Brief communication

Incidence and diagnostic diversity infirst-episode psychosis

Introduction

There has been increasing research into first-episode psychosis (FEP) in recent years, includinglarge studies such as OPUS (an intensive earlytreatment programme for first-episode psychoticpatients), the Swedish parachute project, Aetiologyand Ethnicity in Schizophrenia and other Psycho-ses (AESOP), Lambeth Early Onset (LEO) and theCavan-Monaghan study (1–5). However, most ofthis research has focussed on younger patients

with schizophrenia and there has been little outsideof these age and diagnostic groups. There iswide variation between studies in diagnostic andinclusion ⁄ exclusion criteria. Exclusions from main-stream research may result in a lack of genera-lisability to large numbers of patients in real worldsettings (6). The recent Cochrane review into earlyintervention for psychosis (7) highlighted thechallenges when comparing studies in this area.Despite a comprehensive literature review onlyseven studies were suitable for comparison, with six

Reay R, Mitford E, McCabe K, Paxton R, Turkington D. Incidence anddiagnostic diversity in first-episode psychosis.

Objective: To investigate the incidence and range of diagnostic groupsin patients with first-episode psychosis (FEP) in a defined geographicalarea.Method: An observational database was set up on all patients aged16 years and over presenting with FEP living in a county in NorthernEngland between 1998 and 2005.Results: The incidence of all FEP was 30.95 ⁄ 100 000. The largestdiagnostic groups were psychotic depression (19%) and acute andtransient psychotic disorder (19%). Fifty-four per cent of patients wereaged 36 years and over. Patients with schizophrenia spectrum disorderonly accounted for 55% of cases.Conclusion: This clinical database revealed marked diversity in ageand diagnostic groups in FEP with implications for services andguidelines. These common presentations of psychoses are grossly underresearched, and no treatment guidelines currently exist for them.

R. Reay, E. Mitford, K. McCabe,R. Paxton, D. TurkingtonNorthumberland, Tyne and Wear NHS Trust – PACEOffice, St Georges Park, Morpeth, Northumberland, UK

Key words: first episode; epidemiology; psychoses

Robert Reay, Northumberland, Tyne and Wear NHSTrust – PACE Office, West Wing, St George�s ParkMorpeth Northumberland NE61 2NU, UK.E-mail: robert.reay@ntw.nhs.uk

Accepted for publication October 9, 2009

Significant outcomes

• Pragmatic observational databases help develop a reality based overall clinical picture of a condition.• Broadening the age range and diagnostic spectrum revealed a high incidence of psychosis not

currently covered by Early Intervention Services. Only 46% were within the age criteria for the localEarly Intervention Services.

• Only 55% of patients fell within the schizophrenia spectrum disorder group.

Limitations

• Not all patients may have been included in the study, as some out-patients may have been missed,and some patients may never have presented to services.

• Ninety-nine per cent of patients were Caucasian, limiting generalisability.• ICD-10 was used rather than DSM-IV, reducing diagnostic reliability, but not generalisability.

Acta Psychiatr Scand 2010: 121: 315–319All rights reservedDOI: 10.1111/j.1600-0447.2009.01505.x

� 2009 John Wiley & Sons A/S

ACTA PSYCHIATRICASCANDINAVICA

315

having small sample sizes of between 56 and 83participants. All seven studies used differentinterventions or controls and therefore could notbe pooled for meta-analysis. In an effort to controlconfounding variables, many common groups ofpatients are excluded from FEP studies, such asolder patients, patients with comorbid substancemisuse or patients with affective psychosis. Theseexclusions may lead to research being based ongroups of patients that are not representative ofroutine clinical practice. Simonsen et al. (8)emphasised the importance of identifying all casesof FEP within a defined population to predictoutcomes accurately. FEP studies may also befocussing too exclusively on first-episode schizo-phrenia at the expense of other diagnoses. Baldwinet al. (5) expressed concern that concentratingpurely on first-episode schizophrenia would missthe essential diversity of FEP.The lack of agreement on outcome measures

makes research in the field of mental healthchallenging to undertake (9) and might requiremore unconventional approaches to researchdesigns. Black (10) emphasised the importance ofclinical databases as a research method in all areasof medicine. Databases offer the opportunity tocarry out evaluative research and clinical audit,inform the planning and management of services,and provide individual clinicians with accurateestimates of the outcome of their care, bringingresearch closer to practice.In 1998 the PACE (Population Adjusted Clinical

Epidemiology) process, a prospective pragmaticclinical database was adapted for use in studyingFEP in Northumberland (11).

Aims of the study

The aim of the study was to look at the incidenceand diagnostic diversity of first-episode psychosis,in a county in Northern England, across all agegroups and individual ICD-10 codes, includingschizophrenia spectrum disorder, affective psycho-ses, organic and substance induced psychoses.

Material and methods

Background

The PACE process was developed for use inhaematological oncology (12), creating a registrysystem, detailing the presentation, treatment andoutcome of all reported cases of haematologicalmalignancy within a geographically defined area.The same process was then applied to first episodepsychosis.

Participants

Patients aged 16 and over who presented tosecondary mental health care services in North-umberland (a county in northern England) with adiagnosis of FEP between 1 October 1998 and 30September 2005 were included. There was no upperage limit, and patients with physical health prob-lems and ⁄or co-existing substance misuse problemswere also included. The only exclusion criteria werepresentation outside of the Trust area, past historyof treated psychosis or a diagnosis of dementia atpresentation. ICD-10 diagnostic criteria were used,as this was the classification system used by theconsultant psychiatrists within the Northumber-land, Tyne and Wear NHS Trust. ICD-10 waspreferable to DSM-IV as this may be too narrow inits definition of some psychotic disorders (13).

Procedure

From October 1998 patients with a FEP wereidentified by the consultant psychiatrists in North-umberland, and referred to PACE. Demographicinformation, ICD-10 diagnosis and hospitaladmissions were collected from in-patient andout-patient secondary care medical notes and thedata were entered onto a Microsoft Accessdatabase. Monthly reminders were sent to allconsultant psychiatrists, requesting names ofpatients (both in-patients and out-patients) whohad recently presented with a FEP. PACE staffalso screened the computerised informationsystem for patients who had been dischargedfrom hospital.

Analysis

Patients were grouped according to ICD-10 diag-nostic criteria, including schizophrenia spectrumdisorder (F20–29), affective psychoses (F30–32.3),organic (F6), and alcohol and drug induced psy-chosis (F10.5, F19.5). Incidence rates for allpatients were calculated using the 2001 censusdata for the denominator population (http://www.statistics.gov.uk/census2001). The calcula-tions were repeated for the 16–64 year group tofacilitate comparison with other studies. The pop-ulation of Northumberland in 2001 was 307 190with 195 247 aged 16–64 and 54 038 aged 65 andover.Confidentiality was strictly maintained. Patients

were given unique PACE identification numbersand data were always presented in the form ofgroup outcomes. PACE staff had no contact withpatients and were not involved in their care or

Reay et al.

316

management. The database was set up with the aimof service evaluation as defined by the NationalResearch Ethics service (http://www.nres.npsa.nh-s.uk), and therefore did not require ethicalapproval.

Results

Demographics

Five hundred and forty patients presented with aFEP between 1 October 1998 and 30 September2005. Fifty-seven per cent (308) were male and43% (231) female. Ninety-nine per cent wereCaucasian, reflecting the ethnic make up ofNorthumberland. Forty-six per cent (251) ofpatients were under 36 years of age. Thirty percent (160) were aged between 36 and 64 years ofage. Twenty-four per cent (129) were aged 65 andover. The age range at presentation was 16 to99 years.

Diagnosis

The largest diagnostic group was schizophreniaspectrum disorder with 55% (297) of patients,followed by 29% (156) with affective psychoses(Fig. 1). Drug and alcohol induced psychosesaccounted for 9% (46) and 8% (41) had a diagnosis

of organic psychoses. The most common individualdiagnoses at presentation were psychotic depres-sion (105) and acute and transient psychoticdisorder (103) both with 19% of patients. In theunder 36 age group the most common diagnosiswas acute and transient psychotic disorder with25% (62) of patients. For the 36–64 age grouppsychotic depression was the most common diag-nosis with 27% (43) of patients. In the 65 and overage group the most common diagnoses wereorganic psychosis with 26% (33) of patients andpsychotic depression with 24% (31) of patients.

Incidence (per 100 000 population per year)

The annual incidence of all psychoses (aged 16and over) was 30.95 per 100 000 population peryear [95% confidence interval (CI) 28.34–33.56](Table 1). The incidence of schizophrenia spectrumdisorder (F20–29) was 17.02 (95% CI, 15.08–18.96)and affective psychoses (F30–33) 8.94 (95% CI,7.54–10.34).

Discussion

Most psychosis research has been on FEP andBritish Early Intervention Services. However, wefound that more than half of patients were overthe age of 36 and would not have been included.

Gen

eral

dia

gnos

tic g

roup

at p

rese

ntat

ion

Psychotic depression(F32.3)

Bipolar affectivedisorder (F31)

Manic episode(F30)

Non-organic psychosis(F28)

Schizoaffective disorder(F25)

Acute & transientpsychotic disorder (F23)

Persistent delusionaldisorder (F22)

Schizophrenia (F20)

Drug and alcohol inducedpsychosis (F10.5-19.5)

Organic psychosis (F6)

Percent12.0%10.0%8.0%6.0%4.0%2.0%0.0%

over 6436 to 64under 36

Age Group

Fig. 1. Percentage of patients in eachICD-10 diagnostic group divided byage.

Incidence and diagnostic diversity in FEP

317

Twenty-four per cent of the patients included inthe study were aged 65 and over, showing a clearneed for research and services for older adultswith psychosis. PACE included all FEP ICD-10diagnostic groups and psychotic depression andacute and transient psychotic disorder were themost common diagnoses. As the diagnoses weremade at presentation, it would be expected that alarge proportion of acute and transient psychoticdisorders would change or recover. There are noNICE guidelines for clinicians regarding thesecommon psychotic disorders, but conversely thereis a wealth of information and research onschizophrenia and schizophrenia spectrum disor-ders. Affective and organic psychoses are oftenexcluded from FEP studies, but 45% of ourpatients would have fallen within these over-looked groups.Comparison with other studies in FEP is chal-

lenging as different exclusion criteria have beenused, including age and diagnostic groups. This hasresulted in a somewhat fragmented body ofevidence. We were unable to compare our resultsfully with studies such as OPUS or LEO as theyused DSM-IV diagnostic criteria rather than ICD-10. Baldwin et al. (5) reported one of the fewstudies to include older adults and affective psy-choses and found similar incidence rates to ourown. When we compared our 16–64 year groupwith the AESOP study (3), London had a muchhigher incidence rate than our own, whereasBristol and Nottingham were slightly lower.Most research and resources are targeted at

younger patients with FEP (mostly schizophreniaspectrum disorder). We found wide diversity in ageand diagnostic groups in the present study andbelieve that this needs further investigation.Almost a quarter of the patients were 65 yearsand over (retirement age) and this has majorimplications for other services including Crisis andAssertive Outreach Teams, psychology servicesand for the planning of future services. Patientswith a diagnosis of psychotic depression are amuch overlooked group, with little reference to theillness in either the NICE Guideline for Schizo-

phrenia (14) or the NICE Guideline for Depression(15). Many groups are specifically excluded fromresearch on FEP, which may ultimately lead toinaccurate targeting of resources and consequentunmet needs.Population Adjusted Clinical Epidemiology is a

clinically based, pragmatic, observational, pro-spective database, which has been operating since1998, with data gathered on a continuing basisfrom a defined catchment area with a relativelystable population. Both in-patients and out-patients were included, and diagnoses were madeby consultant psychiatrists according to ICD-10criteria. This design is likely to reflect the trueprevalence and course of the illness in thisgeographical area, and help develop a realitybased clinical picture of the condition. Limitationsof the PACE methodology include the fact thatpatients may have been missed, especially out-patients, and there are no studies into the numberof patients possibly missed. The team tried toensure best possible coverage by talking to con-sultants and secretaries and by screening thehospital computerised data system. When settingup an Early Intervention Service in Northumber-land, the team based their service on our num-bers, which were found to be reasonably accurate.Patients who never presented, or who remained inprimary care will not have been included. North-umberland is overwhelmingly Caucasian so resultscannot be generalised to more ethnically diverseareas.First-episode psychosis is a varied condition in

terms of age and diagnostic groups, but manypatients are routinely excluded from research,leading to an incomplete picture. The value of thePACE methodology is that it has given a clearerpicture of the full spectrum of FEP within anEnglish county, so that unmet needs can beidentified and services targeted and evaluatedmore accurately.

Declaration of interests

None.

Table 1. Incidence by ICD-10 code (16 years andover ⁄ 16–64 years)

ICD-10 diagnosis at presentation

16 years and overincidence per 100 000

per population (95% CI)

16–64 years incidenceper 100 000 per

population (95% CI)

All psychoses (n = 540) 30.95 (28.34–33.56) 30.07 (27.16–32.98)Schizophrenia (F20) (n = 72) 4.13 (3.27–5.08) 4.39 (3.28–5.50)Schizophrenia spectrum (F20–29) (n = 297) 17.02 (15.08–18.96) 17.78 (15.54–20.02)Bipolar (F30–31.9) (n = 51) 3.73 (2.71–4.76) 3.22 (2.27–4.17)Psychotic depression (F32.3) (n = 105) 6.02 (4.87–7.17) 5.41 (6.65–4.18)Affective psychoses (F30–33) (n = 156) 8.94 (7.54–10.34) 8.63 (7.08–10.19)Organic and substance induced (n = 87) 4.99 (3.94–6.03) 3.66 (2.64–4.67)

Reay et al.

318

References

1. Bertelsen M, Jeppesen P,Petersen L et al. Five-year follow-upof a randomized multicenter trial of intensive earlyintervention vs standard treatment for patients with a firstepisode of psychotic illness: the OPUS trial. Arch GenPsychiatry 2008;65:762–771.

2. Cullberg J, Mattsson M, Levander S et al. Treatment costsand clinical outcome for first episode schizophrenia pa-tients: A three-year follow-up of the Swedish �ParachuteProject� and Two Comparison Groups. Acta PsychiatrScand 2006;114:274–281.

3. Morgan C, Dazzan P, Morgan K et al. First episode psy-chosis and ethnicity: initial findings from the AESOPstudy. World Psychiatry 2006;5:40–46.

4. Craig TKJ, Garety P, Power P et al. The Lambeth EarlyOnset (LEO) Team: randomised controlled trial of theeffectiveness of specialised care for early psychosis. BMJ2004;329:1067.

5. Baldwin P, Browne D, Scully PJ et al. Epidemiology offirst-episode psychosis: illustrating the challenges acrossdiagnostic boundaries through the Cavan-Monaghanstudy at 8 years. Schizophr Bull 2005;31:624–638.

6. Bartlett C, Doyal L, Ebrahim S et al. The causes andeffects of socio-demographic exclusions from clinical trials.Health Technol Assess 2005;9:1–152.

7. Marshall M, Rathbone J. Early intervention for psychosis.Cochrane Database Syst Rev 2006;4:CD004718.

8. Simonsen E, Friis S, Haahr U et al. Clinical epidemiologicfirst-episode psychosis: 1-year outcome and predictors.Acta Psychiatr Scand 2007;116:54–61.

9. Thornley B, Adams C. Content and quality of 2000 con-trolled trials in schizophrenia over 50 years. BMJ1998;317:1181–1184.

10. Black N, Barker M, Payne M. Cross sectional survey ofmulticentre clinical databases in the United Kingdom.BMJ 2004;328:1478.

11. Proctor SE, Mitford E, Paxton R. First episode psychosis: anovel methodology reveals higher than expected incidence;a reality-based population profile in Northumberland, UK.J Eval Clin Pract 2004;10:539–547.

12. Proctor SJ, Taylor P. A practical guide to continuouspopulation-based data collection (PACE): a process facil-itating uniformity of care and research into practice. QJM2000;93:67–73.

13. Pillman F, Haring A, Balzuweit S et al. The concordance ofICD-10 acute and transient psychosis and DSM-IV briefpsychotic disorder. Psychol Med 2002;32:525–533.

14. National Institute for Health and Clinical Excellence.Clinical Guideline 1. Schizophrenia: Core Interventions inthe Treatment and Management of Schizophrenia in Pri-mary and Secondary Care. London: HMSO, 2002.

15. National Institute for Health and Clinical Excellence.Clinical Guideline 23. Depression: Management ofDepression in Primary and Secondary Care. London:HMSO, 2004.

Incidence and diagnostic diversity in FEP

319