quality assurance project - usaid assist...with pregnancy-induced hypertension abstract reduces...

Q U A L I T Y

A S S U R A N C E

P R O J E C T

Center for Human Services • 7200 Wisconsin Avenue, Suite 600 • Bethesda, MD 20814-4811 • USA • www.qaproject.org

O P E R A T I O N S R E S E A R C H R E S U L T S

Improving the Quality of Care forWomen with Pregnancy-Induced

Hypertension Reduces Costs in Tver, Russia

April 2002

The Quality Assurance (QA) Project is funded by the U.S. Agency for International Development(USAID), under Contract Number HRN-C-00-96-90013. The QA Project serves countries eligible forUSAID assistance, USAID Missions and Bureaus, and other agencies and nongovernmental organiza-tions that cooperate with USAID. The QA Project team, which consists of prime contractor Center forHuman Services, Joint Commission Resources, Inc., and Johns Hopkins University (including theSchool of Hygiene and Public Health, the Center for Communication Programs [CCP], and the JohnsHopkins Program for International Education in Reproductive Health [JHPIEGO], provides comprehen-sive, leading-edge technical expertise in the design, management, and implementation of qualityassurance programs in developing countries. Center for Human Services, the non-profit affiliate ofUniversity Research Co., LLC, provides technical assistance and research for the design, manage-ment, improvement, and monitoring of health systems and service delivery in over 30 countries.

O P E R A T I O N S R E S E A R C H R E S U L T S

Continued on page ii

Table of Contents

ABBREVIATIONS .................................................................................................... 1

I. INTRODUCTION ............................................................................................... 1

II. THE QUALITY IMPROVEMENT INTERVENTION ............................................. 2

III. EFFECTS ON HEALTH OUTCOMES ................................................................ 3

IV. COST STUDY DESIGN AND METHODS .......................................................... 3

A. Inpatient Costs ............................................................................................. 3

B. Outpatient Costs .......................................................................................... 4

V. COST STUDY FINDINGS.................................................................................. 5

A. Number of PIH Cases .................................................................................. 5

B. Reduction in Direct Costs of Care for PIH Patients .................................... 6

Inpatient care ........................................................................................... 6

Outpatient care ........................................................................................ 7

Direct cost of hospitalization ................................................................... 8

Direct cost of drugs for PIH inpatients .................................................... 8

Direct cost of paraclinical care for PIH inpatients .................................. 9

C. Differences between Hospitals ................................................................. 10

D. Differences by Case Severity .................................................................... 12

VI. ASSUMPTIONS AND POTENTIAL THREATS TO VALIDITY .......................... 13

VII. CONCLUSION AND DISCUSSION ................................................................ 14

ENDNOTES ........................................................................................................... 15

REFERENCES ....................................................................................................... 16

Improving the Quality of Care for Womenwith Pregnancy-Induced HypertensionReduces Costs in Tver, Russia

Abstract

The Quality Assurance Project/Russia implemented a qualityimprovement (QI) demonstrationproject in 1998 at three hospitalsin Tver Oblast, Russia. The projectsought to improve the quality ofcare for women with pregnancy-induced hypertension (PIH), thenthe single largest cause of mater-nal deaths in Tver. Central to theQI effort was the development andintroduction of evidence-basedclinical guidelines for the manage-ment of PIH. The new guidelinesrationalized admission criteria andthe use of drugs, reduced the num-ber of PIH admissions, and calledfor more aggressive treatment ofPIH.

Health outcomes improved follow-ing the introduction of the newguidelines. Complications in new-borns of PIH mothers dropped, noPIH case progressed to eclampsia(which is often fatal), and nomaternal deaths caused by PIHoccurred in the 15 months follow-ing the implementation of the newguidelines.

A before-and-after cross-sectionalcost study at two of the three pilothospitals found that PIH admis-sions decreased by 77 percent inthe six months following introduc-tion of the new guidelines com-pared to the previous six months(from 47 cases before to 11 after).This decrease is consistent with thenew, more stringent guidelines andindicates a high likelihood thatcompliance with the new guide-

Abstract Continued

lines caused the decrease. The coststudy measured PIH-related directcosts for inpatient and outpatientcases. The former included the costof hospitalization (clinical services,food), drugs, lab tests and otherparaclinical services. The latter in-cluded antenatal care-related costsand costs associated with PIH andother conditions: drugs, lab and otherparaclinical care, and medical consul-tations.

Total direct inpatient-related costsdecreased by 86 percent, from about51,000 rubles in the Before group toabout 7,000 rubles in the After group.This is an annualized savings of about118,000 rubles (about US$ 4,720at the time of the study). Direct per-inpatient costs decreased 41 percent.As expected, costs were higher forwomen with more severe PIH, but asubstantial decrease was evident atall severity levels following the intro-duction of the new guidelines, withthe larger decreases occurring inthe more severe cases. Length ofhospital stay for PIH women alsodecreased, on average from 13.5 to11.8 days. Findings also suggest thatoutpatient costs potentially associatedwith PIH care also dropped by roughly13,000 rubles per year, although aninability to separate PIH costs fromregular antenatal care costs makes itdifficult to pinpoint the decrease withcertainty.

Although the number of cases in eachstudy group is small, and numerousassumptions are made in the analy-sis, no substantial threats to the valid-ity of these findings are apparent. Theunusual result of having health out-comes improve at substantially loweroperating costs recommends widerapplication of the demonstrationproject.

Acknowledgements

This paper was written by Hany Abdallah, Olga Chernobrovkina, AnnaKorotkova, Rashad Massoud, and Bart Burkhalter. The Health Depart-ment of the Tver Oblast under the stewardship of Dr. Alexander N.Zlobin, Torzhok Hospital, and the Vyshny Volochyok Hospital greatlysupported and participated in the study.

The staff of Torzhok Hospital and Vyshny Volochyok Hospital signifi-cantly facilitated the coherent and successful execution of the study.Special thanks to Marina Scheglova, Economist with the Tver Fund.The authors also recognize and appreciate the special effort and workof: Dr. Valeria Gerasimova, Obstetrician-Gynecologist (Ob/Gyn), Dr.Andrey Gulin, Ob/Gyn, and Irina Kutilina, Hospital Economist, TorzhokHospital; Dr. Michail Klimov, MD and Head of Perinatal Care, Dr.Natalia Suchova, Ob/Gyn, and Olga Klueva, Hospital Economist,Vyshny Volochyok.

Paula Tavrow and Jolee Reinke from the University Research Co.,LLC,-Center for Human Services (URC-CHS) provided extensivetechnical input. Special thanks to Marc Oliver for his unwaveringencouragement, and Beth Goodrich who edited this paper withdiligence and a ready smile.

While we acknowledge the valuable contributions of the partiesmentioned above, all errors of omission and/or interpretation remainthe sole responsibility of the authors.

Recommended citation

Abdallah, H., O. Chernobrovkina, A. Korotkova, R. Massoud, and B.Burkhalter. 2002. Improving the Quality of Care for Women withPregnancy-Induced Hypertension Reduces Costs in Tver, Russia.Operations Research Results 2(4). Bethesda, MD: Published for theUnited States Agency for International Development (USAID) by theQuality Assurance Project.

About this series

The Operations Research Results series presents the results ofcountry or area research that the QA Project is circulating to encour-age discussion and comment within the international developmentcommunity. If you would like to obtain the larger research report of thisstudy containing all relevant data collection instruments, pleasecontact [email protected].

Improving the Quality of Care Reduces Costs ■ 1

Abbreviations

AHRQ Agency for HealthResearch and Quality,Tver Oblast Departmentof Health

BP Blood pressure

C-section Cesarean section

EBCG Evidence-Based ClinicalGuidelines

Mg SO4 Magnesium sulfate

Ob/Gyn Obstetrician-gynecologist

PIH Pregnancy-inducedhypertension

QA Quality assurance

QI Quality improvement

RFMOH Russian FederationMinistry of Health

SOW Scope of work

USAID United States Agencyfor InternationalDevelopment

USSR Union of Soviet SocialistRepublics

Hany Abdallah, Olga Chernobrovkina, Anna Korotkova,Rashad Massoud, and Bart Burkhalter

Improving the Quality of Care for Womenwith Pregnancy-Induced HypertensionReduces Costs in Tver, Russia

I. Introduction

The Health Committee of theU.S.-Russia Joint Commission onEconomic and TechnologicalCooperation defined the scope ofwork (SOW)1 in 1998 for cooperativeactivities to improve quality. Theactivities included the implementa-tion of a demonstration project toimprove the quality of maternal andchild health services in Tver Oblast.Funded by the U.S. Agency forInternational Development (USAID),the Quality Assurance (QA) Projectprovided technical assistance tothose undertaking the SOW, includ-ing the development of the demon-stration project. A collaboration wasformed and included the Ministry ofHealth of the Russian Federation(RFMOH), the Central Public HealthResearch Institute of the RFMOH,the Tver Oblast Department ofHealth’s Agency for Health Researchand Quality (AHRQ), the U.S.Department of Health and HumanServices, USAID, and the QAProject.

Key decision makers held a plan-ning meeting in Tver that included areview of existing data. Shortlythereafter, pregnancy-inducedhypertension (PIH)2 was selected as

the maternal health priority for thedemonstration project. As the singlelargest cause of maternal deaths inTver, PIH accounted for seven of the25 reported maternal deaths thathad occurred in Tver in the previousthree years (1995–97). Additionally,the diagnosis and management ofPIH was causing much confusionand difficulty. Interestingly, toxemiawas diagnosed in 18 percent ofpregnancies during 1997;3 theworldwide incidence of PIH is 5–6percent of pregnancies (Murray andLopez 1998).

The demonstration project waslaunched at three hospitals in TverOblast: Tver City Maternity Hospital#1, Vyshny Volochyok Hospital, andTorzhok Central District Hospital. Aspart of the improvement in thesystem of care, new guidelines forPIH care based on evidence-basedmedicine were developed andimplemented in these hospitals, andindicators were developed andtracked monthly using time seriescharts to show the effect of the newsystem on the quality of care.Section II describes the qualityimprovement effort and the effect ofthe new evidence-based guidelines.

A cost study was conductedconcurrently, during the spring of2000. It quantified the economicimpact of the improved system ofPIH care. This report documents thecost study’s methodology and

2 ■ Improving the Quality of Care Reduces Costs

findings. Documenting the cost ofquality improvement is important inpart because of the paucity of suchinformation and growing recognitionof the need for evaluations of theeconomic impact of quality improve-ment and evidence-based guide-lines (Eccles et al. 2000; Haycox etal. 1999).

II. The QualityImprovementIntervention

The QA Project/Russia team devel-oped the new system of PIH careusing a quality improvement meth-odology (Massoud et al. 2001;Langley et al. 1996) that includes aframework for clinical qualityimprovement (Batalden and Stoltz1993) and emphasizes the need tointegrate clinical content (subjectmatter knowledge) with the improve-ment methodology. The improvementmethodology consists of four phasesthat integrate a systematic six-stepprocess for developing evidence-based clinical guidelines (EBCG)(Tula Oblast Health Care Departmentet al. 2001; Tver Oblast Health CareDepartment et al. 2001a and b).4

Phase I of the improvement method-ology was to identify the problemneeding improvement. Key decisionmakers had agreed at the Tverplanning meeting to work onimproving the system of care forwomen suffering from toxemia/PIH.They chose three hospitals repre-senting the different types of ruraland urban facilities that providetoxemia/PIH care in Tver to partici-pate in the demonstration project:Tver City Maternity #1, VyshnyVolochyok Hospital, and TorzhokRayon Hospital. A team of thedifferent professional functions

involved in the system of PIH carewas formed at each hospital. Theteams included obstetricians,gynecologists, other physicians,midwives, and nurses. All teammembers were trained in qualityimprovement concepts and methodsbefore embarking on the improve-ment process.

Phase II was to analyze the existingsystem of toxemia/PIH care. Thesystem analysis was done at eachfacility and included the referralsystems (including ambulanceservice) connecting the facilities. Foreach hospital, the teams flowchartedthe entire PIH care system. Nextthey documented existing clinicalpractices at each step in theflowcharted processes, includingclinical diagnostic criteria, referralcriteria, and treatment guidelines.This analysis revealed the differ-ences in practice among thehospitals: inconsistencies in diagno-sis and treatment, as well asorganizational problems. What hadbeen thought to be a standardizedsystem of care proved to havesignificant variation. Importantly, themain strategy for treating PIH was toadmit women to the hospital andthen attempt to control bloodpressure while maintaining thepregnancy. This approach contrastssignificantly from that of the westernworld, which stresses more stringentadmission criteria and more aggres-sive treatment of those admitted.The teams also developed indica-tors to show changes that mightresult from their efforts. Starting withJanuary 1998, the teams collecteddata retrospectively from records athospitals and women’s clinics. Thedata were collected for the totalpopulation (no sampling), aggre-gated monthly, and displayed ontime series graphs. Two differentdatabases were set up, one for data

from hospital records and thesecond for data from the associatedwomen’s clinics.

Phase III of the improvementmethodology was to developinterventions that the teams believedwould yield improvement. First, theteams reviewed the availableevidence-based literature on PIHcare, using special training sessionsabout the existing evidence base inthe literature where appropriate.Second, to update the clinicalcontent in accordance with the bestavailable evidence and to achieveagreement on clinical contentchanges, several clinical review andpolicy dialogue sessions were heldover a period of some six months.These changes adopted recommen-dations from the 10th InternationalClassification of Diseases (ICD-101992). Also adopted was a radicalchange to manage PIH by inducingearly delivery and preventingeclampsia with intravenous magne-sium sulfate (Figure 1). Third, planswere developed to enhance thecapacity of the systems of care toimplement the updated clinicalcontent. New flowcharts were drawnshowing the proposed changes.Fourth, the indicators were reviewedand updated to ensure that theymeasured the clinical and organiza-tional changes introduced.

Phase IV was to test and implementthe changes agreed upon and toshow whether they actually yieldedimprovement. This was accom-plished using the “Shewhart Cyclefor Learning and Improvement,” alsoknown as the Plan-Do-Study-Act(PDSA) Cycle (Shewhart 1931). Theteams pilot tested the proposedinterventions for a few months to seeif they were feasible and worked.Data from the proposed interven-tions convinced the teams that the


Figure 1

Major Differences between Old and New Guidelines for TreatingPIH in Hospitals

Old

Many pregnant women diagnosed with toxemia

Poly-drug therapy

Inpatient management of hypertension

Prolonged pregnancy

New

Fewer pregnant women diagnosed with PIH

Mono-drug therapy (magnesium sulfate: Mg SO4)

Prevention of eclampsia

Early and induced delivery

proposed interventions worked: theteams observed improved care forpatients admitted and managed inthe hospital and saw no complica-tions in those not admitted.

The hospitals began implementingthe new standards during thedevelopment process: Torzhok andVyshny Volochyok hospitals startedfirst, followed by Tver City Maternity.The director of the Tver OblastDepartment of Health and the chiefobstetrician/gynecologist (Ob/Gyn)of the RFMOH officially authorizedthe three demonstration hospitals toimplement the new system of care inSeptember 1999. A conference inlate September officially launchedthe new system and was attendedby all facility administrators, clinicalchiefs, and others involved in theproject (Ethier forthcoming).

III. Effects on HealthOutcomes

Several indicators of health outcomewere investigated at the threedemonstration hospitals. Theseincluded: (1) number of toxemia/PIHcases progressing to eclampsia, (2)number of toxemia/PIH deaths, and(3) number of newborns withcomplications to toxemic/PIH

women. The study team collecteddata on cases progressing toeclampsia and/or resulting in deathover the 19 months after implemen-tation of the new guidelines (Octo-ber 1999–April 2001). Data wereobtained from the monitoring systemdeveloped as part of the new qualityassurance system for PIH care.Because such data were notcollected during the period beforethe new guidelines, data on deliver-ies to PIH women and newborncomplications were obtained fromhospital records during a six-monthperiod before the new system wasimplemented (January 1999–June1999).

The following results were found:

1. No cases progressed to eclamp-sia among the women managedunder the new system of care.This is the most importantindicator reflecting that the newsystem of care is an improve-ment.

2. No deaths occurred among thewomen managed in the newsystem of PIH care.

3. Complications of newborns ofwomen with PIH dropped 60percent.

IV. Cost Study Designand Methods

The cost study used a before-and-after cross-sectional design. Costdata were collected from two of thethree hospitals that participated inthe demonstration pilot: TorzhokHospital and Vyshny Volochyok.Data were collected retrospectivelyfrom medical records over a 15-month period (January 1999–March2000). The Before data werecollected for months 1 to 6 (Janu-ary–June 1999), the new guidelineswere implemented in months 7 to 9(July–September 1999), and theAfter data were collected in months10 to 15 (October 1999–March2000).5 Cost data were not collectedfrom Tver Maternity Hospital be-cause it had delayed implementationof the new guidelines.

A. Inpatient CostsTo estimate inpatient costs, wereviewed medical records for allpregnant women who were admittedwith suspected PIH of any severityto the two hospitals during the studyperiod. Data were obtained only forelements of maternity care that wereexpected to change as a result ofthe new guidelines. Elements thatwere expected to change includeddrugs, laboratory tests, and directcare associated with hospitalization.Hospitalization care includedphysician and nursing time, andfood. The cost of all these elementsis referred to as “direct costs.”

Costs not included because we didnot expect the new guidelines toaffect them included imputed rentfor infrastructure (buildings andequipment), hospital supplies andmaterials, special procedures (suchas C-sections), and hospital staffother than doctors and nurses.


We used the following procedures toestimate the direct cost of the threeelements of inpatient care:

Drugs. First, we examined each PIHpatient’s medical record to deter-mine the number of doses of eachtype of drug she had received.Second, we estimated the averagecost-per-dose of a particular type ofdrug by finding the average pricepaid for that type of drug by eachhospital during the Before and Afterperiods. The average for eachhospital was assumed to be themidpoint between the averageBefore price and the average Afterprice for that hospital. Thus, for eachhospital the cost per dose of aparticular type of drug was assumedto be the same in the Before andAfter groups. Third, the number ofdoses was multiplied by the esti-mated average cost per dose ineach hospital to obtain the total costto the hospital of each type of drugfor each patient.

A special study estimated drugcosts borne by the patient (and/orher family) at Vyshny Volochyok.This information was used to adjusttotal drug costs in both hospitals toinclude costs borne by the hospitaland patient. The adjusted costswere summed across drug types,patients, and hospitals to obtain totaldrug costs in the Before and Aftergroups. The total cost of drugs ineach group was divided by thenumber of PIH patients in the groupto obtain the average drug cost perpatient in each group.

Laboratory tests. Lab test costsinclude the costs of tests andparaclinical services such as bloodexams, urinalyses, and electrocar-diograph readings. The procedureused to estimate the total cost oflaboratory tests for PIH patients andthe average cost per patient was

similar to the procedure used fordrugs, except that no patient-bornecosts were involved. We obtaineddata on the number of tests by typefrom each patient’s medical record;we estimated the average cost pertest in each hospital from hospitalfinancial records; and we calculatedtotal costs by summing across typesof tests and patients. The cost pertest was based on the cost ofmaterials and supplies used toperform the tests and services.

Hospitalization. Hospitalization hererefers to the elements of care thatare directly associated with occupy-ing a bed in the hospital’s maternityward; their cost is roughly propor-tional to the number of beds occu-pied. In this study, we summed thecost of three elements of carethought to be influenced by the newguidelines (physician time, nursingtime, and food) to obtain an indica-tor of the direct cost of hospitaliza-tion. To obtain an estimate of thedirect hospitalization cost per dayper bed, we obtained the annualbudget for the three elements duringthe Before and After periods fromthe financial records of eachhospital. The budget reports yieldeda total annual direct cost of hospital-ization for each hospital and eachperiod. The total direct cost wasdivided by 365 and by the numberof maternity beds in the hospital toobtain the average daily hospitaliza-tion cost per bed.

B. Outpatient CostsAlthough the study team believedthat outpatient costs for PIH patientswould not change appreciably as aresult of the new guidelines, thepotential for questions aboutoutpatient costs led us to collectdata on such costs. While screening

for PIH occurred during regularantenatal care visits both before andafter the implementation of the newguidelines, under the new guide-lines, no special care or attentionwas prescribed for PIH at theoutpatient level. Patients whoshowed signs of slightly elevatedblood pressure (mild hypertension)but who tested negative for proteinin the urine (proteinuria) wereconsidered at risk for developingPIH under the new guidelines andwere typically asked to return to theoutpatient facility within a few daysfor follow up. Any patient with thisprofile without easy or immediateaccess to an emergency facilitywould be referred for admission tothe hospital.

Prior to the implementation of thenew guidelines, treatment at theoutpatient level for pregnant womenconsidered at risk for toxemia/PIHwas more intensive. In the oldsystem, patients typically receivedmore attention, including moremonitoring and treatment (both athome and in the outpatient facility).In addition, some women who wereconsidered at risk also receivedreferrals for hospital admission—upto two or three admissions perpregnancy.

Information about outpatient care forwomen with toxemia under the oldsystem is mainly anecdotal, so it isdifficult to distinguish between carethat was provided as part of regularantenatal care visits and toxemia/PIH care. Consequently, informationon all elements of care received bythe pregnant woman at the outpa-tient level was collected, includingall regular antenatal care. Threeelements of outpatient care wereinvestigated: drugs, paraclinicalservices (mainly lab tests), andmedical consultations (e.g., with the


Ob/Gyn). The cost of each elementwas obtained in the same manner asthe corresponding elements ofinpatient costs. The cost of medicalconsultation was assumed to beproportional to the number ofoutpatient visits, with the cost pervisit equal to the per unit cost usedin the hospital budget reports. Otheroutpatient costs were not estimated,including costs associated withundetected PIH cases, if there wereany, that were not admitted undereither set of guidelines. We assumedthat all of these unmeasured costswere constant in the Before andAfter periods.

The outpatient medical records of allwomen were screened to identifycases with a toxemia/PIH diagnosisin the Before and After periods. Thisincluded women admitted to thehospital, those treated only asoutpatients, and those who werereferred to a hospital as possiblePIH cases but classified as notbeing at risk following referral.Women in the After period who werenot diagnosed with PIH, but whowould have been treated for toxemiaunder the old guidelines, were notincluded as a PIH case. We as-sumed that PIH had no effect on theoutpatient treatment and cost ofcases not identified by this screen-ing. We abstracted and analyzedoutpatient cost data for the PIHcases identified in this screening. Aswith the inpatient analysis, theanalysis of the outpatient dataassumed that the number andseverity of cases was the same inthe Before and After periods, andthat the pattern of regular antenatalcare visits was the same.

A team made up of an Ob/Gynresponsible for treating the admittedwomen and a hospital economistwas responsible for measuring and

allocating costs to the elements ofcare. Data were analyzed usingMicrosoft Access and Excel soft-ware.

V. Cost Study Findings

A. Number of PIH CasesInpatient cases. In the two studyhospitals combined, 47 PIH caseswere admitted in the six months priorto the implementation of the newguidelines and only 11 in the sixmonths after, a 77 percent drop. Thisreduction in admissions indicatesthe degree to which the newguidelines were successfullyimplemented (see Section VI onassumptions and threats to validity).

As shown in Table 1, a largerproportion of the admitted cases inthe After group presented withmoderate to severe cases of PIH (5of 11: 45 percent) than in the Beforegroup (14 of 47: 30 percent). Wehypothesized that the increasedproportion of severe cases wouldincrease the average cost andexamine our hypothesis below.

Outpatient cases. We found 56records of women with PIH who

received outpatient care, 47 in theBefore period and 9 in the Afterperiod (Table 2). Twelve patients (9Before and 3 After) diagnosed withPIH were not hospitalized. These 12included patients with mild forms ofPIH. Note that some women admit-

Table 1

Number of Hospitalized PIH Cases Included in the Cost Study,by Severity and Hospital

Before After Total

Vyshny Vyshny (PercentageSeverity Torzhok Volochyok Total Torzhok Volochyok Total Reduction)

Mild 19 14 33 1 5 6 39 (82%)

Moderate 8 3 11 0 3 3 14 (73%)

Severe 3 0 3 1 1 2 5 (33%)

Total 30 17 47 2 9 11 58 (77%)

Table 2

Number of Cases with PIHSeen as Outpatients,Including Cases Not

Admitted

PIH Patients Seen as Outpatients Before After

Admitted for hospital care 38[1] 6[2]

Not admitted for hospital care 9[3] 3

Total 47 9

Notes:

1. No records of outpatient care were found for 9 ofthe 47 hospitalized women in the hospitalizedBefore group (reported in Table 1). Thesepatients were most likely hospital walk-ins.

2. Similarly, no records were found for 5 of the 11hospitalized patients in the After group.

3. It is a coincidence that in the Before period, thenumber of women admitted for PIH who had notreceived outpatient care (9) equals the numberdiagnosed with PIH receiving outpatient care but

not admitted for PIH.


ted with PIH do not appear to havereceived any outpatient care for PIHor for any other reason: 9 of the 47women admitted with PIH in theBefore period and 5 of the 11women in the After period did notreceive outpatient care.

B. Reduction in Direct Costsof Care for PIH Patients

Inpatient careThe total economic burden oftreating pregnant women admitted tothe hospital for PIH decreasedfollowing the implementation ofclinical guidelines for PIH (Figure 2and Table 3). Total direct inpatientcosts for six months decreased from51,247 rubles in the Before period to7,071 rubles in the After period or 83percent of Before costs in the twohospitals.6 Recall that the directinpatient costs reported here reflectonly the costs likely to change as aresult of the new guidelines and notall the costs incurred by the hospitalto treat PIH patients. Thus thesavings of 44,176 rubles is likely tobe a reasonable estimate of the PIHcosts saved by the two hospitals asa result of the new guidelines, buttotal costs for treating PIH inpatientswould not decrease by 83 percentbecause the total costs includecosts not affected by the guidelines.

The large drop in the cost of PIHinpatient care was driven by twofactors: a 76.6 percent decrease inthe number of patients admitted forPIH, and a 41.0 percent decrease inthe cost per inpatient (Table 3). Thedecrease in average-per-inpatientcost resulted from decreases in allthree elements of inpatient care; thedecreases in drug and hospitaliza-tion costs were much larger than thedecrease in paraclinical servicescosts.

The distribution of inpatient costsamong the three elements did notvary significantly between the twogroups (Table 4). Hospitalizationaccounted for the largest portion ofdirect inpatient costs in both theBefore and After groups (69 percentand 71 percent, respectively). Thevariation in the distribution of costsby case severity is discussed in alater section.

Table 3

Direct Cost of Inpatient PIH Care in Two Tver Hospitals beforeand after New PIH Guidelines[1]

Elements of Care

Hospitalization

Drugs

Paraclinical care (tests)

Total inpatient care

Percentage ChangeBefore (n = 47) After (n = 11) (n decreased 76.6%)

Average Total Average Total Average Totalcost / direct cost / direct cost / directpatient cost[1] patient cost[1] patient cost[1]

744.4 34,987 457.1 5,028 - 38.6% - 85.6%

223.8[2] 10,517 83.7 921 - 62.6% - 91.2%

122.2 5,743 102.0 1,122 - 16.5% - 80.4%

1,090.4 51,247 642.8 7,071 - 41.0% - 86.2%

Notes:

1. All costs are in rubles. Total costs are for a six-month period in the Before and After columns and are subjectto rounding error.

2. Average is based on 46 patients for whom drug data were collected; this average was projected to the 47thpatient for whom data were missing.

Figure 2

Change in Total Cost ofInpatient Maternal Care forPIH Patients (Rubles) over

6-Month Periods

Before After

Test

Drugs

Hospitalization

Million Rubles55

50

45

40

35

30

25

20

15

10

5

0

Table 4

Distribution of the TotalDirect Inpatient Cost of

PIH Treatment in the TwoHospitals Combined

Elements of Care Before After

Hospitalization 69% 71%

Drugs 20% 13%

Paraclinical care (tests) 11% 16%

Total inpatient cost 100% 100%


Outpatient careTable 5 gives total outpatient carecosts for all women treated for PIH(from Table 2) during the studyperiods. The outpatient costs reflectregular antenatal care plus PIH-related and other outpatient carereceived during the six-monthBefore and After study periods. Thecosts of outpatient care for womenwho received such care and wereadmitted for PIH averaged 409rubles per patient before the newguidelines and 334 rubles after. Theaverage per-patient costs of outpa-tient care for women who were nothospitalized dropped from 621rubles to 266. This difference isstatistically significant at the 0.01level. However, the differencesbetween the 409 and 334 figuresand between the 334 and 266figures are not statistically significantusing a 2-tailed t-test. Overall therewas a decrease of 30.7 percent inthe per-patient cost of outpatientcare for patients with a toxemia/PIHdiagnosis.7

Because of the difficulty of separat-ing costs associated with regularantenatal care from costs that arespecifically related to treatment forPIH, it is difficult to project whatproportion of the decrease in thecost of outpatient care is due to theimplementation of the new guide-lines. However, in the absence ofany information on change in theantenatal care program for women8

during the period of the study, andbecause the data on outpatientdrugs and lab tests among the 38admitted patients who receivedoutpatient care in the Before periodshow PIH-related treatment, it islikely that a large part or all of the30.7 percent decrease in per-patientcost was due to the introduction of

the new guidelines. This finding isconsistent with the expected impactof the new guidelines on outpatientcare. In the Before group, womendiagnosed with toxemia receivedmore intensive attention andtreatment as outpatients (includinghome-based care) than those in theAfter group where the new guide-lines redefined the disease classifi-cation for PIH and hence thetreatment provided. In the Aftergroup, only “border” PIH cases, i.e.,those women with elevated bloodpressure (BP) and no protein in theirurine, received any “special”outpatient attention. They weretypically asked to return sooner thantheir next scheduled antenatal visitto receive BP and urine protein tests.The rest, i.e., women diagnosed withPIH including proteinuria during anantenatal visit, immediately becameinpatients.

Based on the above findings andassuming that the cost of PIH-related outpatient care is zero in theAfter group, the cost of PIH-relatedcare at the outpatient level in theBefore group is estimated to be

about 6,500 rubles over the sixmonths in the Before group. This isan annualized saving of 13,000rubles for the two hospitals. Whenthe direct PIH-related outpatientcosts are combined with the inpa-tient costs in Table 3, total costs inthe six-month Before period equal57,747 rubles (51,247+6,500)compared to 7,071 in the Afterperiod. This is an 87.7 percent drop.

The distribution of the cost ofoutpatient care changed onlyslightly following the implementationof the new guidelines (Table 6). Thecost of consultations becamerelatively more important (45 percentof total per patient costs versus 41percent in the Before group), whilethe cost of drug treatments becameslightly less important (from 37percent to 24 percent of total perpatient costs). The relative cost ofparaclinical care did not changefrom the Before to the After period.

Table 5

Total Direct Cost of All Outpatient Care During Six-MonthBefore and After Periods for Patients Diagnosed with PIH

PercentageBefore After Change

Average Number Total Average Number Total AverageCost/ of Direct Cost/ of Direct Cost/

Patient Patients Cost Patient Patients Cost Patient

409.3 38 15,553 334.0 6 2,004 -18.4%

621.0 9 5,588 266.5 3 800 -57.1%

449.8 47 21,142 311.5 9 2,804 -30.7%

Category of patient

Admitted patients

Patients not admitted

Overall average

Notes: Costs include all outpatient costs whether or not related to PIH for women diagnosed with PIH and whoreceived outpatient care in the Before or After period. All figures are in rubles and subject to rounding error.


Table 6

Average Cost per PIH Patient of Outpatient Care,by Element of Care and Group, in Rubles and as a

Percentage of Total Average Cost per Patient

Group Drug Cost Paraclinical Cost Consultation Cost Total Cost

Before (n = 47) 122.5 (27%) 140.7 (31%) 186.6 (41%) 449.8 (100%)

After (n = 9) 75.6 (24%) 95.5 (31%) 140.5 (45%) 311.5 (100%)

Note: The total outpatient care cost by element and group was obtained first, and the average cost per patient byelement and group was then obtained by dividing the total cost for each cell by the number of patients (n) in thatgroup.

Direct cost of hospitalizationTotal PIH inpatient days droppedfrom 635 before the implementationof the new guidelines to 130 after-ward (a 79 percent reduction),largely as a result of the decrease inthe total direct cost of hospitalizationin both hospitals combined from34,987 rubles to 5,028 rubles.9 Theaverage direct cost of hospitalizationper PIH inpatient decreased from744 rubles to 457 rubles, a drop of39 percent (Table 7 and Figure 3).The drop in inpatient days wascaused by the 77 percent reductionin the total number of inpatients anda 13 percent reduction in theaverage length of stay per inpatient

Table 7

Average Direct Cost of PIH Hospitalization per Day and per Stayfor Both Hospitals Combined, before and after New Guidelines

(Rubles)

Before After Percentage Change

Average cost per day 55.1 38.7 -29.8 %

Average days per stay 13.5 11.8 -12.6 %

Average cost per stay 744 457 -38.6 %

Note: The average cost per day is based on the combined average cost of hospitalization in the two hospitals.Vyshny Volochyok has a lower per-day cost than Torzhok. While the average cost per inpatient in each hospitalwas assumed to stay constant during the course of the study, the change in average cost per day reported herereflects the relatively larger proportion of patients being admitted in Vyshny Volochyok hospital in the After groupas compared to the Before group (Table 1).

(from 13.5 to 11.8 days). Length ofstay varies significantly by caseseverity, as is addressed in a latersection.

Direct cost of drugs forPIH inpatientsTo analyze drug costs, we combinedthe cost of drugs provided by thehospital and those purchased by thepatient and her family. The total costdropped from 10,517 rubles in theBefore group to 921 in the Aftergroup (Table 3). This large decreasewas driven by the combination offewer admissions for PIH and lowerdrug costs per inpatient in the After

group. The average per-patient costdropped 63 percent (from 224 rublesbefore to 84 rubles after).

A special study of patient-paidinpatient drug costs at VyshnyVolochyok found that a higherproportion of inpatients in the Aftergroup (i.e., including both hospitals)paid for part of their own drugs thanin the Before group. In the Aftergroup 6 of the 11 inpatients (55percent) paid 59 percent of their

Figure 3

Reduction in the AverageDirect Cost per PIH Inpatient

of Hospitalization, Drugs,and Paraclinical Care forBoth Hospitals Combined

(Rubles)

Medication Tests TotalHospital-ization

Rubles per Patient1200

1000

800

600

400

200

0

BeforeAfter


drug costs, while in the Before group10 of the 47 patients (21 percent)paid 67 percent of their drug costs.The total cost of inpatient drugsborne by the patients represented13 percent of total inpatient drugcosts in the Before group (bothhospitals) and 28 percent in theAfter group (both hospitals).10 Theaverage cost of drugs borne by apatient in Vyshny Volochyok fell 68percent (from 131 rubles Before to42 rubles After: Figure 4). The mostimportant factor in reducing the costto the patient was a 60 percent dropin the average number of drugs paidfor by patients: from five to two.

Fewer types of drugs were usedfollowing the introduction of the newguidelines: 72 different types wereused across all severity levels in theBefore group and 32 in the Aftergroup. The average Before groupinpatient received ten differentdrugs, and the average After groupinpatient received six.

Table 8 illustrates the decreasedpolypharmacy. The six most fre-

Figure 4

Average Inpatient Drug CostBorne by Patients at

Vyshny Volochyok (Rubles)

Before After

131

42

quently used drugs in theBefore group were usedsignificantly less fre-quently or not at all in theAfter group. But otherdrugs (magnesium sulfate[Mg SO4], clofellini, andRinger solution) were usedmore frequently in theAfter group, in accordancewith the new clinicalguidelines. For example,the new guidelinesrecommend using Mg SO4to control convulsions. Inthe After group, a propor-tionately higher percent-age of severe PIH cases—which are the cases mostlikely to need Mg SO4—was admitted. We presentan analysis of costs bycase severity below.

Direct cost ofparaclinical care forPIH inpatientsThe total direct cost of paraclinicalcare including tests decreased from5,743 rubles to 1,122, a drop of 80percent (Table 3). The decrease inthe number of women admitted forPIH accounted for most of thisdecrease. The average direct cost ofall tests per inpatient also de-creased (Table 3 and Figure 3), butnot as dramatically as the costs ofother elements. Direct costsdropped from 122 rubles per patientto 102, a 16 percent decrease.

The average number of tests perinpatient increased by about twofollowing the implementation of thenew guidelines. This pattern held formild and severe PIH, but not formoderate, where the averagenumber of tests per inpatientdropped from nine to five. Several

Table 8

Most Frequently Used Drugs toTreat PIH: Percentage Receiving

Drug, by Group

Percentage of PIH InpatientsReceiving the Drug

Drug Before (n=47) After (n=11)

Papaverine hydrochloride 78% 8%

Glucose (40%) 69% 8%

Methylogobrevin 64% 23%

Dibasol 56% Dis

Novocaine 50% Dis

Diazepam 50% Dis

Magnesium sulfate 42% 46%

Clofellini 30% 31%

Ringer solution 23% 43%

Notes: Frequently prescribed drugs are defined as those given to50% or more of the PIH patients in either or both the Before andAfter groups. “Dis” means discontinued for PIH use under newguidelines, or 0%.

factors contributed to usage,including the type of test, thefrequency of use per patient, andthe unit cost of the test.

As illustrated in Table 9, frequentlyused tests in the Before group werealso frequently used in the Aftergroup, although most dropped off alittle, including electrocardiographs,a relatively expensive test. On theother hand, several tests infre-quently used in the Before groupbecame more popular after imple-mentation, especially relativelyexpensive specialty consultationswith ophthalmologists and thera-pists. The increased use of theproteinuria test was likely the resultof increased monitoring of PIHinpatients according to the newguidelines.


C. Differences betweenHospitals

An analysis of the direct inpatientcosts by hospital found largedifferences between them. Directcosts were substantially higher atTorzhok than at Vyshny Volochyokbefore the new guidelines, and thereduction in direct costs was alsolarger at Torzhok (Table 10). Torzhokhad strikingly larger decreases inboth the number of admissions and

Table 9

Frequently Used and Other Selected Tests and Consultations

Percentage of Patients in Group Receiving Test

Before (n=47) After (n=11)

Most Frequently Used

Urinalysis 100% 100%

Total blood count 98% 45%

Blood platelet and coagulation 91% 73%

Blood biochemistry 85% 82%

Electrocardiograph 42% 27%

Other Selected Tests

Ophthalmologist 2% 45%

Therapist 4% 36%

Proteinuria 2% 18%

Notes: These represent the five most frequently used tests. In Torzhok only, tests for AIDS (moderately expensivecompared to other tests) were administered to 63% of patients in the Before group and 18% of patients in theAfter group.

the average cost per admission,causing higher reductions there.Another difference between the twohospitals was the range of per-patient variation in direct costs inthe Before period, with Torzhokshowing a much larger spread thanVyshny Volochyok (Figure 5).

Although the analysis suffers fromsmall sample sizes at the individualhospital level, the pattern betweenthe hospitals is consistent across

the different factors contributing tocost, thereby according confidencein the results. This hospital-by-hospital analysis brings understand-ing about the cost savings that arelikely to be achieved if the newguidelines are scaled up.

Duration and change in length ofstay varied between hospitals. InTorzhok, average length of stay perpatient dropped from 14.7 daysBefore to 7.0 After. In VyshnyVolochyok it increased from 11.8days Before to 12.9 After. Most of theincrease at Vyshny Volochyok wasdue to an outlier who was admittedin the After period with a moderateseverity level.

In both hospitals, use of certaindrugs appears to have increased infrequency, although the type ofdrugs differed between the twohospitals (Table 11), and the small

Table 10

Comparison of Direct Costs of PIH Inpatients in theTwo Hospitals (Rubles)

Torzhok Hospital Vyshny Volochyok Hospital

Before After Change Before After Change

PIH admissions 30 2 - 93% 17 9 - 47%

Average direct cost 1,342 760 - 43% 618 617 - 0%per PIH inpatient

Total direct cost 40,272 1,520 - 96% 10,503 5,551 - 49%

Figure 5

Range and Average DirectCost per PIH Inpatient

(Before Period; in Rubles)

Torzhok Vyshny Volochyo

Rubles per Inpatient3500

3000

2500

2000

1500

1000

500

0

High-Low Range Average


Table 11

Percentage of PIH Cases Using Frequently Prescribed Drugs,[1] before and after Clinical Guidelines,by Case Severity and by Hospital

Torzhok[2] Vyshny Volochyok[3]

Mild Severe Mild ModerateBefore n = 19 After n = 1 Before n = 3 After n = 1 Before n = 14 After n = 5 Before n = 3 After n = 3

Glucose (40%) 100% 100%

Methylogobrevin 100% 100% 100% 100%

Papaverine hydrochloride 89% 100% 100% Dis 54% 20%

Folliculin 58% Dis

Oxytocin 53% 100% 100% Dis

Dibasol 47% Dis 100% Dis 54% 20%

Novocaine 100% Dis 69% Dis 67% Dis

Magnesium sulfate 100% 100% 23% 60% 100% 67%

Gendevit tablets 100% Dis 0% 40%

Reopolyglucine 100% Dis 0% 40% 67% Dis

Ringer solution 100% 100% 20% 67%

Euphyline 100% Dis 77% 20% 67% 33%

Glucose (5%) 0% 100%

Diazepam 100% 100%

Buxopan 0% 100%

Neostigmine methylsulfate 0% 100% 0% 100%

Relanium 100% Dis

Curantil 67% Dis

Ascorbic acid 67% Dis

Promedol 0% 100%

Glucose (20%) 77% 20% 67% 33%

Notes:

1. Frequently prescribed drugs are defined as those given to 50% or more of the PIH patients in either or both the Before and After groups.

2. This analysis could not be conducted for cases with moderate PIH in Torzhok because no moderate cases were found in the After group.

3. Similarly, no severe cases were found in the Before group in Vyshny Volochyok.

“Dis” means discontinued (from use for PIH under new guidelines) or 0%.


number of cases in the After groupsin each hospital makes before andafter comparisons speculative. InVyshny Volochyok, the drugsclofellini, gendevit tablets,reopolyglucine, and Ringer solutionwere used more frequently in theAfter group, mostly for mild cases,and Ringer solution for moderatecases. In Torzhok, the drugsneostigmine methylsulfate, buxopan,and promedol were used in the Aftergroup for mild cases where they hadnot been used in the Before group.Neostigmine methylsulfate was alsoused in severe cases in the Aftergroup where it had not been used inthe Before group. Use of Mg SO4 forconvulsions did not change atTorzhok. Overall, Vyshny Volochyokused fewer drugs on average andachieved a more significant reduc-tion in number and types of drugsused (50 percent reduction overall innumber and type) than Torzhok.

Unit costs of inputs were generallyhigher at Torzhok than VyshnyVolochyok, owing generally to ahigher cost of living in the Torzhokregion. Torzhok prices were higherfor hospital staff and food, drugs,and clinical tests. For example, thecost of meals per hospital day was10 rubles at Torzhok compared to 6rubles at Vyshny Volochyok. Simi-larly, the most frequently usedtest—urine analysis—cost 5 rublesat Torzhok and 4 at VyshnyVolochyok. The guideline-recom-mended drug Mg SO4 cost 1.42rubles per unit at Torzhok and 0.30rubles at Vyshny Volochyok.

D. Differences by CaseSeverity

Average direct cost per inpatientvaried according to the severity ofPIH in the Before group, as ex-

pected. Table 12 shows that in theBefore group, direct inpatient costsfor PIH were progressively higher inmore severe cases in all threeelements of inpatient care. Theaverage direct costs for mild,moderate, and severe PIH were 888;1,294; and 2,372 rubles, respec-tively. The average costs for mild,moderate, and severe PIH wereroughly equal in the After group(584, 726, and 695 rubles, respec-tively). Clearly, the average directinpatient cost dropped substantiallyfor all severity levels following theintroduction of the new guidelines,with larger decreases in the moresevere cases.

Prior to the study we thought that thenew guidelines might lead to higheraverage per-patient costs becausethe new guidelines recommendedadmitting fewer, higher-severity PIHpatients, who were expected to costmore. The data do not support thisexpectation. No significant costdifference exists under the newguidelines among the severitylevels, and although a greaterproportion of the admitted cases aremore severe than before the newguidelines, costs did not increase(Table 3). These conclusions are

very tentative because of the smallsample sizes.

Moderate and severe cases of PIHhad larger reductions in medicationcost than mild cases (Table 13). Animportant driver of this reductionwas the decrease in the averagenumber of different types of drugsadministered per patient (Table 14).While patients with severe cases oftoxemia/PIH received the largestnumber of different types of drugs,these patients also had the largestreduction in different types of drugsused. Significant reductions alsoappeared in the moderate cases.

Table 13

Average Cost of Medicationper PIH Inpatient for

Both Hospitals Combined,by Case Severity

Mild Moderate Severe

Before 139 349 625

After 74 76 125

Percentage - 47% - 78% - 80%change

Table 12

Average Direct Cost per PIH Inpatient in Rubles,for Both Hospitals Combined by Case Severity

Before After

Mild Moderate Severe Mild Moderate Severen = 33 n = 11 n = 3 n = 6 n = 3 n = 2

Hospitalization 654 810 1,412 430 563 379

Medication 139 349 625 74 76 125

Paraclinical care (tests) 95 146 335 80 86 191

Total 888 1,294 2,372 584 726 695

Element of Care


Reduced polypharmacy was presentat all severity levels (Table 15). Forexample, mild PIH cases reducedusage of all five of the most fre-quently used drugs, from roughlyone out of two patients receiving thedrug Before to one out of three orone out of six receiving it After. Theone exception is Mg SO4, whichincreased from one out of six Beforeto one out of two After, reflectingcompliance with the new guidelines.Similarly, among moderately severecases, the most frequently useddrugs Before were less frequentlyused or discontinued After. All ofthese findings about the polyphar-macy trends are based on verysmall sample sizes.

VI. Assumptions andPotential Threats toValidity

Several necessary assumptionsrelated to factors that potentiallyweaken the study design arediscussed below:

Decline in fertility rate. If thenumber of pregnancies in thepopulation served by hospitalsdeclined between the Before andAfter periods (the population or thefertility rate may have dropped), thedecline may have caused a drop inthe number or severity of PIH cases.Changes in the fertility rate mayhave been due to a long-term trend,seasonality, or monthly fluctuations.In fact, the fertility rate in Russia hasbeen declining since before the fallof the USSR, which mandated anexpanded role for women outsidethe home (Da Vanzo et al. 2000;Zakharov et al. 1996). However, thesize of this trend is very small overthe study period, so its effect on ourfindings is probably inconsequential.

Table 14

Number of Different Types of Drugs Used to Treat PIH by CaseSeverity (Both Hospitals Combined)

Mild Moderate Severe Overall

Total number of different types of drugs used for all patients

Before 55 51 33 72

After 21 15 14 32

Percentage change - 62% - 70% - 42% - 55%

Average number of different types of drugs used per patient

Before 8 14 23 10

After 5 6 8 6

Percentage change - 38% - 57% - 65% - 40%

Table 15

Percentage of PIH Cases Using Frequently Prescribed Drugs inBefore and After Groups, by Case Severity

(Both Hospitals Combined)

Mild PIH Moderate PIH Severe PIHBefore After Before After Before Aftern = 32 n = 6 n = 11 n = 3 n = 3 n = 2

Papaverine hydrochloride 75% 33% 82% Dis 100% 50%

Glucose (40%) 69% 17% 73% Dis

Methylogobrevin 59% 17% 73% 33% 100% 50%

Dibasol 50% 17% 100% Dis

Novocaine 50% Dis 100% Dis

Magnesium sulfate 16% 50% 100% 67% 100% 100%

Reopolyglucine 73% Dis 100% Dis

Relanium 82% 33%

Clofellini 18% 50%

Ringer solution 9% 67% 100% 100%

Euphyline 100% Dis

Glucose 100% Dis

Oxytocin 100% 50%

Diazepan 100% Dis

Notes: Frequently prescribed drugs are defined as those given to 50% or more of the PIH patients in either orboth the Before and After group. “ Dis” means discontinued for PIH use under new guidelines, or 0%.


Seasonality might also be a concernbecause three of the six months inthe Before and After periods are thesame month of the year and threeare not. We cannot estimate thepotential magnitude of a seasonalityeffect, but we do not believe it hadan important effect in this situation.

Decline in the incidence oftoxemia/PIH. If the incidence oftoxemia/PIH per pregnancy declinedduring the study period for somereason, this could account for atleast part of the decline in admittedPIH cases. However, there is noevidence that this is the case. Givencurrent understanding of risk factorsfor PIH in populations—mainlypositive family history, underlyingvascular disease, and first preg-nancy especially among adolescentmothers—no major changesaffecting these risk factors weredetected. Related to this assumptionis the assumption that the incidenceof pregnancies that were not carriedto term in this population has notchanged. No specific data werefound to contend this assumption.

Change in the definition of PIH. Ineffect, the new guidelines changedthe definition of PIH. We havecorrected for this by assuming thatthe same number of suspected PIHpatients presented in the Afterperiod as presented (and wereadmitted) in the Before period for thecost study. This could also influencethe measured health outcomeindicators, because some of thewomen not admitted in the Afterperiod (but who would have beenadmitted in the Before period) mayhave had deliveries with negativehealth outcomes that were notcounted in the After period. As far aswe know, this was not the case, butwe were unable to identify and trackthe care of these women.

Cost of non-included hospitalservices. The cost of other hospitalservices provided to women,specifically the cost of medicalprocedures, such as deliveries andC-sections, was not expected tochange significantly following theimplementation of the guidelines andwas therefore excluded from ourmeasures of cost. Several pointssupport this assumption: (a) the costof a delivery, which will occur in alladmitted cases, is in large partaccounted for in the cost of drugsand hospitalization already calcu-lated in the study for all severitytypes, and (b) the rate of C-sectionsamong women with PIH does notappear to have changed signifi-cantly: one case was observed ineach group.

Other interventions for inpatientcare. We assumed that the ob-served changes in PIH inpatienttreatment patterns were all due tothe new guidelines. We know of noother intervention that influencedPIH treatment. Further, it is plausiblethat the new guidelines caused theobserved changes because theywere implemented with a highdegree of compliance.

Patient-borne costs. We measuredpatient-borne PIH drug costs in onehospital in both the Before and Afterperiods and assumed that the samepattern of patient-borne costsapplied in the second hospital.These costs were added to thehospital-borne drug costs to obtaintotal drug costs. On one hand, thesepatient-borne drug costs overstatethe amount actually paid by thehospital. On the other hand, if thereare other patient-borne costs that wehave not accounted for (e.g.,informal payments to providers forservices and the provision ofmedical supplies and foods by

patients and their families), theycould increase the direct costs,which might alter the reportedsavings.

VII. Conclusion andDiscussion

When the new guidelines for PIHcare were introduced in the pilothospitals in Tver, health outcomesimproved and hospital costsdecreased. This result seems almosttoo good to be true. However, thelogic of the new guidelines supportsthese results. The new guidelinesfocus inpatient resources on fewerand more severe cases of PIH, treatthose cases more aggressively, andrely on systematic outpatient care foridentifying and triaging the PIHcases that need close attention.

Three factors contributed most to thereduction in inpatient costs: reduc-tion in the number of hospitalizedcases, reduction in length ofhospitalizations, and more rationaluse of drugs. While the relativelysmall number of cases analyzed inthe study makes it difficult to drawfirm conclusions, the findingsstrongly suggest that using the newguidelines is associated withsignificant economic benefits.

In the short run, further analysis ofthe effect of using clinical guidelineson costs at the outpatient level maybe valuable. Preliminary resultssuggest that outpatient costs havebeen rationalized following theimplementation of the guidelines,probably resulting in a smalldecrease in costs. However, moreinformation on the patterns of carefor PIH outpatients is needed todetermine the nature and magnitudeof the effects of the guidelines on


Endnotes1 The Scope of Work included the following tasks: (a) Reach

consensus on concepts and terminology in quality of healthcare and publish a glossary, (b) Train a team of senior Russianhealth care professionals in quality improvement, (c) Developtwo demonstration projects that introduce quality improvementmethods in the Russian Federation (one in primary care in TulaOblast; the other in maternal and child health in Tver Oblast),(d) Develop indicators of quality to measure improvements inquality of care, (e) Publish a Russian language qualityimprovement guide, and (f) Disseminate success stories ofquality improvement in Russia.

2 PIH is a complication of pregnancy characterized by elevatedblood pressure and sometimes by protein in the urine andedema after the 20th week of gestation. Its causes are not wellunderstood. During pregnancy, PIH can progress to pre-eclampsia, characterized by headaches, vomiting, impairedvision, abdominal pain, and further increases in bloodpressure, urine protein, and edema. In the most severe cases,pre-eclampsia progresses to eclampsia, characterized byconvulsions. If untreated, eclampsia can result in death frombrain hemorrhage or from failure of the heart, liver, or kidney.Pre-eclampsia is associated with increased perinatal deathsand increased newborn complications, especially if untreated.Risk factors for PIH include first pregnancy especially amongadolescents, overweight, family history, and underlyingvascular diseases including diabetes.

3 “Toxemia” is the diagnostic classification generally used inRussia in place of PIH. Toxemia typically includes all PIH casesplus others. In this paper, the term “toxemia/PIH” refers towomen classified as having toxemia prior to the new guidlinesand women classified as having PIH after the guidelines wereimplemented.

4 These references present examples of how this integratedmethodology was applied in different clinical areas, includingPIH, by the QA Project. The six steps in the EBCG develop-

care. Still, findings are encouragingand consistent with expected resultsof using the new guidelines.

More data and analysis are neededto confidently project the potentialeconomic benefit of the newguidelines to the oblast level or onthe long-term costs. Differences inthe existing and potential patterns oftreatment in different hospitals

should be considered and betterunderstood to make projectionsrealistic. Also, the benefits obtainedfrom improved health outcomes forwomen with PIH and their newbornsremain to be further studied andquantified. For example, additionalinformation on the impact of the newguidelines on reduced drug interac-tions from reduced polypharmacy,on potentially reduced sequelae,

and on future social and economicbenefits would help position thisintervention in relation to otherproposed programs. Analysis offactors that influence care practices,and by extension the extent of useor non-use of guidelines, would alsohelp to strengthen future analyses(e.g., factors related to the charac-teristics of patients, providers, orhospitals).

ment methodology are: (a) study the existing system of care forthe improvement under consideration, (b) document existingclinical practices and content at each step, (c) review evi-dence-based literature on content of care, (d) update clinicalcontent of care, (e) enhance system capability to accommo-date updated content, and (f) review and update indicators tofit updated content.

5 The cost study data were not collected for as long as the“Health Outcomes” data described in Section III.

6 At the time of this study, the exchange rate for $1.00 U.S.ranged from 21.5 to 28.6 rubles.

7 There are two other groups of women who delivered during thestudy periods: those who were hospitalized with PIH but didnot receive outpatient care of any sort, and those who were nothospitalized with PIH and did not receive any outpatient care. Amuch larger proportion of admitted women in the After groupdid not receive outpatient care of any sort (5 of 11 or 45percent) than in the Before group (9 of 47 or 19 percent). Wedo not know the reason for this.

8 Including access to antenatal care services.9 At each hospital the estimated cost per inpatient day is

assumed to be constant throughout the study, and therefore thechange in total direct cost between the Before and Aftergroups is proportional to the change in the number of PIHinpatient days (Total cost = inpatient days x cost per day).However, the average cost per day is not the same in the twohospitals, and therefore the total cost in both hospitalscombined did not drop by the same proportion as the numberof days.

10 Looking at only Vyshny Volochyok patients, total inpatient drugcosts borne by patients represent about 56 percent of totalinpatient drug costs in the Before group and about 30 percentof inpatient drug costs in the After group.


References

Batalden, P.B., and P.K. Stoltz. 1993. A framework for thecontinual improvement of health care: Building andapplying professional and improvement knowledge totest changes in daily work. Joint Commission Journal onQuality Improvement 19(10):424–52.

Da Vanzo, J., and C. Grammich. 2000. Barren ground:Eastern Europe’s transition from communism isn’t theonly factor affecting the region’s demographics.Transitions Online [downloaded January 2001 fromwww.rand.org/popmatters/TOL.html].

Eccles, M., J. Mason, and N. Freemantle. 2000. Develop-ing valid cost effectiveness guidelines: Methodologicalreport from the North of England Evidence BasedGuideline Project. Quality in Health Care 9:127–32.

Ethier, K. Forthcoming. Developing Evidence-BasedStandards for Pregnancy-Induced Hypertension inRussia. Quality Assurance Project Case Study.Bethesda, Maryland: To be published for the U.S.Agency for International Development (USAID) by theQuality Assurance Project.

Haycox, A., A. Bagust, and T. Walley. 1999. Clinicalguidelines – The hidden costs. British Medical Journal318:391–93.

ICD-10. 1992. International Statistical Classification ofDiseases and Related Health Problems (ICD-10) – TenthRevision. Geneva, Switzerland: World HealthOrganization.

Langley, G.L., K.M. Nolan, T.W. Nolan, C.L. Norman, andL.P. Provost. 1996. The Improvement Guide: A PracticalApproach to Enhancing Organizational Performance.San Francisco, California: Jossey-Bass Publishers.

Massoud, R., K. Askov, J. Reinke, L. M. Franco, T.Bornstein, E. Knebel, and C. MacAulay. 2001. AMethodology for Improving Healthcare Quality. QAMonograph Series 1(1). Bethesda, MD: Published forthe U.S. Agency for International Development (USAID)by the Quality Assurance Project

Murray, C.J.L., and A.D. Lopez, eds. 1998. Global Burdenof Disease and Injury Series. Vol. III, Health Dimensionsof Sex and Reproduction. Published by the HarvardSchool of Public Health on behalf of WHO and the WorldBank, and distributed by Harvard University Press.

Shewhart, W. 1931. The Economic Control of Quality ofManufactured Products. New York: D. Van Nostrand Co.Reprinted by the American Society of Quality Control in1980.

Tula Oblast Health Care Department, Central Public HealthResearch Institute of the Ministry of Health of theRussian Federation/The Eurasian Medical EducationProgram – The American College of Physicians, andQuality Assurance Project – University Research Co.,LLC/Center for Human Services, Bethesda, MD, USA.2001. Improving the System of Care for PatientsSuffering from Arterial Hypertension. Financed by theUnited States Agency for International Development.

Tver Oblast Health Care Department, Central Public HealthResearch Institute of the Ministry of Health of theRussian Federation, and Quality Assurance Project –University Research Co., LLC/Center for HumanServices, Bethesda, MD, USA. 2001a. Improving Carefor Women Suffering from Pregnancy-Induced Hyperten-sion. Financed by the United States Agency for Interna-tional Development.

Tver Oblast Health Care Department, Central Public HealthResearch Institute of the Ministry of Health of theRussian Federation, and Quality Assurance Project –University Research Co., LLC/Center for HumanServices, Bethesda, MD, USA. 2001b. (Re)Designingthe System of Care for Neonates Suffering from Respira-tory Distress Syndrome. Financed by the United StatesAgency for International Development.

Zakharov, S.V., and E. Ivanova. 1996. Fertility decline andrecent changes in Russia: On the threshold of thesecond demographic transition. In: J. Da Vanzo and G.Farnsworth (eds.) Russia’s Demographic Crisis. SantaMonica, CA: RAND CF 124 [downloaded January 2001from www.rand.org/publications/CF/CF124/CF124.chap.2.html].

quality assurance project - usaid assist...with pregnancy-induced hypertension abstract reduces...

Documents