peripheral arterial obliterative disease

Peripheral Arterial Obliterative DiseaseCost of Illness in France

Armelle Montron,1 Eric Guignard,1 Alain Pelc1 and Sylvie Comte2

1 Intercontinental Medical Statistics (IMS), Health Economics Department, Nanterre, France2 Department of Pharmacoeconomics, LIPHA Laboratories, Lyon, France

Summary The main purpose of this study, carried out in 1995, was to determine, usingavailable sources, the cost of peripheral arterial obliterative disease (PAOD) inFrance over a 1-year period. This cost-of-illness study was based on a retrospec-tive analysis of the available literature and databases. It involved a description ofepidemiological data and a cost estimate of the different medical resources con-sumed over 1 year. For this latter purpose, a payer perspective was chosen.

Data were extracted from national representative surveys and databases withrespect to morbidity and mortality [from the National Institute of Health andMedical Research (Institut National de la Santé et de la Recherche Médicale;INSERM) and the National Sickness Insurance Fund for Salaried People (CaisseNationale d’Assurance Maladie des Travailleurs Salariés; CNAMTS)], consul-tations, examination tests and drug prescriptions [from the French Medical Auditconducted by Intercontinental Medical Statistics (IMS)], hospitalisations [fromthe Statistical Unit of the Department of Health – Service des Statistiques, desEtudes et des Systemes d’Information (SESI) and the National Public ResearchCentre in Health Economics (Centre de Recherche d’Etude et de Documentationen Economie de la Santé; CREDES)] and related health expenditure fromCNAMTS.

In France, the prevalence of stage II PAOD (Leriche and Fontaine classifica-tion) in 1992 was estimated to be 675 000; 53% of these patients had undergonevascular or bypass surgery. The total annual cost of healthcare (including con-sultations, drugs, laboratory tests, hospitalisation and hydrotherapy) for themanagement of patients with PAOD ranged from 3.9 billion French francs (F) toF4.6 billion (1995 values), depending on the type of hospital considered. 50% ofthis cost was related to hospitalisations and 75% was covered by the CNAMTS.

Although this study was only a partial evaluation and did not take into accountindirect costs or nonmedical direct costs, such as transport and care by healthcareassistants and paramedics, these results may help to establish public health pri-orities and modify clinical practice to favour an earlier diagnosis of PAOD.

ORIGINAL RESEARCH ARTICLE Pharmacoeconomics 1998 Jan; 13 (1 Pt 1): 51-591170-7690/98/0001-0051/$04.50/0

© Adis International Limited. All rights reserved.

It is estimated that 800 000 individuals inFrance experience peripheral arterial obliterativedisease (PAOD).[1] The prognosis of this disease isunfavourable; 20% of affected patients die within5 years and 40% within 10 years of diagnosis.[1]

The 2 main causes of death are coronary artery dis-ease and stroke.

In a recent article, Langley and Coons[2] high-lighted the dearth of economic studies that haveincluded an assessment of the overall costs of treat-ing PAOD. No full analysis of the cost of PAODfrom the perspective of the health system or thetype of care provided (medical, paramedical or hy-drotherapy) is currently available in France. Thetype of care provided depends on the stage of thedisease, the extent of any motor deficiency, riskfactors, concomitant disease, patient age and theability of the patient to perform physical activities.

The aim of this study was to determine the an-nual cost of direct medical care associated withPAOD in France according to the type of care pro-vided, using the major medicoeconomic sourcesavailable.

Methods

General Principles

A retrospective analysis of the available Frenchliterature and databases was performed in 1995,with the aim of evaluating the cost of managingPAOD. The methodology included epidemiologi-cal and economic components. The latter compo-nent comprised an evaluation of the cost of directmedical care over a 1-year period from the perspec-tive of the payers (the National Sickness InsuranceFund and the patients themselves).

Sources of Data

EpidemiologicalEpidemiological data were obtained from sur-

veys conducted by the National Sickness Insur-ance Fund for Salaried People [Caisse Nationaled’Assurance Maladie des Travailleurs Salariés(CNAMTS)],[3] the National Public Research Cen-tre in Health Economics [Centre de Recherche

d’Etude et de Documentation en Economie de laSanté (CREDES)][4,5] and the National Institute forHealth and Medical Research [Institut National dela Santé et de la Recherche Médicale (INSERM)].[6]

The surveys conducted by CNAMTS, CREDESand INSERM were carried out at different timesand, in certain cases, this resulted in an importanttime lag between the published results and ourstudy. The surveys used for our analysis are des-cribed below and were the most recent at the timeof this study.

DiagnosisThe codes used in these surveys to describe

PAOD were those of the 9th International Classifi-cation of Diseases (ICD-9) as determined by theWorld Health Organization (WHO).[7] Codes 440to 448 (diseases of the arteries, arterioles and capil-laries), particularly code 440 (atherosclerosis) andcode 447 (other diseases of the arteries and smallarteries), were used to standardise diagnoses usedacross all parts of the study.

PrevalenceData from the Institut National des Statistiques

des Etudes Economiques (INSEE)-CREDES HealthSurvey were used to estimate the prevalence ofPAOD (personal communication, Dr CatherineSermet, Maître de recherche, CREDES). This na-tional decennial survey provides the annual num-ber of patients per disease for the whole of France.We used the most recent survey, carried out in1991.[4] In addition, a national representative sur-vey conducted by CREDES in 279 French institu-tions from 1987 to 1988 was used to provide dataon institutionalised patients; this survey included arandom sample of 1680 patients aged 80 years.[5]

IncidenceAs certain forms of PAOD may be associated

with minimal or no symptoms, it is difficult to de-termine the incidence of these conditions. Never-theless, it was possible to estimate these figuresusing data from large-scale studies. Data derivedfrom the Framingham and Basel surveys[8] wereused to estimate the incidence of PAOD.

As PAOD is included in the list of 30 chronicdiseases eligible for full sickness insurance coverage

52 Montron et al.

� Adis International Limited. All rights reserved. Pharmacoeconomics 1998 Jan; 13 (1 Pt 1)

in France, most of the patients with this diagnosisare reported to CNAMTS. The last 1990/1991CNAMTS publication was also used to estimatethe incidence of these diseases.[3] However, theselatter results only concerned the beneficiaries ofthe Régime Général de la Sécurité Sociale (i.e.wage earners) for the period 1990 to 1991, repre-senting 80% of the French population.[3,9]

MortalityMortality data for PAOD was provided by an

INSERM publication.[6] INSERM is the French of-ficial body for mortality statistics. This publicationwas based on diagnoses stated on death certifi-cates.

Medical Resource Consumption

Consultations, Drug Prescriptionsand ExaminationsThe French Medical Audit, conducted by Inter-

continental Medical Statistics (IMS) since 1963,was used to determine the number of physicianconsultations and the number and nature of pre-scriptions for drugs, laboratory tests and examina-tions necessary for the management of patientswith PAOD over a 1-year period (March 1993 toFebruary 1994).[10] This medical audit was basedon a national representative sample of 800 medicalpractitioners, selected and stratified according tothe following criteria:• doctor’s age and gender• medical speciality• country, region of work and size of urban unit.

Each doctor participated in the study for 7 con-secutive days each quarter, and collated informa-tion for every patient seen outside an institutionduring this period. The results were presented asthe number of diagnoses noted during the period inquestion and were extrapolated for 1 year to allFrench doctors.

IMS Medical Audits are the main source ofdata on physicians’ prescriptions in 80 countriesthroughout the world. In France, the French Medi-cal Audit is used as a reference by governmentalauthorities to assess patterns of prescribing.

The number of consultations for one diagnosiswere counted as ‘equivalent consultations’. Attribu-

ting a consultation or visit to each diagnosis natu-rally leads to an overestimation of the total numberof consultations reported for all diseases. For ex-ample, in 1994, 306 million consultations or visitswere covered by CNAMTS, and during the sameperiod the French Medical Audit noted an averageof 1.58 diagnoses per consultation.[11] If 1 diagno-sis was considered equivalent to 1 consultation,this would result in health insurance reimburse-ments for 438 million consultations or visits. Thenumber of equivalent consultations per medicalspeciality was therefore determined using the ratioof number of diagnoses for each of the diseasesconsidered : average number of diagnoses per con-sultation. These latter data were collected in theFrench Medical Audit.

HospitalisationThe number of hospitalisations for PAOD was

based on data provided by 2 surveys.The first was a survey of hospital morbidity

conducted by the Statistical Unit of the Departmentof Health [Service des Statistiques des Etudes etdes Systemes d’Information (SESI)] from 1 Octo-ber 1984 to 30 June 1988.[12] This survey was rep-resentative of the annual activity of hospitals at thenational level and included 77 430 hospital stays.It was used to provide a first estimate for the costof care in PAOD. The SESI survey identified thosediseases for which treatment was associated withshort term hospitalisation in public (teaching andnonteaching) hospitals and private establishments.We used this survey to determine the number ofhospitalisations per year for all diseases of the ar-teries. These data were extrapolated to the wholeof France, according to whether short termhospitalisation was provided by medical (generalmedicine, emergency medicine, paediatric, medi-cal intensive care, medical specialty, gynaecolo-gy/obstetrics or oncology) or surgical (general sur-gery, emergency surgery, paediatric surgery,surgical intensive care and surgical specialty) de-partments.

The second was the CREDES national hospitalsurvey,[13] which provided a second estimate of thecost of care in PAOD. This survey, published in

Cost of Peripheral Artery Disease in France 53


1995, was conducted using a national repre-sentative sample of 884 public and private estab-lishments encompassing 4646 patients. It des-cribed the diagnoses of patients hospitalised inacute-care wards (e.g. medical, emergency, gynae-cology/obstetric and surgical departments), in re-habilitation wards (e.g. physiotherapy clinic) andin long-stay wards (e.g. nursing home) for 1 spe-cific day of the year. These data were then extra-polated to the whole of France. The extrapolatedresults were used in our cost-of-illness study byconsidering only the main diagnoses (i.e. those dis-eases leading up to hospitalisation, not theabovementioned ICD-9 codes reported as the asso-ciated diagnosis). For each of ICD-9 codes, the re-sults were distributed among the different depart-ments considered, according to the distributionpattern of hospital stays noted in the SESI hospitalmorbidity survey.[12]

A correction was necessary in order to comparethe results of the 2 surveys. The first survey wasbased on the number of hospitalisations per year,while the second was extrapolated from an analysisof the main diagnoses for 1 specific day. It wastherefore necessary to correct the results of the sec-ond survey to provide data on hospitalisations fora 1-year period and not just for a single day. Thiswas possible using the following calculation:(number of patients � 365)/average duration ofhospitalisation. The average duration of hospitali-sation for diseases of the arteries, as provided bythe SESI hospital morbidity survey and used in thiscalculation, was 13.8 days for hospitalisation in apublic establishment and 12.2 days for hospitalisa-tion in a private establishment.

HydrotherapyHydrotherapy data for patients with PAOD were

derived from the annual statistics for hydrotherapycovered by the National Sickness Insurance Fund(80% of French inhabitants).[14] The latest publica-tion concerned 1990/1991 statistics. Only datafrom spas at which PAOD is specifically treatedwere used (Bains les Bains, Bourbon Lancy andRoyat).

Costing

A cost-of-illness analysis was conducted to pro-vide an estimate of the costs of direct medical careresources required for the management of patientswith PAOD. Nonmedical costs (direct and indirect)were not taken into account in calculating the valueof resources used. The sources used comprised:• cost of physician consultations from agreed na-

tional tariffs[15]

• cost of drug therapy from the IMS Pharmaceu-tical Audit National Drug Price Database[10]

• cost of monitoring and examinations (includinglaboratory tests) from the Agreed National TariffList[15]

• cost of hospitalisation from hospital cost data[16]

• cost of hydrotherapy from CNAMTS data.[14]

Costs related to transport, paramedical and so-cial services were not included.

A payer perspective was used and the resultswere determined separately for the different payers– health insurance and households (patient cover-age) – according to the CNAMTS reimbursementrate applicable on 1 January 1995.

Consultations, Drug Prescriptionsand ExaminationsThe cost of equivalent consultations for general

and specialist physicians was determined accor-ding to the agreed tariffs at 1 January 1995 (F105for a general physician consultation and F145 fora specialist consultation). The rate of reimburse-ment was that of the CNAMTS at 1 January 1995(70%).

The cost of each drug unit was estimated usingthe retail prices listed in the IMS PharmaceuticalAudit National Drug Price Database on 1 January1995. This database is an updated and exhaustivedatabase of all drugs on the French market. Therates of reimbursement used were those ofCNAMTS: 0% for nonreimbursable products, 35%for reimbursable non-essential products and 65%for essential products. The fraction not covered byCNAMTS was treated as a patient expense.

The costs of laboratory examinations were esti-mated according to the Agreed National Tariff List[Union des Caisses Nationales de Sécurité Sociale

54 Montron et al.


(UCANSS) – Nomenclature Générale des ActesProfessionnels], based on the examination ratingand the unit price of the ‘key letter’ prevailing on1 January 1995.[15] For example, in this list, anelectrocardiogram (ECG) is rated K6.5 (key letterK and examination rating 6.5); since the unit priceof key letter K is F12.40, the cost of an ECG isF80.60 (12.40 � 6.5). It was assumed that all labo-ratory tests identified in the French Medical Auditwere carried out in a private medical laboratory.Specialist consultations that may have been re-quired for these examinations or tests were nottaken into account. The reimbursement rate usedwas 60% for biological tests (key letter B) spe-cialised examinations such as ECG or Doppler(key letter K), and radiological examinations ortests (key letter Z), with the exception of radiolo-gical examinations rated over Z50, for which thereimbursement rate was 100%.

HospitalisationThe costs associated with these data were esti-

mated in different ways, according to the differentfinance methods used by each establishment. Forhospitalisation in a public establishment, the typeof hospital ward was taken into consideration byusing the data for the average cost per admissionfrom the group of public teaching hospitals in theParis area (Assistance Publique-Hôpitaux de Paris):general medicine F12 828; paediatrics F12 915;medical intensive care F5528; medical specialisationF12 937; general surgery F12 731; surgical inten-sive care F17 515; surgical specialisation F15 994;and gynaecology/obstetrics F24 673.[16]

The total cost of hospitalisation was obtained bymultiplying the total number of admissions by theaverage cost per admission.

For hospitalisation in a private establishment,the cost per day was estimated by using NationalHealth Expenditure (Comptes Nationaux de laSanté) data and reimbursement statistics from theCNAMTS.[17,18] According to this data source, an-nual healthcare costs for admission to private hos-pitals, covered by all national sickness insurancefunds, amounted to F61.7 billion in 1993. Theamount not covered by national sickness insurance

funds was F7.4 billion (12% of the amount coveredby national sickness insurance funds).[17]

F39.7 billion for admission to private hospitalswas reimbursed by CNAMTS in 1993;[18] the 12%not covered by CNAMTS should be added to thisamount (F4.8 billion). As a result, the total costfor admission to private hospitals was F44.5 bil-lion for households covered by CNAMTS for atotal of 31 278 465 days of hospitalisation. The re-sulting average cost per day of hospitalisation(F1421) was used to evaluate the cost of admissionto private hospitals, regardless of their specialty.As shown by the SESI survey,[12] the approximatecost per admission for a mean stay of 12.2 days inthe private hospitals was F17 336, similar to thecosts per public hospital admission.

HydrotherapyThe cost per hydrotherapy session was deter-

mined from the global costs covered by CNAMTS,including mandatory (medical supervision, treat-ment fees) and additional (accommodation, trans-port fees) benefits. The average cost of treatmentwas estimated using the following formula: globalcost for all treatments/total number of applicationsfor treatment accepted by CNAMTS.

The total cost for the 484 284 treatments coveredby CNAMTS in 1991 was F1.3 billion, equivalent toF2607 per treatment. These costs only correspondedto those indicated and covered by CNAMTS.

Results

Epidemiological

PrevalenceThe prevalence of stage II PAOD according to

the Leriche and Fontaine classification was esti-mated to be 1 to 1.5% for men less than 50 yearsof age and 4 to 6% for men greater than or equal to50 years of age;[19] there were 700 000 to 800 000patients with this disease in France in 1990.[8] Thisresult was confirmed by the 1992 INSEE-CREDEShealth survey which estimated that, for the wholeof France, 673 808 patients experienced PAOD ofwhom 359 000 (53.3%) had undergone orthopae-dic or bypass surgery.



Of the French population aged �80 years livingin institutions, 9.5% experienced artery and capil-lary diseases (mainly PAOD).[5] These diseaseswere described as very serious or of moderate se-verity in 11% and 36% of cases, respectively.

IncidenceThe incidence of stage II PAOD was determined

from the Framingham and Basel studies, whichmonitored normal patients over several years.The annual incidence of the disease in France was80 000 to 90 000 cases,[8] and was dependent onage (the incidence of stage II PAOD increases withage) and gender (men develop PAOD 10 years ear-lier than women), the existence of risk factors(smoking and diabetes mellitus),[20] the existenceof atherosclerosis in other body regions, and thesystolic blood pressure gradient between the ankleand arm.

The results of the CNAMTS survey showed thatdiseases of the arteries, small arteries and capil-laries were responsible for 66 643 requests for fullcoverage of health costs during 1991;[9] this repre-sented almost 11% of the 611 373 requests made intotal. Arteritis of the peripheral arteries (includedin this group of disorders) was the fourth majordisease in terms of full coverage of healthcare costsduring 1991, necessitating 23 498 requests. Thisrepresented 35% of requests for full coverage forvascular disease (arteriopathy).

Full coverage of healthcare costs was more of-ten granted to men than to women, and the numberof beneficiaries increased progressively betweenthe ages of 34 and 69 years for both genders. Dis-eases of the arteries, small arteries and capillaries(WHO DRG codes 440 to 448) were the reasonfor full coverage of healthcare costs for 29 596 and27 820 patients in 1990 and 1991, respectively.[9]

All of these patients received full coverage for thiscondition included in the list of 30 chronic diseaseseligible for full sickness insurance cover.

PrognosisIn the year following the development of clau-

dication, 10% of patients experienced a deteriora-tion in health.[8] Figure 1 illustrates the evolutionof the disease at the stage of intermittent claudica-tion.[19]

MortalityIn 1991, the mortality rate linked to circulatory

diseases was 28.4 per 100 000 men and 26.3 per100 000 women.[6]

PAOD-associated mortality rates of 30, 50 and70% were calculated for 5-, 10- and 15-year fol-low-up periods, respectively.[19] This represents alife expectancy that is reduced by approximately10 years compared with that of the general popula-tion. About half of the deaths were attributable tocoronary artery disease.

100 Stage II patients

25deteriorations

20 to 30 nonfatalcardiovascularcomplicationswithin 5 years

3 majoramputations

30 died within5 years

Localised illnessMorbidity andtotal mortality

15 from stroke or heart attack

5 from another cardiovascular cause

10 from a noncardiovascular cause

Fig. 1. Outcome of 100 patients with symptomatic stage II peripheral arterial obliterative disease (Leriche and Fontaine classification).[19]

56 Montron et al.


Economic Evaluation

The total cost of healthcare in France for dis-eases of the arteries, small arteries and capillarieswas F3.9 billion according to the first estimate ofthe cost of hospitalisation (SESI hospital morbi-dity survey) and F4.6 billion according to the sec-ond estimate of hospitalisation (CREDES nationalhospital survey) [table I]. This was equivalent to0.7% of total healthcare costs for 1993.

Hospitalisation was the largest contributing fac-tor (47 to 55% of the total cost, depending upon theestimate used). The majority of stays were in pub-lic hospitals: 51% (n = 59 884) of total stays ac-cording to the first estimate of hospitalisation, and76% (n = 137 933) according to the second esti-mate.

Drug prescriptions contributed 33 to 39% of thetotal cost. The main therapeutic classes prescribedwere cerebral and peripheral vasotherapeutics(64% of total prescriptions) and systemic venous

therapies (17% of total prescriptions). CNAMTScovered 75% of the total costs.

Discussion

As mentioned previously, no comprehensivecost-of-illness studies on PAOD have been pub-lished.[2] This is the first economic evaluation con-ducted in France to estimate the direct medicalcosts for patients with PAOD from the payer’s pers-pective (CNAMTS and the patient) over a 1-yearperiod.

In France, the prevalence of stage II PAOD(Leriche and Fontaine classification) is estimatedat 675 000. More than 50% of these patients havealready undergone lower-limb vascular surgery.

In this study, we estimated that the annual costof the major medical resources required to care forpatients with PAOD is F3.9 billion to F4.6 billion,depending on the type of hospital. Approximately75% of these costs were covered by CNAMTS, and

Table I. Distribution of the cost of care [French francs (F; 1995 values); F1 = $US0.20] over a 1-year period for patients with diseases of thearteries, arterioles and capillaries

Direct medical costs Number of events(millions)

Cost (F billions)

total (%) proportion covered by:

national health insurance patient

First estimate of the cost of hospitalisation (based on the hospital morbidity study from SESI)[11]

Consultations 3.572 0.398 (10.1) 0.278 0.119

Monitoring examinationsa 1.193 0.097 (2.5) 0.058 0.039

Drug prescriptions 34.124 1.528 (39.0) 0.927 0.601

Hospitalisations 0.117 1.849 (47.1) 1.572 0.277

Hydrotherapy 0.022 0.051b (1.3) 0.051 NA

Total 39.028 3.923 (100.0) 2.886 1.036

Second estimate of the cost of hospitalisation (based on survey by CREDES)[12]

Consultations 3.572 0.398 (8.6) 0.278 0.119

Monitoring examinationsa 1.193 0.097 (2.1) 0.058 0.039

Drug prescriptions 34.124 1.528 (33.1) 0.927 0.601

Hospitalisations 0.181 2.548 (55.1) 2.166 0.382

Hydrotherapy 0.022 0.051c (1.1) 0.051 NA

Total 39.092 4.622 (100.0) 3.480 1.141

a Defined as laboratory tests and exploratory examinations (e.g. doppler and echography).

b The total cost indicated was underestimated as it only corresponds to the expenses covered by CNAMTS. The costs covered by the pa-tient and expenses for hydrotherapy covered by the patient have not been taken into account.

c The total cost indicated was underestimated as it only corresponds to the expenses covered by the national sickness insurance fund.The costs covered by the patient and expenses for hydrotherapy covered by the patient have not been taken into account.

Abbreviation: CREDES = Centre de Recherche d’Etude et de Documentation en Economie de la Santé; NA = not available; SESI = Servicedes Statistiques des Etudes et des Systemes d’ Information.



hospitalisation was the main expense. Dependingon the estimate used, hospitalisation accounted for47 to 55% of all costs for PAOD. This estimate doesnot cover nonmedical direct costs or indirect costs,so it must be considered a minimum estimate.

Data concerning direct medical resources wereobtained from national representative public andprivate surveys and databases. These allowed us todescribe the use of medical resources related toPAOD, in both hospital and ambulatory settings.These sources were the most recent that were avail-able at the time of our study, but some data werealready 4 to 10 years old.

We would like to highlight that no systemati-cally collected data on hospital admission (in eitherpublic or private settings) were available in 1995;therefore, despite the time lag between their com-pletion and the time of our study, these databaseswere the most relevant that we could use. CNAMTSand the Department of Health are the only officialdata providers for total health expenditure at a na-tional level. CREDES and INSERM are the na-tional research centres for health economics andmedical research, respectively, and are mostly pub-licly funded. IMS is a private company that pro-vides medical audits for 80 countries around theworld, collected through national, representativephysicians panels. In France, the IMS Medical Au-dit is considered by governmental authorities to bea reference for assessing prescribing patterns. Noother sources in France are as representative orrelevant to the issue we have described here.

One could argue that we have estimated the costof care for people with PAOD rather than the costof PAOD itself. Obviously even if PAOD is themain diagnosis for hospitalisation, all the costs in-curred during the hospitalisation of the patient arenot necessarily related to PAOD (a part can be theconsequence of the associated comorbidities).However, it is likely that the majority of them are.As the data collected in the French Medical Auditconducted by IMS clearly indicated for which dis-ease a drug, laboratory test or an examination wasprescribed, this problem is of lesser importance forambulatory care costs. So, it is legitimate to con-

sider that these results roughly reflect the cost ofcare for PAOD in France.

The majority of data used in this cost-of-illnessstudy were obtained from the national representa-tive surveys of CREDES, CNAMTS and INSERM.When using different information sources, it is im-portant to explain their respective methodologiesin order to consider the limitations of the data andtheir impact on the results. Such considerations in-clude the availability and quality of the data withrespect to the problem studied, together with theneed to make retrospective, average cost estimates,particularly for public establishments and for pri-vate establishments affiliated to the public sector.Comparison of different sources of the same datatherefore provides a measure of the validity andreproducibility of results.

For this reason, the SESI hospital morbidity sur-vey,[12] which considered short periods of hospi-talisation, and the CREDES national hospital sur-vey,[13] which assessed short, medium and longterm hospitalisation, were used. The results of bothstudies were extrapolated to the whole of France,and the goals of the 2 surveys differed such that theSESI survey provided the number of hospitali-sations per year according to hospital departments,while the CREDES survey described the main diag-noses for an average day in the year.

It was therefore necessary to adjust the CREDESdata in order to compare the results of both surveys.This correction led to a hypothesis concerning me-dium-length stays in hospital. The length of stayused was that reported in the SESI survey. An eval-uation of hospitalisation was conducted for publicestablishments on the basis of the cost of hospi-talisation in the Assistance Publique-Hôpitaux deParis. However, the cost of hospitalisation variesnot only from 1 establishment to another, but alsobetween regions. The average cost per day in privatehospitals was determined using data provided by theNational Health Expenditure (Comptes Nationauxde la Santé) and CNAMTS. For this reason, theaverage cost per day did not include costs coveredby other systems, which encompass approximately20% of the French population.[21]

58 Montron et al.


Considerable progress has been achieved withcost-of-illness studies in France to date, and this ispartly because the main reference institutions haveconducted specific surveys on a regular basis.However, prudence is required when using the re-sults of these surveys as they only provide partialcost estimates. In particular, indirect costs, costs oftransportation and costs of ancillary medical andparamedical intervention were not taken into ac-count. The partial estimate conducted in this studyindicated that the care of patients with PAOD ac-counted for 0.7% of total healthcare costs for 1993in France.

These epidemiological and economic data mayhelp to determine and establish public health pri-orities and also facilitate changes in medical prac-tice by providing a rationale for earlier screeningof PAOD; more than 50% of the total cost waslinked to hospitalisation. They also show that ad-vanced stages (claudication, decubitus pain, necro-sis, etc.) are the most complex to care for and themost expensive because of the high relative contri-bution of the cost of hospitalisation. Earlier detec-tion and management could have a favourable im-pact in limiting the severe consequences of thedisease. Earlier detection of PAOD may be possi-ble using the systolic index, as demonstrated by thepromising results of the recent study by Boccalonand Lehert.[22]

Acknowledgements

This study was funded by LIPHA Laboratories S.A.

References1. Les artériopathies des membres inférieurs : des conséquences

socio-économiques très lourdes. Le Quotidien du Médecin1993; 5157: 16

2. Langley PC, Coons SJ. Peripheral vascular disorders: a pharma-coeconomic and quality-of-life review. Pharmacoeconomics1997; 11 (3): 225-36

3. Mission Statistique Médicale. Exonération du ticket. Paris:Caisse Nationale d’Assurance Maladie des TravailleursSalariés (CNAMTS), 1993

4. Sermet C. De quoi souffre-t-on ? Description et évolution de lamortalité déclarée 1990-1991. Solidarité Santé 1994; 1: 37-56

5. Sermet C. La pathologie des personnes âgées de 80 ans et plusen institution: enquête nationale France 1987-1988. Paris:

Centre de Recherche d’Etude et de Documentation en Econo-mie de la Santé (CREDES), 1992

6. Sources médicales de décès-résultats définitifs France. Paris:Institut National de la Santé et de la Recherche Médicale(INSERM), 1991

7. Organisation Mondiale de la Sante. Classification internatio-nale des Maladies (ICD-9). Geneva: Organisation Mondialede la Sante, 1977

8. Cormier JM, editor. Les artériopathies des membres inférieursI. Impact médecin: les dossier du praticien, 1991: 118

9. Mission Statistique Médicale. Exonération du ticket:modérateur pour affections de longue durée. Paris: CaisseNationale d’Assurance Maladie des Travailleurs Salariés(CNAMTS), 1993

10. Etude Permanente de la Prescription Médicale (EPPM-DOREMA). Nanterre: IMS France, 1994

11. Indicateurs statistiques, résultats 1993. Paris: CaisseNationale d’Assurance Maladie des Travailleurs Salariés(CNAMTS), 1994

12. SESI. Fiches synthétiques par pathologies: enquête demorbidité hospitalière 1985-1987. Paris: Ministère de laSolidarité, de la Santé et de la Protection Sociale, 1990

13. Com-Ruelle L. Les étapes diagnostiques et la maladieprincipale des hospitalisés en 1992. Paris: Centre de Recher-che d’Etude et de Documentation en Economie de la Santé(CREDES), 1995

14. Bulletin juridique: statistiques 1990-1991 sur les cures the-rmales prises en charges par le régime général de sécuritésociale et nomenclature des stations thermales. Paris: Uniondes Caisses Nationales de Sécurité Sociale (UCANSS), 1992

15. Nomenclature générale des actes professionnels et des actes debiologie médicales. Paris: Union des Caisses Nationales deSécurité Sociale (UCANSS), 1995

16. Direction des Finances, Département de Contrôle de Gestion.Les coûts de l’assistance publique: Hôpitaux de Paris. Paris:Assistance Publique – Hôpitaux de Paris, 1994

17. SESI. Comptes nationaux de la santé 1991-1992-1993. Paris:La Documentation Française, 1995

18. Carnets statistiques: le régime général en 1993. Paris: CaisseNationale d’Assurance Maladie des Travailleurs Salariés(CNAMTS), 1994

19. Dormandy J. Le devenir de l’artéritique. Sang, Thrombose etVaisseaux; 1989 ; 1 (5) : 263-6

20. Chanu B. Histoire naturelle des artériopathies athéromateuseschroniques des membres inférieurs. Actualité d’Angeiologie1988; XIII (5) : 111-23

21. Duriez M, Sandier S. Le système de santé en France: Or-ganisation et Fonctionnement. Paris: Centre de Recherched’Etude et de Documentation en Economie de la Santé(CREDES), 1994

22. Boccalon H, Lehert P. Diagnostic précoce de l’artériopathie desmembres inférieurs a l’aide de mesures adaptées à la pratiquegénéraliste: l’index systolique et la perception des pouls. JMal Vasc 1995; 20: 28-37

Correspondence and reprints: Dr Eric Guignard, HealthEconomics Department, I.M.S. France, La Défense‘Bergères’, 345 avenue Georges Clemenceau, TSA 30001,92882 Nanterre CTC Cedex 09, France.



peripheral arterial obliterative disease

Documents