Background and Purpose Cerebrovascular disorders are associated with a high level of morbidity and mortality and call for considerable resources. The objective of this study was to determine from a societal perspective the medical consumption (direct costs) and productivity losses (indirect costs) caused by cerebrovascular disorders in the Netherlands.
Methods This study can be characterized as a cost-of-illness study based on prevalence data. All data gathered refer to 1993. Cerebrovascular disorders are defined according to the International Classification of Diseases, 9th Revision (ICD-9) classification. Data from medical registrations and national statistics have been analyzed. For both direct and indirect costs, volume and cost components are presented. To test the likelihood of the assumptions, a sensitivity analysis was performed.
Results The cost of cerebrovascular disorders in the Netherlands in 1993 amounted to 2.5 billion Dutch guilders, of which 1.9 billion were spent on medical consumption. It was found that direct costs are generated mainly by the long-term care of inpatients (nursing homes and hospitals). The productivity losses were relatively low in comparison with other diseases, probably due to the fact that most patients with cerebrovascular disorders are elderly.
Conclusions More than 3% of the Dutch annual healthcare budget is spent on patients suffering from cerebrovascular disorders. Costs in the future may be influenced by, among other things, demographic changes, new therapies, and cost-reduction programs introduced by the government.
Stroke has a significant social and economic impact on patients, their families, and society as a whole. This is illustrated in the international literature both by COI studies and by case reports. (For an overview of the COI studies and case reports, please contact the first author.) In COI studies, a distinction is made between direct and indirect costs. Direct costs are directly associated with healthcare interventions, ie, medical consumption. Indirect costs are broadly defined as the value of productivity losses caused by absenteeism, disability, or (premature) death.1
The organization of services involved in stroke management is multidisciplinary and can include care provided by GPs, hospitals, rehabilitation centers, nursing homes, and community services. In the Netherlands, no top-down COI study has been performed, although some efforts have been made to quantify the economic consequences of stroke for the Dutch society.2 3 4 5 6 The purpose of this study was to identify and estimate the costs of stroke in the Netherlands, both direct (medical consumption) and indirect (production loss). The study can be typified as a COI study based on prevalence data and focuses on the year 1993.
The means and methods of the study are described in “Methods.” In the following sections, the data, analysis, and results by cost category are described for both the direct and the indirect costs. Finally, in “Results” a comparison is made with other COI studies performed in the Netherlands, and the sensitivity of the data found is discussed.
As the Figure⇓ shows, the first step in a COI study is to define the disease. All data in the present study refer to persons suffering from CVDs (ICD codes 430 to 438) or who died because of it. The method applied here has been used in other prevalence-based COI studies.7 8 The underlying rationale is that direct costs and productivity losses are assigned to the year in which they occur (in this case, 1993). Expected future incomes lost as a result of premature mortality are assigned to the year of death.9
The direct costs are in general based on figures provided by the Financial Statement of Care.10 The Financial Statement of Care is a government publication that is presented to the Dutch Parliament together with the budget. It gives an integrated overview of the financial developments in the healthcare sector during recent years in addition to predictions for the subsequent year. Figures from the National Hospital Institute are used to calculate the costs of hospital care.11 12 13 The costs of drugs are calculated using information from the Health Insurance Executive Board (J.F. Piepenbrink, unpublished data, 1996, and Reference 1414 ). To quantify the indirect costs, the human capital approach is used.7 This approach attempts to estimate the value of lost production, which results from temporary or permanent absence from work, disablement, or premature death. Production losses are estimated using the average salary per employee. Finally, to test the likelihood of the assumptions made when measuring and evaluating the direct and indirect costs, a sensitivity analysis was performed.
Data, Analysis, and Results by Cost Category
Direct Costs Attributable to CVDs
This COI study reflects the course that patients with CVDs follow when they become ill. Patients with stroke generally contact the GP or the emergency department of the hospital. Most of those who visit the GP are referred to more specialized care.15 16 After discharge from the hospital, patients will also visit various types of care providers, such as rehabilitation centers and nursing homes. Furthermore, some patients will be cared for at home by professionals or their own families.
General practitioner. Table 1⇓ shows that there are several studies that have assessed morbidity of stroke as found in general practice.17 18 19 20 21 One can also see that the prevalence rate found for CVDs in these studies varies between 720 and 1018 per 1000. These figures differ slightly because the studies vary regarding population, study period, and methodology. In the remainder of this research, the average figure from these studies is used as a basic assumption, eg, 8.3 per 1000. The total cost of GP practices in 1993 was 2203 million DG.10 This would result in a total of about 18 million DG for GP consultations due to CVDs.
Acute-phase inpatient stay in hospitals. After an attack, most stroke patients are admitted to hospitals.15 22 The National Medical Registration registers 99% of the admissions to general hospitals, university hospitals, or specialized hospitals (ie, rehabilitation centers) using the ICD-9.23 24 In this study, the costs of CVDs for the hospitals are calculated on the basis of both number of discharged patients and number of days spent in the hospital. In 1993, 28 664 patients with CVDs (diagnosis at discharge) were discharged from a hospital, after a total length of stay of 683 270 days. Discharge includes hospital mortality. The diagnosis refers to the primary diagnosis, ie, the diagnosis that was the main reason for admission to the hospital. On the basis of these figures, approximately 1.82% of the admitted patients account for 4.35% of the inpatient days.
To get an impression of the costs incurred by hospitals, we added the costs of care in general hospitals, university hospitals, and rehabilitation centers (specialized hospitals). The resulting total in 1993 was 16 230 million DG (including specialists’ salaries).11 12 13 Taking the overall costs for all hospitals, the costs due to CVDs would be 295.4 million DG based on admission figures and 706.0 million DG based on inpatient days.
We assume that the former calculation is an underestimation, since the average length of stay for patients with CVDs is relatively long (eg, 24 days for CVDs compared with 10 days for all diagnoses). On the other hand, the first days in the hospital are the most expensive; therefore, we assume that the figure based on number of days in hospital is an overestimation. So, an average of 500.7 million DG was reached.
Postacute phase. Stroke patients in the postacute (hospital) phase fall into three main care categories: (1) those who require continuing rehabilitation; (2) those who require continuing supportive care from nursing home care; and (3) those whose condition and social situation permit various kinds of outpatient care, such as home nursing care, outpatient physiotherapy and exercise therapy, outpatient speech therapy, and outpatient drug therapy.
Rehabilitation. Rehabilitation after CVD aims to improve the disturbed function, and in the case of permanent handicaps, to optimize skills and behavior.25 Regarding inpatient care, 35% of the admissions were due to CVDs.26 We assume that the population in 1994 did not differ from that in 1993. The cost of rehabilitation center services in 1993 was 346.3 million DG.11 Regarding the inpatient stay of persons with CVDs, the cost for this group in 1993 would be about 121.2 million DG.11 In this COI study, the costs of rehabilitation center services are included in the calculation of hospital costs.
Nursing homes. Almost all patients who are admitted to a nursing home are registered by the Nursing Home Information System.24 According to these national statistics, on September 30, 1993, 17.6% of the inpatients were diagnosed primarily as having (late sequelae) CVD.27 Of these, 33.8% were inpatients in the somatic department and 2.3% inpatients in the psychogeriatric department. For day care, the figures are more or less similar. On the basis of inpatient figures, the costs of nursing home services were calculated. In 1993, the total cost for nursing homes was 5254 million DG.10 On the assumption that the costs for somatic and psychogeriatric departments are more or less the same, this results in 925 million DG for nursing home services that can be assigned to CVDs.
Outpatient home nursing care. For patients with CVDs, home nursing care encompasses nursing care, hygiene, bathing, reactivation, and rehabilitation. The National Organization of Home Nursing Care does not, however, keep a central register of diseases. The Home Nursing Care Organization in the southern part of the Netherlands (Groene Kruis Heuvelland) provided prevalence rates of CVD. The area covered by this organization includes 209 657 inhabitants, about 1.4% of the Dutch population.28 On December 1, 1995, 12% of the patients of this Home Nursing Care Organization had CVD (H. Habets, unpublished data, 1996). In 1993, 3442 million DG was spent on home care and nursing10 ; therefore, about 413 million DG was spent on home care and nursing related to CVDs.
Outpatient physiotherapy and exercise therapy. After a CVD, patients are often partially paralyzed and may have problems with walking, speech, etc. About 3% of the patients with a CVD are referred to a physiotherapist. The NIVEL published an inquiry on the use of physiotherapy in which a distinction was made between paralysis after CVDs and other consequences after CVDs.29 30 The other consequences of CVDs accounted for 0.5% of the referrals, the paralysis for even fewer referrals. The total outpatient costs of physiotherapy for the Netherlands in 1993 was 1228 million DG.10 Hence, the expenditure of 6.1 million DG in 1993 might be due to CVDs.
In addition to physiotherapy, exercise therapy is well known in the Netherlands.31 This therapy aims to rehabilitate by improving the posture and basic motor skills by means of an active, individual motor learning process. According to the NIVEL, patients with CVDs are not referred to these exercise therapies very often.32
Outpatient speech therapy. In 1993, the NIVEL studied 1761 patients referred to a speech therapist.33 Of those whose referral diagnosis is known, 11.2% were referred because of CVDs. Total outpatient cost of speech therapy in 1993 was 70 million DG.10 Hence, the total cost for speech therapy in 1993 for CVDs was about 7.8 million DG.
Outpatient drug therapy. For various reasons, patients with CVDs may receive medication, frequent examples being antithrombotic, antihypertensive, and antipsychoanaleptic drugs. Extensive information about drug administration is registered by the Institute for Medical Statistics (IMS). Certain drugs are administered for the treatment of CVDs; in this, comorbidity may play an important role. According to the IMS, drugs are usually prescribed by the GP (G. Meurs, unpublished data, 1996). Table 2⇓ presents an overview of the costs of drug treatment for CVDs. The costs are based on a data bank that includes the outpatient use of drugs in a sample of 3.2 million patients covered by the insurance act.14 Extrapolated for the Dutch population as a whole, it is estimated that the outpatient drug cost for CVDs was about 95 million DG.
Indirect Costs Attributable to CVDs
Absenteeism. Patients with CVDs are often absent from work or incapacitated, which leads to productivity losses for society. This study includes only the production loss for paid work because little information is available regarding unpaid work. During absences, individuals receive a pension that is regarded as a transfer payment. This refers to workmen’s compensation payments, which are a cost to the paying party and a gain to the one receiving the pension, and neither a cost nor a gain to society. However, to gain an insight into the total amount of production loss, we used these absenteeism registers.
In the Netherlands, the ill employee is entitled to an absenteeism benefit for the first year of being registered as sick. A problem arises when the medical reason for being on the sick list is examined, especially for short-term absenteeism (up to 7 days); eg, in 9% of the cases, the absentee was not seen by an insurance doctor. This number decreases to 5.7% for long-term absenteeism (≥183 days). Regarding the duration of the illness, we assume that it is likely that patients with CVDs are diagnosed by an (insurance) physician. In 1993, there were 259 081 days of absence under the Sickness Insurance Act due to CVDs, which is about 0.2% of the total days of absenteeism (G. Meurs, unpublished data, 1996). About 56% of the work force is registered under the Sickness Insurance Act by occupational associations.34 The number of days is therefore multiplied by 1.79, resulting in a total number of 462 645 days of absence. The average gross annual income in 1993 for the active participants of the work force aged between 18 and 64 years was about 55 900 DG.35 To obtain the average salary per employee per day, this figure should be divided by 365 (the sick-listing is based on all days, including holidays and weekends). The result is an average daily income of 153 DG. Based on this figure, the total amount of indirect costs due to absenteeism is about 70.8 million DG.
Disability. In the Netherlands, after the 1-year period of absence described above, an employee who fulfills certain conditions may be entitled to a disability allowance because he or she is unable to work. All persons who are entitled to a disability benefit, either according to the Disability Insurance or the General Disability Act, are registered. In 1993, 1 791 596 converted benefit days (0.9%) were due to stroke (J.W. Nool, unpublished data, 1996). Converted benefit days are the number of benefit days corrected for the percentage of disablement and the duration of the disablement, eg, one person who is disabled full-time for the whole year incurs 261 converted benefit days (excluding holidays and weekends).36
The average gross yearly income (including special payments and transfers) in 1993 for persons aged between 18 and 64 years was about 55 900 DG.35 To obtain the average salary per employee per (converted) day, this figure should be divided by 261, giving an average daily income of 214 DG. Thus, the mentioned 791 596 days correspond to 383.4 million DG.
Premature death. Despite a reduction in stroke mortality in the last 30 years, stroke is the third leading cause of death in the Netherlands. Mortality registration by diagnosis in the Netherlands is obligatory, ie, the cause of death has to be registered before the deceased is buried or cremated.24 According to these registrations, 12 907 patients died in 1993 because of stroke, 1130 of whom were 64 years old or younger.37 The number of years of life lost before the age of 65 years from CVDs (ICD-9 430 to 438) in 1993 was 12 595 years.38 One might assume that the years preceding age 65 are the productive years. With the prevalence approach, calculation of mortality costs considers salaries over a lifetime rather than a single year. The rationale for this approach is that if an individual had not died in a given year, that person would have continued to be productive until retirement (65 years); therefore, the present value of future losses is used as an appropriate estimate. However, if these patients had not died, their participation in the working population would not be 100% until retirement (65 years). In the Netherlands, in 1993, 62% of the population had a paying job; this percentage varied with age: 80% in the 25- to 34-year and 26% in the 55- to 64-year age category.39 In this study, the production loss due to mortality was corrected for this participation in employment, resulting in a loss of 5493 production years because of CVDs in 1993.
The economic value of an individual is measured according to salaries. As mentioned above, the average gross yearly income (including special payments and transfers) in 1993 for employed persons aged between 18 and 64 years was about 55 900 DG.35 To value the future production losses due to premature death, these costs need to be discounted. The process of discounting converts a stream of earnings into its present value. With a moderate discount rate of 5%, the costs of mortality due to CVDs would be 112 million DG.
Table 3⇓ shows that the total cost of CVDs for the Dutch society in the year 1993 was about 2532 million DG. The total direct cost amounted to 1966 million DG, which is about 3% of the healthcare budget of 57 781 million spent in 1993.
Comparison With the Results of Other Studies
The results are comparable with previous studies.2 3 4 5 6 Bergman et al2 calculated the healthcare costs of patients that were incident in 1991 and came up with an amount of 4% of the total costs. Koopmanschap et al5 estimated costs of CVDs in a top-down COI study. In this approach, the costs of several types of care are divided by the main disease categories. The total costs for stroke were estimated to be 2.9% of the total costs of diseases in 1988.
A sensitivity analysis usually tests relevant factors to highlight sensitive components of the calculation. In the present study, a sensitivity analysis was considered for those estimates for which no reliable data were available, and informed guesses were made, and furthermore for data for which different value judgements were found. If the results of such manipulations are minor, one can have greater confidence in the results. If the sensitivity analysis produces large changes in the results, then greater caution is necessary when interpreting the data.
All costs in Table 3⇑ are calculated pro rata, ie, we assume that if a certain percentage of patients in a service have CVDs, they will incur proportionally the same percentage of care and the same percentage of costs. However, it may be the case that patients with CVDs require more or less care than patients with other diagnoses. In general, the figures in Table 3⇑ are based on national statistics. This applies to hospitals, nursing homes, absenteeism, work disability, and mortality. Table 3⇑ also displays the costs of each particular service discussed above as a fraction of the total (direct and indirect) costs. The fractions give an indication of the sensitivity values of the various items. Regarding the direct costs, the conclusion will be most responsive to variation in nursing homes, hospitals, and nursing home care. For some of these healthcare facilities (general practices and hospitals), several assumptions were made. With regard to hospital stay, it makes a difference whether one calculates the cost on the basis of number of days in hospital rather than the number of discharged patients (eg, the costs would vary between 295 and 706 million DG). The costs of home nursing care were based on the information provided by one Home Nursing Organization and differ from the results reported by Koopmanschap et al.5 A range of ±5% is chosen to test the likelihood of these figures. Based on this range, the total costs for home nursing care would vary from 241 million to 585 million DG. Finally, if the prevalence of GP consultation ranged from 7 per 1000 (minimum)20 to 10 per 1000 (maximum),18 the corresponding GP costs would vary from 15 to 22 million DG. The cumulative effect of these changes in medical consumption on the total costs, using the minimal and maximal plausible values, would give rise to a ±15% change.
Table 3⇑ also shows that the results of this study are very sensitive to changes in the indirect costs. This study only includes production loss by patients. However, family members and other caregivers sometimes also have to quit their jobs or reduce their working hours. The indirect costs in our study are based on the human capital approach, which regards persons as human resources, and their production loss is valued by salaries. However, this method has several disadvantages. When using this approach, the value of unpaid work (including that of the retired) is not included, and the value of life over and above economic productivity is completely ignored. This may induce an underestimation because the total number of hours of unpaid work is 1 to 1.5 times higher than the total number of hours of paid work.40 Second, it is assumed that the production of the worker is roughly the same as the salary. Third, the assumptions of the human capital approach are only accurate when sick or deceased people are not replaced. This may not be true in periods with high unemployment rates. Recently, a new method, the friction cost method, has been developed that takes the limitations of the human capital approach into account.41 42 The basic principle of this method is that the amount of production lost because of disease depends on the time-span organizations needed to restore the initial production level. For the friction cost method, a lot of additional information is required that is difficult to obtain (eg, one has to measure the real production loss in different jobs and locations over the period needed to find a replacement for a lost worker).
Another issue is that the human capital approach assumes that persons are fully employed. However, only 75% of the employees work more than 35 hours per week.38 Beside these shortcomings, the costs of absenteeism and disablement were computed rather than estimated and are therefore not subject to uncertainty. However, in calculating the cost of mortality, a moderate discount rate of 5% was chosen with regard to the present lifetime earnings. For the sensitivity analysis, the costs of mortality are assessed using 2% and 8% discount rates (eg, the corresponding mortality costs would vary from 71 to 206 million DG).
The purpose of the present study was to assess the burden that stroke places on the Dutch society. Its scope is restricted to the year 1993, and both direct costs (medical consumption) and indirect costs (production losses) are included. The total cost of CVDs is about 2.5 billion DG, of which 1.9 billion are for medical consumption. In accordance with other studies, we found that the costs of medical consumption are mostly generated by inpatient costs (eg, nursing homes and hospital stay). The inpatient costs in hospitals may be high because of the need for waiting lists before discharge to a rehabilitation center or nursing home. In the Netherlands, the government and providers of health care are striving to improve organizational aspects to ameliorate the routing of stroke patients through the healthcare system and to improve the effectiveness and efficacy of care (eg, stroke units).
In health care as a whole, the indirect costs exceed the direct costs. Following Hartunian et al,9 this study shows that the indirect costs for CVDs do not exceed the direct costs. This is because patients with CVDs are older and therefore lose relatively fewer productive years.9
This study estimates the total cost to society of patients suffering from CVDs in the Netherlands. Some items, although they might be important, have been difficult to value in monetary terms. Psychosocial aspects such as pain, grief, and suffering therefore have not been taken into account here. For the analysis, we used the primary diagnosis. It is assumed that all costs can be assigned to CVDs; therefore, comorbidity is not taken into account separately, possibly leading to some overestimation.
Traditionally, COI studies such as this are based on secondary data sources. The estimation of costs of stroke on a national scale is greatly dependent on the quality and precision of the data used. In this study, in addition to national statistics (eg, hospitals, nursing homes, absenteeism, work disability, and mortality) and national surveys (eg, outpatient drug therapy, physiotherapy, exercise therapy, and speech therapy), several other sources have been used to collect data on cost, such as records available through health service suppliers (eg, nursing homes), case records, and continuous morbidity registers (eg, GPs). For these other sources, all results have been extrapolated to a national level, ie, we assume that if a certain percentage is found in a specific patient population, the same percentage of care and costs will apply to the national population. These assumptions may have induced some bias; we therefore performed a sensitivity analysis. As shown in the section on sensitivity analysis, the results of this study may to some extent be subject to uncertainty. The cumulative effect of modifications in direct costs results in a 15% change. In addition, in the Netherlands, no data are available for certain aspects such as prevention, mental health care, and family burden; these have not been included in the analysis. Because of these deficiencies, and because of the outcome of the sensitivity analysis, the results of this study should be treated with some caution.
A number of possible insights can be gained from COI studies such as this one. The most important is that COI studies are another way of highlighting the importance of a particular disease, over and above the more epidemiological estimates of morbidity and mortality.43 It is often the case that ranking disease in terms of economic burden merely mirrors the ranking one would obtain by other methods. In this way, COI calculations can help to determine medical research priorities. However, a word of caution is warranted: COI studies, without other information, do not tell us whether more resources should be devoted to treating the diseases concerned. When deciding on allocation of scarce resources, one must also take into account treatment options available, their cost, and their effectiveness.43
This study describes the situation in 1993. It would certainly be of interest to forecast the development of costs of CVDs in the future. Factors that may influence future costs include demographic changes, new therapies, and cost-reduction programs introduced by the Dutch government. The data of this study could be used, together with prognoses of the trends in CVDs, as the basis for forecasting future costs of health care and social services. Such forecasts would be of particular interest for policy planning.
Selected Abbreviations and Acronyms
|ICD-9||=||International Classification of Diseases, 9th Revision|
|COI||=||cost of illness|
|NIVEL||=||Netherlands Institute of Primary Health Care|
This study is supported in part by a grant from the Fund for Developmental Medicine, Health Insurance Executive Board, the Netherlands. The authors thank Professor Beuls, Professor Groot, Professor Troost, Dr Blaauw, Dr Boiten, Dr Franke, Dr Hupperts, MSc Leffers, Dr Lodder, Antonette Bok, and Mariëlle Goossens for reviewing earlier versions of the manuscript. We would also like to express our gratitude to two anonymous reviewers for comments on this manuscript. Furthermore, we would like to thank the staff of the Medical and Social Information Center of Maastricht University for their help in gathering data, and Mr Berger (Instituut voor Medische Statistiek), Mr Habets (Groene Kruis Heuvelland), Mr Meurs (College van Toezicht Sociale Verzekeringen), Mr Nool (College van Toezicht Sociale Verzekeringen), and Mr Piepenbrink (Ziekenfondsraad) for access to information.
- Received November 28, 1996.
- Revision received March 28, 1997.
- Accepted March 28, 1997.
- Copyright © 1997 by American Heart Association
Bergman L, van der Meulen J, Limburg M, Habbema DF. Cost of medical care after first-ever stroke in the Netherlands. Stroke. 1995;26:1830-1836.
van den Bosch JH, Kardaun JWPF. Diseases of Nervous System Disorders in the Netherlands: an Exploration of the Magnitude and Consequences [in Dutch]. The Hague, Netherlands: Staatsdrukkerij Uitgeverij; 1993.
Koopmanschap MA, van Roijen L, Bonneux L. Costs of Illness in the Netherlands [in Dutch]. Rotterdam, Netherlands: Instituut Maatschappelijke Gezondheidszorg Instituut voor Medische Technology Assessment, Erasmus Universiteit Rotterdam; 1991.
Meyboom-de Jong B, Buis J. Care Following Stroke: a Publication of the Netherlands Heart Foundation Concerning the Consequences of Stroke [in Dutch]. The Hague, Netherlands: Nederlandse Hartstichting Vakgroep Huisartsgeneeskunde Rijksuniversiteit Groningen; 1995.
Evers SMAA, Ament AJHA. Costs of schizophrenia in the Netherlands. Schizophr Bull. 1995;21:141-153.
Tweede Kamer der Staten-Generaal. Financial Statement of Care (FOZ), 1996 [in Dutch]. The Hague, Netherlands: Staatsdrukkerij Uitgeverij; 1995.
National Hospital Institute. Financial Statistics 1993, Specialized Hospitals [in Dutch]. Utrecht, Netherlands: Nederlands Ziekenhuis Instituut; 1994.
National Hospital Institute. Financial Statistics, General Hospitals [in Dutch]. Utrecht, Netherlands: Nederlands Ziekenhuis Instituut; 1994.
National Hospital Institute. Financial Statistics 1993, University Hospitals. Utrecht, Netherlands: Nederlands Ziekenhuis Instituut; 1994.
National Health Service Council (ZFr) Drug Information Project (GIPeilingen). GIPeilingen, Statistics Pharmaceutical Aid, 1994 [in Dutch]. Amstelveen, Netherlands: Geneesmiddelen Informatie Project Ziekenfondsraad; 1995.
Rijkuniversiteit Groningen Vakgroep Huisartsgeneeskunde. Patient With a Cerebrovascular Disease and the General Practitioner [in Dutch]. Groningen, Netherlands: Rijksuniversiteit Groningen, vakgroep huisartsgeneeskunde; 1992.
van der Meer K, Smith RJA, Bremer GJ. Cerebrovascular Diseases Gauged, the Incidence of Cerebrovascular Diseases and Transient Cerebral Ischaemia in the General Practices [in Dutch]. Utrecht, Netherlands: Netherlands Institute of Primary Health Care (NIVEL); 1990.
van den Hoogen HJM, Huygen FJA, Schellekens JWG, Straat JM, van der Velden HGM. Morbidity Figures From General Practice, Data From Four General Practices. Nijmegen, Netherlands: Nijmegen University Department of General Practice; 1985.
van de Lisdonk EH, van den Bosch WJHM, Huygen FJA, Lagro-Janssen ALM. Morbidity Figures From General Practices [in Dutch]. Utrecht, Netherlands: Wetenschappelijke uitgeverij Bunge; 1994.
van der Velden J, de Bakker DH, Claessens AAMC, Schellevis FG. Dutch National Survey of General Practices, Morbidity in General Practices. Utrecht, Netherlands: Netherlands Institute of Primary Health Care; 1992.
Lamberts H. In the Practice of the General Practitioner: Report of the Transition Project [in Dutch]. Lelystad, Netherlands: Meditekst; 1991.
University of Limburg Registration Network of Family Practices. Health Problems and Diagnoses in General Practices [in Dutch]. Maastricht, Netherlands: Rijksuniversiteit Limburg; 1993.
Franke CL. Intracerebral Hemorrhage. Heerlen, Netherlands: De Wever Ziekenhuis, Department of Neurology; 1991. Thesis.
Information Centre for Health Care (SIG). LMR Data on Hospitals [in Dutch]. Utrecht, Netherlands: Information Centre for Health Care; 1993.
Bouter LM, van Dongen MCJM. Epidemiologisch onderzoek, opzet en interpretatie. Houten/Zaventem: Bohn Stafleu Van Loghum; 1991.
Landelijk Coordinatiepunt Niet aangeboren Hersenletsel. Rehabilitation Care for Persons with Non-Congenital Brain Damage [in Dutch]. Utrecht, Netherlands: National Hospital Institute; 1994.
LIvRE. National Information System for Rehabilitation, Yearbook 1994 [in Dutch]. Utrecht, Netherlands: Landelijk Informatiesysteem voor Revalidatie; 1995.
Information Centre for Health Care (SIG). Yearbook Nursing Care 1993 [in Dutch]. Utrecht, Netherlands: SIG zorginformatie; 1994.
Central Bureau of Statistics. Monthly Bulletin of Population Statistics [in Dutch]. Voorburg/Heerlen, Netherlands: Centraal Bureau voor de Statistiek; 1996.
van der Valk RWA, Dekker J, Boschman M. Basic Figures of Outpatient Physiotherapy 1989-1992, Figures From the BEEF Project [in Dutch]. Utrecht, Netherlands: Nederlands instituut voor onderzoek van de gezondheidszorg; 1995.
Roebroeck ME, Hutten JBF, Kerssens JJ, Dekker J. The Magnitude of Physiotherapeutic Care for Several Patient Groups [in Dutch]. Utrecht: Nederlands instituut voor onderzoek van de gezondheidszorg; 1995.
Gisbergen MJWM, Dekker J. Reliability of the diagnosis of impairments and disabilities by exercise therapists. J Rehabil Sci. 1992;5:67-73.
Zuijderduin WM, Dekker J. Exercise Therapy Cesar and Exercise Therapy Mensendieck in the Dutch Health Care [in Dutch]. Utrecht, Netherlands: Nederlands instituut voor onderzoek van de eerstelijnsgezondheidszorg; 1994.
Raaijmakers MF, Dekker J. Speech Therapy in the Dutch Health Care [in Dutch]. Utrecht, Netherlands: Nederlands instituut voor onderzoek van de gezondheidszorg; 1995.
Social Insurance Council. Sickness Insurance Act 1993. Zoetermeer, Netherlands: Sociale Verzekeringsraad; 1994.
Trimp L. Income Distribution 1993 [in Dutch]. Sociaal-economische maandstatistieken; 1994:13-15.
CTSV. Chronicle of Social Insurance [in Dutch]. College van Toezicht Sociale Verzekeringen; 1995.
Central Bureau of Statistics. Mortality by Cause of Death 1993 (A1), Primary Cause of Death in the Netherlands [in Dutch]. Voorburg/Heerlen, Netherlands: Centraal Bureau voor de Statistiek; 1995.
Central Bureau of Statistics and the Ministry of Health Welfare and Sports. Vademecum of Health Statistics of the Netherlands 1995. Voorburg/Heerlen, Netherlands: Centraal Bureau voor de Statistiek; 1995.
Central Bureau of Statistics. Labor Force Survey 1993 [in Dutch]. Voorburg/Heerlen, Netherlands: Centraal Bureau voor de Statistiek; 1994.
Bruyn-Hundt M. The Economics of Unpaid Work. University of Limburg, Netherlands; 1996. Thesis.
Koopmanschap MA, Ineveld van BM. Towards a new approach for estimating indirect costs of disease. Soc Sci Med. 1992;34:1005-1010.