is suspected in thepresence of situs inversus because of its superioradvantage over open appendicectomy. This allowsmore accurate diagnosis and better exploration of theabdomen[4].

Patients with situs inversus reportedly have a normallife expectancy. In the rare instances of cardiacanomalies, the life expectancy is reduced dependingon the severity of the defect.

Our patient was an otherwise normal mother of threeThe diagnosis of situs inversus would probably neverhave been made had she not had appendicitis. It ispossible that many cases of situs inversus have beenmissed and are still being missed especially in thedeveloping world where medical examinations areinfrequently conducted in individuals who are not ill.

Differential diagnosis of lower abdominal pain in thissetting will include disorders of the genitourinary tractsuch as pelvic inflammatory disease, ovarian torsion,and ectopic gestation. Others are disorders of thegastrointestinal tract such as acute Pancreatitis anddiverticular disease

The purpose of this presentation is to report theexistence of this condition in our environment andhighlight some of the difficulties that may arise inmanagement of otherwise straightforward surgicalcondition if its presence is not recognized on time.

REFERENCES1. Jules P. Acute appendicitis in situs inversus; report

of a case. J Med Soc N J 1956;53(12):582-583.

2. Lesser A., Zuckerman H. Acute appendicitis insitus inversus. N Y State J Med 1958;58(1):67-68.

3. Holgersen L.O., Kuehner C.R., Stanley-BrownEG. Acute appendicitis in a child with completesitus inversus. J Pediatr Surg 1970;5(3):379-380.

4. Contini S., Dalla Valle R., Zinicola R. Suspectedappendicitis in situs inversus totalis: an indicationfor a laparoscopic approach. Surg LaparoscEndosc 1998;8(5):393-394.

5. Adeyekun A.A., Onunu A.N., Mazeli F.O.Dextrocardia with situs inversus: a case report. WestAfr J Med 2003;22(4):358-360.

6. Danbauchi S.S., Alhassan M.A. Case report:dextrocardia with situs inversus; two casespresenting differently. Niger Postgrad Med J2002;9(4):248-252.

7. Ofodile F.A., Adeloye A. Acrocephalosyndactylywith hydrocephalus and dextrocardia in a Nigerianchild. J Natl Med Assoc 1982;74(8):800-803.

8. Singhadej P., Sintuvanich A., Vichankaiyakij V.Situs inversus and appendicitis. J Med Assoc Thai1984;67(3):194-200.

9. Prasad V. Left sided appendicitis and situs inversus.Indian J Gastroenterol 1992;11(1):44-45.

10. Carmichael K.A., Gayle W.E., Jr. Situs inversusand appendicitis. South Med J 1979;72(9):1147-1150.


INTRODUCTIONChildhood blindness is increasingly becoming a majorchallenge world wide. It is estimated that a child goesblind some where in the world every minute. This isindeed disturbing in view of the attendant problemsand hopelessness associated with blindness especiallyin children. In view of the importance associated withchildhood blindness the theme of year 2007’s WorldSight Day was Vision in Children.

Childhood blindness describes a number of diseasesand conditions that occur in childhood. Most of themore serious disorders can be prevented or avoided.When they do occur and often in spite of attempts attreatment or if left untreated they can result in lifetimeblindness.

World wide an estimated 1.5 million children arebilaterally blind, most live in the developing world.1,2

The prevalence of childhood blindness variesaccording to the socioeconomic development of thecountry and the mortality rate of the children youngerthan 5 years of age.

Ninety percent of the blind children in the world livein Asia or Africa and 75% of all causes of childhood

blindness in those areas are preventable or treatable.3

For a child who is born blind or who becomes blindthe total number of years of disability are greater thanfor a person who becomes blind later in life. 4 TheWHO’s prevention of blindness programme with theInternational Centre for Eye Health has developed astandard methodology and reporting form to recordthe causes of visual loss in children with the emphasison the identification of preventable or treatable causesof blindness.5

This methodology is now being used in developingand middle income countries. 5,6,7 The prevalence ofblindness in children ranges from an estimated 0.3 per1,000 children in high income countries with low under5 mortality to 1.5 per 1,000 in low income countrieswith high under 5 mortality rates.8,9

Although the actual number of blind children is muchsmaller than the number of adults who are blind, thenumber of blind years resulting from blindness inchildren is also equal to the number of blind years dueto age related cataract.10 The burden of disability interms of blind years in these children represents a major

CAUSES OF BLINDNESS AND VISUAL IMPAIRMENT AT THE SCHOOL FOR THEBLIND OWO, NIGERIA

Correspondence:Dr. Omolase C. OluwoleDepartment of Ophthalmology,Federal Medical Centre,P.M.B 1053,Owo.Ondo State.E mail: [email protected].

C.O. Omolase FWACS, FMCOphth, A.S. Aina M.B.B.S, B.O. Omolase M.B.B.S and E.O. OmoladeM.B.B.S .Department of Ophthalmology, Federal Medical Centre, P.M.B 1053, Owo, Ondo State.

ABSTRACTAim: This study was designed to determine the causes of blindnessamongst the pupils of the School for the blind ,Owo and also identifytreatable causes of blindness in the study population.Methodology: This study was conducted between November andDecember,2007 at the School for the blind, Owo. Ethical clearance wasobtained from the Ethical Committee of Federal Medical Centre, Owoprior to carrying out this study. Sixty two pupils of the School for theblind and blind students in Owo High school were enrolled in this study.The subjects were interviewed and examined by the authors with the aidof WHO’S recording form for blindness and visual impairment inchildren. The data obtained with the study instrument was collated andanalyzed with SPSS 12.0.1.Results: Majority of the subjects: 55 (88.7%) were blind while theremaining 7 (11.3%) had low vision. Lens ranked highest amongst theanatomical sites of visual loss accounting for 24.2%,followed by retina(14.5%),optic nerve (12.9%) and glaucoma (12.9%).The main aetiologicalfactors were intrauterine (37.1%) and hereditary factors (21%). Thecommonest single diagnosis was cataract (21%) followed by glaucoma(12.9%). Twenty seven respondents (43.6%) had treatable causes ofblindness.Conclusion: The lens ranked highest amongst the anatomical sites ofblindness and visual impairment. The main aetiological factors wereintrauterine and hereditary factors. The commonest single diagnosiswas cataract. Less than half of the respondents had treatable causes ofblindness and visual impairment.

Key words: Childhood blindness, visual impairment, cataract, blind school and Nigeria.


social, emotional and economic burden for thechildren, their families, the communities and thenation.11 An estimated 500,000 children become blindeach year, but in developing countries up to 60% arethought to die within a year of becoming blind.8 Blindchildren have a higher death rate than their sightedcounterparts. 12 In view of the above control ofchildhood blindness is considered a high priority withinthe WHO’s Vision 2020 initiative : the right to sight.8,13. This study was designed to establish the causes ofblindness amongst the pupils of the School for theblind, Owo and also identify treatable causes ofblindness in the study population. We hope thatrelevant policy implication will be drawn from thisstudy to contribute to the control of childhoodblindness in this part of the world.

METHODOLOGYEthical clearance was obtained from the EthicalCommittee of Federal Medical Centre, Owo prior tocarrying out this study. The approval of the schoolauthority was sought and obtained beforecommencement of this study. Informed consent wasobtained from each of the respondents who wereeighteen years and above . All the sixty two pupils inthe school for the blind and those in Owo High schoolwere enrolled. WHO’s definition of blindness as bestcorrected visual acuity of 3/60 or less in the bettereye was adopted in this study. A modified type ofWHO’s recording form for blindness and low visionin children was utilized. All the pupils were interviewedwith the aid of the modified form. Informationobtained included biodata of the respondents, whenthe pupil was enrolled in the school, any previoussurgery received for the eye condition. All the pupilswere examined by the authors. Visual acuity wasassessed with snellen E chart for those with severevisual impairment, that is, visual acuity: 3/60 to 6/60while for those whose vision was <3/60 we checkedfor light perception. Anterior and posterior segmentexamination was carried out. Dilated funduscopy wasdone when the need arose. Respondents whose ocularcondition was amenable to treatment were regardedas having treatable causes of blindness. The ones withcauses of blindness that was amenable to primary,secondary and tertiary prevention were regarded ashaving preventable causes of blindness. Those with

treatable causes of blindness or visual impairment werereferred to the eye centre in the community located atFederal Medical Centre, Owo. The chances of therespondents regaining vision following treatment wasused to assess the prognosis for restoration of vision.Those with good prognosis were those who had goodchance of restoration of vision following treatmentwhile those with poor prognosis were not likely toregain vision following treatment. The data obtainedwas collated and analyzed with the aid of SPSSstatistical software version 12.0.1. This study wascarried out in November and December, 2007.

RESULTSSixty-two respondents were interviewed comprisingof thirty-seven in the primary school, eight in specialclass and seventeen in Owo High school where all theblind students who finished from School for the blind,Owo attend. The ages of the respondents rangedbetween 6 and 34 years . Only twenty-eight of thestudy population were aged between 6 and 15 years,the rest were aged between 16 years and 34 years. Therewere 39 males (62.9%) and 23 females (37.1%). Fifty-eight (93.5%) were Christians and four (6.5%) wereMuslims. Fifty-eight (93.5%) respondents wereYorubas, two (3.2%) were Ibos while the other ethnicgroups accounted for the remaining two (3.2%).

Fifty-five (88.7%) respondents were blind while seven(11.3%) had low vision. None of them had normalvision. Most respondents: 51(82.3%) had no historyof surgery while the remaining 11(17.7%) had historyof ocular surgery either in form cataract surgery ortrabeculectomy. Anatomical site of loss of vision: Asdetailed in table 1 the lens ranked highest amongst thesites of loss of vision accounting for 24.2%. The uvealtissue was the least site of loss of vision (4.8%). Theaetiological agents for blindness and low vision is asshown in table 2, most respondents: 23(37.1%) hadintrauterine agents being responsible for blindness andlow vision. However, in 14 respondents (22.6%), theaetiological agent for poor vision was unknown. Asshown in table 3, cataract is the leading cause ofblindness in this study accounting for 21%, followedby glaucoma and optic atrophy each accounting for12.9% respectively. Some of the respondents:27(43.6%) had treatable causes of blindness while 10(16.1%) had preventable causes of blindness while25(40.3%) had non-treatable causes of blindness.


Anatomical site Frequency PercentageWhole globe 12 19.4Cornea 7 11.3Lens 15 24.2Uvea 3 4.8Retina 9 14.5Optic nerve 8 12.9Glaucoma 8 12.9Total 62 100

Table 1: Anatomical sites of visual loss

A et io log i cal fa ct or F reque ncy P ercent ag eH ered it a ry 1 3 2 1 .0I nt rau t erine 2 3 3 7 .1C h ild hoo d 1 0 1 6 .1U n kn ow n 1 4 2 2 .6T rau m a 2 3 .2T ota l 6 2 100

T a b le 2 : A e tio lo gy o f bl in d n es s a n d vi su a li m p a i rm e n t

Prognosis for restoration of vision: Majority of therespondents: 44(71%) had poor prognosis forrestoration of vision while the remaining 18(29%) hadfair prognosis for restoration of vision. None of themhad good prognosis for restoration of vision.

DISCUSSIONThe respondents were predominantly Christians, thisis in keeping with the predominant religion in Owocommunity. It is also not surprising that most of therespondents were Yorubas in view of the fact that thehost community is a Yoruba community. The fact thatsome of the respondents are adults could be due tothe fact that some of them are in the secondary school.The other factor that is likely to have contributed isthat some of them had good vision early in life butlater became blind in the course of childhood therebynecessitating their relocation to the school for the blindfrom other regular schools. Their poor vision couldalso have contributed to their poor performance inschool prior to their enrollment in school for the blind.Most of our respondents never had any form ofocular surgery, this is likely to be related to the factthat a significant proportion had ocular conditions thatwere not amenable to surgical intervention.An overwhelming proportion of the respondents wereblind, this is in keeping with the findings ofAlagaratnam et al. in Edinburg 14 in which 81% of thestudy population had blindness or were severely visuallyimpaired. Kello, et al. also established in a blind schoolstudy in Ethiopia that 94% of the study populationwere blind or visually impaired.10 Our findings is alsoconsistent with that of Hornby, et al. in India in which91.7% of the subjects were classified as been severelyvisually impaired or blind.15

The lens ranked highest amongst the anatomical sitesof vision loss in our study population. This finding isin tandem with another Nigerian study in which lesions

of the lens was the leading anatomical site of visualloss accounting for 30.7% of the study population.11

It is however at variance with the findings of a studycarried out in Ethiopia in which the major anatomicalsite of visual loss was cornea/pthisis (62.4%).10 AnIndian study established the retina as the leading siteof visual loss accounting for 31.1%. 15 The whole globewas identified as the leading site of visual loss in aChinese study accounting for 25.5% of the studypopulation.4 Anatomically the most affected part ofthe eye in a study carried out in Czech Republic wasthe retina accounting for 54.2% of the studypopulation.16 Algaratnam et al. established in Edinburgthat the disease or malformation of the CNS or opticnerve was the commonest cause of visual impairmentand blindness and was present in 50% of their studypopulation.14

Intrauterine and hereditary factors ranked high as theaetiological factors in our study population. Thisfinding is not consistent with an Ethiopian study whichestablished childhood factor as the leading aetiologicalfactor accounting for 49.8%. 10

Vitamin A deficiency and measles were the maincontributory factors. The effective immunizationagainst measles in Nigeria is likely to have contributedto the reduction of childhood factor as the aetiologicalfactor in our study population. Studies in China 4 andIndia 15 however established hereditary factor as beingresponsible for loss of vision in 30.7% and 34.8% ofthe study population respectively. The fact that cataractwas a leading cause of blindness in our studypopulation is consistent with the findings of anotherNigerian study in which cataract accounted for thecommonest single diagnosis (23.5%). 11 Our findingsalso correspond to the findings of other Nigerian 17

and African 18,19 authors. It is likely that some of therespondents in this study were not offered surgery earlyin life prior to their enrollment in school for the blind.Offering them surgery at this stage may be associatedwith poor outcome since the outcome of surgery isoften poor once abnormal foveal functiondevelops.20,21.

This brings to the fore the desirability of promptdiagnosis and treatment of cataract. Community healthworkers could assist by identifying children withleucocoria in the community and promptly refer themto the ophthalmologists for treatment.

Less than half of our study population had nontreatable causes of blindness, this is likely to be relatedto the fact that some of them may have soughttreatment to no avail prior to enrollment at the schoolfor the blind.

This study like other blind school studies even thoughoffered us the opportunity to examine many blindchildren within a short time is subject to selection bias.The other limitation of this study is that less than halfof the respondents were children. A population based


Diagnosis Frequency PercentageCataract 13 21Glaucoma 8 12.9Toxoplasmosis 4 6.5Pthisis bulb i 6 9.7M icrophthalmia 2 3.2L eucoma 6 9.7Retinitisp igmentosa 7 11.3

Optic atrophy 8 12.9Squint 2 3.2Aphakia 2 3.2Anophthalmus 1 1.6M aculopathy 1 1.6Pseudophakia 1 1.6Refractive e rror 1 1.6Total 62 100

Table 3: Diagnosis of ocular condition


study is likely to be more revealing of the causes ofblindness and visual impairment in children.

CONCLUSIONThe lens was the leading anatomical site of blindnessand impairment of vision. Intrauterine and hereditaryfactors were the leading aetiological agents forblindness and visual impairment. Cataract was the singlecommonest diagnosis followed by glaucoma and opticatrophy. Less than half of the study population hadtreatable causes of blindness.

RECOMMENDATIONS1) The challenge of childhood cataract should be

addressed by training community health workersto identify them and promptly refer such cases tothe ophthalmologists. There is also an urgent needto train more paediatrics ophthalmologists.

2) Possible treatment for all children with severe visualimpairment or blindness should be explored priorto enrollment in School for the blind.

3) Effective preventive strategies should be exploredthrough VISION 2020 to eliminate all theavoidable and preventable causes of blindness inchildren.

4) Population based studies should be conducted todetermine the causes of blindness and severe visualimpairment in children.

ACKNOWLEDGEMENTSWe appreciate the authority and teachers of the Schoolfor the blind Owo for their cooperation. Special thanksto the respondents for graciously accepting toparticipate in this study. The support of themanagement of Federal Medical Centre, Owo ishereby acknowledged. We are also grateful to Mr.Olanitori E.B for his contribution to this work. Finallywe glorify God for the insight and the strength tocarry out this work.

REFERENCES1 Foster A., Gilbert C., and Rahi J. Epidemiology

of cataract in childhood: a global perspective. JCataract Refract Surg 1997;23 (suppl):601-604.

2 Wilson M.E., Trivedi R.H. Paediatric cataractin developing world settings In: Wilson M.E.,Trivedi R.H.,Pandey S.K. eds. Paediatric cataractsurgery: Techniques, complications andmanagement. Baltimore, M.D.: LippincottWilliams and Wilkins; 2005: 303-307.

3 Wilson M.E., Pandey S.K. and Thakur J.Paediatric cataract blindness in the developingworld,surgical techniques and intraocular lensesin the new millennium.Br J Ophthalmol 2003;87:14-19.

4 Hornby S.J., Xiao Y., Gilbert C.E., Foster A etal. Causes of childhood blindness in the PeoplesRepublic of China: results from 1131 blindschools students in 18 provinces. Br J Ophthalmol1999;83:929-932.

5 Gilbert C., Foster A., Negrel A.D. et al .Childhood blindness: a new form for recordingcauses of visual loss in children. Bull World Healthorgan 1993;71:485-489. [Medline].

6 Rahi J.S., Sripathi S., Gilbert C.E. and Foster A.Childhood blindness in India : causes in 1318blind school in nine states. Eye 1995;9:545-550[Medline].

7 Eckstein M.B., Foster A. and Gilbert C.E.Causes of childhood blindness in Sri Lanka:results from children attending six schools forthe blind. Br J Ophthalmol 1995;79:633-636.[Medline].

8 World health Organization. Preventing blindness inchildren. Report of a WHO/IAPB Scientificmeeting. WHO/PBL/00.77 Geneva: WHO,2000.

9 Gilbert C.E., Anderton L., Dandona L et al.Prevalence of blindness and visual impairmentin children- a review of available data .OphthalmicEpidemiol 1999;6:73-81. [Medline].

10 Kello A.B. and Gilbert C. Causes of severevisual impairment and blindness in schools forthe blind in Ethiopia.Br J Ophthalmol 2003;87: 526-530.

11 Ezegwu I.R., Umeh R.E. and Ezepue U.F.Causes of childhood blindness: results fromschool for the blind in South Eastern Nigeria. BrJ Ophthalmol 2003;87(1); 20-23.

12 Gilbert C. and Awan H. Blindness in children .BMJ 2003;327(7418): 760.

13 World Health Organization. Global initiative forthe elimination of avoidable blindness . WHO/PBL/97.61. Geneva:WHO,1997.

14 Alagaratnam J., Sharma T.K., Lim C.S. andFleck BW.A survey of visual impairment inchildren attending the Royal Blind School ,Edinburg using the WHO childhood visualimpairment data base. Eye 2002;16(5): 557-561.

15 Hornby S.J., Adolph S., Gothwal V.K., GilbertC.E., Dandona L. and Foster A. Evaluation ofchildren in six blind schools of Andhra Pradesh.Indian J Ophthalmol 2000;48:195-200.

16 Kocur I., Kuchynka P., Rodny S., Barakova D.and Schwartz EC. Causes of severe visualimpairment and blindness in children attendingschools for the visually handicapped in CzechRepublic.Br J Ophthalmol 2001;85:1149-1152.

17 Umeh R.E., Chukwu A., Okoye O. et al.Treatable causes of blindness in school for theblind in Nigeria. J Comm Eye Health 1997;10:14-15.

18 Gilbert C.E., Wood M., Waddel K., et al. Causesof childhood blindness in East Africa: results in491 pupils attending 17 schools for the blind inMalawi, Kenya and Uganda. Ophthalmic Epidemiol1995;2:77-84. [Pubmed].

19 Waddel K.M. Childhood blindness and lowvision in Uganda. Eye 1998;12:184-92.[Pubmed].

20 Lloyd I.C., Goss-Sampson M., Jeffrey B.G., etal. Neonatal cataract: aetiology, pathogenesis andmanagement. Eye 1992;6:184-96. [Pubmed].

21 Robb R.M., Peterson R.A. Outcome oftreatment for bilateral congenital cataracts.Ophthalmic Surg 1992;23:630-56. [Pubmed].

The World of African Neurosurgery

THE TENTH E. LATUNDE ODEKUMEMORIAL LECTUREOCTOBER 29, 1985

BY

ADELOLA ADELOYEMB MS(Lond) FRCP(Edin) FRCS(Eng) FACS FICSFWACS FMCS(Nig) FAS DSc (Hon)Professor of Neurological SurgeryDepartment of SurgeryUniversity of IbadanIbadan, Nigeria.

IntroductionAll of those in this Medical School and TeachingHospital felt a great sense of loss when the newsarrived of the death of Tuesday 20th August 1974, atthe Hammersmith Hospital in Londin, of ProfessorE. Latunde Odeku. He was 47 years old.

Professor Odeku was a much liked personality, whowas totally devoted to the practice of medicine, andhis particular specialty of Neurosurgery, which hepracticed here for 12 years.

Professor Odeku received his basic medical educationin the United States of America, where he also had hisspecialist training in Neurosurgery.

He was appointed a Lecturer/Consultant inNeurosurgery in Ibadan in October 1962, and waspromoted to Senior Lecturer in October 1963. InNovember 1965 he was appointed as Professor inNeurosurgery.

He was Head of the Department of Surgery fromJanuary 1969 to September 1971, and was Dean ofthe then Faculty of Medicine for the period 1968 to1970.

He was Visiting Professor of Neurosurgery at theUniversity of Michigan in October 1970, and wonthe prestigious Alumni Award of Howard University,U.S.A. in 1973. By the time Professor Odeku died, hehas established a fully functional Neurosurgical Unitat Ibadan, the first in Black Africa, and thus carvedfor himself a permanent niche in the history of theUniversity of Ibadan, the Ibadan Medical School, andNeurosurgery in Africa.

In appreciation of his contribution to theNeurosciences, his colleagues and friends at home andabroad launched a Memorial Fund with the objectivesof endowing a Memorial Lecture and a prize inNeurosurgery in his memory. The Council of theUniversity, apart from contributing to the EndowmentFund, also approved the naming of the MedicalLibrary as the E. Latunde Odeku Library, in

recognition of his outstanding academic contributionsto the University.

There have been nine previous Memorial Lectures,given by distinguished scholars from Nigeria andabroad. These previous lecturers included ProfessorR. C. Schneider, of Michigan University, U.S.A.,Professor A. O. Adesola, University of Lagos,Professor Marion Mann of Howard University,Professor F.A.O. Udekwu, of the University ofNigeria, Professor F. D. Martinson, of our ownDepartment of Otorhinolaryngology, ProfessorCharles Easmon of the University of Ghana,Professor B. O. Osuntokun, presently Chief MedicalDirector of U.C.H., Ibadan, and former Dean of ourFaculty of Medicine, Professor L. F. Levy of theUniversity of Zimbabwe, and, last year, by ProfessorOjetunji Aboyade, one-time Head of the Departmentof Economics in Ibadan, and later Vice-Chancellorof the University of Ife.

Tonight, as our Tenth Memorial Lecturer, we haveProfessor Adelola Adeloye. Professor Adeloye is sowell known to us, that it is almost superfluous tointroduce him. Suffice it to say; firstly that ProfessorAdeloye was one of the closest associates of ProfessorOdeku in the Department of Surgery, since he joinedthe Department in 1968 as Lecturer/Consultant.Today, not only is Professor Adeloye himself one ofthe most distinguished Neurosurgeons in Africa, anda former Head of our Department of Surgery, but isalso well known as a medical historian.

I now have great pleasure in calling on ProfessorAdeloye to deliver the 10 th E. Latunde OdekuMemorial Lecture, which he has titled:“THE WORLD OF AFRICAN NEUROSURGERY”

Professor D.G. MontefioreActing Provost, College of MedicineUniversity of Ibadan

29th October 1985Ibadan

CHRONICLES OF MEDICAL HISTORY IN AFRICA 2


The World of African Neurosurgery

by

Prof. Adelola AdeloyeNeurosurgery Unit, Department of Surgery,University College Hospital,University of Ibadan, Nigeria.

The Acting Provost of the College, Deans of Faculties,Acting Director of the Postgraduate Institute ofMedical Research and Training, Head of Surgery,Chairman, Medical Advisory Committee, CollegeSecretary, College Librarians, Visitors from theInternational College of Surgeons, Distinguished Ladiesand Gentlemen.

In January 1985, Dr. Adell Patton Jr, a black Americanauthority of African history from Howard University,Washington, District of Columbia, USA visited me inIbadan. A Fulbright scholarship awarded to him hadbrought him to Freetown, Sierra Leone to continuehis studies in West African history. During ourdiscussions, Adell Patton asked me if I would like to‘re-visit’ some of the grounds covered in An AfricanNeurosurgeon,1 the biographical treatise which I wrotein 1976 on the late Professor E. Latunde Odeku, acelebrated alumnus of Howard University, in whosememory this lecture was endowed. We identified oneor two areas which can profit from special and newemphasis in addition to some errors and inaccuracieswhich those who read the book had kindly broughtto my attention. I subsequently went ahead to interestHoward University in the project. It was about thattime that the invitation came to me to deliver this, theTenth Odeku Memorial Lecture, first, from the thenHead of Surgery, Professor Olajide O. Ajayi, then frommy department of surgery and finally from ProfessorAbiodun Johnson, the Provost of the College ofMedicine, University of Ibadan. For me, the invitationto deliver this lecture is a great honour which I accept,and will always cherish with gratitude.

Over eleven years ago, Professor E. Latunde Odeku(fondly called Lat) died at the Hammersmith Hospital,Ducane Road, London, England, at the comparativelyearly age of 47 years. As a mark of respect for himand in appreciation of his sterling contributions tomedicine and medical education, an annual lecture wasestablished in his memory to be devoted to topicsand subjects relating to neurosciences or medicaleducation or poetry. The first in the series of theselectures was delivered in November 1976 by ProfessorRichard Schneider,2 a neurosurgeon of Ann ArborMichigan, USA. Up till now, nine of these lectures havebeen delivered by six surgeons, a pathologist, aneurophysician and an economist.

Today is the occasion of the Tenth Odeku Lecture. Inpresenting to you “The World of AfricanNeurosurgery”, I intend to share with you, first, someof the less well-known aspects of the life and timesof Odeku and later to examine a evaluate the

development and achievements of neurosurgery inAfrica since the days of Odeku.

Section 1: Another Look at “An AfricanNeurosurgeon”On the occasion of the First Lecture in 1976, the thenVice Chancellor of the University of Ibadan, ProfessorTekena Tamuno who presided at the lecture, launchedOdeku’s biography, An African Neurosurgeon. The bookwas very well received. Inevitably, it was found laterto contain some flaws and omissions in details forwhich the author accepts full responsibility. Here is anopportunity today to correct some of the flaws.

On page 4 of the book, the household of the Odekusat Awe in Oyo State is referred to as Adu-bi-Eiye (jetblack like a bird) erroneously thought to be used todescribed the dark complexion of the Odekus.3 Thecorrect appellation is Adun-bi-Eiye (which refers to thesound of birds).4 The Ninth Odeku MemorialLecturer, Professor Ojetunji Aboyade, the celebratedNigerian economist and native of Awe effected thiscorrection for us.

While still on voices and songs, Odeku’s favouriteoperating theatre melody, ‘Green Green Grass ofHome’ as related on page 33 of his biographicaltreatise, belongs to Tom Jones and not Paul Jones.5

We also learnt from Professor Aboyade that longbefore we knew Odeku as a leading surgeon andmedical educator, the late Professor had betrayedsingular love for our indigenous apala music,6 notablythe type which was popularized by the late HarunaIshola of Ijebu-Igbo. No doubt, the muse in Odekumust have been stirred up by the poetry of HarunaIshola’s music.

A product of the same Awe Community as Odeku,Professor Aboyade in the Ninth Memorial Lecturesupplied other details of the non-scholastic and non-professional side of Odeku and his vast interest in theproblems of his countrymen and women in thebackyards of life. These candid revelations and hometruths about the early life of “brother Latunde” byAboyade helped enormously to produce a welcomebalance in the assessment of the character of Odeku,portrayed, alas, as “a virtual expatriate” by those whowrote about him from later impressions.7

On page 12 of the book reference was made to OgunOlogbon (Ogun, the wise), Odeku’s favourite classmate,as David Ogundipe. The surname should be Ogun,not Ogundipe. They spent six delightful years togetherat Methodist Boys High School (MBHS), Lagos, theirfinal year being 1945. In 1978, David Ogun was theDirector of Nigerian Standards Organization divisionof the Federal Ministry of Industries, at Ladipo Avenue,Ikorodu Road, Lagos. His unit is responsible for theestablishment of national standards and it supervisedour change over to the metric system.

David Ogun in his letter to me, dated February 191978, lamented the premature death of his school mate,


who had blossomed into a “versatile Nigerian medicalgenius”. Ogun touchingly praised his classmate whohad made enormous contributions to medicine in aquite unobstructive manner and without seekingpublicity. Hence, Ogun acclaimed the “excellentbiography” of Odeku, for without it, Latunde wouldhave been like the desert flower “born to blush unseenand waste its sweetness in the desert air”.

The latter quotation which was contained in Ogun’sletter is from the first of the three poems he sent tome as some of the favourite lines of their class ofpoets and lovers of poetry. His second poem ran thus:

“Self-reverence, self-knowledge, self-control.These alone lead life to sovereign power.Yet not for power,Power herself will come uncalled for,But to live by lawActing the law we live by without fear,And since Right is right to follow rightWere wisdom in the scorn of consequence”

One of those who witnessed the birth and beginningof neurosurgery at the University College Hospital,Ibadan was Odeku’s first house surgeon, Dr. PaulOkwudili Chuke, who graduated M.B.B.S. (London)from Ibadan in October 1961. This piece ofinformation is missing from An African Neurosurgeon.Odeku must have exerted a profound influence onhis young houseman, for Paul Chuke who later gainedthe F.R.C.P. (Fellowship of the Royal College ofPhysicians) of Canada is today of Professor ofMedicine at the University of Nigeria, Nsukka, withspecial interest in neurology. Chuke was there inNairobi, Kenya in January 1972 when, with Odeku,his old chief in the vanguard of operations, the PanAfrican Association of Neurological Sciences(PAANS) was formed. At that time, Chuke presentedto the scientific congress the experiences with braintumours from Zambia, a country where he oncefunctioned as the Director of Medical Services.

On page 35 of the book it was stated that the ClinicalSciences Building on t he grounds of the UniversityCollege Hospital, Ibadan “was Odeku’s creation”.8

This is not true. As the Dean of Medicine, Odeku hada lot of achievements, but it must be corrected at thislecture that the Clinical Sciences Building where manyof the Consultants of UCH Ibadan have their officeswas born out of the efforts of Professor OladipupoAkinkugbe, Professor of Medicine, and his ableassistants during his tenure as Dean. The complexcomprising the Clinical Sciences Building, the presentOdeku Medical Library, the Paul Hendrickse LectureTheatre and the Biode building housing thePostgraduate Institute for Medical Research andTraining was developed by Professor Akinkugbeduring his tenure as Dean of the then Faculty ofMedicine, 1970 – 1974. In order to achieve conformitywith the existing architectural structure of the oldbuilding of the teaching hospital, Mr. Pring was brought

in from the firm of Watkins Gray Woodgate andPartners, the original designers of the UniversityCollege Hospital, Ibadan.9 It needed a lot of energy,persuasion, motivating leadership and delicatediplomacy on the part of Dean Akinkugbe to secureunanimity of consent among all concerned to weldthese buildings into the functional complex we havetoday. It was declared opened, as we already know,by the then Head of State, Yakubu Gowon in May1975 when Professor Osuntokun was Dean ofMedicine and Professor H. Oritshejolomi Thomas wasthe Vice Chancellor of the University of Ibadan. Anymisconceptions presented in the early publicationstemmed from the ignorance of the author for whichthis public apology is tendered.

Section 2. Odeku in North AmericaUndergraduate Career in WashingtonMore details about the student career of Odeku inNorth America have been supplied recently by hiscolleagues, teachers and some of his MemorialLecturers who knew him in those days. In 1977, forinstance, I got a letter from Dr. D. M. Yilla of Bo,Sierra Leone after he had read An African Neurosurgeon.

“I can only congratulate you for both yourunselfish attitude and the degree onintellectual honesty shown”, he wrote,…….. “the mistakes I notice in your bookare small and inconsequential”“Latunde Odeku was my contemporaryand we shared a room in Washington whileboth of us pursued our first degree (BSc.)at Howard in 1949-51. I left him in theUnited States and proceeded to pursue mymedical course in Glasgow Universitywhile he opted to stay in America to do hisM.D. I have lost complete touch with himsince 1951.”10

I wrote back in earnest to Dr. Yilla who in his initialletter asked for appointment to consult us on his owndisabling neurological illness. I never heard from himagain. The visit of Professor Marion Mann to Ibadanas the Third Odeku Memorial Lecturer in 1978 andsubsequent contacts, established in person or bycorrespondence with Odeku’s American colleaguesfurnished more information about his undergraduatecareer in Washington.

Odeku went to Howard University in 1947. Afterleaving the Methodist Boys High School, Lagos, heapplied to Edinburgh University for a place in themedical school: he was accepted there, but he couldnot begin his studies until 1948. Anxious to start hisuniversity education without much delay, he acceptedan earlier admission offer from Howard University.11

Odeku entered its College of Liberal Arts in 1947,from where he graduated summa cum laude with theBachelor of Science degree in Zoology in April 1950.We now know that apart from his principal subjectsof Chemistry, Physics and Biology, his subsidiaries atthe College of Liberal Arts were German, History,


Sociology, Phylosophy and a smattering of Survey. InZoology, his teacher was Dr. L. A. Hansborough, anassociate professor; in physics he had Professor FrankColeman. His friend and guardian, Professor WilliamLea Hansberry taught him history and Professor AlainL. Locke lectured him in sociology and philosophy.

Odeku’s performance in the College of Liberal Artswas quite outstanding making A grades in all his subjectsat every examination, except once when he scored Bin organic chemistry. In summing him up at this stageof his education, the Professor of philosophy admittedthat Odeku was “one of the most outstandingstudents” known to him in twenty years, and LeoHansberry of History described him as his “ablest,finest and most promising student in Howard”.12 Itwas not surprising that at the end of Odeku’s coursein Liberal Arts, his four lecturers unanimously andenthusiastically endorsed his application for a place inthe College of Medicine of Howard University.

On January 16, 1950, as part of the preparation toenter the College of Medicine, Odeku hahd to take anaptitude test in which his percentile rating was 780 andthe professional score was 535 to 595. On Friday, May26 1950, at the invitation of Crystal E. Malone,admission Clerk of Howard University, Odekuappeared for an interview before the Dean of theCollege of Medicine, Dr. Joseph L. Johnson.

At the interview which lasted most of the afternoonof that day, Odeku spoke softly, reasoned well, and atthat young age of 22 years, demonstrated a goodpractical philosophy of life. He was interviewed in agroup of four candidates: one white boy called Schoen,two West Indians and himself.

Odeku and Schoen were rated to be outstandingcandidates; the other two were described as “goodbut not equal to Mr. Odeku”.13 The Dean wasconvinced that Odeku had more that just the ability tomake high grades at examinations. In accepting him tostudy Medicine in Howard, the Dean from the outsetrecommended that Odeku should be given every“opportunity for a most comprehensive and thoroughmedical education including Hospital externship duringsummer, followed by residency, specialty training andboard certification before returning to Africa”.14

That was what ultimately happened. Between July andSeptember 1950, we knew that Odeku went to Israelunder the auspices of the Foundation for WorldGovernment. Odeku never forgot Dean Johnson towhom he sent a complimentary card of thanks fromJerusalem, a place which Odeku then described as the“magnificent, historic and eternal city of peace”.15 Hewas late in coming back to United States from Israel,due to some visa problems in Ellis Island. With hisacceptance in Howard University College of Medicine,Odeku conveniently forgot his application for a placein McGill University, Canada. On September 28, 1950,he started his medical education in Howard University.

From that date until June 6, 1954 when he graduatedM.D. from Howard University, Odeku recorded a“superb performance as a student and as anindividual”.16 At the end of his Freshman year, heranked No.1 in a class of 74 students. At the end ofhis Sophomore year, he was No. 3 in a class of 70; inthe Junior year, he was placed 3rd in a class of 71.When he graduated, he was No. 3 in his class of 69and received an award of Honorable Mention for hiswork in the department of Obstetrics andGynaecology. For all these attainments, he was placedon the Dean’s List for scholarstic achievement.

Odeku’s career in the College of Medicine at Howardwas successful in every aspect. According to Robert S.Jackson, who was the Dean of Medicine when Odekuqualified,

“He was at all times polite, cooperative,earnest and conscientious in theperformance of his assignments. He wasof neat appearance and had good rapportwith colleagues and patients. His poise wasgood; he accepted criticisms well, oftenquestioning the reason if he felt it necessary.His mood was consistent, he had an eagerto and assertive temperament and madegood use of past errors.17

Throughout his career, he gave the impression of agood but no subservient follower, an independentthinker and a potential leader.

It was hardly surprising that Odeku was extremelypopular among his colleagues and co-workers manyof who recall his personality of honesty, humility,integrity and dedication to excellence. As a student inHoward, Burke Syphax (now Professor of Surgery atHoward) knew that Odeku was a prodigious worker,even though things seemed to come to him easily.18

He had a special knack for working all day and readingor writing half the night and getting by with 3 to 4hours sleep. According to another mate of his,Professor Jessie Barber, a neurosurgeon in Howard,Odeku kept a little black book in which he had enteredthe names and diagnoses of the patients he hadexamined and worked up. Hence, Odeku could, at ashort notice, present his personal statistics and discussthe problems of patients.19 Such was the very wellorganized personality of this man.

Yet, he had a tremendous zest for life and enjoyed allhis associations and associates. Even in those early days,he betrayed to those who knew him well that strongsense of priorities that was to run throughout his entirelife. His professionalism and concern for health caretranscended politics, polemics and all pettiness. He was,however, singularly uninformative when asked aboutrace and tribes in Nigeria. Sometimes too, he paid littleor no attention to the racial issues in America. How allthese traits affected and influenced his total life is nowcommon knowledge. Marion Mann, the Third OdekuMemorial Lecturer was Odeku’s classmate in Howard.


Mann in his lecture here in 1978 complimented Odekuwhen he confessed that “in our class during almost aquarter century, indeed from the very beginning, theone classmate most liked for his friendliness, mostadmired for his towering intellect, most esteemed forhis extraordinary achievements in medicine andeducational leadership was E. Latunde Odeku”.20

Life as a Young DoctorOdeku’s basic medical education in North Americawas not completed until he had studied and trained inMichigan. In December 1953, before he graduated inHoward, he had applied for his rotating internship tothree hospitals, namely Michael Reese Hospital,Chicago; King’s County Hospital Center, Brooklyn,New York and the University Hospital, Ann Arbor,Michigan. He ultimately went to Ann Arbor for hisrotating internship from July 1954 to June 1955. Heneeded special permission from the United StatesDepartment of Justice, Immigration and NaturalizationService to extend his stay in that country to serve hisrotating internship. His mind kept going back toHoward and in a letter written from MichiganUniversity Hospital to Dean Joseph L. Johnson anddated November 1, 1954, Odeku expressed hisindebtedness and that of his classmates, to HowardUniversity.21

Dear Dean Johnson:For the past few days I have felt the spontaneous urgeto forward a few lines of greetings to my landmarkof medicine, Howard University. The inevitablenecessity of graduation has torn us away from withinthe physical and the protective walls of the College ofMedicine; but our spirit abides with the school forever.Time will not forbear with me to overburden yourinvaluable moments with the many striking details ofthe first few months of this early postgraduate stageof our professional development. However, suffice itto report that my commencement at HowardUniversity has excellently stood me in good stead,practically always. We offer competitions to and takecompetitions from our various selected colleagueswithout any uneasy sense of the least inadequacy; andnot infrequently we have a few gems in the bag thatwe brought with us from Howard to share with thosearound us who so desire. In general, life here hascontinued as expected. We have no earthshakingspectacular discoveries; no particular complaints; noregrets.

Please extend my warmest and most gratefulcompliments to all the professors and the many otherteachers from whom it has been my priceless privilegeto receive that for which no man can ever pay. And

Our pray’rs with thee forever restOur gratefulness unexpress’dOur struggles in the years to comeShall beamn our deeds and crowns to theeIn hopeful thoughts of gratitude

Kindly permit me the free interchange of the singularand plural pronouns that abound continuously above;for I cannot help but express the sincere feelings thatall my classmates wherever they are, share this deepsense of thankful yearning with me. And in the hopefulassurance that the College of Medicine shall ever be abulwark to all of us, and to the innumerable hosts thatwill follow in the countless brighter years ahead, wecheerfully continue this life-long struggle to ourdestinies.

Your pupil as always,

E. Latunde Odeku, M.D.

All these personal attributes and his later success in thefield of medicine and of medical education as ascholar, leading African neurosurgeon, author andhumanitarian won Emmanuel Latunde Odeku theprestigious Charter Day Alumni Award of HowardUniversity in March 1973. In recommending him forthe Award in October 1972, Dean of Medicinespecially cited Odeku’s outstanding record at Howardand his insistence on returning to his native Nigeriaafter qualifying as a neurosurgeon. Odeku was theyoungest of the four alumni honoured on thatoccasion.22 William James Moore, one of the recipientswas 100 years old (born on August 4, 1872) and thenthe oldest living alumnus of Howard University fromwhere he graduated in 1892. He was too old to travelto Washington where he was honoured for hiscontribution in the fields of education and athletics;someone had to receive the award on his behalf. JulianDugas, aged 55, was honoured for his contribution tolaw and public service and Mrs. Dorothy Harris, aged60 years was honoured for her contribution toeducation and community service.

Before becoming a neurosurgeon, we knew thatOdeku came home to Nigeria after his rotatinginternship in Michigan but not before he had taken theL.M.C.C (Licentiate of Medicine) from Canada.According to Professor Fabian Udekwu, the FourthOdeku Memorial Lecturer, those of them whograduated from America in the 50s could not practicemedicine in Nigeria with their American diplomasonly.23 In those days, American Universities wereregarded as something les than cow colleges. Nigeriangraduates of American Medical Colleges were sent toCanada for the L.M.C.C.

After obtaining his Canadian Licentiate, Odeku cameto Nigerian with Dr. Uzoechuna Nwagbo, a Chicagograduate and both of them worked at the LagosGeneral Hospital as house officers. Their “experiencesof disappointments and moments of pride, offrustrating dissatisfactions and hopes” were chronicledin Odeku’s The Things We Suffer, a paper which hecirculated to all the Nigerian students in the UnitedStates.24 Around the time of Nigeria’s politicalindependence, by which time Odeku had become atrained neurosurgeon, the policy of Nigeriangovernment to American education had changed,


thanks, as we shall see later, to the efforts of Sir SamuelManuwa.

Residency Training in MichiganBy June 1956, Odeku was back in Michigan for hisresidency training in neurosurgery. How he came tosecure the position in Michigan was described bySchneider, the First Odeku Memorial Lecturer (Fig.3). As an intern on pathology in 1955, according toDr. Edgar Kahn, Odeku “came on the NeurosurgeryService for only one month. We were impressed withhis intelligence, his kindness to patients and his greatdignity that I told him at the end of the month that ifhe ever wished to make neurosurgery his career, therewould always be a place for him” in Ann Arbor.25

Odeku had applied for residency programme inToronto, Canada. With his acceptance of the sincereand firm offer from Edgar Kahn in Michigan, for thesecond time in his career, Odeku had to abandon plansto seek medical education in Canada.

During his period of training from 1956 to 1960, hewas well-liked by both the permanent and residentstaff. He was kind, easy-going most of the time andmaintained a position of dignity and responsibilitywherever he went. On every available weekend, hewent to Washington to see his family.

There was with Odeku at that time an associate residentfrom South Korea, Dr. Hun-Jae Lee. They sharedadjacent rooms in the house officers’ dormitory. Inthe basement they had facilities to cook and makesnacks. Hun-Jae Lee was fond of cooking octopuson the stove in spite of the bad unsatisfactoryventilation of the basement. That habit later generateda lot of arguments between Odeku and Lee; ultimately,in the interest of peace, the Korean had to give up hisunusual menu of fried octopus.26

Odeku impressed his mentors so much so that he wasgiven permission to visit various universities in NorthAmerica during his residency. It was, however, notuntil his fifth year that he had the opportunity to presenta scientific paper in the United State. In that year, hepresented the paper on “Cervical Spondylosis withneurologic deficit” which was prepared by himselfand Richard Schneider.27 The paper formed asubstantial part of the First Odeku Memorial Lecture.28

Schneider, for ever, feels privileged to have givenOdeku his first break in the presentation of scientificpapers, an exercise in which Odeku later in life becamequite prolific and proficient. Odeku attended manyneurosurgical conferences during his residency and itwas at one of these that he met Laurence Levy,neurosurgeon of Zimbabwe (then Rhodesia) who in1983 gave the Eighth Odeku Memorial Lecture herein Ibadan.29

Odeku’s residency was such an outstanding that inrecognition of his special attributes, he was unanimouslyselected in 1970 by the Neurosurgery Department ofMichigan to deliver the First Edgar Kahn Lecture

established to honour Dr. Kahn after his retirementfrom active neurosurgical practice in Michigan.30 OnOctober 23, 1970, Odeku delivered the Kahn Lectureon “The Personality of Nigerian Neurology”.31 Arecent letter from Richard Schneider, dated October7 1985 carried the sad news of the death of EdgarKahn on August 29, 1985.In recognition of his ability as a writer, the Michiganschool of Neurosurgery invited Odeku to contributea chapter to the second edition of “CorrelativeNeurosurgery” edited by Kahn, Schneider and Crosby.Odeku wrote on “Exotic Lesions of the Brain”.32 Bythe time the book ran into the third edition, Odekuhad died and Schneider had become the senior editor,assisted by Kahn, Crosby and Taren. Odeku wasoffered two University positions in North America.He turned both down and returned to Nigeria to helphis own people. Inspite of difficulties, Odeku quicklysettled at home in Nigeria; the rest of the story of thebeginning of neurosurgery in Ibadan is well known.33

“I am quite busy”, Odeku wrote to his friend andclassmate, Marion Mann in a letter dated May 3 1964,“and neurosurgery is now well established here. Ourprimary limitation, everyday as with you, is Time”.34

Some ten years later, Odeku died on August 20, 1974,having successfully pioneered neurosurgery in Nigeriaand indeed in West Africa.

Section 3: Odeku in IbadanContributions to NeuroscienceWhat Odeku achieved or initiated during his career atthe University of Ibadan and the University CollegeHospital, Ibadan had been documented by those whowrote on him in various places and at different timesand who spoke of him in their capacity as OdekuMemorial Lecturers. These achievements areencapsulated in neurosciences, medical education andpoetry. Very little justice has been done in these lecturesto the contribution of Odeku to poetic andphilosophical literature. While we wait for this tohappen, I endorse the position of Professor Martinson,the Fifth Odeku Lecturer, namely, sincere admirationfor Odeku’s poetic works.35

On the other hand, considerable attention has beengiven to his vast contributions to neurosciences in theFirst and Seventh Lectures delivered by RichardSchneider of Michigan and B. O. Osuntokun ofIbadan, Nigeria.36

It was Odeku’s passionate plea that African scientistsshould patronize and popularize local journals that ledto the appearance of his many early articles in Nigerianand West African medical journals. We know that itwas his “cranio-encephalic trauma”37, an extensivereview article on head injuries that took up the wholevolume 4 of the Journal of the Nigerian Medical Associationin 1967. Among his many “first case in the African”are the “Rubinstein-Taybi syndrome” which Sinnetteco-authored38 and the “Aneurysm of the Great veinof Galen complicated by chronic subdural abscess”which he wrote with Professor O. Ransome-Kuti,39


our present Federal Minister of Health. The diseasesfor which he and his colleagues in Ibadan found andformulated patterns in the African are legion; theseinclude epilepsy, cerebrovascular disorders, headinjuries in general and subdural haematoma inparticular, tumours of the central nervous system andcongenital malformations of the neural axis.40 Someof the diseases he described have become classics.Congenital insensitivity to pain associated with auditoryimperceptions by Osuntokun, Odeku and Luzzatto41

has been named eponymously after the first author,Osuntokun.42 After Odeku’s early case reports of thisdisease, in 1971, Adeloye and Odeku convinced themedical worlds in a detailed description of 18 casesof the identity of congenital subgaleal cyst of theanterior fontanelle.43 In 1978, the congenital disorderbecame known as Adeloye-Odeku disease.44 In that year,similar case reports from Zimbabwe by Glasauer andothers; Glasauer later reported a case in a Negro childin North America.45 It was then suggested that thedisease is probably unique to Africans and those ofAfrican descent. Isolated cases were later reported inwhite Americans, in an Indian and in a Mexican childand later in other non-Africans to challenge thesuggestion made that it is an African disease.46 And sogoes on the story of the congenital dermoid cyst ofthe anterior fontanelle. Today, isolated cases have beenencountered in China and Japan.47,48 Even these reportsstill show that most cases come from African and theAfricans in places and amongst peoples where diseasesare more likely than in others to be under-documented.Ojikutu and others49 and Chaudhari and others50 fromLagos recently reported more cases. Indeed, Olumide’s1980 extensive review of congenital CNSmalformations in the African included 66 cases of thisdisease seen in Ibadan in 14 years.51 Nothing is newunder the sun. Odd swelling of the head were knownin our folklore, as exemplified by Alade in Odunjo’sbook.52 In old medical literature, in 1824, referencewas made to a giant cyst of the anterior part of thehead in a man called Lake who kept the house calledthe Six Bells at Darford, England.53 But by our detailedstudies and writings from Ibadan, we have establishedcharacteristics of this simple but frightening congenitaldisorder, such that it has now attained the dignity of adisease entity and an identity of its own.

Odeku and Medical EducationLet us now assess the contribution of Odeku in therealm of medical education. In my opinion, his greatestachievement was his successful intrusion as an Americangraduate into an established British system to win respectand recognition for American medical graduates.

For a long time, Anglophone West African colonialgovernment did not favour the employment ofAmerican medical graduates. At a meeting of WestAfrican Directors of Education held in Accra, Ghanain February 1951, it was maintained that a first degreefrom an American College accredited in Nigerian didnot automatically confer graduate status on the holder,since the American qualification was regarded as atbest only equivalent to an Intermediate degree of a

British University. By the same token, medical graduatesof United States universities were not admitted to thepermanent register of the United Kingdom. Such wasthe medical educational policy existing in Nigeria whenOdeku qualified.

At the Third Conference of Directors of West AfricanMedical Services held in Ibadan in March-April 1952,it was made clear that this British culture and attitudemust change. Then, because of the urgent need fordoctors, medical departments in West Africa were underconsiderable pressure to employ doctors who werenot eligible for admission into United KingdomRegister. Sir Samuel Manuwa explained at great lengththat Nigeria wanted to admit non-British graduates, ifonly to a temporary register, for practice underconditions prescribed by the Inspector General ofMedical Services.54 His proposal was accepted solelyas an interim measure until an adequate number ofWest Africans with local or United Kingdom degreeswere available. Even the question of automaticadmission to a permanent Register of Americangraduates who had passed the National BoardExamination of the United States was deferred at thatmeeting.

The success of Odeku as a neurosurgeon was one ofthe factors which finally helped the efforts of theadministration led by Sir Samuel Manuwa, to removethe unwarranted prejudice nursed against Americanmedical graduates. By his personal performance andfine example, Odeku helped in a huge way to pavethe way for the unconditional acceptance in Nigeriaof non-British medical graduates. That was not all ofhis contribution. In later years, Odeku became one ofthe pillars and planers of medical education in Nigeria.Below, in his own words, was what he stated in 1972in a letter to Professor Michael Bankole on the planningof health services for Nigeria.

Dear Mike,Thank you for your letter of April 13, 1972. I amsorry for the delay in forwarding these few lines inreply to the letter, with particular reference to theenclosed proposed budget on medical education fornational service in Nigeria.

I think the whole idea is highly commendable and Iwill suggest that the letters to the possible participantsbe carefully re-worded in a cogent manner as to pin-point the specific aims and duties involved in about apage so that there be a careful re-thinking, re-wording,re-organisation of the suggested titles so that thecomplete final list will appear in the order in whichthey are expected to be seen “developmentally” in thebook. The suggested authors could then be placedafter each title and this will not prejudice the positionin which the names come, since what is important willbe the titles themselves.

I would think that a book of this nature cannot bepublished without having more authors involved in it.I will therefore suggest joint authorship of many of


the titles, namely 2 or 3. Experienced educators suchas Professor H. O. Thomas (Lagos), Professor T. A.Lambo, Professor H. C. Kodinlinye, Professor J. O.Mabayoje, Professor I. Audu, Professor F. Dosekun,Professor A. Adesola, Sir Samuel Manuwa, ProfessorA. O. Lucas, Professor M. Ogbeide, Professor G.Onuaguluchi and Professor Nwokolo, - not to mentiona number of others – will have to be included in thelist of possible authors to be considered and fromwhich larger list a final selection should be made.

I think it will be a great credit to the objectivity andfundamental reference of the proposed volume if awider representation of a total experience of healthmatters throughout the Federation is reflected in theauthorship. I hope we will have the opportunity ofdiscussing this work with you when you return toIbadan before a final list is made for circulation amongpossible authors.

With best wishes.

As always.

E. Latunde Odeku,Professor of Surgery

Odeku MemorialsAfter his death, many suggestions were made as tohow best to honour him. In Howard University, thesenior student prize in neurosurgery was re-designatedthe Green-Odeku prize. Scholarship funds for his childrenwas considered not only inadequate, but thought ofas a cold and rather empty gesture, not in consonancewith Odeku’s life philosophy. Establishing a permanentchair in his name in the department of surgery wasconsidered more appropriate.55

Table 1 WINNERS OF THE E. LATUNDE ODEKU PRIZE IN NEUROSURGERY

1976 A. B. O. ADEGBITE 1977 I. AGBAPUONWU 1978 No Examination 1979 W. C. MEZUE 1980 O. O. KUKOYI 1981 I. L. OKEKE

G. U. LEKWAUWA 1982 H. C. J. E. ONYIUKE 1983 O. K. OWOEYE 1984 ALICE B. IKURU 1985 E. T. LAMUREN

Here at home, buildings and roads were to be namedafter him and prizes and lectures were set up in hisname. The Odeku Memorial Lectures formed one of theseenduring memorials; let us examine a couple of thesememorials in some details.

The E. Latunde Odeku Memorial LibraryThe medical library on the grounds of the UniversityCollege Hospital, Ibadan was named after Odeku in a

special release issued on November 20, 1975 by thethen Acting Medical Librarian, Mr. T.A.B. Seriki. Thechristening of the library was performed on therecommendation of the then Faculty of Medicine,headed by Dean Osuntokun, and its subsequentapproval by the Senate of the University of Ibadan.The re-naming of the library was in appreciation ofOdeku’s pioneering work in neurosurgery in Ibadanin particular and in Nigerian in general.

In pursuance of the objective of achieving “Healthfor all by 2000” the World Health Organization (WHO)embarked recently on a policy of developing its HealthLiterature services throughout the world. Eachmember state of the Organization is requested to namea library as the focal point for the purpose. Such a centrewill function by (a) organizing a network of healthscience libraries within its area of influence (b)developing a health literature service and informationnetwork, assisted by WHO, for all categories of healthworkers in the country, (c) indexing all local healthscience journals published, and (d) compiling a nationalbibliography of medicine.

In response to this call, the Federal Ministry of Healthof Nigeria, a few years ago, designated the E. LatundeOdeku Medical Library the National Focal Point notonly for the above purposes, but also to serve as acontinuing education centre for medical libraries inNigeria and West Africa.

What we have at the library today is only part of whatit is design to look like ultimately. It is still growing. Itis yet to go up by another floor. The furtherdevelopment of the library should be placed high onthe priority list of the University of Ibadan and itsteaching hospital. If the E. Latunde Odeku MedicalLibrary must attain its full status, it must beremembered at all times that it is more than a FacultyLibrary; it is the library of the College of Medicine.

Odeku Prize in NeurosurgeryAnother important memorial set up to honour Odekuis the E. Latunde Odeku Prize in Neurosurgery which wasestablished in 1975. The prize has been awarded yearlysince 1976, except in 1978 when as a result of thestrike action by the students, the Department of Surgerycould not conduct the special examination required.

(Table 1). Students in their final year compete for theprize through a written examination in neurosurgery.It is gratifying that the prize winners have beendistinctive students many of whom by theirpostgraduate achievements, have brought considerablecredit to the prize.

Lamuren in 1985 and Miss Alice Ikuru in 1984 werethe best overall students in surgery and both werewinners of the departmental prize in surgery in 1984and 1985. Hilary C. Onyiuke, the 1982 winner andpresently one of our residents in surgery in Ibadan, asan undergraduate gained distinction in Anatomy in 1979and this year, he has passed three primary examinations


for the Surgical Fellowship, first of the West AfricanCollege of Surgeons (April, 1985); next the NigerianCollege of Surgeons of England in July. As a student,he won the certificate of Honour from Mellanby Hall,University of Ibadan, for debating.

Okeke, one of the two prize winners of in 1981 isalso one of our residents in surgery. He has passed thePrimary examination for the Fellowship months agoand should be sitting the Part 1 of the Fellowship ofthe Nigerian Postgraduate College next month (hepassed in November 1985). W.C. Mezue, the prizewinner in 1979 has passed the Part 1 of the Fellowshipexamination resident at the University of NigeriaTeaching Hospital, Enugu.

All these distinguished young doctors should take theircue from the example of the first winner of Odekuprize, Dr. Andrew B. Adegbite, who today is aConsultant Neurosurgeon, practicing in Sudbury,Ontario Canada. It is an interesting coincidence thatOdeku himself had spent part of his vacation in thatregion of Canada in 1972. Indeed, the sabbatical leavewhich Odeku intended to take in 1973 just before hetook ill was to have been spent in Sudbury. WithAdegbite flying the neurosurgery flag in Sudbury, itappears that the spirit of the Old Master is living inone of his most distinguished students.

Andrew Adegbite went to Igbobi College, Lagos from1965 to 1970 and qualified in medicine at the Universityof Ibadan in 1976. In that year, apart from the Odekuprize, he gained the Adeola Odutola prize in medicineand the departmental prize in Preventive and SocialMedicine.

It was Adegbite’s choice to go out of the UniversityCollege Hospital, Ibadan for his pre-registration housejobs which he served in nearby Adeoyo State Hospital,Ibadan. He took care to keep close to us, and it wasthen he betrayed to me his ambition to train inneurosurgery. Then, we started to make plans. Afterhis national service at Port Harcourt School HealthBoard, Adegbite immediately went to Regina GeneralHospital in Saskatchewan, Canada, for one yearresidency in general surgery. By July 1979, he got onthe neurosurgery training programme of the UniversityHospital of Saskatoon. He successfully completed histraining and gained the Fellowship of the Royal Collegeof Surgeons of Canada in December 1983. Beforethen, in March 1981, he had passed the primaryexamination of the American Board of NeurologicalSurgery and completed a short course in micro-neurosurgery in Florida, USA in May 1982.

Section 4. African NeurosurgeryBut the world of African neurosurgery is much olderthan these prize winners. In its slow but steady growthand development, Odeku was the pioneer in his nativeNigeria and one of the pacesetters in Africa. In his lifetime, outside Ibadan here, neurosurgery was establishedin Lagos by the late Colin da Silva in 1968 and in Enuguby Samuel Ohaegbulam in 1974. Outside Nigeria,

Odeku exercised a far-reaching influence throughnational and international associations and societies onthe growth and development of neurosciences,particularly neurosurgery, in various parts of Africa.His teacher and friend, Dr. Richard Schneider summedthis up in the First Odeku Memorial Lecture when herelated the comments of a Cairo Surgeon that “Odekuhas done a lot for neurosurgery not only in Nigeriabut for all of Africa. He has been trying to bring ustogether into a compact Federation of Association ofneurosurgeons such as you have in America or inEurope”.56

Odeku’s pioneering efforts in the formation of theNigerian Society of Neurological Sciences in 1966 andof the Pan African Association of NeurologicalSciences in 1972 have been adequately described byhis biographers and his Memorial Lecturers of 1982and 1983.

What was African neurosurgery like before Odekucame and what is its status today?

When Odeku took up his appointment here in October1962, neurosurgery was actively practiced in NorthAfrica, notably Egypt, and in Southern Africa. Apartfrom Zimbabwe, then Rhodesia, there was noneurosurgery in any part of black Africa.

Let us start our excursion into the world of Africanneurosurgery from Egypt where at least a decadebefore Odeku came to Nigeria, modern neurosurgeryhad become an established discipline. Parts of theMediterranean littorals, like Morocco and Algeria werevisited occasionally by French neurosurgeons. Althoughlong celebrated in antiquity for the practice ofneurosurgery and the treatment of neurologicaldisorders, modern neurosurgery came to Egypt onlyin 1951. It was started in three centres by three Egyptianpioneers, namely Osman Sorour of Cairo University,Samuel Boctor of Alexandria University and IbrahimHigazi of Ein Shams University, Cairo. Sorour whotrained in Sweden was to become the First Presidentof the Pan African Association of NeurologicalSciences in 1972; the Second Vice President of theWorld Federation of Neurological Societies from1973 to 1977 and now an Honorary President of thesame body, for life.

El Banhawy who succeeded Higazi at Ein Shams wasa close friend of Odeku. I recall that Odeku wrotespecifically to El Banhazy in 1971 to win his supportfor the information of an African Neurological Society.A big jump takes us from Cairo in the North to CapeTown in the Republic of South Africa whereneurosurgery had flourished for decades and whichplace has more neurosurgical manpower, materials andsophistication than other parts of the African continent.There are now about 56 neurosurgeons serving the20 million people on the Republic, with neurosurgicalcentres in Durban, Cape Town, Johannesburg andPretoria. The relatively young Durban neurosurgicalcentre in Natal was pioneered by Erasmus; Joubert is


in now in charge. The older centre at Cape Provincewhich is based at the Groote Schurr Hospital providedmaeterials for the review of brain tumours with brainneoplasms in Africa.57 The Professor of neurosurgeryat Groote Schurr, J de Villiers, is the current AssistantTreasurer of the World Federation of NeurosurgicalSocieties.

The centre at Baragwanath Hospital, Johanesburg, inthe Transvaal was established in 1956. It serves theBantu population in the Transvaal.58 Colin Froman whopractised neurosurgery at this hospital was at the 1972Nairobi conference. It needed special appeal to theKenyan government by Odeku and others to bringColin Froman from South Africa to the Nairobiconference of 1972.

Let us go to Zimbabwe, north of the Republic ofSouth Africa, where neurosurgery was started byLaurence Levy at the Godfrey Huggins Medical School,an institution which he had strong links with theUniversity of Birmingham, England. Levy, whodelivered the Eight Odeku Lecture here in 1983 trainedin Canada, USA and England and worked in Hararealone until 1972 when he was joined by Auchterlonie,a Glasgow graduate who trained in neurosurgery atGroote Schuur Hospital, Cape Town. For many years,Levy and his team provided neurosurgical coveragefor Malawi and Zambia.

Neurosurgery in Odeku’s TimeThese centres were in existence in Africa before Odekucame to Nigeria in 1962. The attainment of politicalindependence by many countries in Africa positivelyinfluenced the development of neurosurgery on thecontinent. Free from the shackles of foreigndomination, the newly independent countries startedto extend and modify the scope and style of theirmedical services. New medical school were established,old one were expanded, novel disciplines and boldideas were introduced into medical curricula as thesecountries reached out to new frontiers of medicine. Itwas in such a setting that neurological surgery was bornin Sudan, Tunisia, Algeria and in Nigeria and othercentres were set up in Egypt. In the latter country, thenew neurosurgery centres were set up in the MaadiMilitary Hospital; at the Religious University of AlAzhar and at Tanta, midway between Alexandria andCairo. It is not surprising that today, Egypt has about45 neurosurgeons who belong to the Egyptian Societyof Neurological Surgery, an association which togetherwith similar societies in Iran, Lebanon and Turkeyconstitute the Near East Neurosurgical Society.

The background information on how neurosurgerywas introduced to Nigeria was provided by ProfessorFabian Udekwu in the Fourth Odeku MemorialLecture which he gave in 1979. According to him,Kenneth Dike, the Principal of University College,Ibadan saw through the colonial mentality which uptill that time discouraged the practice of neurosciences;he went ahead to establish the Unit of Psychiatry,neurology and neurosurgery with the assistance of

Rockefeller Foundation.59 Odeku himself later salutedthe imagination of the principal officers of theUniversity of Ibadan when he attributed the“Beginnings of Neurosurgery in Ibadan” to “thedynamic programme of the Vice Chancellor Dr.Kenneth Dike and the vigilant outlook of the newDean of the medical school, Professor J.C.Edozien.”60

When Odeku died in 1974, neurosurgery had expandedits coverage considerably in Africa.61 Centres wereestablished in Libya (by Reddy), Sudan (by Abu Salihin 1971), Kenya (by Ruberti in 1967), Uganda (IanBailey in 1969), Zaire (by Groenbaek in 1967), Senegal(by Courson in 1967), Ghana (by Mustaffah in 1969)and Zambia (Zohrabian 1971). Other pioneers wereAcquaviva (Morocco), Bettaieb (Tunisia) and Abadain Algeria.

Courson who brought neurosurgery to Dakar, Senegalwas succeeded by Cournil and later, Alliez.Neurosurgery was formally introduced to Ghana, acountry with a population of 14 million, by JamesMustaffah in 1969. He was trained in England andstarted neurosurgery at the Korle Bu Hospital beforehe came to Accra, all grossly obvious intracranialhaematomas, subdural in particular, were subjected toburr hole drainage by general surgeons.

In Kenya, the earliest neurosurgical procedures werecarried out at the Kenyatta National Hospital in the1940s by J. F. Jarvin, the Chief consultant of head andneck surgery.62 He operated on some congenitalmalformations. His successor and chief of the Headand Neck Surgery Unit, Peter Clifford, also operatedon cases of hydrocephalus and brain tumours, afterspending a few months with Pennybacker of Oxford.The Italian, Renato Ruberti, a graduate of Padua,brought neurosurgery to Kenya in 1967, setting up inprivate practice in Nairobi. He had visited Kenya beforeon hunting safaris and became enchanted by thecountry. Ruberti consulted at the Kenyatta Nationaland the private Aga Khan Hospital. Fellow Italianneurosurgeons came out to assist Ruberti for shortspells; Dr. Poppi in 1970 and Dr. Carmaganani in 1971.In 1972, Ruberti set up his private Nairobi NeurologicalCentre at the Nairobi Hospital.

The Head and Neck Surgery Unit of Kenyatta NationalHospital attended to neurosurgical emergencies untilin 1972 when a Kenyan who trained in neurosurgeryin Canada, J. Nabwangu, joined the staff. He soonbecame disillusioned and returned to Canada within ayear. To the rescue of neurosurgery came ProfessorHarold Paxton of Portland, Oregon who spent a year’ssabbatical leave in Kenya in 1973.

Neurosurgery after OdekuThe growth of neurosurgery on the continent continuedafter 1974. More centres were founded and in somecounties, Africans started to take charge of existingneurosurgical units. The neurosurgical manpower inEgypt and South African has increased. In Kenya, late


in 1974 Dr. Jawahar Dar of India joined he Facultyof Medicine and took charge of the neurosurgery unitthere; Dar trained under Tandon at the All-IndiaInstitute of Medical Sciences, New Delhi. Darestablished a Division of Neurosurgery at the KenyattaNational Hospital with 25 beds for electiveneurosurgery while keeping the head injuries in thegeneral surgical wards.63

Africans who trained in neurosurgery appeared on thescene, starting with G.M. Sande who joined the unit in1976 the year the Kenyatta National Hospital started afive-year training programme in neurosurgery.64 Sandewas the secretary of the Pan African Association ofNeurological Sciences (PAANS) from 1981 to 1983.After 3 years training in Kenya, Sande spent a year inBelfast and Glasgow. Dr. M. Kahwa of Uganda laterjoined Dar’s Unit. Kahwa had started neurosurgicaltraining in Kampala where he wrote a dissertation oncompound depressed fractures for his Master ofMedicine degree (M.Med) of Makerere University. Thepolitics of Uganda forced him out of his country; hewanted to come to Nigeria but soon settled forneighbouring Kenya. Two young Kenyans, Dr. A.Maingi in 1979 and Dr. J. Mwangombe in 1982 laterjoined the Kenyatta Neurosurgical Unit for training.Dr. Maingi died in a road traffic accident beforecompleted his training. Professor Dar left Kenya in1982 and returned to India where he is now aConsultant Neurosurgeon at the G.B. Pan Hospital,New Delhi.

In the private sector in Kenya, Dr. Ruberti was joinedin 1977 by Dr. M. Risso, another Italian who 2 yearslater set up his own private practice in Nairobi, Kenya.The Neurological Society of Kenya was formed in1972.

In Uganda, the African Jovan Kiryabwire took overthe unit left by Ian Bailey of Belfast who now lives atLisburn, Northern Ireland. Kiryabwire started out asa general surgeon and later trained in neurosurgeryunder Professor Valentine Logue at the NationalHospital for Nervous Diseases at Queens Square,London. Dr. M. Kakwa, after 4 years training in Kenya,went back to Kampala as a Consultant Neurosurgeon.In Zaire, following the earlier suggestion by Dr. R.Wirth, one time Dean of the Faculty of Medicine atthe University of Lovanium (now the NationalUniversity of Zaire) that neurosurgical services bedeveloped in that country, a Neuro-psycho-pathologyCentre was established in Kinshasa University in 1973by Dr. G. Dechef, for the integrated practice ofneurosciences. The Centre has modern diagnostic toolsin neurology and neurosurgery. Africans training inneurosciences start at this centre; those in neurosurgerycomplete their tutelage either at the Free University ofBrussels under Professor Jean Brihaye with the financialassistance of the Belgian Ministry of Foreign Affairsor in the United States under Professor JosephRansohoff of New York University. Such anarrangement produced Dr. Shako Djhnga, MD ofBrussels who returned to Kinshasa in 1974 to practice

neurosurgery in his native land. He did not actually“start absolutely from square one” because Groenbaekhad tackled some diseases of the craniospinal axis inthat country years before. However, Shako’s commentis true that “to be able to make the necessary transitionto his home environment, it is crucial for the ThirdWorld neurosurgeon to be a “jack-of-all-trades”, to be aG.P. of neurosurgery, trained not merely inneuropathology, neuro-ophthalmology and neuro-anaesthesiology. The Third World neurosurgeon mustmake a tremendous intellectual effort compared tohis more fortunate colleagues.65.

Back home here in West Africa, Dr. Kanga, a nativeof Ivory Coast started off in the University Hospital,Abidjan, in 1974. He even offered services of aperipatetic practitioner to the neighbouring country ofLiberia. Kanga soon left to set up in private practice.Dr. Cournil who was once in Dakar is now the onlyneurosurgeon in the University Hospital in Abidjan.

In 1978, a West African of Sierra Leone, Dr. Ulric Jonesintroduced neurosurgery to Connaught Hospital,Freetown. Jones, a graduate in medicine from Durham,England, trained in neurosciences in Japan at the NihonMedical School of Tokyo. Jones recently left SierraLeone to take up a job in Gambia.

In Ghana, the teaching hospital complex in Accra hasa neurosurgery ward with about 70 beds, under Dr.James Mustaffah. The Ghanaian, Fredua Agyeman,who trained in Switzerland, came home early in 1980,but soon returned to Europe. In September 1980, Dr.Buenor Puplampu, another Ghanaian and M.D. ofToronto University returned to Accra and set up hislittle neurosurgery unit at the Military Hospital, Accra.66

Puplampu had his neurosurgical training at Wayne StateUniversity in Detroit and later went into private practicein San Jose in northern California for years before hereturned to Ghana. With him today is Dr. M.I. Iddrissuwho trained in Germany and who practices at theKorle By Hospital.

In Dakar , the Senegalese African, ProfessorMohammed Gueye took over the neurosurgerydepartment of Fann Hospital after Alliez left. Gueyeis assisted by two Africans on the unit, namely, Dr.Seydou Baidane and Dr. Youssoupha Sakho.

More recently in Nigeria, in 1981 Dr. Nnamdi Ibe, aproduct of European Neurosurgery returned homeand set up in private practice in Owerri. Two yearsago, Peter O. Binitie, FRCS started neurosurgery inJos and only this month, Dr. Anthony Ige who trainedin neurosurgery with us here in Ibadan, the firstneurosurgery product of the Nigerian NationalPostgraduate Medical College, started his neurosurgeryunit at the University of Ife, Ile-Ife. It is also expectedthat Dr. Ellams who trained in West Germany willsoon start his neurosurgery unit at the Ahmadu BelloUniversity Hospital, Zaria. Dr. S. O. Onabanjo, ourSenior Registrar at UCH Ibadan in 1978-80, is now


Consultant Neurosurgeon at the Ring Road StateHospital, Ibadan.

Africa in World NeurosurgeryThe picture of the present status of neurosurgery inAfrican can be made more complete by reference tothe figures contained in World Directory ofNeurological Surgeons. In 1985, 11,654, were listedon the register of the World Federation. With anestimated world population of 4,722,000,000 (4,7billion) there is now one neurosurgeon to 450,000people. This is still less by about 50 percent than theestimated one neurosurgeon to 200,000 people whichis the basic neurosurgical manpower requirements ofthe world.

In Africa, where the shortage is still more acute, therewas in 1985 a surgeon population ratio of oneneurosurgeon to 20,950,000. Review of the figuresduring the last 15 years shows a small but steadyimprovement in neurosurgical manpower on thecontinent. In Nigeria, the shortage is even more acute,with one neurosurgeon to 8 million people.67 Whatthen is the future of neurological surgery in Africa?

Future of Neurosurgery in AfricaWe agonize today about the dwindling professionalismin medicine in general and in specialist surgery inparticular. Materials to work with diminish graduallyleading to a fall not just in the number of operationsperformed but also in morale and job satisfactionamong surgeons. Academic performance is alsoaffected. As financial constraints limit research activities,the scientific papers meant for publication lacksophistication and those ingredients which characterizethe growth frontiers of medical knowledge. These days,attendance at international conferences by manypractitioners on the African continent is a punishmentand an apology, since many do not appear tounderstand the language of the current technologicaladvances in medicine.

In Nigeria, the recent call in certain quarters for a de-emphasis of specialist surgery, exemplified byneurosurgery, ostensibly to promote primary or basichealth care is another blow to the sinking hope ofimprovement in specialist surgery in the country.68

Clearly, tertiary medicine is already very much part ofour health care delivery and the presence of manyskilled specialists emphasizes its established position.We know from Aboyade, the Ninth Odeku Lecturerthat Odeku himself believed in basic health care forthose of us in the backyards of life. Other OdekuMemorial Lecturers harped on the subject. MarionMann described primary health care programme asthe most significant development in health planningduring the sixties. Fabian Udekwu devoted a gooddeal of his lecture to primary health care. “First andforemost”, he said, “we must plan for the health ofthe community and not the individual. Secondly, wemust get the community involved in its own health”.69

During this fortnight in Nigeria, the Nigerian chapter

of the International College of Surgeons is holding aseminar on the role of the surgeon in primary healthcare delivery.

Neurosurgery in Africa needs a lot of input andassistance from primary health care. Thus, head injuriesin children which result from falls from staircases andhigher floors at home; injuries to our teenagedpedestrians and to vehicle passengers who travel incars without safety belts and falls among those climbtrees for fun or on duty, can all be prevented. Hugebrain tumours which are basically benign are oftenneglected and allowed to progress to cause blindnessor attain unmanageable, inoperable dimensions; suchlate presentations can also be prevented. These tumoursneed early detection at the basic health care unit butthey also need prompt evaluation and treatment at thesecondary and tertiary tiers of our health care deliveryif the whole patient must be treated.

Odeku himself predicted in his “Beginnings ofNeurosurgery” that the “various difficulties ofneurosurgery in Africa are by no means confined toits beginning but will continue and grow into the yearsahead”. How true. Adesola in the Second MemorialLecture described the dilemma and paradox ofacademic surgery in the developing countries wherehealth planners consider subspecialities as toosophisticated, but at the same time reject as “unpatrioticthe practice of referring patients with complicatedsurgical problems to Europe or North America forspecialist treatment”.70

The answer is not confrontation but cooperation withand education of our health planners and of ourcolleagues inside and outside government. We all needto remember that primary health care needs othersystems of medical care vice versa; that our resourcesshould be managed in such a way that the Africanpatient with measles or malabsorption or meningiomaget properly treated in his or her own milieu and thatwhat specialized medicine and surgery want are basictools, not necessarily sophisticated equipments, to justifyits existence and to function meaningful and withoutpretentions. The presence in Africa today of manymedical specialists makes the latter need even morecompelling. These campaigns are not to be conductedby individuals alone, but by national and internationalsocieties of neurosciences throughout the length andbreadth of Africa.

In Nigeria, the recent move by government to havespecialist hospitals functioning as centres of excellencemay well solve some of the dilemmas of specialistsurgery in this country. The concept of “centres ofexcellence” should be given a good chance. It will not,and should not matter what area of excellence isassigned to a centre for development. True excellencecannot evolve in isolation since other disciplinesinevitably interrelate to any chosen area of medicine.A well-developed centre of excellence can onlypromote overall improvement in the total health caredelivery in that area.


Neurosurgery in NigeriaI do not know what the University College Hospital,Ibadan, will get in the allocation of centres ofexcellence. But what I know, without being immodest,is that the contributions of Ibadan to neurosurgery inNigeria, in Africa and in the world, begun by Odekuand sustained by his successors, has been quitesubstantial. In recognition of all these, it is high timefor us to have a full-fledge department of neurosurgeryhere in Ibadan if this surgical specialty must grow anddevelop. The department can be created by theUniversity of Ibadan and its teaching Hospital or beendowed by philanthropy as is done elsewhere in theworld.

African Neurosurgeons AbroadThere is another issue: the problem of Africanneuroscientists trained abroad and staying away fromhome. In Africa, the slow improvement inneurosurgical manpower can be quickened if the manyAfricans trained abroad return home to practice. Sofar, the number of neurosurgeon produced by localtraining programmes in West and East Africa are veryfew. On the other hand, many more have completedtheir training abroad. The zero-sum, saturated societyin the developed countries which trained these Africansdoes not encourage them to stay.71 their host countrieswant them to go back home to Africa.

Their countries of origin, unfortunately, sometimesreject them as expensive super-specialists. Many of themdo not want to return home because of the political,economic and professional uncertainties of Africa.Some come home, genuinely desirous to work, butsoon, like the proverbial Andrew of Nigeria, check outand return before the doors of employmentopportunities close behind them in Europe and Africa.A few stayed for years after which, exhausted incommitment or disenchanted or confronted togetherwith their young ‘foreign’ families, by the harshpragmatic realities of life quietly return or give up activeneurosurgery. All these variations in adjustment andcommitment by African neurosurgeons trained abroadput Odeku’s pioneering efforts in Nigerianneurosurgery into bolder and better perspective andclearly underlines his unique place in the World ofAfrican neurosurgery.

Concluding RemarksYet, the neurosurgery practised today in Nigeria is notwhat Odeku practised in the sixties or dreamt of inthe seventies. In the early days of Odeku in Ibadan, inthe immediate post-independence years, everybody wasinfested with the fever of nationalism and the call forparticipation by all in nation building and in mouldingthe destiny of our renascent Africa. In that crusade,Odeku was in the vanguard of the development ofneurosciences. Then, we were well motivated; then,there was job satisfaction; foreign organizations andfoundation came in with generous investment inmedical education and research in our new world.There was a lot of movement in those years because

we all believed, and the world thought so too, that wewere going somewhere.

Two decades later, these hopes of achievement andthe sense of mission are still to be fulfilled. Themovements of those early years have either stoppedor slowed down, or become wobbly or uncertain,taken us one step forward and sometimes two stepsbackwards.

In 1974 Odeku planned to go away to Canada; histwo or three years leave would have taken him intothe beginnings of years of economic slump in Nigeria.At the end of his stay, he would have been faced withsome options. He could draw on his old resource ofdedicated service and return home to help our ailingmedical service. He could argue that having successfullyestablished neurosurgery, he could retire to Canada.There could be other options. The question is, whichway would he have turned?

Whatever the answer, by then and as shown in hismemorial lectures, Odeku’s place in the building ofneurosurgery in Africa had been established. What anobituary columnist wrote of Wilfrid Card of Glasgowin the British Medical Journal of February 9, 1985described Odeku very well:

He will be remembered by his friend forhis unfailing good nature and sense of fun;his immense integrity and loyalty; by hiscolleagues for his vision and intellect andhis unflagging enthusiasm for theadvancement of medicine; by his patientsfor compassion and unremitting care; byhis students for the clarity of his teachingand encouragement; and by the scientificcommunity for his elegant pellucidwritings.72

To all these can be added the compliment that Odekuwill be remembered by Africa and the world for histowering stature in neurosurgery.

With his many merits, personal attributes and nationaland international achievements, Odeku embodiedmodesty, charity and unhurried grace. Above all these,like John Bunyan, he had an inner strength because hewas contented with what he had, “be it little or much”,sustained by the Christian belief that:

“Little here, and hereafter blissIs best from age to age”

AcknowledgementsI am indebted to the following in the preparation ofthis lecture: Drs. Adell Patton Jr, Richard Schneider,Marion Mann, Jesse Barber and other friends in NorthAmerica; all past Odeku Memorial Lecturers; theDepartment of Surgery and the College of Medicine;Medical Illustration and Biomedical CommunicationDepartments of the University of Ibadan and the


College of Medicine, members of my family, myfriend and colleagues and my secretaries and otherinvaluable assistants.

References1. Adeloye, A. (1976) E. Latunde Odeku: An

African Neurosurgeon. Ibadan University Press,143 pp.

2. Schneider, R.C. (1976) “Cervical Spondylosisand Cervical Spinal Stenosis”. The First E. LatundeOdeku Memorial Lecture. Ibadan University Press,16 pp.

3. op. cit. Reference 1, page 4.4. Personal communication with Professor Ojetunji

Aboyade5. op. cit. Reference 1, page 33.6. Aboyade, Ojetunji (1984). “In the Backyards of

life: Challenges of a National Health Policy”. TheNinth E. Latunde Odeku Memorial Laecture,October 30, 1984.

7. ibid.8. op. cit. Reference1, page 35.9. Communication from O. O. Akinkugbe to A.

Adeloye, November 30, 1976.10. Communication from Dr. D.M. Yilla, Prince

Williams Street, Bo, Sierra Leone to A. Adeloye,January 16, 1977.

11. Communication from Mrs. Jill Odeku, LinfordAvenue, Newport Pagnell, Bucks, 1976.

12. Personal communication with Marion Mann,Dean of Medicine, Howard University,Washington, 1978.

13. ibid.14. ibid.15. ibid.16. ibid.17. ibid.18. Syphax, Bo to Jesse Barber, May 22, 1985.19. Barber, Jesse on E. Latunde Odeku, June 27,

1985.20. Mann, Marion (1978) “A Decade of Changes in

American Medical Education”. The Third E.Latunde Odeku Memorial Lecture, October 1978,Ibadan University Press, page 1.

21. Odeku, E.I, Ann Arbor, Michigan, to DeanJoseph L. Johnson, Howard University,November 1, 1954.

22. Charter Day Convocation: Commemorating theone hundred and sixth anniversary of the foundingof Howard University, Washington, D.C. 1973.

23. Udekwu, F.A.O. (1979) “Medical Education inNigeria: a look at the future”. The Fourth E.Latunde Odeku Memorial Lecture, November1979, 12 pages.

24. op. cit. reference 1, pages 91-118.25. op. cit. reference 2, page 1.26. Personal Communication with Professor Richard

C. Schneider, October 7, 1985.27. Odeku, E.L. and Schneider, R.C. (1966). Cervical

spondylosis with neurologic deficit. MichiganMedicine, 65, 533-544.

28. op. cit. reference 2.

29. Levy, Laurence (1983) “Society, Doctors andLiability”. The Eight E. Latunde Odeku MemorialLecture, November 1, 1983.

30. op. cit. reference 26.31. Odeku, E.L. (1972) “The Personality of

Nigerian Neurology”, Ibadan University Press.25pp.

32. Odeku, E.L. (1969) Chapter 15 in CorrelativeNeurosurgery (2nd Edition), Charles C. Thomas,Sprignfield, pp 266-278.

33. Odeku, E.L. (1965). Beginnings of Neurosurgeryat the University of Ibadan, W. Afr. Med. J. 14,85-98.

34. Odeku, E.L. University of Ibadan to Evelyn andMarion Mann, Washington, D.C., May 3, 1964.

35. Martinson, F.D. (1980) “Oto-rhino-laryngology inthe medical curriculum and its contributions toneurosciences”. The Fifth E. Latunde OdekuMemorial Lecture, November 1980, IbadanUniversity Press, page 1.

36. Osuntokun, B.O. (1982) “The Mustard seed: E.Latunde Odeku and the growth of neurosciencesin Africa” The Seventh E. Latunde OdekuMemorial Lecture, December 1982, IbadanUniversity Press, 63 pages.

37. Odeku E. L. (1967) Cranio-encephalic trauma.Its evaluation and management. J. Nig. Med. Ass.4, 5-47.

38. Sinnette, C. and Odeku, E.L. (1968). TheRubinstein-Taybi syndrome. The first case in anAfrican child. Clin. Paediat, 7, 488-492.

39. Odeku, E.L., Ransome-Kuti, O and Elam, H.P.(1967) Aneurysm of the Great Vein of Galencomplicated by chronic subdural abscess. W. Afr.Med. J. 16, 109-113.

40. op. cit. reference 36, page 12.41. Osuntokun, B.O., Odeku, E.L. and Luzzatto,

L. (1986) Congenital pain asymbodia and auditoryimperceptions. J. Neurol. Neurosurg. Psychiat. 31,291-6

42. op. cit. reference 36, page 12.43. Adeloye, A. and Odeku, E.L. (1971). Congenital

subgaleal cysts over the anterior fontanelle inNigerians. Arch. Dis. Childh. 46, 95-98.

44. Mohanty, S; Clezy, J.K.A. and Adeloye, A. (1978)Dermoid cysts of the anterior fontanelle (Adeloye-Odeku disease). J. Neurosurg. 48, 627-8.

45. Glassauer, F.E., Levy, L.F., Auchterlonie, W.C.(1978). Congenital inclusion dermoid cyst of theanterior fontanelle. J. Neurosurg. 48, 627.8.

46. Adeloye, A. (1980) Congenital malformations ofthe craniospinal axis in Ibadan, Nigeria. ZeitschriftKinderchirurgie 31, 327.

47. Adeloye, A. Neurosurgery in Africa. IbadanUniversity Press.

48. Kanamaru, K. and Waga, S. (1984) Congenitaldermoid cyst of the anterior fontanelle in aJapanese child. Surg. Neurol. 21, 287-90.

49. Ojikutu, N.A. and Mordi, V.P.N. (1980)Congenital inclusion dermoid cyst located overthe region of the anterior fontanelle in adultNigerians J. Neurosurg. 52, 724.


50. Chandhari, A.B., Ladapo, F, Mordi, VPN,Choudhury, KJ and Nassem, A (1982) Congenitalinclusion cyst of the subgaleal space. J. Neurosurg.56, 540-4.

51. Olumide, A. A. (1980). Congenital malformationof the central nervous system in the African. SearaMed. Neuro-cir 9, 19-28.

52. Odunjo, J.F. Alawiye, Iwe Kerin Eko, page 66,Longmans, Nigeria.

53. op. cit. reference 46.54. Third Conference of Directors of West Africa

Medical Services at Ibadan, Nigeria, 31st Marchto April 1, 1952 in C.O. 554/541/62879 – PublicRecords Office, London, pp 48-50.

55. op. cit. reference 12. Stuart Scheiner to MarionMann, Sept. 13, 1974; Marion Mann to StuartScheiner, Sept. 24, 1974.

56. op. cit. reference 2, page 2.57. Louis, S. and Keet, P.C. (1964) A survey of

supratentorial gliomas and meningiomas. S. Afr.Med. J. 38, 949.

58. Froman, C. and Lipschitz, R. (1970).Demography of tumours of the central nervoussystem among the Bantu African population ofthe Transvaal (South Africa). J. Neurosurg. 32, 660.

59. op. cit. reference 23, page 3.60. op. cit. reference 33.61. Adeloye, A. (1982) Neurosurgical Education in

Africa. Neurosurgery 10, 130-132.

62. Dar, Jawahar (1985). Perspectives in InternationalNeurosurgery: Neurosurgery in Kenya.Neurosurgery, 12, 241.

63. Dar. J. (1980) Specialty training-neurosurgery: Itsplace in a developing country. Medicom 2, 241-2.

64. Djhnga, Shako (1983). Perspectives inInternational Neurosurgery: neurosurgery in Zaire.Neurosurgery, 13, 95.

65. Puplampu, Buenor (1983). Perspectives inInternational Neurosurgery: neurosurgery 12, 241.in Ghana. Neurosurgery, 12, 241.

66. Adeloye, A. (1983). Perspectives in InternationalNeurosurgery: Neurosurgery in Nigeria.Neurosurgery 13, 333.

67. Ransome-Kuti, O. (1978). ‘New Health CareScheme Soon’ in Daily Sketch, Nigeria. September20, 1985, page 1.

68. op. cit. reference 23, pages 9-10.69. Adesola, A.O. (1978) “An Academic Surgeon’s

Dilemma”. The Second E. Latunde OdekuMemorial Lecture, Ibadan University Press, page9.

70. Bergland, R.M. (1981). Neurosurgery in a zero-sum society. Neurosurgery 8, 280-2.

71. British Medical Journal, February 9, 1985, page474.


DIAGNOSIS AND TREATMENT OF DRUG-RESISTANT TUBERCULOSIS

Neil W. Schluger, MDCopyright 2000 Up-to-date®, Inc. •(800) 998-6374•(781) 237-4788

Tuberculosis remains one of the leading causes ofmorbidity and mortality throughout the world. Itsmanagement has become more complex because ofincreased resistance to commonly used antituberculosisdrugs.

This article will review the diagnosis and therapy ofpatients infected with drug-resistant strains of M.tuberculosis. Treatment of drug-resistant tuberculosiscan be difficult, and may necessitate the use of second-line drugs or resectional surgery. Therefore,management of patients with resistant disease shouldonly be undertaken by, or in very close consultationwith, experts in this area. Good patient outcomesdepend upon a rapid and accurate diagnosis, and theinstitution, administration, and monitoring of propertherapy 1.

Definitions – There are a series of definitions usedin describing the different types of drug-resistanttuberculosis:

• The term “drug-resistant tuberculosis” refers tocases of tuberculosis caused by an isolate ofMycobacterium tuberculosis which is resistant to oneof the first-line antituberculosis drugs: isoniazid,rifampin, pyrazinamide, ethambutol, or streptomycin.

• Multidrug-resistant tuberculosis (MDR-TB) iscaused by an isolate of M. tuberculosis which is resistantto two or more of the first-line chemotherapeuticagents, usually isoniazid and rifampin.

• Primary drug-resistance is said to occur in a patientwho has never received antituberculosis therapy.

• Secondary resistance refers to the development ofresistance during or following chemotherapy, for whathad previously been drug-susceptible tuberculosis.

DiagnosisThe diagnosis of drug-resistant tuberculosis dependsupon the collection and processing of adequatespecimens for culture prior to the institution of therapy.Sputum cultures are positive in 85 to 90 percent ofcases of pulmonary tuberculosis, and every attemptshould be made to collect adequate material beforetreatment is initiated. If a patient cannot produce anexpectorated sputum sample, sputum induction shouldbe performed. If an adequate sample is still notproduced, bronchoscopy may provide diagnosticmaterial for culture in a significant number of cases,although the risks of nosocomial transmission ofpotentially drug-resistant tuberculosis must be weighedagainst the benefits of the procedure 2. In patients withexclusively extrapulmonary disease, samples of

involved tissue (eg, lymph nodes, bone, blood) shouldbe obtained.

Susceptibility testing for first and second-line agentsshould be performed at a reliable reference laboratory.Improvements in culture techniques and moleculardiagnosis (eg, PCR) permit more rapid identificationof antibiotic resistance than was possible when solidmedium alone was used for mycobacterial growth 3-5.

A careful history must be obtained from every patientwith tuberculosis before treatment begins, and whiledrug susceptibility data are pending. Clinical andradiographic features on presentation are not alteredin comparison with drug-susceptible disease, but severaldemographic and historical features should raise thesuspicion of drug-resistant tuberculosis. These include:

• Previous treatment for active tuberculosis,particularly if therapy was self-administered

• HIV infection • Contact with a case of drug-resistant tuberculosis • Failure to respond to empiric therapy, particularly

if adherence to therapy has been documented

In cases in which prior antituberculosis therapy hasbeen prescribed, the physician must obtain a completerecord of all previous cultures, drug susceptibilitytesting results, and treatment.

Chemotherapy of Monoresistant DiseaseIf drug-resistant disease is suspected on historical orclinical grounds, therapy should be instituted with atleast two new drugs which the patient has not receivedpreviously, and to which community isolates aresensitive. Treatment of these infections in HIV-infectedpatients is discussed in detail elsewhere.

Effective therapy for drug sensitive and isoniazidmonoresistant tuberculosis is associated with very high(>95 percent) success rates, defined by bacteriologicand clinical response, and low relapse rates (<5percent), defined as recurrence of culture-positivedisease after completion of therapy. The trialsdiscussed below were all performed in HIV-negativepersons; however, similar responses can be expectedin HIV-infected patients, although many expertschoose to prolong therapy for 3 months in this setting.

Isoniazid monoresistance – Tuberculosis resistant toisoniazid (INH) should be treated with a rifamycin(rifampin or rifabutin), pyrazinamide, and ethambutolfor six to nine months, or four months after cultureconversion 6. This recommendation is based upon trialsconducted by the Hong Kong Chest Service/BritishMedical Research Council, which demonstratedsuccess rates of 95 to 98 percent with this type ofregimen among 107 patients with INH-resistant disease7. Some experts consider continuing isoniazid in thesetting of “low-level” INH resistance, ie, resistant toa concentration of 0.2 µg/mL but sensitive to 2.0µg/mL.


DISEASE REVIEW

Some experts recommend extending treatment to 12months for HIV-infected patients who do not converttheir sputum cultures and clinically improve during thefirst 2 months of treatment 6,12.

• Isoniazid and ethambutol given for 18 months, insome cases supplemented by streptomycin for the first3 to 6 months 13-15. We prefer to supplement withstreptomycin as noted when this regimen is used.

Of the 14 mutant RNA polymerase alleles whichconfer resistance to rifampin, 9 also confer high levelresistance to rifabutin 16. Approximately 25 percentof rifampin-resistant clinical isolates are sensitive torifabutin, which appears as effective clinically asrifampin in short-course regimens for patients withdrug-sensitive disease 17-19. However, no data areavailable regarding short-course, rifabutin-containingregimens in patients with rifabutin-sensitive, butrifampin-resistant disease. Therefore, its use in thissetting cannot be recommended at the present time.

The New York City Department of Health hasrecommended that a quinolone (ciprofloxacin 750 mgBID or ofloxacin 400 mg BID) be added to thisregimen for the duration of therapy 8. While thisapproach may be advisable in patients with HIVinfection or an unusually high burden of organisms,no randomized, controlled data demonstrateimproved outcomes with this approach. It has notbeen our practice to add a quinolone in this setting.

Rifampin monoresistance – Rifampin monoresistancemost often occurs in HIV-positive patients andrepresents an uncommon but increasingly frequentclinical problem 9-11. Because rifampin is the cornerstoneof all 6-month regimens, resistance to this drug requiresprolongation of treatment. Treatment regimens arebased on large trials conducted before the introductionof rifampin, in which success rates of greater than 95percent were documented with prolonged treatment.Acceptable regimens include:

• Streptomycin, isoniazid, and pyrazinamide giventogether for 9 months. This is the shortest durationregimen with good efficacy for use in rifampinmonoresistance, and it is our preferred regimen.


Pyrazinamide monoresistance – Single drug-resistanceto pyrazinamide requires a nine month regimen ofisoniazid and rifampin. This combination is well studied,and has a greater than 96 percent success rate in largetrials 20-22.

Monoresistance to other agents – Single drug resistanceto ethambutol, streptomycin, or second line agents isof little clinical significance. Patients can still be treatedwith the standard short course regimen of two monthsof isoniazid, rifampin, and pyrazinamide followed byfour months of isoniazid and rifampin, which yieldssuccess rates of around 97 percent 23.

Chemotherapy of Multidrug-Resistant DiseaseThere have been few controlled trials of the therapyof multidrug-resistant tuberculosis; treatmentrecommendations are based upon experience inpatients who had been previously treated and relapsed,as well as prevailing practices at referral centers forMDR cases 24. The key principle underlying treatmentof multidrug-resistant tuberculosis is to always add atleast two additional drugs to which the isolate issusceptible. Suggested treatment regimens are listed inTable 1, and the dosages of second-line agents aregiven in Table 2.

When resistance is present to two or more of the fourfirst-line oral drugs, most experts recommend that aparenteral aminoglycoside (streptomycin, kanamycin,amikacin, or capreomycin) and a quinolone(ciprofloxacin, ofloxacin, levofloxacin, or sparfloxacin)be added. Treatment with parenteral agents is usuallygiven for six months, and cures rates are high (in the85 percent range) for MDR-TB regimens that includethese two classes of drugs. Treatment in HIV-positivepatients should be continued until 24 months afterculture conversion; follow-up visits every four monthsfor an additional 24 months are recommended tomonitor for relapse 6.

Aminoglycosides – The injectable aminoglycosides mayall cause nephrotoxicity and eighth cranial nervedamage; renal function should be assessed regularly,and audiometry performed on a monthly basis. Theaminoglycoside is usually given daily or either two orthree times per week. Cross-resistance betweenkanamycin and amikacin is common, but is rarelyobserved among the other agents in this class. Patientswho cannot tolerate repeated intramuscular injectionsmay receive intravenous therapy through an indwellingcatheter.

Quinolones – None of the quinolones is approvedby the Food and Drug Administration for use againsttuberculosis, but some authors attribute the improvedprognosis of MDR-TB over the past decade largelyto the administration of these agents [25,26]. Quinoloneantibiotics have considerable in vitro activity againstM. tuberculosis and have become widely used in thetreatment of MDR-TB 27. However, there are fewuseful data from controlled trials to define the exactrole of these drugs in the treatment of tuberculosis.

Sparfloxacin is the most active quinolone against M.tuberculosis in vitro, followed by levofloxacin,ofloxacin, and ciprofloxacin. Complete cross-resistance

within this class of drugs is generally present, so thatresistance to one quinolone implies resistance to all. Asa group, the quinolones are usually well tolerated, withfew treatment-limiting adverse effects exceptphotosensitivity reactions (reported in up to 8 percentof patients taking sparfloxacin) 28,29.

Other second-line agents – A number of second-linedrugs are less active against tuberculosis and haveconsiderable adverse effects, making them difficult touse.

• Ethionamide is a nicotinic acid derivative withstructural similarities to isoniazid. It can cause significant



gastrointestinal upset, hepatitis, neurologic reactions,and hypothyroidism. One mechanism of isoniazidresistance, mutations in the inhA gene, also is associatedwith ethionamide resistance.

• Para-aminosalicylic acid (PAS) is one of the oldestantituberculosis drugs, and inhibits the growth of M.tuberculosis by interfering with folate metabolism. Itis difficult to tolerate secondary to nausea, vomiting,and diarrhea, but a recently approved enteric-coatedformulation may have fewer gastrointestinal sideeffects.

• Cycloserine has been associated with a variety ofadverse psychological effects, including psychosis,anxiety, and severe depression. Pyridoxine (vitamin B6)should be co-administered (at a dose of 50 mg forevery 250 mg of cycloserine) to reduce the risk ofneurotoxicity.

• Aerosolized interferon-gamma has been used asancillary treatment in patients with MDR-TB who arefailing multiple agent chemotherapy. One open labeltrial consisted of five patients with smears and culturespositive for pulmonary MDR-TB despite at least fivemonths of observed appropriate therapy; the patientswere treated with recombinant human interferon-gamma (500 µg via aerosol given three times per weekfor one month) [30]. Two months after endingtreatment, which was well tolerated, acid-fast bacilliwere cleared from the sputum, body weight stabilizedor increased, and the size of cavitary lesions wasreduced. Interferon-gamma is normally produced byCD4+ T lymphocytes and serves to activate alveolarmacrophages; it may be a useful adjunctive therapy inpatients with MDR-TB who are otherwise notresponding to treatment.

Adherence to TherapyCompliance with therapy is crucial, but is particularlydifficult for patients with MDR-TB because regimensare prolonged and employ drugs with considerableadverse effects. For these reasons, directly observedtherapy (DOT) is mandatory for all patients with drug-resistant tuberculosis. It insures compliance, therebyeliminating the major cause of treatment failure 31.Regular observation also allows collection of sputumsamples, which can be used to provide an objectiveassessment of clinical response.

Serum drug levels – Therapeutic drug level monitoringis not routinely performed. However, low serum levelsof antituberculosis drugs have been reported inpatients with overt malabsorption and in some patientswith HIV infection even in the absence of clinicalmalabsorption 32,33. For this reason, monitoring serumdrug levels may be useful in patients who do notrespond clinically to a directly observed, seeminglyappropriate regimen 34.

Surgical TherapyMost patients with MDR-TB respond to appropriatechemotherapy. However, patients with sputum cultures

positive for longer than three months despiteappropriate therapy or with isolates resistant to all ofthe first-line oral agents have a worse prognosis. Thesepatients may benefit from surgical intervention35.

Patients with localized pulmonary disease which canbe completely removed at operation are most likelyto benefit from surgery 35,36. Drug therapy is often givenfor 1 to 3 months before surgery to try to reduce thebacteriologic burden, but there are no firm guidelinesor data about this issue. Resection alone should not beconsidered curative; patients with MDR-TB shouldcontinue the best possible drug regimen for 18 monthsafter surgery.

Prevemtive Therapy for Contacts of Drug-Resistant TuberculosisThere is obviously no way to determine whether ornot a person with a positive tuberculin test has beenexposed to a case of drug-resistant tuberculosiswithout a careful contact investigation, and the role ofthe health department is vital in this regard. After athorough contact investigation, an estimate should bemade both of the likelihood of recent infection witha resistant strain and the risk of the contact developingactive tuberculosis. These factors should then beweighed against the risks of preventive regimens ofunknown benefit before therapy is prescribed 37.

Exposure to isoniazid monoresistant tuberculosis –Contacts of INH mono-resistant tuberculosis can betreated for two months with the combination of arifamycin (rifampin or rifabutin) plus pyrazinamide 6.A four to six month regimen of a rifamycin alone isacceptable in patients who cannot toleratepyrazinamide; we favor six months of therapy in thissetting.

Exposure to rifampin monoresistant tuberculosis –Contacts of patients with rifampin monoresistanttuberculosis can be treated with the usual isoniazidregimen given to patients with exposure to drug-sensitive organisms.

Exposure to multidrug-resistant tuberculosis –Treatment of contacts to MDR cases is difficultbecause there are no published data concerning thecomposition, duration, or efficacy of preventiveregimens for MDR-TB. Potential regimens which havebeen suggested (based primarily on animal studies)include pyrazinamide and ethambutol, or pyrazinamideand a quinolone, with drugs given in the doses to treatactive disease, for 6 to 12 months.

Prevention of Multidrug-Resistant TuberculosisThe best way to prevent MDR-TB is by promptinstitution of appropriate therapy with efforts toguarantee adherence to therapy. Proper infectioncontrol will also limit spread to others. However, theremay be instances in which ongoing contact with casesof multidrug-resistant tuberculosis on a regular basisis unavoidable. This may the case for selected healthcare workers in facilities which treat significant numbers


of MDR-TB patients. In these instances, considerationmay be given to vaccination with Bacille Calmette-Guérin [38]. This vaccine is felt to be roughly 50 percentprotective when given to children, and is most usefulin preventing meningitis and disseminated disease. Itshould be stressed that the use of BCG vaccine in theprevention of MDR-TB is not well established andshould be employed only as a measure of last resort,since the ability to perform screening tuberculinexaminations will be lost after the vaccine isadministered.

References1. Goble, M., Iseman, M.D., Madsen, L.A., et al.

Treatment of 171 patients with pulmonarytuberculosis resistant to isoniazid and rifampin. NEngl J Med 1993; 328:527.

2. Schluger, N.W., Rom, W.N. Current approachesto the diagnosis of active pulmonary tuberculosis.Am J Respir Crit Care Med 1994; 149:264.

3. Heifets, L. The mycobacteriology laboratory. ClinChest Med 1997; 18:35.

4. Telenti, A., Imboden, P., Marchesi, F., et al. Direct,automated detection of rifampin-resistantMycobacterium tuberculosis by polymerase chainreaction and single-strand conformationpolymorphism analysis. Antimicrob AgentsChemother 1993; 37:2054.

5. Williams, D.L., Spring, L., Gillis, T.P., et al.Evaluation of a polymerase chain reaction-baseduniversal heteroduplex generator assay for directdetection of rifampin susceptibility ofMycobacterium tuberculosis from sputumspecimens. Clin Infect Dis 1998; 26:446.

6. Centers for Disease Control and Prevention.Prevention and treatment of tuberculosis amongpatients infected with human immunodeficiencyvirus: Principles of therapy and revisedrecommendations. MMWR Morb Mortal Wkly Rep1998; 47:1.

7. Five-year follow-up of a controlled trial of five 6-month regimens of chemotherapy for pulmonarytuberculosis. Hong Kong Chest Service/BritishMedical Research Council. Am Rev Respir Dis1987; 136:1339.

8. Clinical policies and protocols. Chest Clinics, Bureauof Tuberculosis Control, New York CityDepartment of Health.

9. Sandman, L., Schluger, N.W., Davidow, A.L.,Bonk, S. Risk factors for rifampin-monoresistanttuberculosis. A case-control study. Am J Respir CritCare Med 1999; 159:468.

10. Lutfey, M., Della-Latta, P., Kapur, V., et al.Independent origin of mono-rifampin-resistantMycobacterium tuberculosis in patients with AIDS.Am J Respir Crit Care Med 1996; 153:837.

11. Nolan, C.M., Williams, D.L., Cave, M.D., et al.Evolution of rifampin resistance in humanimmunodeficiency virus-associated tuberculosis.Am J Respir Crit Care Med 1995; 152:1067.

12. British Medical Research Council. Controlled trialof 6-month and 9-month regimens of daily andintermittent streptomycin plus isoniazid plus

pyrazinamide for pulmonary tuberculosis in HongKong. The results up to 30 months. Am Rev RespirDis 1977; 115:727.

13. Bobrowitz, I.D., Robins, D.E. Ethambutol-isoniazid versus PAS-isoniazid in original treatmentof pulmonary tuberculosis. Am Rev Respir Dis1967; 96:428.

14. Bobrowitz, I.D. Ethambutol compared tostreptomycin in original treatment of advancedpulmonary tuberculosis. Chest 1971; 60:14.

15. Bobrowitz, I.D. Ethambutol-isoniazid versusstreptomycin-ethambutol-isoniazid in originaltreatment of cavitary tuberculosis. Am Rev RespirDis 1974; 109:548.

16. Grassi, C., Peona, V. Use of rifabutin in thetreatment of pulmonary tuberculosis. Clin InfectDis 1996; 22(Suppl 1):S50.

17. Schwander, S., Rusch-Gerdes, S, Mateega, A, etal. A pilot study of antituberculosis combinationscomparing rifabutin with rifampicin in thetreatment of HIV-1 associated tuberculosis. Asingle-blind randomized evaluation in Ugandanpatients with HIV-1 infection and pulmonarytuberculosis. Tuber Lung Dis 1995; 76:210.

18. McGregor, M.M., Olliaro, P., Wolmarans, L., etal. Efficacy and safety of rifabutin in the treatmentof patients with newly diagnosed pulmonarytuberculosis. Am J Respir Crit Care Med 1996;154:1462.

19. Gonzalez-Montaner, L.J, Natal, S., Yongchaiyud,P., Olliaro, P. Rifabutin for the treatment of newly-diagnosed pulmonary tuberculosis: a multinational,randomized, comparative study versus Rifampicin.Rifabutin Study Group. Tuber Lung Dis 1994;75:341.

20. British Thoracic and Tuberculosis Association:Short-course chemotherapy in pulmonarytuberculosis. Lancet 1975; 1:119.

21. Stead, WW, Dutt, AK. Chemotherapy fortuberculosis today. Am Rev Respir Dis 1982;125:94.

22. Dutt, AK, Jones, L, Stead, WW. Short-coursechemotherapy for tuberculosis with largely twice-weekly isoniazid-rifampin. Chest 1979; 75:441.

23. Combs, DL, O’Brien, RJ, Geiter, LJ. USPHSTuberculosis Short-Course Chemotherapy Trial 21:effectiveness, toxicity, and acceptability. The reportof final results. Ann Intern Med 1990; 112:397.

24. Iseman, M.D. Treatment of multidrug-resistanttuberculosis. N Engl J Med 1993; 329:784.

25. Salomon, N., Perlman, D, Friedman, P, et al.Predictors and outcome of multidrug-resistanttuberculosis. Clin Infect Dis 1995; 21:1245.

26. Park, M.M., Davis, A.L., Schluger, N.W., et al.Outcome of MDR-TB patients, 1983-1993.Prolonged survival with appropriate therapy. AmJ Respir Crit Care Med 1996; 153:317.

27. Peloquin, C.A. Quinolones and tuberculosis. AnnPharmacother 1996; 30:1034.

28. Berning, S.E., Madsen, L., Iseman, M.D.,Peloquin, C.A. Long-term safety of ofloxacin andciprofloxacin in the treatment of mycobacterial

infections. Am J Respir Crit Care Med 1995;151:2006.

29. Goa, K.L., Bryson, H.M., Markham, A.Sparfloxacin. A review of its antibacterial activity,pharmacokinetic properties, clinical efficacy andtolerability in lower respiratory tract infections.Drugs 1997; 53:700.

30. Condos, R., Rom, W.N., Schluger, N.W. Treatmentof multidrug-resistant pulmonary tuberculosiswith interferon-gamma via aerosol. Lancet 1997;349:1513.

31. Schluger, N., Ciotoli, C., Cohen, D., et al.Comprehensive tuberculosis control for patientsat high risk for noncompliance. Am J Respir CritCare Med 1995; 151:1486.

32. Peloquin, C.A., Nitta, A.T., Burman, W.J., et al.Low antituberculosis drug concentrations inpatients with AIDS. Ann Pharmacotherapy 1996;30:919.

33. Sahai, J., Gallicano, K., Swick, .L, et al. Reducedplasma concentrations of antituberculosis drugs

in patients with HIV infection. Ann Intern Med1997; 127:289.

34. Peloquin, C.A. Therapeutic drug monitoring ofthe antimycobacterial drugs. Clin Lab Med 1996;16:717.

35. Treasure, R.L., Seaworth, B.J. Current role ofsurgery in Mycobacterium tuberculosis. AnnThorac Surg 1995; 59:1405.

36. Pomerantz, M., Brown, J. The surgicalmanagement of tuberculosis. Semin ThoracCardiovasc Surg 1995; 7:108.

37. CDC. Management of persons exposed tomultidrug-resistant tuberculosis. MMWR -Morbidity & Mortality Weekly Report 1992;41(RR-11):61.

38. CDC. The role of BCG vaccine in the preventionand control of tuberculosis in the United States.A joint statement by the Advisory Council for theElimination of Tuberculosis and the AdvisoryCommittee on Immunization Practices. MMWR- Morbidity & Mortality Weekly Report 1996;45(RR-4):1.


INAUGURAL LECTURE ONMEDICAL STATISTICS: A MICROSCOPE

FOR HEALTH AND DISEASE

BY

PROF. E. A. BAMGBOYEDEPARTMENT OF EPIDEMIOLOGY,

MEDICAL STATISTICS ANDENVIRONMENTAL HEALTH (EMSEH)

COLLEGE OF MEDICINE,U.C.H., IBADAN

The Vice-Chancellor, Deputy Vice-Chancellors (Administrationand Academic), The Registrar, The Librarian, Provost, Collegeof Medicine, Dean of the Faculty of Public Health, Dean ofthe Postgraduate School, Dean of other Faculties and ofStudents, Distinguished Guests, Ladies and Gentlemen.

I am grateful to God Almighty for the honour andprivilege to be selected to deliver the 2005/2006inaugural lecture on behalf of the Faculty of PublicHealth. The feeling of honour and privilege stems fromthe fact that if the Faculty of Public Health had notbeen created in 2002, I may not have had theopportunity of giving an inaugural lecture beforeretirement despite attaining the position of Professor14 years ago.

This inaugural lecture is the fourth from the Facultyof Public Health and second from the Departmentof Epidemiology, Medical Statistics andEnvironmental Health (EMSEH). The first lecture wasby Prof. J.D. Adeniyi from the Department of HealthPromotion and Education, the second was fromDepartment of EMSEH by Prof. M.C.K. Shridhar, aProfessor of Environmental Health and the third wasby Prof. O.O. Keshinro of the Department of HumanNutrition while the fourth one is also from theDepartment of EMSEH by a Professor of MedicalStatistics. I can therefore predict with 95% confidencethat the next inaugural lecturer from the Departmentof EMSEH will be by a Professor of Epidemiology.

When I received the e-mail from the Dean of theFaculty of Public Health, Prof. Olaolu Akinyele that Ihad been nominated to give the 2005/2006 inaugurallecture on behalf of the Faculty I readily accepted. Iwas then at the John Hopkins University School ofPublic Health as a visiting scholar. Since I was to senda topic immediately, I started to ponder on the topicand that night I had a dream in which I received thetitle of today’s inaugural lecture which is “MedicalStatistics, a Microscope for Health and Disease”.

What Really is Medical Statistics, One May Ask?Mr. Vice-Chancellor sir, as I ponder on this title manydays later, I realized that Medical Statistics as the Scienceof counting health and disease, making sense out ofthose counts and foretelling what lies ahead is actuallya microscope into health and disease. Previously,

Medical Statistics was only concerned withenumerating those dying in relation to those living, butthere has been a paradigm shift as facts and impressionsin Medicine are now expressed in numerical forms.This includes laboratory investigations, radiologicalinvestigations and clinical observations. In short, themethods by which these numerical facts on health anddisease are collected, manipulated, sieved, summarizedand interpreted, making use of the wholearmamentarium of statistical theories is the subjectknown as Medical Statistics (Chiang, 1985, Zellen,1983).

The World Health Organization defined health “as astate of complete physical, mental, spiritual and socialwell-being and not only or merely the absence ofdisease”. Therefore, the question that comes to mindis: How does one know who satisfies this definitionof health? In other words, who is healthy? Or can weassume that anyone who has not sought help fromany health care service is healthy? The answer is NO.Or else how do you explain the case of a man whohad never complained of any sickness in the last 5years or so, but suddenly dropped dead after seeingoff a visitor. We all know that many people sufferfrom one kind of ailment or the other in silence partlybecause they do not know that they are sick. The truthof the matter is, there is a subtle difference betweenhealth and ill-health that cannot be seen with nakedeye and this suggests that we need a microscope toidentify the ailment. Medical Statistics utilizes relevant,reliable and valid information to identify the significanceof any difference at a known probability and has beenthe most useful microscope in this respect. Thus, thedemographic information, medical history, clinical signsand symptoms collected to aid the diagnosis of anydisease is germane to sound statistical manipulation.Population Medicine informs that diseases or deathsdo not occur at random. They are cause, though notby witchcraft nor the ‘worst’ of the ‘wicked’.

Mr. Vice-Chancellor sir, this inaugural lecture has beencarefully designed to focus not only on how bestmedical statistics has acted as a microscope to detecthealth and disease of people in the community but onthe factors casing the disease. Perhaps some of us whoare familiar with Epidemiology as the study of thefrequency, distribution, determinants and deterrent ofdiseases in human populations would wonder ifMedical Statistics is not taking up that definition. Sufficeto say that this microscope (medical statistics) is usedfor both the community diagnosis of ill-health andthe diagnosis of disease of an individual. Both publichealth and disease and of course information isknowledge and knowledge is power.

Medical Statistics and EpidemiologyMedical Statistics as a discipline in Public Health andMedicine represents one of the major intellectualadvances of the twentieth-century. It has developedto initiate and strengthen the collaboration ofstatisticians and health professionals (particularlyphysicians) in any medical or health research. Suffice

FEATURE ARTICLES


to say that a medical statistician is at first amathematician who adapts his knowledge to the studyof disease. Although the study of disease could besaid to be mainly in the allied discipline ofEpidemiology, the epidemiologist and the statisticianwork together to the extent that it is permissible touse the words “Statistical” and “Epidemiological”interchangeably.

The epidemiologist is first all (in most cases) skilled inmedicine, but adopts a mathematical approach to thestudy of disease in the population. The workingrelationship of medical statisticians and epidemiologisthas resulted in the statistician become a betterepidemiologist and the epidemiologist a betterstatistician both using numbers and their logic tounravel the mystery behind diseases. According toProfessor Healy of the London School of Hygieneand Tropical Medicine, “the difference betweenepidemiologists and statisticians is a subtle one…..as itis one emphasis” (Healy, 1977: Zeger, 1991). Whileepidemiologist is mainly interested in the results ofinvestigations, irrespective of the methods by whichthese results were obtained, medical statisticians arealso interested in results but equally in the methodologyof obtaining them. The medical statistician, becauseof his mathematical background often develops newand modified techniques for handling specific medicaldata or results in a novel application of methodsdeveloped in another field.

The University of Ibadan, since 1968, as employed amedical statistician with background in mathematicalstatistics in its medical school and this has improvedtremendously, the quality of medical research ascollaboration between the medical statistician andmedical doctors become a routine affair. Thedevelopment and relevance of the discipline of medicalstatistics as a component course in medicine is suchthat there is a full department with its complement ofmedical statisticians in Ibadan. It is not only thatmedical statistics is part of the curriculum in bothundergraduate and postgraduate medical education,but people now acquire postgraduate degrees inmedical statistics and statistical epidemiology from theUniversity.

Medical Statistics and Community Diagnosis ofHealthPublic Health is concerned with the health of the entirepopulation as a whole, but the clinician is interestedonly in the treatment of that individual who hasapproached him because of ill-health. Public Healthstrategies are largely designed to prevent ill-health,identify and treat promptly those with ill-health, thusavoiding any serious consequences. As an illustrationof the main focus of public health and also of clinicalmedicine, public health tends to find out why manysick people from, say Oja-Oba in Ibadan come to thehealth centre with diarrhea diseases and a very negligiblenumber come from, say Bodija with this condition. Inthis scenario, the clinician’s primary duty is to treat casesof diarrhea that come to his clinic on a 1-1 basis, but

the public health person traces the people to theircommunity to find out possible risk factors present inthe areas they have come from. If he later discoversthat people from Oja-Oba take their drinking waterfrom a source polluted by Ogunpa River and thosefrom Bodija from well-treated pipe borne water, hecan associate the higher prevalence of diarrhea withthe source of drinking water, more so if the differencein the proportion with diarrhea between the twogroups is statistically significant. In fact, this is howJohn Snow linked cholera to contaminated water andthousands of people are being prevented from thisdeadly disease by providing them with good portablewater. One clear difference between the impact ofthe strategies of clinical medicine and preventivemedicine is the number of people affected at a time.While the clinicians may see an average of 10 patientsa day in a medical out-patient clinic, a preventive actionmay affect and save the lives of several millions ofpeople.

Medical Statistics also develops mechanisms to findthe causes of diseases or at lease locate modifiableand non-modifiable risk factors, whetherenvironmental, behavioral, occupational hazards andother factors that could pre-dispose to the occurrenceof diseases. Imagine seeking for risk factors of diseasesin a population as large as Nigeria with over 130 millionpeople. This is like fishing in the dark and MedicalStatistics has provided the microscope for makingcorrect community diagnosis of ill-health andexamining if those who are apparently healthy mayprobably be so because they have not shown any signsor symptoms of any disease or that they are trulyhealthy. If they are healthy, the microscope examinesmeasurement of such factors that make them healthy.The diseases are classified into communicable and non-communicable diseases for ease of understanding andpresentation. This is further classified into seventeencategories ranging from infectious and parasitic diseasesto accidents and ill-defined conditions.

The development of statistical indices of health suchas rates, with the four components – numerator (no.of events, of interest), denominator (population at riskof the event), timeframe (the period the events occurs)and the standard form in which it is expressed hasfound immense use in determining health and diseasestatus. The size or extent or magnitude of a disease orany health problem in the community can be measuredby the prevalence rate. But the rapidity at whichdiseases occur in the population on the other hand ismeasured by the incidence rate. The census, a purelystatistical exercise that is routinely conducted every tenyears in any developed nation provides sources of thedenominator for most of our rates. The collectionand analysis of records of vital events such as births,marriages and deaths have been used to assess fertility,nuptiality, mortality and growth of human population.Thus, the ability to measure is paramount (sine qua non)to understanding health and disease and this has beenfacilitated by the science of medical statistics. In factas far back as 1889, Lord Kelvin had written that


“when you can measure what you are speaking about,and express them in numbers, you know somethingabout it; but when you cannot measure it, when youcannot express them in numbers, your knowledge isof a meager and unsatisfactory kind” (Oldham, 1968).

Medical Statistics as a Microscope for IndividualDiagnosisMedical Statistics as a microscope for individualdiagnosis of disease, its prognostic factors and besttreatment options was resisted at its earlier debut inclinical medicine. The notion of clinical doctors is that“patients are different and therefore cannot besubjected to statistics”. A clinician even said that if hehas done a study and requires statistics, then he knowshe has done a bad study. But can anyone in thisaudience tell if the person sitting next to him or her ishealthy? Do you know if a healthy looking man hasAIDS? Or can you say if that well-built man next toyou can be hypertensive? Is the woman on your leftside suffering from diabetes? The truth is that there isno window to look into the system or heart of manto identify a healthy man, or to know if a drug willcure a disease, or know the cause of a disease, or knowwho will die in the next minute demand more thanwhat can be seen with naked eye. The naked eye ofthe doctor cannot tell if the social habit of a patient isresponsible for that damage in the liver. And neithercan the naked eye of the doctor nor that of the nursetell if the patient admitted for diabetic care will survivethe next 3 days. Neither could the naked eye see theexact diagnosis of the illness of a peasant farmer froma village in Ekiti State who suddenly jumped on theroad, claiming himself as the president of the USAand asking people and relatives around him to takecommands from him.

Mr. Vice-Chancellor sir, God created man to bedifferent in many respects from each other and oneclear evidence is that no two fingerprints of the over6.5 billion people in the world are the same. Theseinherent biological variations in man and the differencesin characteristics of individuals would have made itmore difficult to find the truth of the actual factorscausing or affecting ill-health and diseases’ worstoutcome – death. But God so loved the world thatHe created the science of medical statistics and medicalstatisticians to sort out the variations in man and diseaseand thus able to identify the true causes or risk factorsof ill-health and death to specified degrees ofconfidence. But it was initially difficult for clinicians toimbibe medical statistics as a reliable and validmicroscope to diagnose diseases.

However, by and large, the concept of medicalstatistics realized ironically as far back as the 16th centuryby some people has now become the norm ratherthan the exception in medical practice. Leonardo DaVinci who lived in the 16th century (1452 – 1519) said“No human investigation can be called true sciencewithout passing through statistical test”. Medicine hasbecome real science. The clinical issues of patientdiagnosis, prognosis of disease and treatment have

benefited from medical statistics. Diagnosis andassessment of prognostic factors depend largely onpatients’ data on physical characteristics, laboratoryparameters, radiological information, clinical andpersonal characteristics. The 3-stage process of science,observations, building up a model to describe orexplain the model and using the model to predictfuture observations or events have found a strongplace in medicine particularly in communicable diseasewhere epidemics are dealt with. The use of models indescribing any epidemic and forecasting the spread intime if nothing is done is the telescopic use of MedicalStatistics.

Normal Range of ValuesThe commonly used concept of normal range ofvalues for separating healthy people from those whoare diseased has been facilitated by the microscopicnature of medical statistics, starting from its use of“mean plus or minus two standard deviations” basedon the assumption of a Gaussian distribution forcontinuous variables. The diagnosis of most medicalailments such as diabetes, hypertension, anemia and soon depend much on the normal range of values ofbiochemical parameters, blood pressure readings andhematological indices obtained from the patients. Thusthe doctor can classify as hypertensive with a highdegree of confidence, a 50 year old man with threeconsecutive blood pressure readings of 180/100, 178/95, 181/100. Or as diabetic, a middle-aged womanwith fasting blood sugar of 180mg/dl; and randomblood sugar of 200mg/dl and a young man with CD4count of 150 as suffering from AIDS. Also, the doctorcan say that a woman with a white blood cell countof 320,000 per ml may be suggestive of myeloidleukaemia or a child with a temperature of 39.3degrees as having a fever. The doctor can also diagnosea 20 year old girl with PCV of 10% as severely anemic.

But the so-called concept of normal range of valuesoften used, is developed on the normal distributiontheory, an important aspect of medical statistics thathas been very useful as a microscope to sort out patientsinto normal and abnormal groups or healthy anddisease group.

Using the properties of the model, the parameterswere calculated from observed values in an apparentlyhealthy population. Subsequently, confidence intervalsfor which admissible values with 95% confidence areallowed are obtained and consequently the constructionof the normal range of values. It suffices to say thatthis simple mathematical model is the bedrock ofparametric tests which have found immense use todayin the testing of hypotheses. Clinicians have drunk sodeep of the statistical concept of hypothesis testing tohave developed a p-value syndrome to an extent thatthey have felt that any research in Medicine cannot bepublished without an accompanying p-value. This maynot necessarily be so as some studies are purelydescriptive and do not test any hypothesis.


Evaluation of TreatmentsAlso, apart from identifying health and disease, MedicalStatistics or the Medical Statistician has always acted asa policeman in medical practice. We are in a world ofgreat advancement in computer technology and thepractice of medicine has become evidence based. Thepatients in the hospital want to know the treatmentplan for them, at least to have psychological satisfaction.Astute observations of regularities of events often callfor clinical or epidemiological studies. Mr. A had thisproblem, received drug A and was relieved. But Mr.B had the same problem, was given drug A but didnot improve. Medical Statistics provides the path ofexplanation; first, it asks if the disease condition issimilar in both groups in terms of onset before seekingtreatment and then finds out if the drug works in theacute or chronic phase, or whether confounders interms of the persons of Mr. A and B have beenremoved. Thus, clinicians are reminded that themeasurements collected are subject to variations andinterpretations and can be marred by confoundingfactors if these are not properly controlled for. Factorsmasking the true diagnosis or the true efficacy of anytherapy are sorted out and removed through the useof Medical Statistics, the powerful microscope forhealth and disease.

This microscope is not only towards health and diseasebut also to allied areas where the truth is needed. Suchis the performance of students in the universities andtheir admission criteria. My work in this area has beenin examining the University of Ibadan as a case study.This effort resulted in the present admission guidelinesfor Post-UME screening referred to as the IbadanModel. (Bamgboye 1981; Bamgboye et al., 2001; Kaleand Bamgboye, 1984; Screening Committee, 2005)

Medical Statistics as a Microscope for the TruthI was privileged to attend the 2002/2003 Universitylecture delivered by Professor Oladele Kale, an iconin Public Health and in particular filariasis, thefoundation Dean of the Faculty of Public Health whowas my Head of Department (HOD) at several pointsin the life of the Department of Preventive and SocialMedicine, including the period of ‘distress’. Heemphasized the need to seek the truth always in thatlecture titled “Nigeria in Distress! A Trilogy of theNation’s Health Status”. The apostle John said in John41; “Dear friends, do not believe every spirit, but testthe spirits to see whether they are from God, becausemany false prophets have gone out into the world”.Indeed, in a similar fashion, many false (unscientific)researchers have flooded the health field and havealways capitalized on the psychology of the people toinflict untold story on them. I have watched severaltimes on my television, and advertisement of a magicdrug (panacea) that can treat typhoid, hypertension,appendicitis and any other disease (just name it) with asingle dose!

Professor Kale used the Abalaka claim of aprophylactic and therapeutic cure for HIV/AIDS as acase study. This is a well known case circulated on the

pages of newspaper a few years ago. Because of thedeadly nature of AIDS and the fact that no cure isknown, it is normal to draw attention to anyone whoannounces he has found a cure. No doubt, amicroscope is needed to ascertain the minute detailsof the truth of any claim to a cure of AIDS. Andindeed, if you ask me the sample size required to assessthe efficacy of a drug that could cure AIDS, I wouldsay one person. But then the one person must indeedbe diagnosed to truly have AIDS and all the facts mustbe available.

As usual, the Abalaka drug issue was reviewed byProfessor Kale, and many people including myselfended up disappointed at the methods of investigationby all the parties concerned. But the Abalaka issue is atip of the iceberg in the multitude of malpracticesinvolving the inappropriate use of Medical Statisticsthat pervades our health service system. This is oftenattributed to inadequate knowledge of the real healthproblem, the root cause of which is the poor datagathering system, method of data analysis, and poorinterpretation of findings. This is the situation in whichany nation will find itself when there is a lack of reliableempirical evidence to aid decision making and thosetrained to provide the mechanism to generate thisreliable data are either not available or are very few. Itis not surprising however, for one to grope in the darkwhen that instrument to provide light and see minutestissues as a microscope is missing. Our Lord Jesus Christsaw the crowd following Him at one stage and hadpity on them and declared in Mathew 937; “TheHarvest is plentiful but the workers are few”. For themultitude of health problems that need research allover the world and in particular in Nigeria, there arefew medical statisticians.

The advancement of computer technology and theavailability of statistical packages in the open market,with piracy going on unchecked, have given birth to anumber of do-it-yourself statisticians. These peoplewho often ‘hit their thumbs with hammer’ in theirpractice do more damage than I will cause if I attempta brain surgery. There is no amount of formaleducation a health policy maker can acquire that hewill be able to draw up meaningful policies based onwrong data or subjective reasoning. We all know thatthe bane of our health problems in Nigeria today isthe lack of reliable health data and lack of professionalmedical statistician in the health system team. If thiswas not so, the Nigeria Institute of PharmaceuticalResearch and Development (NIPRD) would have hadat least one medical statistician on their team, whowould have advised the director investigating Abalakaclaim based on two patients, that they needed to includea control patient probably on a placebo and see thenatural course of the disease.

In clinical trials, a great deal of attention is directed atthe patient and they are much aware of this andsubjects’ condition can change because of this and notnecessarily because of any treatment. This situation,known as the Hawthorne effect cannot be separated


from the treatment effect in uncontrolled trials as carriedout in NIPRD. God is wonderful; there is the naturalcourse of things and medical statisticians have studiedthis situation to conclude obedience to the “Law ofregression to the mean”. This explains why cliniciansrequire a repeat of a laboratory test of an unexpectedlyabnormally high value and usually, the second test resultis closer to normal values. Anyway, a cure for a diseaseas deadly as AIDS may not attract any large samplesize as I said before, a sample size of one is enough ifpeople can truly attribute the cure of AIDS disease tothe ‘curative agent’ alone. The statistical methods forthe evaluation of treatments of diseases date back tothe early life of man. In the Holy Bible (Judges 636-40),Gideon wanted to know if God would indeed supporthim in the task He put before him. Then he carriedout a trial by putting a fleece of wool on the floor andasking God to make the fleece only wet and the eartharound it dry. It happened, but he reversed hisallocation of subject units and asked God to makethe fleece of wool dry and make the surrounding earthwet. This happened and he was convinced beyondany reasonable doubt that God would support him.This design can be taken as the beginning of the cross-over design often used in clinical trials today.

Let me tell you the story of the blind man and thedumb man brought to Jesus. The dumb man wasdemon-possessed and could not talk and everyone inthe crowd knew this and so also the blind man. Theydeliberately brought them to Jesus, probably to testHis power. The blind man regained his sight and Jesusdrove the demon out and the dumb man also regainedhis speech. According to Matthew 933, “…………thecrowd was amazed and said nothing like this has everbeen seen in Israel”. In spite if the several miraclesusing the same therapy, many people were still skepticalabout Jesus and His therapy. The medical statistician isalways a skeptical scientist.

Now, about Abalaka’s drug trial, I reproduce in table1 the NIPRD result with permission from my formerHOD. The changes may simply be the whole conceptof regression to the mean.

Table 1: NIPRD Result on Abalaka’s Drug TrialViral load (copies/ml)

After two weeks of immunotherapyNIPRD/HIV/JAV/001NIPRD/HIV/JAV/0028.2 x 105 (12/1/2000)1.6 x 105 (12/1/2000)2.4 x 103 (25/1/2000)1.6 x 103 (25/1/2000)

The medical statistician would have suggested the needfor a control group and asked many other questionsif he were on the research team. What is the viral loadof the AIDS patient who never took Abalaka’s drugs?Are the patients truly comparable or do they satisfythe same criteria for inclusion? Any ambiguity in thedefinition of AIDS? Have they used the same selectioncriteria or mechanism for counting the viral load? Arethere copying errors, transcription errors, measurementerrors? And so on. Apostle Paul said in John 8 verse32; “if we know the truth, the truth shall set us free”.

Since the naked eye cannot see the truth in health anddisease and we must know the TRUTH, we give Gloryto God for the birth of the discipline of MEDICALSTATISTICS to serve as a MICROSCOPE to tellthe truth about health and disease.

Let me at this point appreciate Professor Bamidurowho defined Statistics in his inaugural lecture of 7th

July 2005 “as the science to detect the TRUTH”. Severalstrategies have been put in place by man (preventive,curative) to ensure that everyone is healthy. But theorganization of these strategies has benefitedimmensely from the use of the science of MedicalStatistics. That the psychiatrics is able to define variablesassociated with maniac depression, and the surgeonable to measure the degree of burns or the probabilityof non-spread of bone cancer if the affected part isamputated, is due to the use of medical statistics as amicroscope which has always been used on observeddata to separate the chaff from the wheat and indeedthe diseased from the healthy. Not that alone, it hasbeen used to identify the risk factors of diseases.Medical Statistics has made it possible to associate withheart attack, the case of a 55 year old prosperous-looking man who alighted from his new MercedesBenz Jeep which he drove by himself from his stationof about 200km away, entered his father’s compoundsafely and suddenly slumped in the process ofprostrating before his father and was declared dead afew minutes later! Or else the witches in the familycould have been held responsible. The amount ofstatistical jargon found in any medical journal today isan eloquent testimony to the use of this science tounderstand health and diseases.

Some other Specific Applications of StatisticalConcepts to MedicineControl of Communicable DiseasesSeveral hypotheses based on quantitative data havebeen evaluated through the application if mathematicalmodels. The earlier area of mathematical applicationto the development of medicine was in the area ofcommunicable diseases. It will only be important tomention a few in this lecture. Following the successfulfitting of the normal curve by William Farr in 1840 toquarterly data on deaths from smallpox, furtherapplication was for the description of epidemics.Notable is the work of Ronald Ross on thetransmission of malaria by female anophelesmosquitoes. Simple mathematical assumptions alreadypostulated by Hammer in 1906 that the net rate ofspread of infection is proportional to the product ofsusceptibility and the density of infections ofindividuals, and that they are expressed in a discreettime was all he needed to establish the popularthreshold theory for malaria, but after the modificationof the model to a continuous time frame (Bailey 1957,1975). The only snag in the application of mathematicalmodels to disease and health is the lack of empiricaldata and this has sometimes defeated the wholeexercise.


In fact, a look at literature reveals that less than 10%of published mathematical models in medicine havepractical applications due to lack of empiricalobservations. But more data-oriented studies ofdiseases using mathematical theories are coming up.For example, the current public health policy on diseasessuch as onchocerciasis and malaria has benefited fromthe mathematical theories developed that werenecessary and sufficient to recommend effectivemethods of controlling the sources of infection. Inthe earlier stage of my career, I was introduced to thecontrol of onchocerciasis (river blindness) and myhumble contribution was in the statistical analysis ofthe drug trial. The analysis examined the relative efficacyof Banocide and Suramin Injections on the disease(Bamgboye 1977). Studies later revealed thatmathematical models can be applied to determine thepoint of intervention for effective control and I wasinvited to Ouagadougou by WHO to participate butthe load of work in Ibadan prevented me fromcontinuing with them. However, today, the modeldeveloped by other medical statisticians who finallyworked on the research has led to the effective controlof the disease.

Decision making in medicine has benefited frommathematical statistical theories appropriately facilitatedby Baye’s Theorem. It has been discovered that, noinvestigation can be called true science without passingthrough mathematical tests. Thus, the reliability andvalidity of most diagnostic tests commonly used inmedicine today have been developed throughmathematical concepts of statistics. For example, theprobability that a patient has a particular disease if thediagnostic test is positive given the prevalence of thedisease in the population and the false positive rate,can only be calculated by statistical methods.

Designs of various studies in medicine have benefitedthrough mathematical knowledge. Initially,Mathematicians or Statisticians were used for technicalservice in the analysis of data but in recent times, thenotion has changed and the collaborative role ofstatisticians in the development of treatment strategiesin medicine is now well recognized. Observationalstudies are more common in medicine than realexperiments. But the medical domain is beginning torealize that the only way to find out what will happento a complex system as in medicine is to disturb thesystem. Several statistical methodologies have beendeveloped to bring out the scientific values of thepopular observational studies, particularly in identifyingrisk factors of diseases or in studies of disease etiology.

The present approach to the control of communicabledisease has emerged through the applications ofmathematical and statistical theories to the descriptionof the process of acquiring the disease. This has led tothe suggestion that, immunization, control ofunanimated vehicles of infection, control of insectsand other living vectors and appropriate therapy couldbe basic procedures that could subjugatecommunicable diseases.

A lack of proper understanding of the host – parasiterelationship which expresses the genetic make-up ofthe system has slowed down the control of manyinfectious diseases. However, a model has beendeveloped that is essential to describe, predict, andexplain our health situations. That this is possible hasbeen through mathematical reasoning and logic. A clearunderstanding of the transmission process of epidemichas thus benefited from mathematical methods. Anotherbasic reasoning is the phenomenon developed by ReedFrost in 1928 and further explained by Greenwood in1931. They both stipulate that the probability of a givenepidemic chain is the product of probability termsfrom different binomial distributions. This is thepopulation chain binomial model which has foundimmense use in the description and prediction of therates of infections and has led to the telescopic valueof medical statistics for predicting future occurrencesof epidemics based on observed facts now, if nothingis done.

A major contribution was the work of the first Ph.D.candidate produced in Medical Statistics in thisUniversity, Dr. Bidemi Oyindamola Yusuf, who undermy supervision applied mathematical models tounravel the problem of anti-malaria drug resistancewhich permeates the population (Yusuf, 2005). Thisthesis developed a mathematical framework to explainand describe the population dynamics of drugresistance, using stability analysis and simulation. Itrevealed that it was more beneficial to prevent drugresistance than to cure it.

My ContributionsMr. Vice-Chancellor sir, let me at this point talk brieflyabout the role of the medical statistician and mycontributions to my profession. It is a common mistaketo assume that medical statistician need only beconcerned with the analysis of already collected medicaldata. This is of course the notion of many otherworkers about medical statisticians working in a medicalmilieu as in the College of Medicine. It is no doubtthat the medical statistician plays a major role as a dataanalyst but he/she does equally well in study designand conduct. One of the duties of a medical statisticianis collaboration and his/her role in such a situation isto ensure that both protocol design and interpretationof study and trial findings conform to sound principlesof scientific investigation that will remind us that thecause of humans’ health and disease is such that cannotbe easily seen with the naked eye.

Apart from collaborative research, a good medicalstatistician, as a scientist especially one who works inthe academia has his own special area where he shouldmake novel contributions. I will now describe where Ihave made novel contributions as an individual or incollaboration with others. This will also include myrole in the Development of Medical StatisticsEducation in the University of Ibadan.


My Contribution as a Collaborative ScientistEver since I joined Professor Ayeni at the Universityof Ibadan in 1975, I have used medical statistics as theinstrument to search for the truth of health and disease,a task tantamount to a blind man looking for a blackcat in a dark room. The first aspect of my contributionwas through collaborative research and to date, I havecollaborated with 177 medical persons or researchersto solve or throw light on a number of medicalproblems or issues both in Nigeria and Saudi Arabia.It is interesting that I have only 8 non-medical peopleas collaborators. I have also been involved in 27national health studies both in Nigerian and SaudiArabia and over 5,000 on-the-spot statisticalconsultations averaging one per day.

This aspect of my work brought some confusion tosome naïve scientists saddled with the task of reviewingmy publications for professorship as to whether theyshould recommend me for Professor of ‘Obstetricsand Gynecology? Pediatrics? Medicine?Ophthalmology? Psychiatry? Demography? MedicalEducation? Or Pathology?’ (Name any area ofMedicine except Surgery which has just begun to haveits doses of the use of medical statistics). And this issimply because of my numerous scientific publicationswith collaborators in these specialties. But mostimportantly, what is medical about the professionalmedical statistician that I am? This is that the medicalstatistician has a fair amount of medical knowledge,medicine being the subject matter of his field. But themedical statistician is no more a doctor manqué thanhe is a mathematician manqué. Since he is involved incollaborative work, his collaboration shall not bewasted if the medical member of the research teamtakes time to explain things taught to medical studentsin their first pre-clinical years to him. The newprofessional medical statistician will like to learn amongother things, in an apprentice fashion on the job, someof the medical knowledge, no one is allowed to adviceon a subject he does not understand. This implies thatthe medical statistician obtains enough materials on anymedical topic brought to him for statistical advice andhe is expected to have a fair knowledge of this beforeoffering any advice.

My collaborative role in the application of statisticalmethodology to the design, analysis and interpretationof medical studies underscores the good proportionof my publications in various specialties of Medicine.As is well known, my collaborators are medical peopleand I am happy to say that my collaborative researchwork is not limited to the shores of Nigeria but to faraway Saudi Arabia where, during my initial short stayof about 3 years at their first College of Medicine,King Saud University, Riyadh (although I had anaggregate romance close to 12 years), I was able toproduce 30 publications. At this point Mr. ViceChancellor, I like to say that the credit of authorship,according to the International Committee of MedicalJournals, should be based and it is based on:

(1) substantial contributions to conception anddesign of a study, or acquisition of data, oranalysis and interpretation of data;

(2) drafting the article or revising it critically forimportant intellectual content; and

(3) final approval of the version to be published.

Authors should meet conditions 1, 2 and 3, Irecommend considering these criteria of assessingcontributions in multiple author publications, in theUniversity of Ibadan, new guidelines for the promotionof academic staff.

I am bold to say that in all my collaborative work,regardless of my position in the authorship list, I metall 3 criteria. In Pediatrics, working in collaborationwith others in Saudi Arabia, a standard growth chartfor monitoring the growth of children was developedbased on a national cross-sectional study to replacethe American Reference Standard which was foundnot suitable for Saudi Arabia children (al-Frayh andBamgboye 1992, 1995; al-Nuaim and Bamgboye, 1995,1996; al-Sekait et al., 1992). The curves were drawnusing the Spline Polynomial Regression Method and itis used in all clinics in Saudi Arabia to monitor thegrowth of children. Another contribution in Paediatricswas the identification of factors influencing lineargrowth in children (al-Fawaz and Bamgboye 1994),the results of which have helped in health promotionprogrammes. So also was the study on the determinantsof faltering growth in Saudi Children (Bamgboye andal-Nahedh, 2003). Other studies included ‘BronchialAsthma in Saudi Arabia’ and the determination of therisk factors of Coronary Artery Disease in Nigerianand Saudi Arabia (Taylor et al., 1996; al-Nozha et al.,1994). Another area was in the determination of visualloss in Saudi Arabia (Badr and Bamgboye 1990, 1992).Studies on the causes of adult morbidity also provideda good benchmark for planning preventive actions (al-Balla et al., 1993; Bahakim et al., 1993).

In Nigeria, basic data on morbidity showing the typeand pattern of disease in our children received myattention (Bamgboye and Familusi, 1990). So also, wasthe utilization of health services in the hospital(Bamgboye and Jegede, 1987; Jegede et al 1990, al-Shammari et al., 1994), (Bamgboye and Shoge, 1987;Bamgboye, 1985). This concept has been of help topolicy makers examining the burden of disease. Alsohospital administrators benefited from studies onwaiting time of patients in the hospital before seeing adoctor (Bamgboye, et al., 1992; Bamgboye and Jarallah1994). I have also collaborated with colleagues inPsychiatry to study health problems of the elderly(Baiyewu et al., 1997; al-Shammari et al., 1994).Indicators of morbidity in workers, using sicknessabsence records, have also been analyzed to producerelevant morbidity statistics both in Nigeria and SaudiArabia (Erinosho and Bamgboye, 1998; Bamgboyeand Adeleye, 1992; Bamgboye et al., 1994; al-Shammariet al., 1994).


Another area of my contribution is in the developmentof health service research in Nigeria. The availabilityof reliable morbidity and mortality data serve asevidence of good database for better planning ofhealth service and effective health policy. In this respect,I have been involved in the national health informationsystem and the development of techniques forestimating the burden of disease in Nigeria, (Taylorand Bamgboye 1979; Taylor et al., 1992). Part of mycontribution has been to a number of national studiescarried out by the Federal Ministry of Health withsupport from numerous foreign donors. These includeNational Health Surveys for the determination of levelsof reproductive health including HIV/AIDS (FederalMinistry of Health, 2005), the Evaluation of Preventionof Mother to Child Transmission of HIV/AIDS inNigeria (FMOH, 2005) and Sentinel Surveys for theEstimation of the Prevalence of HIV in Nigeria(FMOH, 2006). The later report has informed thepresent scaling up of the programmes and refinementsin the logistics. A national behavioral surveillance surveyof high risk groups of HIV/AIDS in Nigeria has justbeen concluded and the report is yet to be published.The information from the various surveys benefitingfrom the science of Medical Statistics has been usefulto policy makers in determining the current health anddisease status of Nigerians for effective health serviceplanning and development of strategies to aid thecampaign against HIV/AIDS.

Let me also add that I have collaborated withcolleagues in the Department of Obstetric andGyneacology to contribute intellectually to reproductivehealth research. The objectives of our Lord ascontained in Genesis 1 verse 28 which says: “Be fruitfuland multiply and replenish the earth….” can be realizedwith good healthy reproductive practices. Studies onthe growth patterns of the fetus at different gestationalages in a normal obstetrics population have producedappropriate techniques for measurements of fetalgrowth and standards for evaluation (Osinusi andBamgboye, 1987; Marinho and Bamgboye, 1987). Thestudies on the evaluation of ante-natal services in SaudiArabia have helped to develop a sound policy on thenumber of antenatal visits and the provision ofmaternity services in general (al-Nasser et al., 1991, 1992and 1994). Other contributions were in the area ofthe determinants of abortion and reproductivepotentials following abortions as well as ectopicpregnancies. The information has been a useful guidefor health education and obstetric practices (al-Nuaimet al., 1995; Adelusi et al., 1998).

My Contribution to the Science of MedicalStatisticsMy Ph.D. thesis on the Methodology of IndirectTechniques is one of my major contributions to MedicalStatistics. The outcome of this Ph.D. degree work isnow a central item in the United Nations Manual Xfor Indirect Demographic Methods (U.N. Manual X,Chapter IV, 1983). The methodology developed underthe supervision of Professor Brass, the father ofMedical Demography is one of the great topics in

courses on Indirect Techniques today. It is known asthe Brass-Bamgboye technique for locating time pointsfor adult mortality estimation from retrospectivesurvivorship data (Brass and Bamgboye, 1981;Bamgboye 1983). It was a great delight when ProfessorKen Hill, the greatest living icon in IndirectDemographic Techniques at the John HopkinsUniversity in April this year introduced the topic in alecture on Indirect Methods and declared it as one ofthe greatest contribution to indirect methods formortality estimation. He further challenged me to havea closer look at the methodology for possiblerefinements (if necessary) in the light of new availabledata. Fortunately, we now have a programme for aPh.D. in Medical Demography in my department andone of our promising, young and talented lecturers inMedical Statistics, Odunayo Akinyemi has turned hismicroscope in this area to search for more truth inpursuit of his Ph.D.

The Indirect Method of Estimating Mortality(Medical Demography)For the benefit of the audience, let me talk briefly onIndirect Demography Techniques. The numerator formortality indices can be obtained directly from recordsof deaths collected routinely through the vitalregistration system and its denominator from censusfigures (Bamgboye, 1983). It suffices to mention thelack of traditional demographic data in developingnations where the organization of national census orthe establishment of vital registration systems continuesto be a problem (Ayeni, 1971). The childhoodsurvivorship data obtained from simple questions formother on total children ever born and numberssurviving are known to provide good estimates ofinfants and childhood mortality (Brass and Hill, 1973).So also, reports on orphan hood and the ever-widowed have provided acceptable adult-mortalitylevels (Hill, 1977). These estimates are obtained fromsimple questions such as ‘Is your father alive? Is yourmother alive? Is your spouse alive? And so on requiringYes or No answer and included in single rounddemographic surveys. The underlying assumption ofthe methodology has been that mortality changes inthe past can be ignored or that mortality remainsconstant. My novel contribution to this issue is theexamination of this assumption and development ofa mathematical model that locates the time in the pastto which the estimates of mortality levels from theseretrospective survivorship reports refer (Brass andBamgboye, 1981; Bamgboye 1983). My contention isthat mortality in developing countries cannot and isnot constant but changes. In the light of many economicdevelopment and social strategies it should be declining.However, this decline is assumed to be on the logitscale.

These methods have been developed using soundmathematical theories such as the Taylor’s series theoryused to link cohort life tables with period life tables.The development of the model was based on severalassumptions of fertility and mortality regularities andpopulation models utilizing theories of integral calculus,


power series and the logic models. The evaluation hasresulted in the equation for the time location (T) in thepast given as:

xWa = proportion of the group in the age group(a, a+x)

= Integral signs at which persons enter the system(a) and appropriate lengths of exposure respectively

(The system may be marriage or orphanhood)

((a+x),T) = Life table survivorship to a+x at timeperiod T

This model, developed for estimating time locationof estimates obtained from retrospective survivorshipreports may look complex. And this underscoresresearchers’ view that while a physician is said to makean analysis of a complex illness, a statistician makesone ill with complex analysis and this may explain whya physician is held in higher esteem than the medicalstatistician. But the interest in the medical field is moreof the application and results rather than the complexmethodology.

My interest in the development of mathematicalmodels had encouraged me to estimate indicators forhealth and diseases for countries with poor and deficientdata. The work has facilitated my estimating fertilityand infant mortality levels for Saudi Arabia and SouthWest Nigeria (Bamgboye 1989, 1990, 1993; Al-Nasserand Bamgboye, 1992).

A disease surveillance and notification system formonitoring the expanded programme onimmunization was also developed in the 1908s while Iwas working as consultant to the UNICEF office inNigeria through the establishment of sentinel sites(Bamgboye, 1988).

Development of Medical Statistics EducationThe science of Medical Statistics was at its infancy whenI joined the then Department of Preventive and SocialMedicine as there was only person in the unit. Thiswas the first Professor of Medical Statistics in Africaand one of the 3 pioneers of the programme ofMedical Statistics at the University of London, Schoolof Hygiene and Tropical Medicine – the renownedProfessor Olusola Ayeni. A statistically significantinformation is the fact that I was in the 7th set ofstudents that drank from the fountain of MedicalStatistics at the London School of Hygiene andTropical Medicine. Suffice to say that the growth ofMedical Statistics in Ibadan today was master-mindedby the doyen of Public Health worldwide – our ownProfessor Adetokunbo Lucas whose dynamism andbrilliance has led to what we witness today. Myexperience has shown that Medical Statistics is not anattractive subject to the medical student. Perhaps they

do not understand the doctrine it is preaching at thetime it is introduce to them. Therefore, my knowledgeof the psychology of students to mathematics, andparticularly of medical students and others in the alliedhealth professions who had thought the onlymicroscope they need is to read slides on patients fromthe pathology and hematology laboratories diagnoseshas led me to find simple innovative ways to teach theMedical Statistics course. A simple textbook wasproduced based on the earlier approach by ProfessorLucas (Bamgboye, 2004). The approach haspopularized the Medical Statistics course and this hasincreased the interest of students and staff in a subjecthitherto regarded as relatively difficult.

Design of Medical ExperimentsAnother area of my contribution is in the design andanalysis of various medical experiments. It is importantto note that properly designed and well analyzedclinical trials have resulted in bringing to lightappropriate prophylactic and therapeutic treatmentsof major diseases. Many treatments or procedurestoday have emerged through experiments aptlydesigned using the mathematical approach. To cometo this stage has not been a smooth journey as thestatistician suggests elements of randomization toavoid bias and at times recommends blindness as ameasure to avoid both subject and assessments bias.The physician holds a contrary view that he has toknow and be in control of the drugs or treatmentsgiven to his patients, and does not subscribe to chancein determining the dosages or treatments.

The picture today is that no treatment or drug can beintroduced into the market without undergoing clinicaltrials heavily policed by medical statisticians. In fact,for drugs developed in other countries and which haveundergone clinical trials in those countries, the regulationbody in Nigeria, NAFDAC requires a repeat of suchclinical trials using Nigerian subjects. To this end, ourdrug companies are often testing new drugs for thetreatment of urological problems, chronic diseases suchas hypertension and diabetes. My major contributionin this regard has been either in the design or statisticalanalysis of data collected during these drug trials. Forexample, one major contribution was in the studydesign and statistical analysis of multi-centre drug trialin Niprisan, a drug for the control of sickle-cell disease(SCD) in Nigeria. This drug was recently launched inAbuja with the trade name Nacosan (Wambebe et al.2001; Bamgboye, 2006).

Other clinical trials I have participated in include – justto mention a few:

(1) Anastrazole trial in the treatment of Breast Cancer;(Bamgboye, 2006).

(2) Open label, flexible dose escalation study to assessthe efficacy and safety of Doxazosin (Cardura) inpatients with benign prostatic hyperplasia;(Bamgboye, 1999).

ℓ

T1 = ∫x ∫a xWa (ℓ((a+x)T)/ℓ(a,T)(xLa-xℓ(a,T)dadx

∫x ∫a xWa (ℓ((a+x)T)/ℓ(a,T)-ℓ((a+T), T)dadx

∫x ∫a


(3) The effectiveness and safety of an individualizedSymbicort Turbuhaler maintenance dosingregimen (Symbicort Asthma Control Plan) versusSymbicort Turbuhaler given as standard regulartwice daily therapy – The Symbicort AdjustableMaintenance Study (SAM), (Bamgboye, 2001).

(4) A multi-center open label assessment of theefficacy and safety of Atorvastatin (Lipitor) in thetreatment of Types IIA & IIB Hyperlipo-proteinemia (Bamgboye, 2003); and

(5) Viagra for erectile dysfunction.

This microscope is not only towards health and diseasebut also to allied areas where the truth is needed. Suchis the performance of students in the university andtheir admission criteria. My work in this area has beenin examining the University of Ibadan as a case study.This effort resulted in the present admission guidelinefor Post-UME screening referred to as the IbadanModel, (Bamgboye, 1981; Bamgboye et al., 2001; Kaleand Bamgboye, 1984; and Ibadan Post-UME ScreeningCommittee, 2005).

The Growth of Medical Statistics in the Collegeof MedicineMr. Vice Chancellor sir, this inaugural lecture wouldnot be complete if I fail to speak on the growth ofMedical Statistics in the College of Medicine. The storywould have been “sweeter in the mouth” of ProfessorAyeni, who created the path for me to follow but hedid not have the benefit of an inaugural lecture beforehe retired from this university in 1985. A look at therecords of inaugural lectures in the University revealeda deficient data system as we can only estimate thatonly 200 professors have given since 1949. The lecturethat I am giving today is the 35th of all inaugural lecturesfrom the Faculty/College of Medicine which hasproduced 17% of all inaugural lectures ever-given inthe university. As at today, there are 57 professors inthe College of Medicine.

Medical Statistics in the Training ofUndergraduate Medical StudentsAs earlier mentioned, Professor Adetokunbo Lucas,the doyen of Public Health introduced medical statisticsinto the medical curriculum of the University of Ibadanand taught the course himself until 1969. Because ofhis understanding of the multi-disciplinary nature ofpublic health, he recruited the first professional medicalstatistician, Professor Olusola Ayeni in 1968. Thereafter,Professor Lucas established a special unit of medicalstatistics and realizing its uniqueness to medical researchand practice, located the unit in the first floor of thenew Clinical Science Building housing the other playersof the medical profession (Surgery, Medicine,Pediatrics and Obstetrics and Gyneacology). Theexperiences as told by the ‘Mungo Park’ of medicalstatistics in Africa was that of ignorance of this newdiscipline that found its root in medicine and medicalresearch. The Medical Statistics course was taught tomedical students for only 8 hours as part of a coursein the pre-clinical years but was initially never examined.

When the senate finally approved that the course beexamined under Biochemistry, Professor Ayeni toldme of a story of the Head of Biochemistry settingquestions on Biostatistics, a course he taught as amedical statistician. And unfortunately, the question waslike a sentence with neither a verb nor a predicate. Ittook the poorest student to point out this mistake. Itcame to the open and the senate queried the professorof Biochemistry who replied that the senate whenapproving the inclusion of Biostatistics as an examinablecourse in the medical curriculum did not specify whoshould set the questions. There are no two ‘Lucases’when we talk of public health in University of Ibadan.However, all researchers in medicine at that timebenefited from the only medical statistician in the first5 years that they clamoured for another one. Topicsand terms like Chi-Square (X2), Regression, Correlation,P-Value, Ramdom Sample, t-test, ANOVA, havedeveloped deep roots in medicine as they form acompulsory part of the undergraduate medicalcurriculum. But my experience has shown that medicalstatistics is not an attractive subject to any medical orpara-medical student. There was in fact a dialogue withsome medical students who wanted to know fromthe Professor of Medical Statistics why they need tostudy medical statistics at all. The professor answered;“because statistics save life.” And the students asked, “how?”The professor answered, “statistics keep all the idiots out of themedical or nursing profession.”

Postgraduate Training in Medical StatisticsOne major contribution to knowledge as a Professorof Medical Statistics is the development of the scienceof medical statistics. Mr. Vice-chancellor sir, I amhappy to announce that there has been dramaticprogress in this area since 1996/1997 academic sessionwhen a sub-department of Epidemiology, MedicalStatistics and Environmental Health (EMSEH) wascarved out of the Department of Preventive and SocialMedicine for administrative convenience. Perhaps itmay be illuminating to give some backgroundinformation on how we came about a sub-department.

The Birth of EMSEHIt all started when I was appointed as the convener ofthe 4-man committee set up by the then Provost –Professor B. Osotimehin who recommended that PSMshould divide into 6 departments for the academicgrowth of Public Health. It was not an easy task butwe managed to get 3 sub-departments of whichEpidemiology, Medical Statistics and EnvironmentalHealth (EMSEH) was one. Unfortunately, I was inSaudi Arabia when the sub-departments were created.

When I finally returned to Nigeria in 1996, to assumecoordination of the sub-department of EMSEH, thesituation was like a Yoruba saying which means that“who comes last meets the watery part of the soup”.A sub-department, scattered, with not enough officespace, poor complement of non-teaching staff, twotypist, no secretary, no office for the HOD, and aboveall, no specific academic programme except the MBBS


held on to by all members of the College, the onlyMPH in Environmental Health programme still beingcontested, and so on, was all that was inherited. It is tothe Glory of God that today, we are not a sub-department but a full department in the Faculty ofPublic Health with well-defined academic programmes.

Postgraduate EducationThe first step was to initiate the development ofpostgraduate programmes and because of the staffstrength and mix then – four epidemiologists, twomedical statisticians and one environmental healthspecialists, we decided to start with a Master’s degreeprogramme in Epidemiology and Medical Statisticsand we relied largely on part-time lecturers. Incidentally,before my final return to the Department in 1996,Professor Ayeni and I, had discussed the possibilityof a Masters Programme in Biostatistics and he assuredme of the support of WHO through its specialprogramme on Human Reproductive Health Researchand Development in which he was the Chief MedicalStatistician, and with his technical support and that ofProfessor Oluwole Akande, who was Project Managerof HRP in the WHO at that time, the proposal for anM.Sc. degree in Biostatistics was approved in principle.

The WHO support for this course which includesscholarship for some students, building capacity tostrengthen the academic staff and enhance computerfacilities, library and communication equipment, wasimplemented in the 1998/1999 academic session. Sinceits inception, the programme has trained 103professional biostatisticians and epidemiologists, bothmale and female, at the Master’s degree level (see Table2). Of these, eleven were foreign students from otherAfrican countries such as Ghana, Kenya andCameroon. Four others have already completed theirPh.D. degrees, three in Epidemiology and one inMedical Statistics, two of whom were supervised byme.

Table 2: Distribution of Students’ Enrolment forM.Sc. Biostatistics Programme since inception

Session Male Female Total 1998/99 1 4 5 1999/00 6 2 8 2000/01 8 5 13 2001/02 13 4 17 2002/03 11 1 12 2003/04 16 1 17 2004/05 10 1 11 2005/06 14 6 20 Total 79 24 103

WHO Evaluates of M.Sc. Biostatistics CourseThe course was evaluated in July 2006 by a team ofexperts from the World Health Organization andwondered how the thin staff on ground could dosuch a wonderful job. They interviewed some of thestudents and read a few of their projects. Theyconcluded that the quality and scope of the Master’s

programme is comparable to any such programme inthe World. The evaluation team recommendedcontinued financial WHO support for the course forat lease a further 2-3 years based on the high level ofperformance.

Other Master’s ProgrammeAt the instance of the request of the WHO throughthe Federal Ministry of Health another Master’s degreeprogramme in Filed Epidemiology was developedand approved in the 2002/03 academic session in theDepartment. In the 2003/04 academic session, yetanother Master’s Course Program, the MPH in MedicalDemography, the first of its kind in any AfricanUniversity was developed and approved for theDepartment. This course however also received sometechnical support from the Melinda and Gates Institutefor Reproductive Health based at the John HopkinsUniversity, USA. Thus today, to the Glory of God,we have 6 Master’s degree programmes in theDepartment.

Mr. Vice Chancellor sir, we need more staff to copewith the heavy academic postgraduate run by ourDepartment. The impact of which has helped researchoutput in the College of Medicine, Federal Ministryof Health and other medical and health agencies inNigeria.

CONCLUSIONThis lecture has revealed the science of medical statisticsas the microscope to understand health and disease.The diagnostic criteria for numerous diseases have beendeveloped using sound statistical theories with thedominant contribution of probability and multivariateanalytic technique such as discriminant analysis.

Any medical statistician often gets worked up aboutother people’s activities so that he can make (a numberof times) some of them his own and assume fullresponsibility. One particular attribute of a medicalstatistician is his being at home with mathematics, havingthe ability to read the more mathematical literature, tofollow mathematical argument and use mathematicaltools with some confidence. Every mathematical topicshould be of potential interest to a medical statisticianbut what he needs is the relatively elementary bit.However, the methodologically-oriented statistician isable to make the fullest use of his non-medical basictraining and by the generality of his approach, may besuccessful in developing new and modified techniquesthat may be of value far outside the immediate contextthat prompted the works. He should however notleave the task of extracting information from real datato the epidemiologist. Both Medical Statistician andEpidemiologists should foster a symbolic relationshipto be more productive. This is the basis of thetremendous strength of the Department ofEpidemiology and Medical Statistics from our currentDepartment of EMSEH.

Sadly enough, the practice gaining unprecedentedground in collaborative research between our medical


researchers and colleagues in foreign countries is thatNigerian is being turned to a mere collecting site,whereby the blood specimens as well as data collectedin some major studies (particularly on HIV/AIDS),are sent directly to the foreign partners overseas. Inmost cases the researcher in Nigeria cannot even accessthe data after it has been entered into the computer. Imake bold to say that this is not a collaborative research!

In any collaborative research, every professional is anequal member of the research team, and so the besttime to involve every member, particularly the medicalstatistician is at the planning stage. This is so that hecan contribute technological suggestions about theconduct of the research which may greatly increase itsimpact or reduce its cost; and he/she can contribute agood deal more than this if he/she has an adequateand broad knowledge of the subject matter, which isoften recommended.

There is no amount of statistical message that can bringthe truth out of a badly planned study. But a studydeveloped according to statistical science can becorrected for appropriateness of statistical methodsof analysis of data if discovered later. From myexperience, Medicine contains its full quota ofdominant personalities and if one of these canpersuade a medical statistician to be processing his dataand be providing him with the answers he expects, hewill certainly turn the statistician into his technician.

This inaugural lecture has shown beyond reasonabledoubt that a medical statistician as a professionalunderstands the real needs of his clients or employer,often very much better than the client himself, and hasthe ability and status to persuade the client to recognizehis own interest and to abandon his less useful flightsof fancy. He has a professional skill to recommendfrom a range of known and trusted techniques, thoseresearch methods that in the given circumstance willachieve the required effect at minimum cost andinconvenience to the client; and finally, he has theprofessional integrity to resign his post or commissionif his recommendations are not accepted. This is thetrue collaboration which I have found myselfdemonstrating as a medical statistician in the medicalmilieu.

Thus, if the chief medical director of any hospital orthe director of a medical research group needssomeone to do routine data analysis, they can hire awell trained statistician or buy a computer; just likewhat Paul told Corinthians in his first letter in Chapter7 verse 8 that “they do well”. But if they want a realmedical statistician, they need to pitch their expectationsat a realistic level. In this regard, the medical statisticianon their staff can regard him/herself as a member ofa group of professionals which cuts across narrowspecialties of medicine, that he can give seminars,organize workshops and take part in discussions atwhich his special problem can be seen in their widestatistical context. As one of my professors at theLondon School of Hygeine and Tropical Medicine

said, “the medical statistician is thus a generalist whosespecial gift is precisely her lack of specialization”. Thisis the level of collaboration that has a promise. Becausein this regard, the medical statistician is to be welcomedin the health or medical domain as an equal colleaguewith the right of entry, not a cap-in-hand visitor.

The bulk of our health problems in this country canbenefit more from a population based approach inwhich the members of the ‘club’ of Public Health areendowed with the necessary expertise. The faculty ofPublic Health in the College of Medicine, Universityof Ibadan can change the poor health situation inNigeria if all members of the team come togetherand throw away what I label professional intolerance.The physician in the group should not arrogate theadministrative leadership role exclusively to himselfand see others as support staff. The faculty of PublicHealth should live to the expectations of the foundingfathers and develop to a School of Public Health.

The Federal Ministry of Health should set aside 3%of all allocations to medical statistics. This will help toexplain and to assess how remaining 97% is disbursedand spent.

Today’s inaugural lecture calls for celebration as PSMhas developed to Faculty of Public Health and myDepartment of EMSEH for which I am head is almostbecoming as large as the old PSM. The Departmentof EMSEH has been a melting point for professionalmedical statisticians, epidemiologists andenvironmental health specialists. Our products are nowfound in Ghana, Cameroon, Kenya and in virtually allmedical research institutions and agencies in Nigeria.

AcknowledgementsI wish to thank the Almighty God who has made itpossible for me to be what I am, especially to standbefore you today to give an inaugural lecture, which Ihad thought has eluded me, as those who becameProfessors 10 years after me have since given theirs.

I will like to thank my parents, Elder & Mrs. JacobDada Bamgboye of blessed memory who thought itfit to send me to school despite the fact that they hadno formal education. I also thank my Uncle, Mr.Emmanuel Ayorinde Bamgboye who exposed me tothe University of Ibadan since 1958 when I lived withhim to complete my primary school at the AbadinaSchool. I appreciate my late maternal uncle, Chief AyoOmidiji whom I followed to Ibadan in 1958 and whobought me my first suit in 1966 when I left secondaryschool and late Chief Titus Oke who coached me athome for the A-levels in Pure and Applied Mathematicsin 1967, working sometimes late in the night in spiteof his busy schedules.

I am most grateful to Professors M.I. Iro and OpeyemiOla who first introduced Medical Demography to meas a Subset of Medical Statistics in 1973. I thank theUniversity of Ibadan for employing and allowing meto travel overseas when it mattered to have Master’s


and Doctorate Degrees. In this respect, I want to thankProfessor E.O Akande specially, who as Dean of theFaculty of Medicine in 1978, recommended theapproval of a leave of absence with pay for me andsecond day approved the recommendation as ActingVice-Chancellor. I appreciate my supervisor andmentor, the late Professor William Brass and thecontributions of Professors Basia Zaba, John Blacker,John Simon, and M.J.R Healy and the entire facultymembers in the London School of Hygiene andTropical Medicine between 1976 and 1983.

I thank Professor Olusola Ayeni immensely, forbringing happiness to the Department when it mattered,by developing a Master’s programme in Biostatisticsand for facilitating the financial support for theprogramme and I also appreciate his intellectual reviewof the manuscript of today’s inaugural lecture. Iappreciate highly, late Professor Benjamin Osuntokunfor his mentoring, Professor M.T Shokunbi for makinggood his promise in facilitating the creation of theFaculty of Public Health and support of ProfessorI.F .Adewole, former Provost of the College ofMedicine for the growth of EMSEH.

I thank Professor Adefunke Oyemade whoencouraged me to travel to Saudi Arabia, ProfessorOlu Longe, who submitted my CV to the departmentof Family and Community Medicine, King SaudUniversity and the then Head of Department there,Prof. Shabrawy Alli, one of the leading Public HealthPhysicians in the world (Africa) who recommendedinstant appointment for me on seeing my CV.

I thank my Dean, Professor Akinyele who gave methe opportunity of an inaugural lecture and read throughthe manuscript for constructive contributions, and Ithank all the members of staff of my Department.Permit me to mention Professor J.D Adeniyi, a gentleand honest man to the core, who has inspired me onmany occasions in my academic pursuit.

I thank my sister, Mrs. Florence Ogundare Awayinofor her kindness to me at all times. I also appreciateMr. Thompson Akinwunmi for accommodating mein his 10B, Repton Court in Hommerton area duringmy very first visit to London in 1976. My uncle, Mr.Joseph Omidiji, is also appreciated. He has alwaysstayed with me in times of need.

You will agree that you need to be spiritually healthyaccording to WHO to be healthy and so I have a fullcompliment of members of my church from C.A.C.Oke-Agbara, Prophet M.O. Olowere and his wife,our District Superintendent, Pastor Adams and theDistrict Evangelist, Pastor Farayola and a host of otherPastors including Pastors Gbuyiro, Omidiji,Oluwasanmi, and Faturoti, as well as Elders andDeaconesses. I thank them all. I also thank the Presidentof C.A.C worldwide, Pastor (Dr.) E.H.L Oluseye andmembers of the governing council of Joseph AyoBabalola University (JABU) headed by the Chairman,Professor A.M. Imevbore and other members of staff.

My appreciation goes to the Alaaye of Efon-Alaaye,Oba (Dr.) E.A Aladejare who is visiting this hall forthe 2nd time in a month; his Olori and other sons anddaughters of Efon Alaaye present. I appreciate andthank my friend of 43 years, 10 months and 2 daysfor his care and encouragement in turbulent anddifficult times of life. He is on a special assignment inBenue State but had to be here today for this specialoccasion. I am talking of no less a person thanProfessor James Oladipo Kolawole, a Professor ofPolitical Science and current Deputy Vice-Chancellor,University of Ado-Ekiti. I am grateful to our seniorbrother, Mr. Rufus Kolawole for the 201 pounds hegave me in 1981 when University of Londondemanded some money I did not have before I couldbe allowed to defend my Ph.D. thesis. I appreciate mytwo siblings, Mr. Ezekiel Adebayo Bamgboye and Dr.Isaac Adeleke Bamgboye.

Finally, I will like to thank my better half, Mrs. EbunAyodele Bamgboye who has just completed her Ph.D.in Virology and currently Chief Technologist, IMRAT,College of Medicine, for her ability to cope with thefamily situation when I am away on several occasionsor come home late from the office.

Lastly and not the least, I will like to pay tribute to mychildren, Miss Eunice Adeola Bamgboye, the careerlady and understanding daughter, my ever smiling son,Ebenezer Adebola Bamgboye and the baby of thefamily, Dr. Eniola Adetola Bamgboye, who typed asignificant proportion of the manuscript for thisinaugural lecture.

Thank you and God bless you all.

REFERENCESal-Balla S.R., Bamgboye E.A., al-Rasheed R., Sekiat

M. A. and Balla M. (1993) Pattern of adultadmissions into the medical wards of King KhaledUniversity Hospital, Riyadh, Saudi Arabia. Saudi Med.J. 14(3): 225-229.

al-Balla S.R., Bamgboye E.A., Sekiat M.A. and BallaM. (1993) Causes of morbidity in elderlypopulation of Saudi Arabia. J. Trop. Med. Hyg. 96:157-162.

al-Fawaz, Bamgboye E.A. and Al-Eissa Y.A. (1994)Factors influencing linear growth in Saudi Arabianchildren aged 6-24 months. J. of Trop. Peadiatric:40:235-23.

al-Frayh A.S. and Bamgboye E.A. (1995) Acomparison of length, weight and OFC of Saudifull-term newborns with the KACTS and MOHstandard growth charts. Paediatrics review andCommunications – An Int. J.: 8:193-201.

al-Frayh A.S. and Bamgboye E.A. and Mousa M.A.(1993) The physical growth chart of pre-schoolSaudi Arabian Children. Ann. Saudi Med.: 13(2): 15-19.


al-Frayh A.S. and Bamgboye E.A (1992) The growthpattern of Saudi Arabian pre-school children toNCHS/CDC reference population J. R. Soc. Health:113(5): 234-239.

al-Nuaim A.R. and Bamgboye E.A. (1995) Acomparison of the anthropometry of Saudi Arabiamale children aged 6-11 years with the NCHS/CDC reference population J. Medical Principles andPractice. 7:96-103.

al-Nuaim A.R., Bamgboye E.A. and Al-Herbish A.,(1996) The pattern of growth and obesity in SaudiArabian male school children . Int. J. of Obesity andRelated Metabolic Disorders 20(1): 1000-1005.

Ayeni O. (1971), Guidelines for the establishment ofvital registration system in nigeria. Medical Journal(1) pp 214-217.

Badru I. and Bamgboye E.A. (1990). Survey of eyedisease and visual loss in the Eastern province ofSaudi Arabia. Report submitted to KKESH, Riyadh,Saudi Arabia. 1-93.

Bailey, N.T.J. (1957). The mathematical theory ofepidemics, London: Griffin.

Bailey, N.T.J. (1975). The mathematical theory ofinfectious diseases and its applications. London:Griffin.

Baiyewu O., Bella A.F., Adeyemi J.O., BamgboyeE.A., and Jegede R.O., (1997), Attitude to agingamong different groups in Nigeria. Int. J. Aging andHuman Dev. 44(4); 2883-2892.

Health problems and socio-demographic findings inelderly Nigerians. Afr. J. Med. Sci. 26, 13-1.

Bamgboye, E.A. (1997). Sentinel Surveillance Systemfor EPI target disease in Nigeria. Report submittedto UNICEF (Lagos, Nigeria). pp. 1-57.

Comparative study of the performance of studentsadmitted to study medicine and science at thepreliminary examination of the University ofIbadan, 1973-1981.

The demographic characteristics of discharge anddeaths in a University Teaching Hospital in SouthWestern Nigeria. Nig. Med. Journal. 16(1&2): 139-146. 1985.

Sentinel surveillance system for EPI target diseases inNigeria. Report submitted to submitted UNICEF(Lagos, Nigeria). pp. 1-57.

Levels and trends in infant and under five year oldmortality in South Western Nigeria. Reportsubmitted to USA Agency for CCCD (Lagos,Nigeria). 1-56.

Dating estimates of infant and childhood mortalityfrom child survivorship data with applications. GenuLXV (10): 25-37.

Estimates of infants and childhood mortality levelsand trends in rural and urban South Western Nigeria.Res. For. Dev. 10(1&2): 74-84.

A Statistical report on open label, flexible doseescalation study to assess the efficacy and safety ofDoxazosin (Cardura) in patients with benignprostatic hyperplasia, submitted to Pfizer.

The effectiveness and safety of an individualizedsymbicort Turbuhaler maintenance dosing regimen(Symbicort Asthina Control Plan) versus symbicortTurbuhaler given as standard regularly twice takingtherapy. THE SYMBICORT ADJUSTABLEMAINTENANCE STUDY submitted to Astra-Zeneca.

A Statistical report on a multi-center open labelassessment of the efficacy and safety of atorvastatin(Lipitor) in the treatment of types IIA and IIBhyper-lipoproteinemia submitted to Pfizer.

A companion of Medical Statistics. Published byFOLBAM Health Research and Data ManagementCenter. Ibadan.

A statistical report on the prevalence of dyslipidemiain hypertensive and diabetic Nigerian patients.

A statistical report of a multi-centre prospectiverandomized trial of tamoxifen versus anastrozoleas first line hormonal treatment in post menopausalpatients with advanced breast cancer in Nigeria.Submitted to Astra, Zeneca

Bamgboye, E.A. and Adeleye A.I. (1992). Sicknessabsenteeism in a Nigerian Teaching Hospital, EastAfr. Med. J.: 69(8): 40-4.

Bamgboye, E.A. and Al-Nahedh., (2003). Factorsassociated with growth faltering in children fromrural Saudi Arabia. Afr. Med. Med. Sci.

Bamgboye E.A. and Familusi J.B. (1990). Mortalitypattern at the children emergency room. UniversityCollege Hospital, Ibadan, Nigeria. Afr. J. Med. Sci.19(2): 127-132.

Bamgboye E.A. and Jarallah J. (1994). Long waitingoutpatients: A target for health education. Patterneducation and counseling. 23:49-54.

Bamgboye E.A. and Jegede R.O. (1987). Aprospective study of the pattern of utilization ofmental health services of a Nigerian UniversityCollege Hospital. Afr. J. Med and Med. Sci. 16(1):27-32.

Bamgboye E.A. and Shoge F.A. (1987). An analysisof in-patient statistics in a Nigerian UniversityCollege Hospital. West Afr. J. Med. (6 & 4): 225-229.

Bamgboye E.A. and Zoaka A. (2005). A StatisticalEvaluation of Prevention of Mother Transmissionof HIV/AIDS submitted to the Federal Ministryof Health, NASCAP, Abuja.

Bamgboye E.A., Erinoso H.O. and Ogunlesi A.(1992). The waiting time at the children emergencyroom of the University College Hospital, Ibadan.Nig. Paediatr. 19(1):9-14, 1992.


Bamgboye E.A., Ogunnowo B.E, Badru O.B. andAdewoye O.E. (2001). Students admissiom gradeand their performance at the Ibadan University Pre-clinical MBBS examinations. Afr. J. Med. Sci. Vol.30, pg. 207-211.

Bamiduro T.A. (2005). Statistics and search for thetruth: a Biometrician’s view. An Inaugural lecturegiven at the University of Ibadan.

Brass W. and Bamgboye E.A. (1981). The time locationof reports of survivorship estimates of maternal,paternal and the ever-widowed. Centre for PopulationStudies, London School of Hygiene and Tropical Medicine,Occasional paper 81-1:1-18.

Erinoso B.O. and Bamgboye E.A. (1988). Sicknessabsenteeism in a Nigeria Polytechnic. Afr. J. Med.and Med. Sci. 17(1) 57-61.

Federal Ministry of Health (2005). National AIDS andSTI Control Programme: National Situation andResponse Analysis of the Health Sector Responseto HIV/AIDS in Nigeria.

National AIDS and STI Control Programme: NationalHealth Sector Strategic Plan for HIV/AIDS inNigeria 2005-2009. Abuja, Federal Ministry ofHealth

Healy M.J.R. (1977). Does medical statistics exist? Alecture given at London School of Hygiene andTropical Medicine London.

Hill K. (1977). Estimating adult mortality levels frominformation on widowhood population students3. 75-84.

Jegede R.O., Ohaeri J.U., Bamgboye E.A. andOkunade A. O. (1990). Psychiatric morbidity in ageneral outpatient clinic. West Afr. J. Med. 3: 177-187, 1990.

Jewell N.P. (1990). Some statistical issues in studiesof epidemiology of AIDS. Statistics in Medicine.9:1387-1416

Kale O.O. and Bamgboye E.A. (1984). Survey ofshot essays type MBBS examination questions inpreventive and social medicine at the University ofIbadan. 1975-1979. West Afr. J. Med. 3(1):25-30.

Taylor G.O. and Bamgboye E.A. (1979). Serumcholesterol and diseases in Nigeria. American Journalof Clinical Nutrition 32. 2540-2545.

Taylor G.O., Bamgboye E.A., Oyediran A.B.O.O.,and Longe O. (1992). Serum creatinine andprediction formula for creatinine. Afr. J. Med. andMed. Sci. 11:75-81.

Taylor G.O., Oyediran A.B.O.O., Bamgboye E.A.,Afolabi B.M. and Osuntokun B.O. (1996). Profileof some risk factors for coronary heart disease in adeveloping country: Nigeria. Afr. J. Med. and Med.Sci. 25:341-346.

United Nations Manual X (1983). Indirect Techniquesfor Demographic Estimation. 145-147

Wambebe C.O.N., Bamgboye E.A., Badru O.B.,Khamofu I. I., Momoh J. A., Ekpeyong M.U.B.S.,Njoku O.S., Bamgboye E. A., Nasipuri R..N., KunleO.O., Okogun J.I., Enwerem M.N., Audam J.G.,Gamaliel K.S., Obodozie O.O., Samuel B., FojuleG., Ogunyale O., (2001). Efficacy of Niprisan inthe preventive management of sickle cell patients:Current Therapeutic Research: 62. 1:26-43.

Yusuf, O.B. (2005). The epidemiology, populationdynamics and mathematical modeling ofantimalaria drug resistance in Ibadan. Ph.D.Dissertation, University of Ibadan.

Zeger, S. (1991). Statistical reasoning in epidemiology.American Journal of Epidemiology, 134: 1062-66.

Zelen, M. (1983). Biostatistical science as a discipline:A look into the future (with discussion). Biometrics,827-37.


IntroductionThe increasing volume of information published inbiomedical literatures poses an enormous challengeto evidence-based health care and scientific discoveries.It is common for important issues in medical researchto be addressed in several studies. The idea ofsummarizing a set of studies is not new in medicalliterature; review articles have long had an importantrole in helping practitioners keep up to date and makesense of the many studies on any given topic. Metaanalysis goes a step further by using statisticalprocedures to combine the results of several studies.

DefinitionSeveral definitions exist in the literature. However,Glass who developed the technique defined it as thestatistical analysis of a large collection of analysis results for thepurpose of integrating the findings 1. Another author definedit as a statistical procedure that integrates the results ofseveral independent studies considered to be“combinable.”2. According to Medical SubjectHeadings (MESH), meta-analysis is a quantitativemethod of combining the results of independentstudies (usually drawn from published literature) andsynthesizing summaries and conclusions which maybe used to evaluate therapeutic effectiveness, plan newstudies, etc., with application chiefly in the areas ofresearch and medicine 3. Well conducted meta-analysesallow a more objective appraisal of the evidence thantraditional narrative reviews, provide a more preciseestimate of a treatment effect, and may explainheterogeneitybetween the results of individual studies.Ill conducted meta-analyses, on the other hand, maybe biased owing to exclusion of relevant studies orinclusion of inadequate studies4. Misleading analysescan generally be avoided if a few basic principles areobserved.This review article discusses these principles,along with the steps in performing meta-analysis. Itconcludes by highlighting the future and role of meta-analysis in medical discoveries.

Evolution in MedicineThe first meta-analysis was a combination of studies(with small sample sizes) of typhoid vaccineeffectiveness performed by Karl Pearson in 19045 inan attempt to overcome the problem of reduced

statistical power. Although meta-analysis is widely usedin epidemiology and evidence-based medicine today,a meta-analysis of a medical treatment was notpublished until 1955. The term was coined by Glass1

in 1976, even though some meta-analytic methods havebeen in use for almost fifteen years in Education andPsychology. The concept made its way into medicineas researchers began to incorporate the idea. At theonset, the concept was not popular among medicalscientists until mid 1980s when a group of cliniciansand statisticians at Oxford University initiated theprocess of giving it scientific prominence. The Oxfordgroup took the approach of gathering all studies,published and unpublished, and excluding those thatused different endpoints. They focused on studies oftherapeutic issues6. Their conclusions were applied byothers to clinical practice, to further ascertain validity.By 1985, there was a book on statistical methods formeta-analysis 7. In addition, there was a 1985publication, “Findings for Public Health from Meta-analysis” which clarifies the difference between meta-analysis and traditional literature review. Since then, thetechnique has grown in leaps and bounds withapplication in different areas of medicine and otherspecialties. Its statistical methodology is always reviewedand constantly improved to accommodate new realities.For the results of a meta-analysis to be meaningful, agreat deal of thought and planning are needed.Protocols for the reporting of meta-analysis resultswere developed for Randomized Clinical Trials (RCTs)(Quality of Reports of Meta-analysis [QUOROM])8

and observational studies (Meta-analysis ofObservational Studies in Epidemiology [MOOSE])9.These guidelines were developed to provide properprocedures for conducting a meta-analysis and tostandardize the methods of reporting it. Using these 2protocols as a guide, the steps necessary to perform ameta-analysis include the following: (1) define theresearch question, (2) perform the literature search, (3)select the studies, (4) extract the data, (5) analyze thedata, and (6) report the results.

Define the Research QuestionA meta-analysis begins with a question. Commonquestions addressed in meta-analyses are whether one

META-ANALYSIS: THE WAY FORWARD IN MEDICAL DISCOVERYAkinyemi J.O. MSc (Medical Statistics), B Tech (Comp. Sc.)

Correspondence:Mr. J.O. AkinyemiDepartment of Epidemiology,Medical Statistics and Environmental Health,Faculty of Public Health,College of Medicine,University of Ibadan, Ibadan, NigeriaE-mail: [email protected], [email protected]: +234-8055676325


treatment is more effective than another or if exposureto a certain agent will result in disease. Before beginningan analysis, the investigators need to define the problemor question of interest. The investigators should alsohave a good understanding of the problem and thesubject matter10. The study population baseline data(e.g., age, race, gender, diagnosis, length of illness), thestudy outcomes, treatment or intervention, and typeof study designs to be used (e.g., restricted to RCTsor include observational studies such as prospectiveor retrospective studies) also should be defined11.

Perform the Literature SearchOnce the research question has been defined, asystematic search of the literature can begin. This is acritical step in the meta-analysis and often the mostdifficult part. The initial search of the literature shouldbe broad so that as many studies as possible aregathered. During the selection phase, some of the initialstudies will be weeded out using the inclusion criteria.

The literature search begins with searching electronicdatabases of published studies such as MEDLINE,EMBASE, CINAHL, etc. MEDLINE is maintainedby the National Library of Medicine and contains morethan 13 million citations dating back to 196612.EMBASE is a database produced by the publisherElsevier BV and contains data from 1974 to thepresent13. Although EMBASE and MEDLINEoverlap in their coverage of the literature, EMBASEhas better coverage of European journals14. CINAHLcovers literature related to nursing and allied healthfrom 1982 to the present15. The researchers shouldsearch more than just MEDLINE to ensure acomprehensive search. For example, a report foundthat approximately only half of all RCTs presented asabstracts are subsequently published on MEDLINE16.It is necessary to use other sources to access many ofthese unpublished studies. A good source forunpublished clinical trials is the Cochrane CentralRegister of Controlled Trials, which is a database ofcontrolled trials. The database was set up to provide asource of data for systematic reviews and containsmore than 300,000 references to RCTs17. Othersuggestions for locating studies include searchingreference lists from the gathered reports, manuallysearching journals with lists of abstracts presented atmeetings, or searching on the Internet. Contactingexperts in the field or networking with colleagues alsocould be a source of studies, although this mode ofdata gathering is seldom used.

Select the StudiesOnce the literature search is complete, it is time toselect which studies to include in the meta-analysis. Theinclusion and exclusion criteria for studies need to bedefined at the beginning, during the design stage ofthe meta-analysis. Factors determining inclusion in theanalysis are study design, population characteristics, typeof treatment or exposure, and outcome measures 18.The inclusion and exclusion criteria should be part ofthe meta-analysis protocol. One should keep track ofthe studies included and excluded at each step of the

selection to document the process. The QUOROMguidelines for reporting a meta-analysis request thatinvestigators provide a flow diagram of the selectionprocess8. The flow diagram lists the number of studiesexcluded and included at each stage of the selectionprocess and the reasons for exclusion. The selectionprocess involves reviewing the titles and abstracts ofall articles identified through the literature search.

Many of the studies will be excluded at this stage basedon the exclusion criteria. The remaining studies will beread to determine their suitability for inclusion. Thevalidity of a meta-analysis depends on the quality ofthe studies included, and an assessment of quality is anecessary part of the process. The researcher wants toinclude as many studies as possible, but reduce thenumber of studies with low quality data; however,restricting the meta-analysis to only perfect studies mayleave the researcher with little data19. A variety ofchecklists and scales have been developed to assessquality in RCTs20. Checklists provide guidelines as towhat should be reported in an RCT, whereas scalesare a way of quantifying the level of bias in an RCT.For example, a scale will assign a score based on aspecific characteristic of an RCT (e.g., presence ofadequate concealment of patient assignment totreatment groups), but a checklist does not assign scores.Although quality needs to be assessed in some way,caution should be used when using these scales orchecklists20, 21. The reasons for the inclusion of items ina scale or checklist often are not given and the scoreassigned to scales can be arbitrary20. The relativelyimprecise scoring scheme in some of the scales maychange the results of a meta-analysis21.

There are several options available to deal with studyquality once it has been ascertained. A cut-off valuefor the quality score can be used to exclude or includestudies19, 22. Another choice is to use the quality scoresto weight study results in the analysis. MOOSEreporting guidelines, however, recommend using asensitivity analysis rather than weighting for qualityscores9. Sensitivity or subgroup analysis allowscomparisons between studies of different quality22. Forexample, studies can be separated into high versus lowquality and then the meta-analysis can be repeated foreach group. Results then can be compared betweenthe 2 groups. A method that is being used increasinglyis meta-regression. Quality scores or some measureof study quality (e.g., assignment to a treatment group)are entered into a regression model as an explanatoryvariable19. This method allows the researchers toestimate the effect of quality on the results of the meta-analysis.

Extract the DataThe type of data to be extracted from each studyshould be determined in the design phase and astandardized form is constructed to record the data.Examples of data commonly extracted include studydesign, descriptions of study groups (e.g., number ineach group, age, gender), diagnostic information,treatments, length of follow-up evaluation, and


outcome measures. Two independent reviewers willbe instructed on the appropriate data to collect. Forexample, how will age be recorded on the abstractform? Will the standard deviation or standard errorbe used in the analysis? If data are missing, they shouldbe recorded on the form. If too much data aremissing, the study may need to be excluded. It isrecommended that the reviewers be blinded to theinvestigators’ names but it is not essential7, 23. Data tobe extracted are identified before beginning the meta-analysis to avoid data dredging or a fishing expedition.The difficulty with data extraction is that studies oftenuse different outcome metrics, which make combiningthe data awkward. The data should be converted to auniform metric for pooling. For example, datareported may be continuous (e.g., blood pressure) orbinary (e.g., high blood pressure vs low blood pressure).A meta-analysis estimating the effect of a medicationon blood pressure may find some studies reportingblood pressure as a continuous outcome whereas inother studies the outcome is reported only as high orlow blood pressure. In this case it would be necessaryto convert continuous blood pressure measurementsinto categories of high or low blood pressure tostandardize the data into one format. Similarly, somestudies report the standard deviation and others reportstandard error. Again, it is necessary to convert oneinto the other to make the data uniform. Although itis difficult to resist combining the data, if combiningdata is not possible because different metrics are usedthen it is best to leave the analysis as a systematic review.

Analyze the DataA statistician who is familiar with meta-analysis shouldbe consulted to help plan this type of project and toparticipate in analyzing the data. Detailed instructionsfor data analysis exist 19, 24 – 26. A meta-analysis calculatesa weighted average of the study effect that is pooledfrom the selected studies. The weight is directlyproportional to the precision of the effect estimateand usually the inverse of the variance (square of thestandard error) of the effect estimate19. Therefore,larger studies will have more influence over thesummary estimate than smaller studies23. A summaryestimate is calculated by multiplying each study’s weightby its effect estimate and adding these values together.This sum then is divided by the sum of the studyweights. There are 2 statistical models used in a meta-analysis: fixed effects and random effects. The fixed-effects model assumes that the true effect of treatmentis the same for every study. Because there is noheterogeneity between study results, only within-studyvariability is taken into account. Given the degree ofvariation or heterogeneity among studies, thisassumption may be unreasonable. The random-effectsmodel is often more realistic because it assumes thatthe true effect estimate for each study does vary. Sourcesof variation may include differences in patientpopulation or treatment methods. The random-effectsmodel will produce an estimate with wider confidenceintervals, but the summary estimates for both modelswill be similar if there is not a great deal ofheterogeneity among studies. A statistical test for

heterogeneity can be used, but this test has low statisticalpower in most cases19. Power refers to the ability of astatistical test to reject the hypothesis being tested (nullhypothesis) when it is false. The null hypothesis statesthat there is no heterogeneity or variation among thestudies. Low power for the heterogeneity test meansthat we are unable to reject the null hypothesis of noheterogeneity even when important heterogeneity exists.For example, the studies used in the meta-analysis mayin reality vary considerably, but the low power makesthe heterogeneity test non-significant. This would leadthe researcher to the incorrect conclusion that theamount of variation among the studies is low. Thebest choice may be to always use the random-effectsmodel or to use both models and compare the results.Statistical packages are available to calculate summaryestimates using either model. If heterogeneity can beexplained, then it should be included in the model.For instance, we may observe that some of the variationin studies can be explained by gender. In that case,separate summary estimates can be calculated for malesand for females. Or, meta-regression models can beused to explain heterogeneity, but a large number ofstudies are needed when investigating multiple effects.

Report the ResultsDetailed guidelines for the reporting of meta-analysesfor RCTs were described in the QUOROM statement8.Similar guidelines were developed for observationalstudies by the MOOSE group9. These articles shouldbe consulted during the design phase to ensure thatthese reporting procedures are used and that properdata are collected and presented in the report. Similarto a research report, a meta-analysis report shouldinclude a title, abstract, and introduction, and methods,results, and discussion sections. The title should identifythe report as a meta-analysis. The introduction shouldindicate the clinical question of interest, the hypothesisbeing tested, the types of treatment or exposure beingstudied, the study designs to be included, and adescription of the study population. The methodssection should describe the literature search, specificallythe databases used, and if the search was restricted inany way (e.g., English language only). The selectionprocess for articles, quality assessment, methods ofdata abstraction, and synthesis also should be describedin this section. The results section should include a flowchart of studies included in each step of the selectionprocess, a figure displaying the results from eachindividual study such as a forest plot, results ofheterogeneity testing, overall summary statistic and its95% confidence interval, and results of a sensitivityanalysis and meta-regression, if performed. Forsensitivity analysis, several features of a meta-analysiscan be altered to assess the robustness of the results,such as excluding questionable or unpublished studies.The sensitivity analysis may include an analysis weightedby a quality score for each study. The discussion sectionshould summarize the key findings and identify possiblesources of bias and heterogeneity. A forest plot, thefigure with the effect estimate from each study andtheir associated confidence intervals along with the


summary estimate, is an important part of the report.Studies can be grouped by size or by other studycharacteristics such as year of publication. It allowsthe reader to observe the heterogeneity of the studiesincluded. If the confidence intervals for effectestimates are not overlapping, indicating a great dealof study variation, a meta-analysis may not beappropriate. In this case, it is necessary to explore thereasons for the variation among the studies, which maylead to the discovery of associations between the studydesign or patient groups and the study outcome.

A funnel plot is used as a way to assess publicationbias in a meta-analysis. The funnel plot is a scatter plotof each study’s effect estimate (e.g., odds ratio or meandifference) on the x-axis against a measure of the study’sprecision on the y-axis 19, 27. The overall sample sizecan be used on the y-axis but often an inverse of thestandard error is used28. If a publication bias is notpresent, the plot will resemble an inverted funnel. Largestudies should have smaller variation and therefore amore precise effect estimate, whereas small studiesshould have larger variation and therefore a less-preciseestimate. One expects the effect estimates for smallstudies will have wider scatter at the bottom of theplot and larger studies will have less scatter at the topof the plot. If small studies with negative or null resultstend not to be published, one would see asymmetryin the funnel plot from the left bottom of the plotcontaining few or no data points. On the other hand,if fewer studies with non-statistically significant oddsratios were included in the literature search, it wouldresult in an asymmetric funnel plot. Funnel plots canbe inspected visually but interpretation can differ fromperson to person. Statistical tests such as the rankcorrelation test developed by Begg and Mazumdar29

are available to assess the symmetry of the plot. Thecorrelation test, however, should be used with cautionin small meta-analyses because the power of the testdepends on the number of studies included29. In asmall meta-analysis (25 studies), the correlation test willhave low statistical power so a non-significant test willnot rule out bias in the literature search.

A meta-analysis is a statistical method of combiningresults from multiple studies to determine the overallimpact of a treatment or exposure. If performedproperly, using the steps above, a meta-analysis can bea powerful research tool. Although meta-analyses areconsidered to have the highest level of evidence andare cited more often than other study designs, thereare still lingering questions regarding its validity whencompared with well-conducted clinical trials 30, 31. Theresults from a meta-analysis can be used to plan a largeRCT to test a treatment effect; however, a reportcomparing the results from meta-analyses andsubsequent RCTs found only fair agreement32. Thedistrust of meta-analyses and the lack of agreementwith RCTs do not imply that the meta-analysis shouldbe abandoned. It does, however, point out thelimitations and biases involved with the meta-analysisand shows the need for conducting a thoughtful, wellplanned meta-analysis with minimal bias. When

performed appropriately, however, a meta-analysis canlend evidence to many of the difficult decisions cliniciansmade in their daily practice.

Recent Development and Future of Meta-analysisIn a review of recent developments in meta- analysis,Stutton et al33 noted that there is considerable researchactivity in the field of meta-analysis. Meta-analysismethodologies are being developed for concepts suchas prospective meta-analysis, meta-analysis ofindividual patient data, etc. There are meta-analysistechniques for complex evidence synthesis whichinvolve models that incorporate evidence on multipleparameters and/or that specifically model data fromdifferent study designs. There are developments inmeta-analyses of studies on effects of interventions,aetiology, diagnosis and screening. A recent publicationof recommendations for reporting tumor markerprognostic studies was reputed to be a good initiativethat will aid the meta- analysis of such studies in thefuture34.

There are notable developments in software for meta-analysis as well. However, due to the fact that meta-analysis of summary data needs a unique set of analysistools, the large developers of general statistical softwarehave been reticent about providing the requiredroutines. Fortunately, users have developed collectionsof macros, e.g. for SAS35–37 and, most comprehensively,for STATA38. Stand-alone packages have also beendeveloped, the most sophisticated of which is probablythe commercial Comprehensive Meta Analysis39. TheCochrane Collaboration software, RevMan 40,continues to be developed and a new freelydownloadable Excel add-in MIX41 offers excellentfunctionality and educational content for those on atight budget. Sutton et al33 reported that they foundMetaDiSc42 very useful for carrying out the specializedanalyses required for diagnostic tests meta-analysesThere is little doubt that the development of meta-analytic techniques will continue into the future. Multipletreatment comparisons will receive greater attentionas it has potential to address the relative benefits ofcompeting treatments, and address questions such asthe probability that a particular treatment is superiorto all the alternatives.

CONCLUSIONThe potential of meta-analysis for discovery wasdemonstrated in the recent discovery of ten new genesrelated to human growth which was published in thelatest issue of Nature Genetics43. There is no doubtthat meta-analyses have many positive attributes. Busyphysicians have difficulty keeping abreast of the hugevolume of medical literature, and some may notpossess the analytic skills to resolve the often non-definitive or conflicting findings. Meta-analysis providesan attractive solution to this problem. By examiningthe totality of data available about an issue, meta-analysis can identify inadequacies in existing data andpoint to areas of needed research, reduce the potentialfor erroneous findings occurring by chance, and more


accurately define the benefit and possible adverseeffects of management strategies44. In fact, bycombining smaller datasets, meta-analysis may establishunrecognized outcomes, may provide evidence ofstatistical significance where it was previously absent,or may eliminate any possible bias in individual studies.Although, there are weaknesses such as publicationbias, citation bias, etc; in spite of these, meta-analysishas a promising future for biomedical research anddevelopment.

REFERENCES1. Glass G.V. Primary, secondary and meta-analysis

of research. Educational Researcher. 1976; 5:3 –8.

2. Huque M.F. Experiences with meta-analysis inNDA submissions. Proceedings of theBiopharmaceutical Section of the AmericanStatistical Association 1988; 2:28-33.

3. National Library of Medicine. National Institutesof health. U.S. Department of Health and humanServices. Medical Subject headings – annotatedalphabetic list. Bethesda, MD: The Library, 1989;1-40

4. Egger M., Davey Smith G., Schneider M. andMinder C.E. Bias in meta-analysis detected by asimple, graphical test. BMJ 1997; 315:629-634.

5. Pearson K. Report on certain enteric feverinoculation statistics. British Medical Journal 1904;3: 1243–1246.

6. Yusuf S., Collins R. and Peto R. Intravenous andintracoronary fibrinolytic therapy in acutemyocardial infarction: overview of results onmortality, reinfarctions and side-effects. Eur HeartJ. 1985; 6; 556 – 585.

7. Hedges L.V., Olkin I. Statistical methods ofmeta-analysis. New York: Academic Press, 1985.

8. Moher D., Cook D.J., Eastwood S., Olkin I.,Rennie D. and Stroup D.F. Improving the qualityof reports of meta-analyses of randomizedcontrolled trials: the QUOROM statement.Quality of reporting the meta-analyses. Lancet.1999; 354:1896–1900.

9. Stroup D.F., Berlin J.A., Morton S.C., Olkin I.,Williamson G.D. and Rennie D. et al. Meta-analysisof observational studies in epidemiology: aproposal for reporting. Meta-analysis ofObservational Studies in Epidemiology(MOOSE) group. JAMA 2000; 283:2008 –2012.

10. Bailar J.C. The practice of meta-analysis. J ClinEpidemiol 1995; 48:149 –157.

11. Chung C.K., Burns P.B., Kim M.H. A PracticalGuide to Meta-Analysis. The Journal of HandSurgery. 2006. 31A; 10; 1671-1678.

12. National Library of Medicine. MEDLINE factsheet. Available at: http://www.nlm.nih.gov/pubs/factsheet/ medline.html.

13. Science and Technical Information Network(STN) database summary sheet EMBASE.Available at: http://www.cas.org/ONLINE/DBSS/embasess.html.

14. Egger M., Davey Smith G. Principles of andprocedures for systematic reviews. In: Egger M.,

Davey Smith G., Altman D.G., eds. Systematicreviews in health care: meta-analysis in context.2nd ed. London: BMJ Publishing Group, 2001:26.

15. CINAHL Information Systems. Available at:www.cinahl.com.

16. Scherer R.W., Dickersin K., Langenberg P. Fullpublication of results initially presented inabstracts. A meta-analysis [erratum: JAMA1994;272:1410]. JAMA 1994; 272:158–162.

17. United States Cochrane Center. Available at: http://www.cochrane.us/faq.htm.

18. Berman N.G., Parker R.A. Meta-analysis: neitherquick or easy. BMC Med Res Method. 2002; 2:10–18.

19. Sutton A.J., Abrams K.R., Jones D.R. Anillustrated guide to the methods of meta-analysis.J Eval Clin Pract 2001;7:135–148.

20. Moher D., Jadad A.R., Nichol G., Penman M.,Tugwell P. and Walsh S. Assessing the quality ofrandomized controlled trials: an annotatedbibliography of scales and checklists. Control ClinTrials 1995; 16:62–73.

21. Juni P., Witschi A., Bloch R. and Egger M. Thehazards of scoring the quality of clinical trials formeta-analysis. JAMA 1999; 282:1054 –1060.

22. Jadad A.R., Moher D., Klassen T.P. Guides forreading and interpreting systematic reviews: II.How did the authors find the studies and assesstheir quality? Arch Pediatr Adolesc Med 1998;152:812– 817.

23. Egger M., Smith GD, Phillips AN. Meta-analysis:principles and procedures. BMJ 1997; 315:1533–1537.

24. Hedges L.V., Olkin I. Statistical methods formeta-analysis. London: Academic Press, 1985.

25. Wolf F.M. Meta-analysis: quantitative methods forresearch synthesis. Newbury Park: SagePublications, 1986.

26. Fleiss J.L. The statistical basis of meta-analysis.Stat Methods Med Res 1993; 2:121–145.

27. Egger M., Davey Smith G., Schneider M. andMinder C.E. Bias in meta-analysis detected by asimple graphical test. BMJ 1997; 315:629–634.

28. Sutton A.J., Duval S.J., Tweedie R.L., AbramsKR and Jones DR. Empirical assessment of effectof publication bias on meta-analyses. BMJ2000;320:1574 –1577.

29. Begg C.B., Mazumdar M. Operatingcharacteristics of a rank correlation test forpublication bias. Biometrics 1994;50: 1088–1101.

30. Horwitz R.I. Large-scale randomized evidence:large, simple trials and overviews of trials:discussion. A clinician’s perspective on meta-analyses. J Clin Epidemiol 1995; 48:41– 44.

31. Davey Smith G., Egger M. Meta-analysis.Unresolved issues and future developments. BMJ1998; 316:221–225.

32. LeLorier J., Gregoire G, Benhaddad A, LapierreJ, Derderian F. Discrepancies between meta-analyses and subsequent large randomized,controlled trials. N Engl J Med 1997; 337: 536–542.


33. Sutton A.J., and Higgins P.T.J. Recentdevelopments in meta-analysis. Statist. Med. 2008;27; 625–650.

34. McShane L.M., Altman D.G., Sauerbrei W.,Taube S.E., Gion M., Clark G.M., for theStatistics Subcommittee of the NCI-EORTICWorking Group on Cancer Diagnosis. Reportingrecommendations for tumor marker prognosticstudies (REMARK). Journal of the NationalCancer Institute 2005; 97:1180–1184.

35. Wang M.C., Bushman B.J. Integrating Resultsthrough Meta-analytic Review Using SAS(R)Software. SAS Institute Inc.: Cary, NC, U.S.A.,1999.

36. Kuss O., Koch A. Meta-analysis macros for SAS.Computational Statistics & Data Analysis 1996;22:325–333.

37. Van Houwelingen H.C., Arends L.R., StijnenT. Advanced methods in meta-analysis:multivariate approach and meta-regression.Statistics in Medicine 2002; 21:589–624.

38. Sterne J.A.C., Bradburn M.J., Egger M. Meta-analysis in Stata. In Systematic Reviews in Health

Care, Egger M., Davey Smith G., Altman D.G.(eds). BMJ Publishing Group: London, 2001; 347–369.

39. Borenstein M., Hedges L., Higgins J., RothsteinH. Comprehensive Meta-analysis, Version 2.Biostat: Englewood, NJ, 2005.

40. RevMan Analyses [computer program]. TheCochrane Collaboration: Oxford, England.

41. Bax L., Yu L-M., Ikeda N., Tsuruta H. and MoonsK.G.M. Development and validation of MIX:comprehensive free software for meta-analysis ofcausal research data. BMC Medical ResearchMethodology 2006; 6:50.

42. Zamora J., Muriel A., Abraira V. Meta-DiSc forwindows: a software package for the meta-analysisof diagnostic tests. XIth Cochrane Colloquium,Barcelona, 2003 (Available http://www.hrc.es/investigacion/metadisc.html).

43. News-Medical.Net. Discovery of ten new genesrelated to human growth (available at www.news-medical.net accessed on 19/09/08).

44. DeMaria A.N., Meta-analysis. JACC. 2008.52; 3;237–238.


P values and confidence intervals are reported in almostall scientific writings and are used in interpreting resultsof statistical analysis. It is usual for medical researchersand other investigators to ask questions such as ‘Is theresult significant?’ or ‘what is the p value?’ Manyclinicians worry when they carry out statistical analysisand there are no significant results. This article describessome facts about the p value and confidence intervals.

The reporting of p values and confidence intervalsusually follows hypothesis testing or significance testing.Most scientific investigations involve the testing ofhypotheses. These are formal procedures for testingwhether findings from the investigations arecompatible with a so called null hypothesis. Hypothesesrefer to statements concerning the situation beinginvestigated which are usually stated as two mutuallyexclusive options; a null hypothesis and an alternativehypothesis. The null hypothesis is a statement of noassociation between variables or no difference in meansof groups while the alternative hypothesis states thatthere’s a difference or an association. The interests ofmedical researchers are varied and research questionsresult in statement of hypotheses. Examples of suchquestions are: Is there a significant difference inproportion of low birth weight babies delivered tomothers with single and multiple pregnancies? ; Is therea difference in effects of three antiretroviral drugs onreduction in viral load? ; Is there a correlation betweenbody mass index and systolic blood pressure; or Isthere a difference in reduction in blood sugar betweena standard hypoglycemic and a new drug. The nullhypothesis for the last study objective will be ‘There isno correlation between body mass index and systolicblood pressure’. The use and interpretation of p valuesand confidence intervals will now be discussed.

There are different definitions of the ‘p value’. Perhapsthe most popular definition is ‘The probability ofobtaining a value as extreme or more extreme as foundin the study if the null hypothesis were true’1. Put moresimply, it can be defined as the probability that theobserved result is due to chance alone2. An importantpoint to note in these definitions is the use of phrases‘found in the study’ and ‘observed result’. The p valueonly tells us whether what we have observed – whichis usually obvious- is statistically significant. This is animportant point to note. For example in a study whichexamined the difference in prevalence of low birthweight deliveries between singleton and multiple

pregnancies, the figures for the prevalence could havebeen 12.5% for multiple pregnancies and 3.6% forsingletons. All the statistical jargon about p value andconfidence intervals do not negate the fact that theproportion of low birth babies delivered to motherswith multiple pregnancies is higher than the proportionfor singletons. Hence from the initial descriptive statisticsused to summarize variables such as proportions andmeans we have an idea of the results of our study butthe statistical significance is what the p value helps to‘endorse’. The interpretation of p values is based onreference to a particular cut off for the probability orthe so called level of significance which is conventionallyset at 0.05. Hence p values less than this number aresignificant while those above are not.

The confidence interval gives the range of values withinwhich we are reasonably confident that the populationparameter lies2. The parameter here could be differencein means or proportions of two groups or it couldbe a measure of association between two variables.The most commonly reported interval is the 95%confidence interval. A way to think about the conceptof confidence intervals is to imagine that the studywas repeated about a thousand times. About 95% ofthe different possible results obtained will lie in thisinterval. Alternatively we can say that we are 95%confident that the true population value of what weare estimating in our study lies within the interval. Thecriterion for judging an interval as significant or notdepends on the presence of a null value. The null valuerefers to the value of the test statistic when the nullhypothesis is true. In the earlier example on differencein mean reduction in blood sugar between a standardhypoglycemic and a new drug, the null hypothesis isthat there is no difference in the mean reduction inblood sugar by the two drugs or that the difference inthe means between the two groups being comparedis zero. The null value here is zero and any intervalcomputed for the difference in the means whichincludes zero is not significant. Another set of studydesigns involves investigation of relationships wherebytwo variables typically an exposure (or risk factor) e.g.alcohol intake and an outcome such as liver diseaseare being related. The appropriate measure ofassociation between these variables is the odds ratioand the null value- that is when there is no relationshipbetween alcohol intake and liver diseaseis 1. Hence aconfidence interval including 1 will not be a significantinterval. A third scenario is if the variables being

P - VALUE AND CONFIDENCE INTERVALS – FACTS AND FARCESAdedokun B.O.

Correspondence:Adedokun B. O.LecturerDepartment of Epidemiology and Medical Statistics,College of Medicine,University of Ibadan.


investigated are both numeric, say the relationshipbetween maternal body mass index (in kg/m2) andbabies’ birth weight (in kg) where the measure ofassociation here is the correlation coefficient. The nullvalue here is zero and any interval for the correlationcoefficient between body mass index and birth weightincluding zero will not be significant.

As a guide to interpreting confidence intervals fordifference in means, when the lower and upper limitsare both positive or both negative - depending ondirection - then the difference is significant. Also forodds ratios when the upper and lower limits are bothdecimals or both whole numbers then we have asignificant result.

It is worthy of note that the p value and confidenceinterval are two equivalent methods of interpretingresults of a statistical analysis. Whether or not we havea significant result can be determined from the p valuebased on whether it is less than 5% or not; or theconfidence interval based on whether the null value


lies within the interval. The width of the confidenceinterval and the size of the p value are related, thenarrower the interval, the smaller the p value. Howeverthe confidence interval gives valuable information aboutthe likely magnitude of the effect being investigatedand the reliability of the estimate. Larger sample sizeswill give narrower and hence more reliable intervals.In conclusion confidence intervals should be thepreferred means of interpreting results of statisticalanalysis, because in addition to evaluating the role ofchance it reflects the degree of variability and thesample size.

References1. Bamgboye EA 2002. A Companion of Medical

Statistics, 1st edition. Folbam publishers, Ibadan

2. Kirkwood BR, Sterne JAC 2003 Essential medicalstatistics, 2nd edition Blackwell Publishing Ltd,London

INTRODUCTIONThe medical journals are replete with P values and testsof hypotheses. It is a common practice among medicalresearchers to quote whether the test of hypothesisthey carried out is significant or non-significant andmany researchers get very excited when they discovera “statistically significant” finding without reallyunderstanding what it means. Additionally, whilemedical journals are florid of statement such as:“statistical significant”, “unlikely due to chance”, “notsignificant,” “due to chance”, or notations such as, “P> 0.05”, “P < 0.05”, the decision on whether to decidea test of hypothesis is significant or not based on Pvalue has generated an intense debate amongstatisticians. It began among founders of statisticalinference more than 60 years ago1-3. One contributingfactor for this is that the medical literature shows astrong tendency to accentuate the positive findings;many researchers would like to report positive findingsbased on previously reported researches as “non-significant results should not take up” journal space4-7.

The idea of significance testing was introduced by R.A.Fisher, but over the past six decades its utility,understanding and interpretation has beenmisunderstood and generated so much scholarlywritings to remedy the situation3. Alongside thestatistical test of hypothesis is the P value, whichsimilarly, its meaning and interpretation has beenmisused. To delve well into the subject matter, a shorthistory of the evolution of statistical test of hypothesisis warranted to clear some misunderstanding.

A Brief History of P Value and SignificanceTestingSignificance testing evolved from the idea and practiceof the eminent statistician, R.A. Fisher in the 1930s.His idea is simple: suppose we found an associationbetween poverty level and malnutrition among childrenunder the age of five years. This is a finding, but couldit be a chance finding? Or perhaps we want to evaluate

whether a new nutrition therapy improves nutritionalstatus of malnourished children. We study a group ofmalnourished children treated with the new therapyand a comparable group treated with old nutritionaltherapy and find in the new therapy group animprovement of nutritional status by 2 units over theold therapy group. This finding will obviously, bewelcomed but it is also possible that this finding ispurely due to chance. Thus, Fisher saw P value as anindex measuring the strength of evidence against thenull hypothesis (in our examples, the hypothesis thatthere is no association between poverty level andmalnutrition or the new therapy does not improvenutritional status). To quantify the strength of evidenceagainst null hypothesis “he advocated P < 0.05 (5%significance) as a standard level for concluding thatthere is evidence against the hypothesis tested, thoughnot as an absolute rule’’ 8. Fisher did not stop there butgraded the strength of evidence against null hypothesis.He proposed “if P is between 0.1 and 0.9 there iscertainly no reason to suspect the hypothesis tested. Ifit’s below 0.02 it is strongly indicated that the hypothesisfails to account for the whole of the facts. We shallnot often be astray if we draw a conventional line at0.05’’ 9. Since Fisher made this statement over 60 yearsago, 0.05 cut-off point has been used by medicalresearchers worldwide and has become ritualistic touse 0.05 cut-off mark as if other cut-off points cannotbe used. Through the 1960s it was a standard practicein many fields to report P values with the star attachedto indicate P < 0.05 and two stars to indicate P <0.01. Occasionally three stars were used to indicate P< 0.001. While Fisher developed this practice ofquantifying the strength of evidence against nullhypothesis some eminent statisticians where notaccustomed to the subjective interpretation inherent inthe method 7. This led Jerzy Neyman and EgonPearson to propose a new approach which they called“Hypothesis tests”. They argued that there were two

P – VALUE, A TRUE TEST OF STATISTICAL SIGNIFICANCE? A CAUTIONARY NOTETukur Dahiru (MBBS), FMCPH, Dip. HSM (Israel)

All Correspondence to:Dr. Tukur DahiruMBBS, FMCPH, Dip HSM (Israel)DEPT OF COMMUNITY MEDICINEAHMADU BELLO UNIVERSITY,ZARIA, NIGERIA.Email:[email protected]

ABSTRACTWhile it’s not the intention of the founders of significance testingand hypothesis testing to have the two ideas intertwined as ifthey are complementary, the inconvenient marriage of the twopractices into one coherent, convenient, incontrovertible andmisinterpreted practice has dotted our standard statisticstextbooks and medical journals. This paper examine factorscontributing to this practice, traced the historical evolution ofthe Fisherian and Neyman-Pearsonian schools of hypothesistesting, exposed the fallacies and the uncommon ground andcommon grounds approach to the problem. Finally, it offersrecommendations on what is to be done to remedy the situation.

Dept. of Community Medicine, Ahmadu Bello University, Zaria, Nigeria.


types of error that could be made in interpreting theresults of an experiment as shown in Table 1.

Table 1.Errors associated with results of experiment.

The outcome of the hypothesis test is one of two: toreject one hypothesis and to accept the other. Adoptingthis practice exposes one to two types of errors: rejectnull hypothesis when it should be accepted (i.e., thetwo therapies differ when they are actually the same,also known as a false-positive result, a type I error oran alpha error) or accept null hypothesis when it shouldhave rejected (i.e. concluding that they are the samewhen in fact they differ, also known as a false-negativeresult, type II error or a beta error).

What does P value Mean?The P value is defined as the probability under theassumption of no effect or no difference (nullhypothesis), of obtaining a result equal to or moreextreme than what was actually observed. The P standsfor probability and measures how likely it is that anyobserved difference between groups is due to chance.Being a probability, P can take any value between 0and 1. Values close to 0 indicate that the observeddifference is unlikely to be due to chance, whereas a Pvalue close to 1 suggests no difference between thegroups other than due to chance. Thus, it is commonin medical journals to see adjectives such as “highlysignificant” or “very significant” after quoting the Pvalue depending on how close to zero the value is.

Before the advent of computers and statisticalsoftware, researchers depended on tabulated valuesof P to make decisions. This practice is now obsoleteand the use of exact P value is much preferred.Statistical software can give the exact P value and allowsappreciation of the range of values that P can take upbetween 0 and 1. Briefly, for example, weights of 18subjects were taken from a community to determineif their body weight is ideal (i.e. 100kg). Using student’st test, t turned out to be 3.76 at 17 degree of freedom.Comparing tstat with the tabulated values, t= 3.26 ismore than the critical value of 2.11 at p=0.05 andtherefore falls in the rejection zone. Thus we rejectnull hypothesis that ì = 100 and conclude that thedifference is significant. But using an SPSS (a statisticalsoftware), the following information came when thedata were entered, t = 3.758, P = 0.0016, meandifference = 12.78 and confidence intervals are 5.60and 19.95. Methodologists are now increasinglyrecommending that researchers should report theprecise P value. For example, P = 0.023 rather than P< 0.05 10. Further, to use P = 0.05 “is an anachronism.It was settled on when P values were hard to computeand so some specific values needed to be provided in

tables. Now calculating exact P values is easy (i.e., thecomputer does it) and so the investigator can report(P = 0.04) and leave it to the reader to (determine itssignificance)”11.

Hypothesis TestsA statistical test provides a mechanism for makingquantitative decisions about a process or processes.The purpose is to make inferences about populationparameter by analyzing differences between observedsample statistic and the results one expects to obtainif some underlying assumption is true. This comparisonmay be a single observed value versus somehypothesized quantity or it may be between two ormore related or unrelated groups. The choice ofstatistical test depends on the nature of the data andthe study design.

Neyman and Pearson proposed this process tocircumvent Fisher’s subjective practice of assessingstrength of evidence against the null effect. In its usualform, two hypotheses are put forward: a nullhypothesis (usually a statement of null effect) and analternative hypothesis (usually the opposite of nullhypothesis). Based on the outcome of the hypothesistest one hypothesis is rejected and accept the otherbased on a previously predetermined arbitrarybenchmark. This bench mark is designated the P value.However, one runs into making an error: one mayreject one hypothesis when in fact it should be acceptedand vise versa. There is type I error or á error (i.e.,there was no difference but really there was) and typeII error or â error (i.e., when there was differencewhen actually there was none). In its simple format,testing hypothesis involves the following steps:

1. Identify null and alternative hypotheses.2. Determine the appropriate test statistic and its

distribution under the assumption that the nullhypothesis is true.

3. Specify the significance level and determine thecorresponding critical value of the test statisticunder the assumption that null hypothesis is true.

4. Calculate the test statistic from the data.Having discussed P value and hypothesis testing,fallacies of hypothesis testing and P value are nowlooked into.

Fallacies of Hypothesis TestingIn a paper I submitted for publication in one of thewidely read medical journals in Nigeria, one of thereviewers commented on the age-sex distribution ofthe participants, “Is there any difference in sexdistribution, subject to chi square statistics”? Statistically,this question does not convey any query and this is oneof many instances among medical researchers(postgraduate supervisors alike) in which test ofhypothesis is quickly and spontaneously resorted towithout due consideration to its appropriateapplication. The aim of my research was to determinethe prevalence of diabetes mellitus in a ruralcommunity; it was not part of my objectives todetermine any association between sex and prevalence

Resultof experimentThe truth

Null hypothesis true Null hypothesis false

Reject null hypothesisAccept null hypothesis

Type I error rate(α) Power = 1-βCorrect decision Type II error rate (β)


of diabetes mellitus. To the inexperienced, thiscomment will definitely prompt conducting test ofhypothesis simply to satisfy the editor and reviewersuch that the article will sail through. However, theresults of such statistical tests becomes difficult tounderstand and interprete in the light of the data. (Theresult of study turned out that all those with elevatedfasting blood glucose are females). There are severalfallacies associated with hypothesis testing. Below is asmall list that will help avoid these fallacies.1. Failure to reject null hypothesis leads to its

acceptance. (No. When you fail to reject nullhypothesis it means there is insufficient evidenceto reject)

2. The use of á = 0.05 is a standard with an objectivebasis (No. á = 0.05 is merely a convention thatevolved from the practice of R.A. Fisher. There isno sharp distinction between “significant” and “notsignificant” results, only increasing strong evidenceagainst null hypothesis as P becomes smaller.(P=0.02 is stronger than P=0.04)

3. Small P value indicates large effects (No. P valuedoes not tell anything about size of an effect)

4. Statistical significance implies clinical importance.(No. Statistical significance says very little aboutthe clinical importance of relation. There is a biggulf of difference between statistical significanceand clinical significance. By statistical definition atá = 0.05, it means that 1 in 20 comparisons inwhich null hypothesis is true will result in P < 0.05!.Finally, with these and many fallacies of hypothesistesting, it is rather sad to read in journals howsignificance testing has become an insignificancetesting.

Fallacies of P ValueJust as test of hypothesis is associated with somefallacies so also is P value with common root causes, “It comes to be seen as natural that any finding worthits salt should have a P value less than 0.05 flashing likea divinely appointed stamp of approval’’12. Theinherent subjectivity of Fisher’s P value approach andthe subsequent poor understanding of this approachby the medical community could be the reason why Pvalue is associated with myriad of fallacies. Thirdly, Pvalue produced by researchers as mere ‘’passports topublication’’ aggravated the situation 13. We were earlieron awakened to the inadequacy of the P value in clinicaltrials by Feinstein 14,

“The method of making statistical decisions about‘significance’ creates one of the most devastating ironiesin modern biologic science. To avoid usual categoricaldata, a critical investigator will usually go to enormousefforts in mensuration. He will get special machinesand elaborate technologic devices to supplement hisold categorical statement with new measurements of‘continuous’ dimensional data. After all this work ingetting ‘continuous’ data, however, and after calculatingall the statistical tests of the data, the investigator thenmakes the final decision about his results on the basisof a completely arbitrary pair of dichotomouscategories. These categories, which are called

‘significant’ and ‘nonsignificant’, are usually demarcatedby a P value of either 0.05 or 0.01, chosen accordingto the capricious dictates of the statistician, the editor,the reviewer or the granting agency. If the leveldemanded for ‘significant’ is 0.05 or lower and the Pvalue that emerge is 0.06, the investigator may be readyto discard a well-designed, excellently conducted,thoughtfully analyzed, and scientifically importantexperiment because it failed to cross the Procrusteanboundary demanded for statistical approbation.

We should try to understand that Fisher wanted tohave an index of measurement that will help him todecide the strength of evidence against null effect. Butas it has been said earlier his idea was poorlyunderstood and criticized and led to Neyman andPearson to develop hypothesis testing in order to goround the problem. But, this is the result of theirattempt: “accept” or “reject” null hypothesis oralternatively “significant” or “non significant”. Theinadequacy of P value in decision making pervades allepidemiological study design. This head-or-tailapproach to test of hypothesis has pushed thestakeholders in the field (statistician, editor, revieweror granting agency) into an ever increasing confusionand difficulty. It is an accepted fact among statisticiansof the inadequacy of P value as a sole standardjudgment in the analysis of clinical trials 15. Just ashypothesis testing is not devoid of caveats so also Pvalues. Some of these are exposed below.

1. The threshold value, P < 0.05 is arbitrary. As hasbeen said earlier, it was the practice of Fisher toassign P the value of 0.05 as a measure of evidenceagainst null effect. One can make the “significanttest” more stringent by moving to 0.01 (1%) or lessstringent moving the borderline to 0.10 (10%).Dichotomizing P values into “significant” and “nonsignificant” one loses information the same way asdemarcating laboratory finding into normal” and“abnormal”, one may ask what is the differencebetween a fasting blood glucose of 25mmol/L and15mmol/L?

2. Statistically significant (P < 0.05) findings are assumedto result from real treatment effects ignoring thefact that 1 in 20 comparisons of effects in whichnull hypothesis is true will result in significant finding(P < 0.05). This problem is more serious whenseveral tests of hypothesis involving several variableswere carried without using the appropriate statisticaltest, e.g., ANOVA instead of repeated t-test.

3. Statistical significance result does not translate intoclinical importance. A large study can detect a small,clinically unimportant finding.

4. Chance is rarely the most important issue. Rememberthat when conducting a research a questionnaire isusually administered to participants. Thisquestionnaire in most instances collect large amountof information from several variables included inthe questionnaire. The manner in which the questions


where asked and manner they were answered areimportant sources of errors (systematic error) whichare difficult to measure.

What Influences P Value?Generally, these factors influence P value.1. Effect size. It is a usual research objective to detect a

difference between two drugs, procedures orprogrammes. Several statistics are employed tomeasure the magnitude of effect produced by theseinterventions. They range: r2, ç2, ù2, R2, Q2, Cohen’sd, and Hedge’s g. Two problems are encountered:the use of appropriate index for measuring the effectand secondly size of the effect. A 7kg or 10 mmHgdifference will have a lower P value (and more likelyto be significant) than a 2-kg or 4 mmHg difference.

2. Size of sample. The larger the sample the more likelya difference to be detected. Further, a 7 kg differencein a study with 500 participants will give a lower Pvalue than 7 kg difference observed in a studyinvolving 250 participants in each group.

3. Spread of the data. The spread of observations in adata set is measured commonly with standarddeviation. The bigger the standard deviation, themore the spread of observations and the lower theP value.

P Value and Statistical Significance: AnUncommon GroundBoth the Fisherian and Neyman-Pearson (N-P) schoolsdid not uphold the practice of stating, “P values ofless than 0.05 were regarded as statistically significant”or “P-value was 0.02 and therefore there wasstatistically significant difference.” These statements andmany similar statements have criss-crossed medicaljournals and standard textbooks of statistics andprovided an uncommon ground for marrying the twoschools. This marriage of inconvenience furtherdeepened the confusion and misunderstanding of theFisherian and Neyman-Pearson schools. Thecombination of Fisherian and N-P thoughts (asexemplified in the above statements) did not shed lighton correct interpretation of statistical test of hypothesisand p-value. The hybrid of the two schools as oftenread in medical journals and textbooks of statisticsmakes it as if the two schools were and are compatibleas a single coherent method of statistical inference 4, 23,

24. This confusion, perpetuated by medical journals,textbooks of statistics, reviewers and editors, havealmost made it impossible for research report to bepublished without statements or notations such as,“statistically significant” or “statistically insignificant”or “P<0.05” or “P>0.05”.Sterne, then asked “can weget rid of P-values? His answer was “practicalexperience says no-why? 21”

However, the next section, “P-value and confidenceinterval: a common ground” provides one of thepossible ways out of the seemingly insoluble problem.Goodman commented on P–value and confidence

interval approach in statistical inference and its abilityto solve the problem. “The few efforts to eliminate Pvalues from journals in favor of confidence intervalshave not generally been successful, indicating that theresearchers’ need for a measure of evidence remainsstrong and that they often feel lost without one”6.

P Value and Confidence Interval: A CommonGroundThus, so far this paper has examined the historicalevolution of ‘significance’ testing as was initiallyproposed by R.A. Fisher. Neyman and Pearson werenot accustomed to his subjective approach andtherefore proposed ‘hypothesis testing’ involving binaryoutcomes: “accept” or “reject” null hypothesis. This,as we saw did not “solve” the problem completely.Thus, a common ground was needed and thecombination of P value and confidence intervalsprovided the much needed common ground.

Before proceeding, we should briefly understand whatconfidence intervals (CIs) means having gone throughwhat p-values and hypothesis testing mean. Supposethat we have two diets A and B given to two groupsof malnourished children. An 8-kg increase in bodyweight was observed among children on diet A whilea 3-kg increase in body weights was observed on dietB. The effect in weight increase is therefore 5kg onaverage. But it is obvious that the increase might beless than 3kg and also more than 8kg, thus a range canbe represented and the chance associated with thisrange under the confidence intervals. Thus, for 95%confidence interval in this example will mean that ifthe study is repeated 100 times, 95 out of 100 thetimes, the CI contain the true increase in weight.Formally, 95% CI: “the interval computed from thesample data which when the study is repeated multipletimes would contain the true effect 95% of the time.”

In the 1980s, a number of British statisticians tried topromote the use of this common ground approachin presenting statistical analysis16, 17, 18. They encouragedthe combine presentation of P value and confidenceintervals. The use of confidence intervals in addressinghypothesis testing is one of the four popular methodsjournal editors and eminent statisticians have issuedstatements supporting its use 19. In line with this, theAmerican Psychological Association’s Board ofScientific Affairs commissioned a white paper, “TaskForce on Statistical Inference”. The Task Forcesuggested,

“When reporting inferential statistics (e.g. t - tests, F -tests, and chi-square) include information about theobtained ….. value of the test statistic, the degree offreedom, the probability of obtaining a value asextreme as or more extreme than the one obtained[i.e., the P value]…. Be sure to include sufficientdescriptive statistics [e.g. per-cell sample size, means,correlations, standard deviations]…. The reporting ofconfidence intervals [for estimates of parameters, forfunctions of parameter such as differences in means,and for effect sizes] can be an extremely effective way


of reporting results… because confidence intervalscombine information on location and precision andcan often be directly used to infer significance levels”20.

Jonathan Sterne and Davey Smith came up with theirsuggested guidelines for reporting statistical analysisas shown in the box21:

Box 1: Suggested guidance’s for the reporting ofresults of statistical analyses in medical journals.

1. The description of differences as statisticallysignificant is not acceptable.

2. Confidence intervals for the main results shouldalways be included, but 90% rather than 95% levelsshould be used. Confidence intervals should notbe used as a surrogate means of examiningsignificance at the conventional 5% level.Interpretation of confidence intervals should focuson the implication (clinical importance) of the rangeof values in the interval.

3. When there is a meaningful null hypothesis, thestrength of evidence against it should be indexedby the P value. The smaller the P value, the strongeris the evidence.

4. While it is impossible to reduce substantially theamount of data dredging that is carried out,authors should take a very skeptical view ofsubgroup analyses in clinical trials and observationalstudies. The strength of the evidence forinteraction-that effects really differ betweensubgroups – should always be presented. Claimsmade on the basis of subgroup findings shouldbe even more tempered than claims made aboutmain effects.

5. In observational studies it should be rememberedthat considerations of confounding and bias areat least as important as the issues discussed in thispaper.

Since the 1980s when British statisticians championedthe use of confidence intervals, journal after journalare issuing statements regarding its use. In an editorialin Clinical Chemistry, it read as follows,

“There is no question that a confidence interval forthe difference between two true (i.e., population) meansor proportions, based on the observed differencebetween sample estimate, provides more usefulinformation than a P value, no matter how exact, forthe probability that the true difference is zero. Theconfidence interval reflects the precision of the samplevalues in terms of their standard deviation and thesample size …..’’22

On the final note, it is important to know why it isstatistically superior to use P value and confidenceintervals rather than P value and hypothesis testing:

1. Confidence intervals emphasize the importance ofestimation over hypothesis testing. It is moreinformative to quote the magnitude of the size ofeffect rather than adopting the significant-nonsignificant hypothesis testing.

2. The width of the CIs provides a measure of thereliability or precision of the estimate.

3. Confidence intervals makes it far easier to determinewhether a finding has any substantive (e.g. clinical)importance, as opposed to statistical significance.

4. While statistical significant tests are vulnerable to typeI error, CIs are not.

5. Confidence intervals can be used as a significancetest. The simple rule is that if 95% CIs does notinclude the null value (usually zero for difference inmeans and proportions; one for relative risk andodds ratio) null hypothesis is rejected at 0.05 levels.

6. Finally, the use of CIs promotes cumulativeknowledge development by obligating researchersto think meta-analytically about estimation,replication and comparing intervals across studies25.For example, in a meta-analysis of trials dealingwith intravenous nitrates in acute myocardialinfraction found reduction in mortality ofsomewhere between one quarter and two-thirds.Meanwhile previous six trials 26 showed conflictingresults: some trials revealed that it was dangerousto give intravenous nitrates while others revealedthat it actually reduced mortality. For the six trials,the odds ratio, 95% CIs and P-values are: OR =0.33 (CI = 0.09, 1.13, P = 0.08); OR = 0.24 (CI =0.08, 0.74, P = 0.01); OR = 0.83(CI = 0.33, 2.12, P= 0.07); OR = 2.04 (CI = 0.39, 10.71, P = 0.04);OR = 0.58 (CI = 0.19. 1.65; P = 0.29) and OR =0.48 (CI = 0.28, 0.82; P = 0.007). The first, third,fourth and fifth studies appear harmful; while thesecond and the sixth appear useful (in reducingmortality).

What is to be done?While it is possible to make a change and improve onthe practice, however, as Cohen warns, “Don’t lookfor a magic alternative … It does not exist” 27.

1. The foundation for change in this practice shouldbe laid in the foundation of teaching statistics:classroom. The curriculum and class room teachingshould clearly differentiate between the two schools.Historical evolution should be clearly explained soalso meaning of “statistical significance”. Theclassroom teaching of the correct concepts shouldbegin at undergraduate and move up to graduateclassroom instruction, even if it means this teachingwould be at introductory level.

2. We should promote and encourage the use ofconfidence intervals around sample statistics andeffect sizes. This duty lies in the hands of statistics


teachers, medical journal editors, reviewers and anygranting agency.

3. Generally, researchers, preparing on a study areencouraged to consult a statistician at the initial stageof their study to avoid misinterpreting the P valueespecially if they are using statistical software fortheir data analysis.

REFERENCES1. Goodman SN. P value hypothesis and likelihood:

implications for epidemiology of a neglectedhistorical debate. Amer Journ Epidemiol 1993;137: 485-96.

2. Lehmann El. The Fisher, Neyman-Pearsontheories of testing hypothesis: one theory or two?Journ Amer Stat Assoc 1993; 88:1242 – 9

3. Goodman SN. Toward evidence-based medicalstatistics: the P-value fallacy. Ann intern Med 1999;130:995-1004.

4. Berking JA, Begg CB, Louis JA. An assessmentof publication bias using a sample of publishedclinical trials. Journ Amer Stat Assoc. 1989; 84:38-92.

5. Easterbrook PJ, Berking JA, Gopalan Rand Mathews DR. Publication bias in clinicalresearch. Lancet 1991. 337:887 – 892

6. Dickerson K, Min YI and Meinert CL. Factorsinfluencing publication of research results: follow-up of applications submitted to two institutionalreview boards. Journ Amer Med Assoc 1992;263:374 -378

7. Stern JAC, Smith GD. Sifting the evidence-whatis wrong with significance tests? Br Med Journ2001; 322:226-231.

8. Fisher RA. Statistical methods for researchworkers. London Oliver and Boyd, 1950:80.

9. Bakan D. The test of significance in psychologicalresearch. Psychology Bulletin 1960, 66: 423-437

10. Greenwald AG, Richard G, Richard 3H andDonal G. Effect sizes and P value: what shouldbe reported and what should be replicated?Psychophysiology 1996:33:175 – 183

11. Wainer H, Robinson DH. Shaping of the practiceof null hypothesis significance testing. EducationalResearcher 2002:32:22 -30.

12. Jekel JF. Should we stop using the P value indescriptive studies? Paediatrics 1977; 60:124 – 25.

13. Mainland D. Statistical ritual in clinical journals:is there a cure?-I. Br Med Journ 1984; 288:841-43

14. Feinstein AR. Clinical biostatistics. St Louis: CVMosby; 1977: 66-70

15. Diamond GA, Forrester JS. Clinical trials andstatistical verdicts: probable ground for appeal.Ann Intern Med 1983; 98:385-394

16. Altman DG, Gore SM, Gardner MJ and PocockSJ. Statistical guidelines for contributors to medicaljournals. Br Med Journ 1983; 286: 1989 – 93.

17. Gardner MJ, Altman DG. Confidence intervalsrather than P-value: estimation rather thanhypothesis testing. Br Med Journ1986; 292:746-750.

18. Gardner MJ, Altman DG. Statistics withconfidence – confidence intervals and statisticalguidance. BMJ Books, London 1989.

19. Iacobucci D. On P value. Editorial. Journal ConsRes. 2005

20. American Psychological Association. PublicationManual, 5th Ed, Washington, DC: AmericanPsychological Association.

21. Sterne JAC, Davey Smith. Suggested guidelinesfor the reporting of results of statistical analysis inmedical journals. Br Med Journ 2001; 322:226-231.

22. Eugene KH. On P value and confidence intervals(Why can’t we P with more confidence). Editorial.Chin Chem 1993; 39(6):927 – 928.

23. Royal R. Statistical Evidence: a likelihood primer.Monographs on statistics and applied probability#71. London: Chapman and Hall; 1997.

24. Gigerenzer G, Swijtink Z, Porter T, Daston L,Beatty J and Kruger L. The Empire of Chance.Cambridge, UK: Cambridge Univ Pr; 1989.

25. Thomas B. What future quantitative socialscience research could look like: confidenceinterval for effect sizes. Educ Research 2002;31:25-32.

26. Yusuf S, Collins R, Mac Mahon, Peto R. Effectof intravenous nitrates on mortality in acutemyocardial infarction: an overview ofrandomized trials. Lancet 1988; 1:1088-92

27. Cohen J. Things I Have Learned (So Far).Amer Psychologist 1990; 45:997-1003.


Abbreviations: ACE-inhibitors: angiotensin-convertingenzyme inhibitors, AF: atrial fibrillation, CT-head:computerized tomography scan of the head, DCGH: DerbyCity General Hospital, DRI: Derbyshire Royal Infirmary,HMG-CoA: 3-hydroxy-3-methylglutaryl-coenzyme-A, MRI:magnetic resonance imaging, NHS: National Health Service,RCP: Royal College of Physicians, TIA: transient ischemicattack.

INTRODUCTIONCerebrovascular disease refers to focal neurologicaldeficit of vascular origin. This spectrum of clinicaldisorders ranges from a transient ischemic attack (TIA)which lasts less than 24 hours to a stroke which lastslonger than 24 hours. Approximately 150 000 casesof stroke are diagnosed annually in the population ofEngland and Wales with 65% of cases occurring inthose over the age of 751. Cerebrovascular diseaseremains a leading cause of morbidity and mortality

worldwide particularly among the elderly 2.Furthermore, strokes remain one of the mostcommon reasons for the acute admission of elderlypatients3. In a large population-based study ofapproximately 20 000 people in Copenhagen, age, sex,household income, smoking, systolic blood pressure,diabetes mellitus, serum cholesterol, ischaemic heartdisease and atrial fibrillation were identified assignificant risk factors for stroke4. In multi-racialcommunities, cerebrovascular disease has a higherincidence in the black population5. The largestlongitudinal study reported in the UK suggests thatdespite a reduction in the age and gender-specificincidence rates of stroke over the past 25 years anddisability from stroke, the overall disease burden hasbeen stable due to the increasing elderly population6.This suggests that given the further projected increasein the elderly population over the forthcoming years,more effort needs to be devoted to the effective

RETROSPECTIVE AUDIT OF THE ACUTE MANAGEMENT OF STROKE IN TWODISTRICT GENERAL HOSPITALS IN THE UK.

O.O. Faluyi, PhD, J.A. Omodara, MSc, K.H. Tay, M.B.B.S. and K. Muhiddin, PhD, FRCP.Medical Directorate, Derby NHS Hospital Foundation Trust, Derby, UK

Correspondence:Dr. O.O. Faluyi,Edinburgh Cancer Centre,Western General Hospital,Edinburgh EH4 2XU, UKe-mail: [email protected]

ABSTRACTBackground: There is some evidence to suggest that the standard of acutemedical care provided to patients with cerebrovascular disease is a majordeterminant of the eventual outcome. Consequently, the Royal College ofPhysicians (RCP) of London issues periodic guidelines to assist healthcareproviders in the management of patients presenting with stroke.Objective: An audit of the acute management of stroke in two hospitalsbelonging to the same health care trust in the UK.Method: Retrospective review of 98 randomly selected case-notes of patientsmanaged for cerebrovascular disease in two acute hospitals in the UK betweenApril and June 2004. The pertinent guidelines of RCP (London) are highlightedwhile audit targets were set at 70%.Results: 84% of patients presenting with cerebrovascular disease had a strokerather than a TIA, anterior circulation strokes were commonest. All patientswith stroke were admitted while those with TIAs were discharged on the sameday but most patients with TIA were not followed up by Stroke specialists.Most CT-imaging of the head was done after 24 hours delaying thecommencement of anti-platelets for patients with ischaemic stroke orneurosurgical referral for haemorrhagic stroke. Furthermore, there was a lowrate of referral for carotid ultrasound in patients with anterior circulationstrokes. Anti-platelets and statins were commenced for most patients withischaemic stroke while diabetes was well controlled in most of them. However,ACE-inhibitors and diuretics such as indapamide were under-utilized forsecondary prevention in such patients. Warfarin anti-coagulation was under-utilized in patients with ischaemic stroke who had underlying chronic atrialfibrillation. While there was significant multi-disciplinary team input,dysphagia and physiotherapy assessments were delayed. Similarly, occupationaltherapy input and psychological assesment were omitted from the care ofmost patients.Conclusion: Hospital service provision for the management of cerebrovasculardisease needs to provide appropriate specialist follow up for patients with TIA,prompt radiological imaging and multi-disciplinary team input for patientswith stroke. Furthermore, physicians need to utilize appropriate anti-hypertensives and anti-coagulation more frequently in the secondary preventionof stroke.

Keywords: Stroke, anti-platelets, anti-coagulation, carotid stenosis, secondary prevention


ORIGINAL ARTICLES

management of cerebrovascular disease to reducedisability and mortality from this condition.

In the management of stroke, treatment given in theacute phase and the secondary preventive measuresinstituted appear to be the major determinants ofoutcome7. Therefore, the Royal College of Physicians(RCP) of London issues periodic guidelines to assistphysicians with the management of stroke8. Some ofthese guidelines specify standards of carerecommended in the acute phase. This audit wascarried out to determine if the management ofpatients presenting at two District General Hospitalswith emergency services in the UK in 2004 conformedto the RCP guidelines.

MATERIALS AND METHODS50 patients managed for cerebrovascular disease at eachof the two National Health Service (NHS) AcuteHospitals in Derby - Derby City General Hospital(DCGH) and Derbyshire Royal Infirmary (DRI)between April 1 and June 30, 2004 were randomlyselected by a computer of the records department ofthe hospitals. Case notes for 48 patients were retrievedat DCGH while the case notes for 50 patients wereretrieved at DRI. Of the 98 patients, 46 were male, 52were female (Age range: 46 to 98). Clinical syndromesof stroke were determined based on a physicalexamination on admission and CT-head findings. Anaudit pro-forma was designed based on the standardsof care recommended by the RCP (London) in 2004and used for data collection from the case notes by 4doctors in postgraduate training in medicine (2 fromeach hospital). The data was collated and analyzed at ajoint meeting. Audit standards were set at 70%. Wherepatients died, they were excluded from the analysis.However, as the guidelines and good clinical practicespecify that the steps in the management of patientswith stroke should be well documented, where therewas no documentation to suggest that the stipulatedstandard of care had been met, the case was treatedas if the standard of care had not been met. Ethicalapproval for the project was sought from the hospitalaudit committee and the results have been submittedto the same committee.

Summary of RCP Guidelines Relevant to thisStudyAcute Management(1) All patients presenting with a TIA should be

followed up by Stroke Specialists (defined by theRCP as a Neurologist or a Geriatric Physician) ina Neurovascular Clinic.

(2) All patients presenting with stroke should beadmitted and have a complete neurologicalexamination.

(3) All patients presenting with stroke should have acomputerized tomography scan (CT-scan) of thehead within 24 hours.

(4) All patients with ischaemic stroke (diagnosed onCT-head) should have additional anti-plateletscommenced in the absence of contra-indications.

(5) No patient should receive anti-platelets prior to aCT-head confirmed diagnosis of ischaemicstroke.

(6) All patients with suspected intracranialhaemorrhage on CT-head should have an urgentneurosurgical opinion sought.

(7) All diabetic patients with stroke should have theirblood glucose levels maintained at less than - 10mmoll-1 while on admission.

Secondary Prevention(1) All patients with ischaemic stroke should receive

HMG-CoA reductase inhibitors (statins) as soonas possible (where serum cholesterol > 3.5 mmoll-1).

(2) All patients with ischaemic stroke should have pre-admission anti-hypertensives continued onadmission.

(3) In the absence of contra-indications, within twoweeks of an ischaemic stroke, all patients shouldbe commenced on an angiotensin-convertingenzyme inhibitor (ACE-inhibitor) with or withouta thiazide or indapamide.

(4) In the absence of contra-indications, within twoweeks of an ischaemic stroke, all patients withchronic atrial fibrillation should be commencedon warfarin anti-coagulation.

(5) All patients with anterior circulation ischaemicstroke should have a Doppler ultrasound scan ofthe internal carotid arteries and end-arterectomywithin two weeks is recommended in the presenceof critical internal carotid artery stenosis.

Multi-disciplinary Care(1) All patients presenting with stroke should be

assessed for dysphagia with a 50 ml water swallowtest on admission by an appropriately trainedprofessional.

(2) All patients presenting with stroke should have adysphasia assessment by a Speech and Languagetherapist as soon as possible.

(3) All patients admitted with stroke should have aweekly multi-disciplinary team review while onadmission.

(4) All patients presenting with stroke should have aphysiotherapy assessment within 72 hours ofadmission.

(5) All patients presenting with stroke should have anoccupational therapy assessment prior to dischargeto assess for adaptations or equipment requiredat home for safety after discharge.

(6) All patients should have a psychological assessmentwithin a month of a stroke to check for symptomsof anxiety or depression which could otherwiseremain undetected.

RESULTSFollow-up of Patients Presenting with TIA82 of 98 patients (83.7%) were diagnosed with strokewhile the rest were diagnosed with TIA. Of patientspresenting with TIA, only 38% were followed up byStroke Specialists after discharge from the hospital


upon resolution of their symptoms. The others werefollowed up by other medical specialists.

The rest of the audit results are on patients presentingwith stroke.

Acute Management of StrokeOf the 82 patients with stroke, 75 (91.5%) werediagnosed to have ischaemic stroke followingexclusion of haemorrhage on a CT-head scan (Table1). Cases of haemorrhagic stroke accounted for only8.3% of cases with the majority of those being due toanterior fossa haemorrhage (Table 1). Most of thepatients with stroke had partial anterior circulations

strokes which accounted for 53.7% of the totalnumber of cases (Table 1). Conformity with RCPguidelines on some relevant aspects of investigationand acute care are summarized in Table 2. With regardsto admissions and the timing of CT-head scans, theentire population of patients presenting with stroke (n= 82) was the reference population. With regards tocommencement of anti-platelets, the total number ofpatients in which haemorrhage was excluded on a CT-head scan was the reference population (n = 75). Withregards to the management of haemorrhagic stroke,the number of patients diagnosed with haemorrhagicstroke on a CT-head scan was the reference population(n = 7). All the patients presenting with stroke were


Clinical syndrome (N = 82) Proportion (%)

Partial anterior circulation stroke 44 (53.7)

Total anterior circulation stroke 11 (13.4)

Posterior circulation stroke 11 (13.4)

Lacunar stroke 10 (12.2)

Anterior fossa haemorrhage 6 (7.1)

Posterior fossa haemorrhage 1 (1.2)

Subarachnoid haemorrhage 0 (0)

Table 1: Clinical Syndromes - The proportion of patients presenting with stroke at DCGH and DRI whohad the respective clinical syndromes of stroke. Total number of patients admitted with stroke: N=82 (1patient had combined partial anterior circulation and lacunar features on CT-head).

A u d it P a ra m e te r P rop o rtio n (% ) N u m b er of p a tie n ts re c ord s tha t w e re a v a ila b lefor a s se ssm e n t (R )

A dm iss ion to h os pita l 8 2 /8 2 (1 0 0 ) R = 8 2 (A ll p a tien ts pre sen ting w ith s tro k e)

C T-h ead w ith in 2 4 ho u rs 3 8 /7 9 (4 8 .1 ) R = 7 9 (3 d ied w ith in 2 4 hou rs b efore the y cou ldhav e h ead s cans and w ere e xcluded )

A nt i-p la t e let the rapy fo risch aem ic s t rok e w ith in 2 4hour s

2 9 /7 5 (3 8 .6) R = 7 1 (dr ug c hart s m iss ing in 4 { in cluded inana lys is} , haem orr hag ic st rok e in 7 w ho w ereex cluded)

A nt i-p la t e let the rapy fo risch aem ic s t rok e co m m ence dafte r C T-h ead

5 9 /7 5 (7 8 .6) R = 6 8 (T im e of head sca n no t d ocum ented in 3and d rug chart s m iss ing in 4 { included in ana lys is} ,hem orrh agic s t rok e in 7 w ho w ere ex cluded )

Ur gen t n eu ros u rg ica l opin io nfor p atien ts w ithhaem orr hagic s t rok e o n C T-head

7 /7 ( 1 0 0 ) R = 7 (T he no o f pat ient s w ith haem orrh agic s tr ok eon CT -head )

Table 2: Emergency Care - Some audit parameters of emergency care for patients presenting with stroke atDCGH and DRI (Audit targets = 70%). Total number of patients presenting with stroke: N=82.

admitted and a neurosurgical opinion was sought forall the patients with haemorrhagic stroke diagnosedon a CT-head scan. In contrast, less than 70% ofpatients had a CT-head scan done within 24 hours oranti-platelet therapy commenced for ischaemic strokewithin 24 hours (Table 2). Notably, more than 70%of patients had anti-platelets commenced only afterthe exclusion of intra-cranial haemorrhage on headCT-scan (Table 2).

Secondary Prevention of StrokeThe group of patients with ischaemic stroke (n = 75)were the target of medical measures at secondaryprevention of stroke. In such patients, with regards tothe continuation of anti-hypertensives on admission,the control of blood glucose levels in diabetic patientsand treatment with simvastatin, RCP guidelines wereadhered to in more than 70% of patients (Table 3). Incontrast, as regards the utilization of thiazides/indapamide and ACE-inhibitors in secondaryprevention of ischaemic stroke, RCP guidelines wereadhered to in less than 70% of patients (Table 3).Similarly, warfarin prophylaxis against embolism inchronic atrial fibrillation and carotid dopplerultrasound scans to detect critical internal carotid arterystenosis in anterior circulation strokes were done inless than 70% of patients (Table 3).

Multi-disciplinary Care of Patients with StrokeRCP guidelines were adhered to in less than 70% ofthe patients in any of the areas of multi-disciplinarycare assessed (Table 4). With regards to dysphagia,physiotherapy assessments and weekly MDT meetings,all the patients presenting with stroke except for thosethat died within a week (n = 79) were the referencepopulation (3 patients died within a week and they alldied within 24 hours of admission). The referencepopulation for a speech and language therapist’sassessment (n = 43) were all the patients with strokeand dysphasia. The patients with stroke that were aliveon discharge (n = 68) were the reference groupexpected to have had an occupational therapist andpsychological assessment. However, 2 patients diedafter they had been assessed by an occupationaltherapist but prior to discharge from hospital.

DISCUSSIONThis audit project was designed to assess theperformance of services for the management ofstroke in two hospitals which jointly provide emergencyservices for a community of half a million peopleand is probably representative of hospitals of a similarnature in the UK. For simplicity, we have presentedthe results from both hospitals together as they servethe same population and belong to the same National


Audit Parameter Proportion (%) Number of patients records that were availablefor assessment (R)

Continuation of pre-existinganti-hypertensives onadmission

62/82 (75.6) R=82 (No of patients presenting with stroke)

Control of diabetes mellitus(glucose < 10 mmoll-1 ) onadmission

10/14 (71.7) R=14 ( No of diabetic patients)

On thiazide within twoweeks or after discharge

13/62 (21.0) R=62 (75 patients presented with ischaemic stroke,13 of them died in-hospital and were excluded)

On ACEI within two weeksor after discharge (patientson other anti-hypertensiveswere excluded)

29/62 (46.8) R=62 (75 patients presented with ischaemic stroke,13 of them died in-hospital and were excluded)

On simvastatin within twoweeks or after discharge

46/62 (74.2) R=75 (No of patients with ischaemic stroke, 13 ofthem died in-hospital and were excluded)

On warfarin for chronicatrial fibrillation (AF)

4/20 (20.0) R=20 (6 patients with AF had contra-indications towarfarin documented and were excluded)

Carotid duplex scan within 2weeks for ischaemic anteriorcirculation stroke

7/54 (13.2) R=54 (3 patients who died within 24 hours wereexcluded)

Table 3: Secondary Prevention - The proportion of patients presenting with stroke at DCGH and DRIwho had various secondary preventive measures instituted within two weeks or prior to discharge (Audittargets = 70%). Total number of patients presenting with stroke: N = 82.

Health Service (NHS) Trust with common patientmanagement approaches. From these results, it can beseen that the vast majority of patients admitted withcerebrovascular disease presented with stroke ratherthan a TIA. This may be a reflection of the fact thatindividuals with short lasting TIAs rather than strokesare less likely to present as acute admissions. Majorityof patients with TIAs were not followed up by Strokespecialists as suggested by the guidelines but by otherphysicians. This situation arose primarily due to thedischarge of such patients under the care of the on-call consultant rather than a stroke specialist. Such atrend may affect the quality of care received by suchpatients as not all other physicians are familiar withcurrent concepts in the management of stroke. Over90% of strokes were ischaemic in nature with partialanterior circulation strokes being the commonest. Thisobservations is similar to that of a larger Danish studyof over 20 000 patients where 89.4% of strokes wereischaemic9. Incidentally, no case of subarachnoidhaemorrhage was observed in this study.

It was satisfying to note that all patients with strokewere admitted and that anti-platelets were startedpromptly after the CT-diagnosis of ischaemic strokein over 70% of patients. The International Stroke Trialwhich recruited approximately 20 000 patientsdemonstrated the beneficial effect of aspirinadministration within 48 hours in reducing morbidityand mortality from stroke10. Similarly, all the patientswith a CT-diagnosis of haemorrhagic stroke werediscussed with neurosurgeons and anti-platelets werenot usually started prior to a CT scan of the head.However, one major shortfall in these hospitals was adelay in CT-head scans with most patients having their

scans done over 24 hours after admission. It appearsas if delayed scans delayed the commencement ofanti-platelets in most patients. At the present time,selected patients with ischaemic stroke can be treatedwith thrombolysis in specialized centres11. However,this modality of treatment was not in use in the Trustat the time of the audit. The introduction ofthrombolysis in the routine management of stroke inthese hospitals would require even more prompt headimaging as a CT or MRI within 3 hours is usuallyrequired12. Therefore, the timing of head CT-scans isdefinitely an area for improvement in strokemanagement in these hospitals. A case can be madefor starting anti-platelets prior to a CT of the head onthe grounds that the vast majority of strokes areischaemic. However, this could be detrimental to thepatients with haemorrhagic stroke as it could potentiallyaggravate intra-cranial haemorrhage. A careful analysisof the potential benefits and risks of such an approachis therefore necessary but prompt head CT-scanningwould make this unnecessary.

As regards the secondary prevention of stroke, it isnoteworthy that over 70% of patients on anti-hypertensives had these continued, while diabetesmellitus which was present in 17% was well controlledon admission and statin therapy was judiciouslyinitiated in ischaemic stroke in over 70% of patients.On the other hand, despite the results of thePROGRESS study suggesting their benefit in thesecondary prevention of ischaemic stroke13, ACE-inhibitors and diuretics such as indapamide were under-utilized in secondary prevention. Furthermore, anti-coagulation with warfarin in patients with chronic atrialfibrillation which reduces mortality two and a half fold


Audit parameter Proportion (%) Reference subgroup of patients(R)

Dysphagia assessment (bythe water swallow test) onadmission

16/79 (20.3) R=79 (3 were too drowsy to beassessed on admission and wereexcluded)

Physictherapy assessment(within 72 hours)

28/79 (35.4) R=79 (3 died within 72 hours andwere excluded)

Dysphasia assesment (byspeech and languagetherapist) during admission

25/43 (58.1) R=43 (patients with stroke anddocumented dysphasia)

Occupational Therapistassessment prior to discharge

36/70 (51.4) R=70 (12 died prior to assessmentduring admission and wereexcluded)

Weekly multi-disciplinaryteam review on admission

52/79 (65.8) R=79 (3 died prior to assessment ina weekly MDT and were excluded)

Psychological assessmentwithin a month or scheduledafter discharge

3/68 (4.4) R=68 (14 died prior to dischargeand were excluded)

Table 4: Multi-disciplinary Care - Audit of different aspects of the multi-disciplinary care of patientspresenting with stroke at DCGH and DRI. Total number of patients presenting with stroke: N=82.


within a year9 appears to be largely over looked.Admittedly, for a few patients there were genuineconcerns about the risk of falls and intracranialhaemorrhage on warfarin, but even after the exclusionof these, anti-coagulation appears not to have beencontemplated in many patients. This observationbecomes more worrying when we consider that 32%of the patients presenting with stroke had chronic atrialfibrillation.

Carotid endarterectomies in patients with severeinternal carotid artery stenosis (> 70%) reduces thefrequency of recurrence of anterior circulation strokesfor over 8 years after the initial event14. Very few patientsmanaged in these hospitals had duplex scans within 2weeks of their stroke which suggests that urgentendarterectomies were not a priority in themanagement plan. This observation would also beworrying, given that anterior circulation strokes werethe most common clinical syndrome and severe carotidartery stenosis was a potentially reversible factor whichcould predispose to further cereborvascular events.However, there may be genuine concerns over thesurgical risk of some patients particularly the olderones15, which could discourage physicians fromconsidering endarterectomies. This was not explicitlystated for any of the patients.

Dysphagia is common among patients with stroke withprevalence ranging from 37 to 78% estimated bydifferent techniques16. Prompt assessment of dysphagiawould prevent aspiration and malnutrition. The 50mlwater swallow test is accepted as one of the mostreliable ways of assessing dysphagia but was under-utilized in these hospitals17. In a similar manner, theproportion of our patients assessed properly fordysphasia was below audit targets. Furthermore, audittargets were not met for physiotherapist assessmentwithin 72 hours which may affect optimization ofresidual neurological function. Meta-analyses suggeststhat patients who have community occupationaltherapy have improved personal and extended activitiesof daily living18 after discharge. These hospitals utilizedoccupational therapists inadequately in the rehabilitationof stroke patients. Despite the fact that patients withpost-stroke depression have a significantly increasedmortality19, less than 5% of our patients had apsychological assessment within a month of having astroke. This raised the possibility that psychologicalneeds were not adequately catered for. Given that theproportion of patients with stroke who were reviewedat weekly multi-disciplinary team meetings nearly hitaudit targets, it appears as if the major problem withmulti-disciplinary input was prompt involvement ofrelevant professionals to assess dysphagia,physiotherapy needs, dysphasia and psychological needs.Stroke Units provide a unique opportunity to centralizethe care for stroke patients and may potentially helpfocus appropriately trained personnel and resourceson the care of patients with stroke. To this effect, astroke unit has been recently introduced in one of thehospitals in which these patients were managed. Furtherstudies within these hospitals and others would assess

how the introduction of stroke units and reflectionon findings such as those reported here will influencethe management of patients presenting with featuressuggestive of cerebrovascular disease.

CONCLUSIONStroke management services should increase thepriority of head CT scans and carotid dopplerultrasounds which could significantly influencemanagement. Furthermore, physicians should beencouraged to consider ACEI, diuretics such asindapamide and warfarin in the secondary preventionof ischaemic stroke where indicated. Finally, promptmulti-disciplinary involvement in the assessment andmanagement of stroke patients right from the timeof admission should be encouraged.

REFERENCES1 Website of the Office for National Statistic of

England and Wales: http://www.statistics.gov.uk/articles/hsq/HSQ12stroke.pdf (last accessed 12February, 2008).

2 Kleindorfer C., Broderick J, Khuory J, et al., (2006)The unchanging incidence and case-fatality ofstroke in the 1990s: a population-based study.Stroke, 37(10): 2473-8.

3 Ogun S.A., Ojini F.I., Ogungbo B., et al., (2005)Stroke in south-west Nigeria: a 10-year review.Stroke, 36(6): 1120-2.

4 Boysen G., Nyobe J., Appleyard M., et al., (1988)Stroke incidence and risk factors in Copenhagen,Denmark. Stroke, 19(11): 1345-53.

5 Kissela B., Schneider A., Kleindorfer D., et al.,(2004) Stroke in a biracial population: the excessburden of stroke among blacks. Stroke, 35(2): 426-431.

6 Rothwell P.M., Coull A.J., Giles M.F., et. al., (2004)Change in stroke incidence, mortality, case-fatality,severity, and risk factors in Oxofordshire, UK from1981 to 2004 (Oxford Vascular Study). Lancet,363(9425): 1925-1933.

7 Zweifler R.M. (2003) Management of acutestroke. South Med Journal, 96(4): 380-385.

8 The website of Royal College of Physiciansof London: http://www.rcplondon.ac.uk/pubs/books/stroke/stroke_guidelines_2ed.pdf (Lastaccessed 21.03.07.).

9 Andersen K.K. and Olsen T.S. (2007) Reducedpost-stroke mortality in patients with stroke andatrial fibrillation treated with anticoagulants: resultsfrom a Danish quality-control registry of 22 179patients with ischemic stroke. Stroke, 38(2): 259-263.


10 International Stroke Trial CollaborativeGroup. (1997) The International Stroke Trial (IST):a randomised trial of aspirin, subcutaneousheparin, both, or neither among 19435 patientswith acute ischaemic stroke. Lancet, 349(9065):1569 – 1581.

11 Khaja A.M. and Grotta J.C. (2007) Establishedtreatments for acute ischemic stroke. Lancet,369(9558): 319-330.

12 Szabo K., Lanczik O. and Hennerici M.G. (2005)Vascular diagnosis and acute stroke: what, whenand why not ? Cerebrovascular Disease: Suppl 2: 11-18.

13 PROGRESS Collaborative Group (2001)Randomised trial of a perindopril-based bloodpressure-lowering regimen among 6 105individuals with previous stroke or transientischemic attack. Lancet, 358(9287): 10331-1041.

14 Barnett H.J., Taylor D.W., Eliasziw M., et al.,(1998) Benefit of carotid endarterectomy inpatients with symptomatic moderate or severestenosis. North American Symptomatic CarotidEndarterectomy Trial Collaborators. New EnglandJournal of Medicine, 339(20): 1415-1425.

15 Miller M.T., Comerota A.J., Tzilinis A. et al.,(2005) Carotid endarterectomies in octogenerians:does increased age indicate ‘high risk’. NatureClinical Practice in Cardiovascular Medicine, 2(8): 382-383.

16 Martino R., Foleyl N. and Bohgal S. (2005)Dysphagia after stroke: Incidence, diagnosis andpulmonary complications. Stroke, 36(12): 2756-2763.

17 Martino R., Pron G. and Diamant N. (2000)Screening for or pharyngeal dysphasia in stroke:insufficient evidence for guidelines. Dysphagia, 15(1):19-30.

18 Walker MF, Leonardi-Bee J., Bath P et al., (2004)Individual patient data meta-analysis ofrandomized controlled trials of communityoccupational therapy for stroke patients. Stroke,35(9): 2226-2232.

19 Williams L.S., Ghose S.S., Swindle R.W., et al.,(2004). Depression and other mental healthdiagnoses increase mortality risk after ischemicstroke. American Journal of Psychiatry, 161(6): 1090-1095.

SYNOPSIS OF MEDICAL STATISTICSFrom Wikipedia, the free encyclopedia

Sample sizeThe sample size of a statistical sample is the numberof observations that constitute it. It is typically denotedn, a positive integer (natural number). Typically, all elsebeing equal, a larger sample size leads to increasedprecision in estimates of various properties of thepopulation. This can be seen in such statistical rules asthe law of large numbers and the central limit theorem.Repeated measurements and Replication ofindependent samples are often required in measurementand experiments to reach a desired precision.

A typical example would be when a statistician wishesto estimate the arithmetic mean of a continuousrandom variable (for example, the height of a person).Assuming that they have a random sample withindependent observations, then if the variability ofthe population (as measured by the standard deviationó) is known, then the standard error of the samplemean is given by the formula:

It is easy to show that as n becomes large, this variabilitybecomes very small. This yields to more sensitivehypothesis tests with greater statistical power andsmaller confidence intervals.

Central Limit TheoremThe central limit theorem is a significant result whichdepends on sample size. It states that as the size of asample of independent observations approachesinfinity, provided data come from a distribution withfinite variance, that the sampling distribution of thesample mean approaches a normal distribution.

Estimating ProportionsA typical statistical aim is to demonstrate with 95%certainty that the true value of a parameter is within adistance B of the estimate: B is an error range thatdecreases with increasing sample size (n). The value ofB generated is referred to as the 95% confidenceinterval.For example, a simple situation is estimating aproportion in a population. To do so, a statisticianwill estimate the bounds of a 95% confidence intervalfor an unknown proportion.The rule of thumb for (a maximum or ‘conservative’)B for a proportion derives from the fact the estimatorof a proportion, , (where X is thenumber of ‘positive’ observations) has a (scaled)binomial distribution and is also a form of samplemean (from a Bernoulli distribution [0,1] which has amaximum variance of 0.25 for parameter p = 0.5).So, the sample mean X/n has maximum variance 0.25/n. For sufficiently large n (usually this means that weneed to have observed at least 10 positive and 10negative responses), this distribution will be closelyapproximated by a normal distribution with the samemean and variance.

Using this approximation, it can be shown that ~95%of this distribution’s probability lies within 2 standarddeviations of the mean. Because of this, an intervalof the form

will form a 95% confidence interval for the trueproportion.

If we require the sampling error å to be no larger thansome bound B, we can solve the equation

to give us

So, n = 100 <=> B = 10%, n = 400 <=> B = 5%, n= 1000 <=> B = ~3%, and n = 10000 <=> B = 1%.One sees these numbers quoted often in news reportsof opinion polls and other sample surveys.

Extension to other casesIn general, if a population mean is estimated using thesample mean from n observations from a distributionwith variance ó², then if n is large enough (typically>30) the central limit theorem can be applied to obtainan approximate 95% confidence interval of the form

If the sampling error å is required to be no larger thanbound B, as above, then

Note, if the mean is to be estimated using P parametersthat must first be estimated themselves from the samesample, then to preserve sufficient “degrees offreedom,” the sample size should be at least n + P.

Required Sample Sizes for Hypothesis TestsA common problem facing statisticians is calculatingthe sample size required to yield a certain power for atest, given a predetermined Type I error rate á. A typicalexample for this is as follows:Let X i , i = 1, 2, ..., n be independent observationstaken from a normal distribution with mean ì andvariance ó2 . Let us consider two hypotheses, a nullhypothesis:

and an alternative hypothesis:

for some ‘smallest significant difference’ ì* >0. This isthe smallest value for which we care about observinga difference. Now, if we wish to (1) reject H0 with aprobability of at least 1-â when Ha is true (i.e. a powerof 1-â), and (2) reject H0 with probability á when H0is true, then we need the following:If zαis the upperαpercentage point of the standardnormal distribution, then

H0:μ = 0

Ha:μ = μ *


α

and so

‘Reject Ho if our sample average ( ) is more than

is a decision rule which satisfies (2). (Note, this is a 1-tailed test)

Now we wish for this to happen with a probability atleast 1-â when Ha is true. In this case, our sampleaverage will come from a Normal distribution withmean ì*. Therefore we require

Through careful manipulation, this can be shown tohappen when

where Ö is the normal cumulative distributionfunction.

Stratified Sample SizeWith more complicated sampling techniques, such asStratified sampling, the sample can often be split upinto sub-samples. Typically, if there are k such sub-samples (from k different strata) then each of themwill have a sample size ni, i = 1, 2, ..., k. These ni mustconform to the rule that n1 + n2 + ... + nk = n (i.e. thatthe total sample size is given by the sum of the sub-sample sizes). Selecting these ni optimally can be donein various ways, using (for example) Neyman’s optimalallocation.

According to Leslie Kish,[1] there are many reasons todo this; that is to take sub-samples from distinct sub-populations or “strata” of the original population: todecrease variances of sample estimates, to use partlynon-random methods, or to study strata individually.A useful, partly non-random method would be tosample individuals where easily accessible, but, wherenot, sample clusters to save travel costs.In general, for H strata, a weighted sample mean is

with

The weights, W(h), frequently, but not always, representthe proportions of the population elements in thestrata, and W(h)=N(h)/N. For a fixed sample size, thatis n=Sum{n(h)},

which can be made a minimum if the sampling rate

within each stratum is made proportional to thestandard deviation within each stratum: nh / Nh = kSh.

An “optimum allocation” is reached when the samplingrates within the strata are made directly proportionalto the standard deviations within the strata and inverselyproportional to the square roots of the costs perelement within the strata:

or, more generally, when

Statistical PowerThe power of a statistical test is the probability thatthe test will reject a false null hypothesis (that it will notmake a Type II error). As power increases, the chancesof a Type II error decrease. The probability of a TypeII error is referred to as the false negative rate (â).Therefore power is equal to 1 ” â.

Power analysis can either be done before (a priori) orafter (post hoc) data is collected. A priori power analysisis conducted prior to the research study, and is typicallyused to determine an appropriate sample size toachieve adequate power. Post-hoc power analysis isconducted after a study has been completed, and usesthe obtained sample size and effect size to determinewhat the power was in the study, assuming the effectsize in the sample is equal to the effect size in thepopulation.

Statistical tests attempt to use data from samples todetermine if differences or similarities exist in apopulation. For example, to test the null hypothesisthat the mean scores of men and women on a test donot differ, samples of men and women are drawn,the test is administered to them, and the mean scoreof one group is compared to that of the other groupusing a statistical test. The power of the test is theprobability that the test will find a statistically significantdifference between men and women, as a function ofthe size of the true difference between those twopopulations. Despite the use of random samples, whichwill tend to mirror the population due to mathematicalproperties such as the central limit theorem, there isalways a chance that the samples will appear to supportor refute a tested hypothesis when the reality is theopposite. This risk is quantified as the power of thetest and as the statistical significance level used for thetest.

Statistical power depends on:· the statistical significance criterion used in the test.· the size of the difference or the strength of the

similarity (that is, the effect size) in the population.· the sensitivity of the data.


A significance criterion is a statement of how unlikelya result must be, if the null hypothesis is true, to beconsidered significant. The most commonly usedcriteria are probabilities of 0.05 (5%, 1 in 20), 0.01(1%, 1 in 100), and 0.001 (0.1%, 1 in 1000). If thecriterion is 0.05, the probability of the difference mustbe less than 0.05, and so on. One way to increase thepower of a test is to increase (that is, weaken) thesignificance level. This increases the chance of obtaininga statistically significant result (rejecting the nullhypothesis) when the null hypothesis is false, that is,reduces the risk of a Type II error. But it also increasesthe risk of obtaining a statistically significant result whenthe null hypothesis is in fact true; that is, it increases therisk of a Type I error.

Calculating the power requires first specifying the effectsize you want to detect. The greater the effect size, thegreater the power.

Sensitivity can be increased by using statistical controls,by increasing the reliability of measures (as inpsychometric reliability), and by increasing the size ofthe sample. Increasing sample size is the mostcommonly used method for increasing statistical power.Although there are no formal standards for power,most researchers who assess the power of their testsuse 0.80 as a standard for adequacy.

A common misconception by those new to statisticalpower is that power is a property of a study orexperiment. In reality any statistical result that has a p-value has an associated power. For example, in thecontext of a single multiple regression, there will be adifferent level of statistical power associated with theoverall r-square and for each of the regressioncoefficients. When determining an appropriate samplesize for a planned study, it is important to considerthat power will vary across the different hypotheses.

There are times when the recommendations of poweranalysis regarding sample size will be inadequate. Poweranalysis is appropriate when the concern is with thecorrect acceptance or rejection of a null hypothesis. Inmany contexts, the issue is less about determining ifthere is or is not a difference but rather with getting amore refined estimate of the population effect size.For example, if we were expecting a populationcorrelation between intelligence and job performanceof around .50, a sample size of 20 will give usapproximately 80% power (alpha = .05, two-tail).However, in doing this study we are probably moreinterested in knowing whether the correlation is .30 or.60 or .50. In this context we would need a muchlarger sample size in order to reduce the confidenceinterval of our estimate to a range that is acceptablefor our purposes. These and other considerations oftenresult in the recommendation that when it comes tosample size, “More is better!”

Funding agencies, ethics boards and research reviewpanels frequently request that a researcher perform apower analysis. The argument is that if a study is

inadequately powered, there is no point in completingthe research.

Student’s t-testA t-test is any statistical hypothesis test in which thetest statistic has a Student’s t distribution if the nullhypothesis is true. It is applied when sample sizes aresmall enough that using an assumption of normalityand the associated z-test leads to incorrect inference.

UseA t-test is any statistical hypothesis test in which thetest statistic has a Student’s t-distribution if the nullhypothesis is true. It is applied when sample sizes aresmall enough that using an assumption of normalityand the associated z-test leads to incorrect inference.Among the most frequently used t tests are:

· A test of the null hypothesis that the means of twonormally distributed populations are equal. Given twodata sets, each characterized by its mean, standarddeviation and number of data points, we can use somekind of t test to determine whether the means aredistinct, provided that the underlying distributions canbe assumed to be normal. All such tests are usuallycalled Student’s t tests, though strictly speaking thatname should only be used if the variances of the twopopulations are also assumed to be equal; the formof the test used when this assumption is dropped issometimes called Welch’s t test. There are differentversions of the t test depending on whether the twosamples are• unpaired, independent of each other (e.g., individuals

randomly assigned into two groups), or• paired, so that each member of one sample has a

unique relationship with a particular member of theother sample (e.g., the same people measured beforeand after an intervention, or IQ test scores of ahusband and wife).

If the calculated p-value is below the threshold chosenfor statistical significance (usually the 0.10, the 0.05, or0.01 level), then the null hypothesis which usually statesthat the two groups do not differ is rejected in favorof an alternative hypothesis, which typically states thatthe groups do differ.

· A test of whether the mean of a normally distributedpopulation has a value specified in a null hypothesis.

· A test of whether the slope of a regression linediffers significantly from 0.

Once a t value is determined, a p-value can be foundusing a table of values from Student’s t-distribution.

Assumptions· Normal distribution of data, tested by using anormality test, such as Shapiro-Wilk andKolmogorov-Smirnov test.· Equality of variances, tested by using either the Ftest, the more robust Levene’s test, Bartlett’s test, orthe Brown-Forsythe test.


An example of a repeated measures t-test wouldbe if one group were pre- and post-tested. (Thisexample occurs in education quite frequently.) If ateacher wanted to examine the effect of a new set oftextbooks on student achievement, (s)he could test theclass at the beginning of the year (pretest) and at theend of the year (posttest). A dependent t-test wouldbe used, treating the pretest and posttest as matchedvariables (matched by student).

CalculationsIndependent one-sample t-testThis equation is used to compare one sample mean toa specific value

Where s is the grand standard deviation of the sample.n is the sample size. The degrees of freedom used inthis test is n ” 1.

Independent two-sample t-testEqual sample sizes, equal variance

This equation is only used when both:

· the two sample sizes (that is, the n or numberof participants of each group) are equal;

· it can be assumed that the two distributionshave the same variance.

Violations of these assumptions are discussed below.

The t statistic to test whether the means are differentcan be calculated as follows:

where

[1]

Here is the grand standard deviation (orpooled standard deviation), 1 = group one, 2 = grouptwo. The denominator of t is the standard error ofthe difference between two means. For significancetesting, the degrees of freedom for this test is 2n ” 2where n = # of participants in each group.

Unequal Sample Sizes, Equal VarianceThis equation is only used when it can be assumedthat the two distributions have the same variance.(When this assumption is violated, see below.) The tstatistic to test whether the means are different canbe calculated as follows:

· Samples may be independent or dependent,depending on the hypothesis and the type ofsamples:o Independent samples are usually two randomly

selected groups

o Dependent samples are either two groupsmatched on some variable (for example, age) orare the same people being tested twice (calledrepeated measures)

Since all calculations are done subject to the nullhypothesis, it may be very difficult to come up with areasonable null hypothesis that accounts for equalmeans in the presence of unequal variances. In the usualcase, the null hypothesis is that the different treatmentshave no effect — this makes unequal variancesuntenable. In this case, one should forgo the ease ofusing this variant afforded by the statistical packages.See also Behrens-Fisher problem.

One scenario in which it would be plausible to haveequal means but unequal variances is when the ‘samples’represent repeated measurements of a single quantity,taken using two different methods. If systematic erroris negligible (e.g. due to appropriate calibration) theeffective population means for the two measurementmethods are equal, but they may still have differentlevels of precision and hence different variances.

Determining TypeFor novices, the most difficult issue is often whetherthe samples are independent or dependent.Independent samples typically consist of two groupswith no relationship. Dependent samples typicallyconsist of a matched sample (or a “paired” sample)or one group that has been tested twice (repeatedmeasures).

Dependent t-tests are also used for matched-pairedsamples, where two groups are matched on aparticular variable. For example, if we examined theheights of men and women in a relationship, the twogroups are matched on relationship status. This wouldcall for a dependent t-test because it is a paired sample(one man paired with one woman). Alternatively, wemight recruit 100 men and 100 women, with norelationship between any particular man and anyparticular woman; in this case we would use anindependent samples test.Another example of a matched sample would be totake two groups of students, match each student inone group with a student in the other group based onan achievement test result, then examine how mucheach student reads. An example pair might be twostudents that score 90 and 91 or two students thatscored 45 and 40 on the same test. The hypothesiswould be that students that did well on the test mayor may not read more. Alternatively, we might recruitstudents with low scores and students with high scoresin two groups and assess their reading amountsindependently.

μ0.


where

[2]

Note that the formulae above are generalizations forthe case where both samples have equal sizes (substituten1 and n2 for n and you’ll see).

is the unbiased estimator of the variance ofthe two samples, n = number of participants, 1 =group one, 2 = group two. n ” 1 is the number ofdegrees of freedom for either group, and the totalsample size minus 2 (n1 + n2 “ 2) is the total numberof degrees of freedom, which is used in significancetesting.

The statistical significance level associated with the tvalue calculated in this way is the probability that, underthe null hypothesis of equal means, the absolute valueof t could be that large or larger just by chance—inother words, it’s a two-tailed test, testing whether themeans are different when, if they are, either one maybe the larger (see Press et al, 1999, p. 616).

Unequal Sample Sizes, Unequal VarianceThis equation is only used when the two sample sizesare unequal and the variance is assumed to be different.See also Welch’s t test. The t statistic to test whetherthe means are different can be calculated as follows:

where

Where s2 is the unbiased estimator of the variance ofthe two samples, n = number of participants, 1 =group one, 2 = group two. Note that in this case, isnot a pooled variance. For use in significance testing,the distribution of the test statistic is approximated asbeing an ordinary Student’s t distribution with thedegrees of freedom calculated using:

This equation is called the Welch-Satterthwaite equation.Note that the true distribution of the test statistic actuallydepends (slightly) on the two unknown variances: seeBehrens-Fisher problem.

This test can be used as either a one-tailed or two-tailed test.

Dependent t-testThis equation is used when the samples are dependent;that is, when there is only one sample that has beentested twice (repeated measures) or when there aretwo samples that have been matched or “paired”.

For this equation, the differences between all pairs mustbe calculated. The pairs are either one person’s pre-test and post-test scores or between pairs of personsmatched into meaningful groups (for instance drawnfrom the same family or age group: see table). Theaverage (XD) and standard deviation (sD) of thosedifferences are used in the equation. The constant ì0 isnon-zero if you want to test whether the average ofthe difference is significantly different than ì0. Thedegree of freedom used is N-1.

ExampleA random sample of screws have weights

30.02, 29.99, 30.11, 29.97, 30.01, 29.99Calculate a 95% confidence interval for thepopulation’s mean weight.

Assume the population is distributed as N(μ, ó2).

The samples’ mean weight is 30.015 with standarddeviation of 0.0497. With the mean and the first fiveweights it is possible to calculate the sixth weight.Consequently there are five degrees of freedom.

Example of matched pairs

Pair Name Age Test

1 Jon 35 250

1 Jane 36 340

2 Jimmy 22 460

2 Jessy 21 200

Example of repeated measures

Number Name Test 1 Test 2

1 Mike 35% 67%

2 Melanie 50% 46%

3 Melissa 90% 86%

4 Mitchell 78% 91%


We can lookup in the table that for a confidence rangeof 95% and five degrees of freedom, the value is2.571.

i.e.

If we sampled many times, our interval would capturethe true mean weight 95% of the time; thus, we are95% confident that the true mean weight of all screwswill fall between 29.96 and 30.07

Alternatives to the t testRecall that the t test can be used to test the equality ofthe means of two normal populations with unknown,but equal, variance.· To relax the normality assumption, a non-parametricalternative to the t test can be used, and the usual choicesare:

o for independent samples, the Mann-Whitney U testo for related samples, either the binomial test or the

Wilcoxon signed-rank test· To test the equality of the means of more than twonormal populations, an Analysis of Variance can beperformed· To test the equality of the means of two normalpopulations with known variance, a Z-test can beperformed

HistoryThe t statistic was introduced by William Sealy Gossetfor cheaply monitoring the quality of beer brews(“Student” was his pen name) [3]. Gosset was astatistician for the Guinness brewery in Dublin, Ireland,and was hired due to Claude Guinness’s innovativepolicy of recruiting the best graduates from Oxfordand Cambridge to apply biochemistry and statistics toGuinness’ industrial processes[3]. Gosset published thet test in Biometrika in 1908, but was forced to use apen name by his employer who regarded the fact thatthey were using statistics as a trade secret. In fact,Gosset’s identity was known not only to fellowstatisticians but to his employer — the company insistedon the pseudonym[4] so that it could turn a blind eyeto the breach of its rules.

Today, it is more generally applied to the confidencethat can be placed in judgments made from smallsamples.

Most spreadsheet programs and statistics packages,such as DAP, gretl, R and PSPP includeimplementations of Student’s t-test.

Meta-analysisIn statistics, a meta-analysis combines the results ofseveral studies that address a set of related researchhypotheses. The first meta-analysis was performed byKarl Pearson in 1904, in an attempt to overcome the

problem of reduced statistical power in studies withsmall sample sizes; analyzing the results from a groupof studies can allow more accurate data analysis.

Although meta-analysis is widely used in epidemiologyand evidence-based medicine today, a meta-analysisof a medical treatment was not published until 1955.In the 1970s, more sophisticated analytical techniqueswere introduced in educational research, starting withthe work of Gene V. Glass, Frank L. Schmidt andJohn E. Hunter.

The online Oxford English Dictionary lists the firstusage of the term in the statistical sense as 1976 byGlass. The statistical theory surrounding meta-analysiswas greatly advanced by the work of Nambury S.Raju, Larry V. Hedges, Harris Cooper, Ingram Olkin,John E. Hunter, Jacob Cohen, and Frank L. Schmidt.

Uses in Modern ScienceBecause the results from different studies investigatingdifferent independent variables are measured ondifferent scales, the dependent variable in a meta-analysis is some standardized measure of effect size.To describe the results of comparative experimentsthe usual effect size indicator is the standardized meandifference (d) which is the standard score equivalentto the difference between means, or an odds ratio ifthe outcome of the experiments is a dichotomousvariable (success versus failure). A meta-analysis canbe performed on studies that describe their findingsin correlation coefficients, as for example, studies ofthe correlation between familial relationships andintelligence. In these cases, the correlation itself is theindicator of the effect size.

The method is not restricted to situations in whichone or more variables is defined as “dependent.” Forexample, a meta-analysis could be performed on acollection of studies each of which attempts to estimatethe incidence of left-handedness in various groups ofpeople.

Researchers should be aware that variations in samplingschemes can introduce heterogeneity to the result,which is the presence of more than one intercept inthe solution. For instance, if some studies used 30mgof a drug, and others used 50mg, then we wouldplausibly expect two clusters to be present in the data,each varying around the mean of one dosage or theother. This can be modelled using a “random effectsmodel.”

Results from studies are combined using differentapproaches. One approach frequently used in meta-analysis in health care research is termed ‘inversevariance method’. The average effect size across allstudies is computed as a weighted mean, whereby theweights are equal to the inverse variance of eachstudies’ effect estimator. Larger studies and studies withless random variation are given greater weight thansmaller studies. Other common approaches includethe Mantel Haenszel method and the Peto method. A


free Excel-based calculator to perform MantelHaenszel analysis is available at: http://www.pitt.edu/~super1/lecture/lec1171/014.htm. They also have afree Excel-based Peto method calculator at : http://www.pitt.edu/~super1/lecture/lec1171/015.htm

Cochraine and other sources provide a usefuldiscussion of the differences between these twoapproaches.

Q : Why not just add up all the results across studies ?

Answer : There is concern about Simpson’s paradox.Note, however that Mantel Haenszel analysis and Petoanalysis introduce their own biases and distortions ofthe data results.

A recent approach to studying the influence thatweighting schemes can have on results has beenproposed through the construct of gravity, which is aspecial case of combinatorial meta analysis.

Modern meta-analysis does more than just combinethe effect sizes of a set of studies. It can test if thestudies’ outcomes show more variation than thevariation that is expected because of sampling differentresearch participants. If that is the case, studycharacteristics such as measurement instrument used,population sampled, or aspects of the studies’ designare coded. These characteristics are then used aspredictor variables to analyze the excess variation inthe effect sizes. Some methodological weaknesses instudies can be corrected statistically. For example, it ispossible to correct effect sizes or correlations for thedownward bias due to measurement error orrestriction on score ranges.

Meta analysis leads to a shift of emphasis from singlestudies to multiple studies. It emphasises the practicalimportance of the effect size instead of the statisticalsignificance of individual studies. This shift in thinkinghas been termed Metaanalytic thinking.

The results of a meta-analysis are often shown in aforest plot.

WeaknessesA weakness of the method is that sources of bias arenot controlled by the method. A good meta-analysisof badly designed studies will still result in bad statistics.Robert Slavin has argued that only methodologicallysound studies should be included in a meta-analysis, apractice he calls ‘best evidence meta-analysis’. Othermeta-analysts would include weaker studies, and adda study-level predictor variable that reflects themethodological quality of the studies to examine theeffect of study quality on the effect size. Anotherweakness of the method is the heavy reliance onpublished studies, which may increase the effect as it isvery hard to publish studies that show no significantresults. This publication bias or “file-drawer effect”(where non-significant studies end up in the deskdrawer instead of in the public domain) should be

seriously considered when interpreting the outcomesof a meta-analysis. Because of the risk of publicationbias, many meta-analyses now include a “failsafe N”statistic that calculates the number of studies with nullresults that would need to be added to the meta-analysisin order for an effect to no longer be reliable.

Forest PlotFrom Wikipedia, the Free EncyclopediaA forest plot is a graphical display that shows thestrength of the evidence in quantitative scientific studies.It was developed for use in medical research as ameans of graphically representing a meta-analysis ofthe results of randomized controlled trials. In the lasttwenty years, similar meta-analytical techniques havebeen applied in observational studies (e.g.environmental epidemiology) and forest plots are oftenused in presenting the results of such studies also.

Although forest plots can take several forms, they arecommonly presented with two columns. The left-handcolumn lists the names of the studies (frequentlyrandomized controlled trials or epidemiologicalstudies), commonly in chronological order from thetop downwards. The right-hand column is a plot ofthe measure of effect (e.g. an odds ratio) for each ofthese studies (often represented by a square)incorporating confidence intervals represented byhorizontal lines. The graph may be plotted on a naturallogarithmic scale when using odds ratios, so that theconfidence intervals are symmetrical about the meansfrom each study. The size of each square is proportionalto the study’s weight in the meta-analysis. The overallmeta-analysed measure of effect is represented on theplot as a vertical line. This meta-analysed measure ofeffect is commonly plotted as a diamond, the lateralpoints of which indicate confidence intervals for thisestimate.

A vertical line representing no effect is also plotted. Ifthe confidence intervals for individual studies overlapwith this line, it demonstrates that at the given level ofconfidence their effect sizes do not differ from noeffect. The same applies for the meta-analysed measureof effect: if the points of the diamond overlap theline of no effect the overall meta-analysed result cannotbe said to differ from no effect at the given level ofconfidence.

Forest plots date back to at least the 1970s, althoughthe first use in print may be 1996.[1] The name refersto the forest of lines produced. In September 1990,Richard Peto joked that the plot was named after abreast cancer researcher called Pat Forrest and the namehas sometimes been spelt “forrest plot”.[1]


acanthosis nigricans in the head and neck regionaipmed.org/publications/june2008edition/vol. 6 no.1...

Documents