The demographic impact on ambulatory pharmaceutical expenditure in Belgium

Health Policy 45 (1998) 1 – 14

The demographic impact on ambulatory pharmaceutical expenditure in Belgium Rudy Van Tielen *, Frank Peys, Jean Genaert Belgian Institute of Health Economics, Rue Leon Theodorstraat 102. 1090 Brussels, Belgium Received 3 February 1998; accepted 21 April 1998

Abstract It has become customary in our society to attribute the rise in health care expenditure to a large extent to the aging of the population. The Belgian Institute of Health Economics (BIGE-IBES) finalised a study to measure the impact of the demographic trend on public spending for ambulatory pharmaceuticals. The results show that the changes in age structure as such have a relatively small impact on health care expenditure. Over the period 1986 – 1996 an annual growth of + 0.73% was observed, while total public spending showed an annual average increase of +6.2% (in constant francs 1996). This first static demographic impact will become +0.75% per year for the next decade 1996 – 2006. It is, however, possible to calculate the extent to which per capita expenditure envolved differently according to age group: over 1986–1996 expenditure for the elderly was rising more sharply than for the younger age groups. This leads to a second dynamic demographic impact which will cause an annual growth rate of +1.91% for the next decade, according to the working hypothesis put forward. This growth rate is to be compared with the authorized budgetary increase of +1.5% per year (inflation excluded), according to the EU-Maastricht quotes…, and it does even not take into account the other increasing factors as technological innovation and epidemiological needs. © 1998 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Belgium; Pharmaceutical expenditures; Age structure; Ambulatory care

* Corresponding author. Tel.: +32 2 4211504; fax: + 32 2 4211509. 0168-8510/98/$19.00 © 1998 Elsevier Science Ireland Ltd. All rights reserved. PII S0168-8510(98)00026-8

2

R. Van Tielen et al. / Health Policy 45 (1998) 1–14

1. Introduction Economic experts often refer to the ‘grey revolution’ to explain the growing demand for health care resources. The impact of population aging on health care expenditure requires special attention in all OECD countries [1–5]. Aging implies the increase in the number as well as the proportion of elderly people. For the purposes of this article, we will be talking about the aging population as a whole. In order to somehow control health care expenditure, successive governments over the last two decades have worked out all sorts of rationalization plans of which we do not yet know the impact on individual and collective consumption patterns. Up to now, the major parameter to justify the investments made in health care is the life expectancy. Paradoxically, politicians and health care officials will be more concerned about the future of the social security system as this indicator improves [6 – 8]. However, when we want to examine the impact of the age factor on public spending, we find that the official Belgian statistics do not provide any data per age group. They mention criteria such as TIP (normal insured person) and VIPO (widows, disabled, pensioners and orphans). Within the VIPO group, certain categories are entitled to preferential reimbursement for certain health care services. With such a classification, the notion of ‘age’ remains rather vague. A particular status will depend on the claimant and his situation on the labour market. Many elderly (over 60) are still classed as TIP because they still have ample professional or other income, while in the VIPO group we also find young people such as widows and orphans. Other methods will therefore need to be found to assess the impact of the age factor on health care expenditure in general and on pharmaceutical expenditure in particular. That is why this study will use an ‘indirect’ method to determine health insurance expenditure on ambulatory pharmaceuticals (i.e. excluding magistral preparations), divided according to age group. This enables to analyze the demographic impact on expenditure over the period 1986–1996. The study also tries to project this impact for the decade 1996 – 2006.

2. Methods The total INAMI – RIZIV (National Sickness Insurance Office) reimbursement figures for ambulatory pharmaceuticals are published each year. However, no distribution is given according to age group. We used the statistics of IMS Belgium, which give an estimate of drug sales in retail pharmacies on the basis of a representative panel of Belgian wholesale pharmacies. In our calculation model we took account of the type of drug and its specific reimbursement level, the retail price of the drug, the patient’s share, the share of TIP and VIPO (with or without preferential status), and the estimated quantities sold (by IMS Belgium). In this way we recalculated the INAMI-RIZV reimbursements according to product and therapeutic class. These calculations were made for 1986 and 1990. We experienced

R. Van Tielen et al. / Health Policy 45 (1998) 1–14

3

some difficulty calculating the reimbursement of certain pharmaceuticals. Some are reimbursed only for certain specific indications (reimbursable pharmaceuticals with an ‘F status’). Compliance with those indications is verified by a consulting doctor of the sickness fund of the patient. Nevertheless, we were able to determine a coefficient for each of the pharmaceuticals in question. The IPhEB (Belgian Institute of Pharmacoepidemiology) supplied the reimbursement figures per therapeutic class for 1993 and 1996. These figures were provided by the tarification offices of the member retail pharmacies (61% of the Belgian pharmacies in 1993 and around 80% in 1996). We validated our model with the figures of the IPhEB. The results were compared and adjusted accordingly. Nevertheless, the methodology and data used do not enable us to work with confidence intervals. Since pharmaceutical expenditure is directly dependent on the number of doctor’s visits at which a medical prescription is issued, we used another IMS Belgium database [9]. This gives an idea of the number of prescriptions per age group as well as of the indications for which they were issued. The panel of doctors changes regularly. Once we have a definite idea of the estimated total number of prescriptions per therapeutic class, the estimated total cost per therapeutic class and the distribution of the prescriptions according to age group, it is relatively easy to determine the weighted cost per age group and per therapeutic class. An extrapolation of the sum of the weighted cost per age group and per therapeutic class will then give an idea of the distribution of the total public spending on ambulatory pharmaceuticals according to age group. When we then divide the resulting amounts by the number of persons statistically present in each age group, we get the per capita expenditure (not per patient). The population data come from the INS-NIS (National Statistical Office) [10]. To calculate the demographic impact a methodology is used which has been applied in several OECD countries for some 15 years already [6,8,11,12]. First we calculated a static demographic impact in which (all parameters being equal) no account is taken of changes in either the supply or demand structure. In this way we are able to determine the annual variation in expenditure on reimbursable pharmaceuticals, caused only by changes in the age structure, and assuming that the expenditure per age group remains constant in relation to the reference year (in this case 1996). In order to evaluate the static demographic impact (growth and aging of the population) on public spending on pharmaceuticals, we used the following index [11,13]: DEMOGRAPHIC INDEX S(POPi,96EXPphar.inamii,96) = (Index 1) S(POPi,86EXPphar.inamii,96) The reference year is 1996 POPi,96 =total population in age group ‘i’ in year 1996.

R. Van Tielen et al. / Health Policy 45 (1998) 1–14

4

POPi,86 =total population in age group ‘i’ in year 1986. EXPphar.inamii,96 =per capita public spending on pharmaceuticals in age group ‘i’ in 1996. The impact of population aging on ambulatory pharmaceutical expenditure between 1986 and 1996 will therefore depend on both the magnitude of the changes in age structure and the division of expenditure according to age group. That is why we used a second index to isolate the impact of population aging while keeping the population size constant. OVERALL POPULATION AGING S(%POPi,96EXPphar.inamii,96) = S(%POPi,86EXPphar.inamii,96) (Index 2) The calculation of these two indices allow us to determine the (static) effect of demographic changes. The expenditure recorded in 1996 serves as a cost reference for the different estimates. After having analyzed the variations in the total and average (per capita) expenditure per age group, we made the second parameter of the above-mentioned indices vary, i.e. the per capita expenditure per age group. In this way we were able to estimate the more relevant dynamic demographic impact on future expenditure [6,12]. The dynamic demographic impact is calculated in the same way as index 1, though taking into account a ‘capitalization’ factor (expenditure differential growth) proper to each age group. Before analyzing the demographic impact on the expenditure under consideration, the overall picture of the trend in the total public spending on ambulatory pharmaceuticals is given. Fig. 1 shows this expenditure for the years in question. An increase in expenditure by a factor of 2.5 is observed, which corresponds to a 147% increase over 10 years (in current francs), or 9.4% per year. The increase in constant francs (1996) amounts to 86.6%, or 6.4% per year, for the same period. By applying the above-mentioned indices we can quantify the demographic impact on the increase in total expenditure on pharmaceuticals reimbursed in the ambulatory sector.

3. Results

3.1. The demographic impact on public spending on ambulatory pharmaceuticals: definition of the ‘static’ factor First an estimate is made of the combined contribution of the absolute aging and the growth of the population, in terms of annual percent, to the rise in public spending over different periods between 1986 and 1996, as well as of the projection for the next 10 years (1996 – 2006). Table 1 shows the annual variations in public spending by distinguishing the aging of the population as a whole and its growth. As was to be expected, these

R. Van Tielen et al. / Health Policy 45 (1998) 1–14

5

Fig. 1. Trend in public spending on ambulatory pharmaceuticals, 1986 – 1996.

figures show that the effect of the shift in the different age groups is far more significant than the population growth. The impact of the aging of the population on public spending on pharmaceuticals over the period 1986 – 1996 is 5.3%, which corresponds to an average of 0.52% per year. The impact of the age factor can be compared to the real increase in public spending on pharmaceuticals. The variation in this expenditure (constant francs 1996), per capita, over the same period came to + 86.6%, or an average of 6.2% per Table 1 Annual variations in total public spending on pharmaceuticals, attributable to the growth and aging of the population, 1986–1996 Period

Annual population growth (%)

Annual aging impact (%) (Index 2)

Overall annual demographic impact (%) (Index 1)

1986–1990 1990–1993 1993–1996 1996–1999 1999–2003 2003–2006

0.05 0.45 0.17 0.25 0.20 0.15

0.64 0.50 0.40 0.45 0.63 0.54

0.69 0.95 0.57 0.71 0.83 0.69

1986–1996 1996–2006

0.21 0.20

0.52 0.55

0.73 0.75

Sources: IMS Belgium, IPhEB, National Statistical Office, BIGE-IBES calculations.

6

R. Van Tielen et al. / Health Policy 45 (1998) 1–14

year. This means that a number of factors other than population aging are responsible for most of the increase in public spending. The overall annual demographic impact was +0.73% for the previous decade and is calculated at +0.75% for the next decade: these are the static demographic factors.

3.1.1. What role could the general aging of the population play in future expenditure? Table 1 also shows the projected annual increase in public spending, taking into account the changes in age structure over the period 1996–2006. These projections are obtained through linear extrapolations whereby the per capita expenditure in 1996 is multiplied by the estimated number of persons in each respective age group from 1996 to 2006. The results show that public funding for the reimbursement of pharmaceuticals will continue to increase as a result of aging by an annual percentage which is barely higher than what we have known in the past decade: 0.55% compared with 0.52%. From Table 1 we can conclude that the impact of the aging of the population as a whole on public funding for the reimbursement of pharmaceuticals is fairly moderate. This view is entirely consistent with the findings of similar European studies [9,11,14,15]. However, the modest impact of overall aging on the expenditure under consideration does not mean that the relative aging of a population composed of increasing numbers of elderly people is not significant. The total expenditure is conditioned by the number of persons physically present in each age group as well as by the fact that expenditure on pharmaceuticals for the elderly is rising more rapidly than for the other age groups. 3.2. Public spending on pharmaceuticals per age group: definition of the ‘dynamic’ demographic factor for the period 1986 – 1996 All calculations described so far are based on the hypothesis that the relative expenditure per age group corresponds to that calculated for 1996 (reference year) and that it would remain constant in time. However, the historical data for the period 1986 – 1996 show significant variations in relative expenditure according to age group. Fig. 2 gives a detailed overview of the total expenditure per age group. Whether the figures are expressed in constant francs or in current francs, the trend in expenditure remains fairly spectacular. Table 2 gives a detailed survey of the trend in total public spending on pharmaceuticals, per age group, between 1986 and 1996. The total expenditure for the over-65s is clearly increasing faster than the trend observed in Fig. 1. On the other hand, for the younger age groups a remarkably smaller increase in percentage terms can be seen than the increase in total expenditure. Figs. 3 and 4 show the relative share of each age group in the total population and the relative share of each age group in the total public spending on pharmaceuticals.

Fig. 2. Total public spending on pharmaceuticals per age group (constant francs 1996).

R. Van Tielen et al. / Health Policy 45 (1998) 1–14 7

R. Van Tielen et al. / Health Policy 45 (1998) 1–14

8

Table 2 Comparison of the trend, in percentage terms, in total public spending on reimbursable ambulatory pharmaceuticals, 1986–1996 Current francs Variation (%) 1986–1996

Constant francs 1996 Annual variation (%)

Variation (%) 1986 – 1996

Annual variation (%)

0 – 4 years 5 – 11 years 12 – 19 years 20 – 29 years 30 – 39 years 40 – 54 years

+103 +72 +64 +95 +119 +141

+7.3 +5.6 +5.1 +6.9 +8.1 +9.2

+53 +30 +24 +47 +65 +82

+4.3 +2.6 +2.2 +3.9 +5.1 +6.1

55 – 64 years 65 – 79 years 80 years and older

+125 +215 +163

+8.4 +12.1 +10.1

+70 +138 +98

+5.4 +9.1 +7.1

Global average

+147

+9.4

+86

+6.4

Source: BIGE-IBES calculations.

We see at once that in 1986 the over-55s represented 25.4% of the population, but accounted for 54.7% of the total expenditure. In 1996 that same group represented 26.8% of the population and generated around 60.5% of the total expenditure! The age group between 65 and 79 was the only one to witness a relatively steady increase over the 10-year period. Contrary to what one might expect, the relative share of expenditure by the over-80s has remained remarkably stable in the course of time. This would mean that the very old present a different consumption pattern to the other age groups. When we analyze the trend in expenditure per individual of each age group, we get the trends shown in Table 3 and Fig. 5. As with the total expenditure, the rates of increase are above the average for the older age groups. The age bracket under 55 shows an average per capita trend of +4.9% per year, and the over-55s + 6.8% per year, which is a difference of 1.84% per year (in constant francs 1996). There are many factors, all interdependent, that can explain the differences in trend of the expenditure as shown in Table 3, but the differential of 1.84% seems to us to be specifically related to the age factor and can considered as a ‘dynamic’ demographic factor.

3.3. Assessment of the dynamic demographic impact for the period 1996 – 2006 The results of Table 3 have urged us to reconsider more carefully the projections for the next 10 years. In our calculations we had linearly extrapolated the demographic impact by multiplying the per capita expenditure for 1996 by the estimated number of persons in each age group in the future. This was the static

R. Van Tielen et al. / Health Policy 45 (1998) 1–14

9

demographic impact. However, it would be more correct to have the per capita expenditure vary as well, since a clear difference was observed according to age group in the annual growth in expenditure. In order to assess this dynamic demographic impact, we focused our attention on the trend in per capita expenditure for the persons under 55 on the one hand and those over 55 on the other. By way of sensiti6ity test, we will put forward the hypothesis that the per capita expenditure will rise by + 1.84% per year for the age groups over 55, up to the year 2006. The figure for the age groups under 55 will be kept constant. The used formula and the results of this dynamic demographic impact are shown in Table 4. As opposed to the static demographic impact which was determined early on in this study ( +0.75% per year for the next decade), the dynamic demographic impact gives a totally different picture of the trend in ambulatory pharmaceutical expenditure: +1.91% per year for the next decade, according to our hypothesis.

3.3.1. How can the difference be explained between the static demographic impact of +0.75% per year between 1996 and 2006, and the dynamic demographic impact of +1.91% per year for the same period? According to several international studies, this discrepancy should be explained by a ‘generation effect’, which is primarily related to the evolution in consumer behaviour of the different generations in a society. The elderly appear to have benefited particularly from the per capita growth in national wealth. Their attitude to health care has been carefully studied by the French. For the same morbidity rate, our parents seem to have taken the habit of looking after themselves better

Fig. 3. Share in percentage terms of each age group in the total population.

10

R. Van Tielen et al. / Health Policy 45 (1998) 1–14

Fig. 4. Relative weight of each age group in the total public spending on pharmaceuticals.

than our grandparents, who in turn were already looking after themselves better than our great-grandparents. The technological progress in medicine certainly does not diminish the cost of medical care, for several reasons. New drugs do not displace the existing ones. Moreover, medical progress reveals new needs which

Fig. 5. Public spending on pharmaceuticals per capita: 1986 – 1996 (constant francs 1996).

+85 +47 +118 +129 +104 +105 +144 +163 +128 +142

+6.3 +3.9 +8.1 +8.6 +7.4 +7.5 +9.3 +10.1 +8.2 +9.2

+40 +11 +65 +73 +54 +55 +84 +98 +72 +83

Annual variation (%) Variation (%) 1986 – 1996

Constant francs 1996

a

Source: BIGE-IBES calculations. The decrease is due to the fall in birth rate during the period 1938 – 1945.

0–4 years 5–11 years 12–19 years 20–29 years 30–39 years 40–54 years 55–64 years 65–79 years 80 years and older Overall average

Variation (%) 1986 – 1996

Current francs

Table 3 Comparison of the trends in public spending on pharmaceuticals per capita, 1986 – 1996

+3.4 +1.1 +5.1 +5.6 +4.4 +4.5 +6.3 +7.1 +5.6 +6.2

+0.9 +1.5 −2.8 −1.6 +0.7 +1.6 −0.7a +1.8 +1.4 +0.21

Annual variation (%) Annual demographic trend (%)

R. Van Tielen et al. / Health Policy 45 (1998) 1–14 11

12

R. Van Tielen et al. / Health Policy 45 (1998) 1–14

Table 4 Trend in public spending on ambulatory pharmaceuticals according to our hypothesis for the period 1996–2006 (constant francs 1996) 1996

Total public spending Growth of total population Annual population growth Impact of aging over the period 1996–2006 Annual impact of agingb Total demographic impact 1996–2006 Annual demographic impactc

49.9 billion

2006 Hypothesisa 60.2 billion +2.1% +0.2% +18.7% +1.71% +20.8% +1.91%

Source: IBES-BIGE calculations. a Our hypothesis assumes that the annual per capita spending of the over-55s increases, each year, +1.84% faster than the under-55s per capita spending. b The dynamic impact of aging is determined in the same way as index 2. The over-55s spending per capita in the numerator is simply multiplied by a ‘capitalization’ factor equal to the difference in the observed trend in percentage terms between the under-55s and the over-55s, per capita, over the period 1986–1996. See below. c Same remarks as (1), but for index 1 that becomes: DI=

S(Popi(B55 year), 2006*EXPpharm.inamii(B55 year), 1996)+S(Popi(\55 year), 2006*EXPpharm.inamii(\55 year), 1996*(1+0.0184)10) S(Popi,1996*EXPpharm.inamii 1996).

were either unknown or incurable before, and which automatically create a new demand.

4. Discussion and conclusion The Belgian population is only growing very slowly year by year. Even if this phenomenon remains of a fairly modest scale, the level of expenditure is yet certain to rise. Moreover, the demographic structure of the population clearly shows that the proportion of so-called elderly people is increasing in relation to the other age groups. This trend also brings about changes in the consumption pattern of drugs and health care provisions in general. The government has already taken a whole series of measures - mostly with a view to Belgium’s membership of the European Monetary Union—to contain health care expenditure. One of the main objectives was to limit the annual growth in health care expenditure to 1.5% per year, excluding inflation. Given this reality, i.e. a natural demographic population growth versus the government objective of cost containment, we thought it was a good idea to study and even measure the impact of the demographic trend on public spending on ambulatory pharmaceuticals. The results show that the changes in age structure as such have a relatively small impact on health care expenditure. Over the period 1986–1996 we observed an

R. Van Tielen et al. / Health Policy 45 (1998) 1–14

13

annual growth in expenditure of + 0.73%, while the total per capita public spending, all age groups taken together, for reimbursable ambulatory pharmaceuticals showed an annual average increase of + 6.2% (in constant francs 1996). It therefore seems that other factors, far more difficult to quantify, cause a rise in expenditure. Scientific and technological developments, advances in epidemiology, better prevention etc. are all factors that weigh on the health care budget. When we try to examine the actual demographic impact on the trend in public spending on ambulatory pharmaceuticals, we can identify two effects. The first, which we called the static demographic effect, will provoke an annual + 0.75% increase for the next ten years (1996–2006). However, we were able to calculate the extent to which per capita expenditure evolved differently according to age group: expenditure for the elderly is rising more sharply than for the younger age groups. This led us to identify a second effect. This second effect, which we called the dynamic demographic impact, causes an annual growth rate of + 1.91%, according to the working hypothesis put forward. In other words, we may conclude that the demographic factor alone will already provoke a rise in health care expenditure which cannot be met by the maximum norm currently proposed by the authorities. The static demographic impact (+ 0.75%) by itself will already account for 50% of the authorized budgetary increase. The problem becomes even more pressing when the dynamic effect (+ 1.91%) is taken into account. In this study, we have only limited our discussion to the annual demographic effect without relating this factor to the main concern of policymakers: the current and future rise in health care costs, but in the future, some questions could be raised. To give a sense of the relative importance of our results, we have to mention these questions. Given the growing life expectancy observed in most western countries, the improvement of public health and also the progress of medical technology, it seems logical that the elderly will use more health care services in general and pharmaceuticals in particular. If we can expect a better health status for the elderly, it does not imply necessarilly less health care spending in the future as it will be extended over a longer period. In other words, it is unlikely that the improvement of the health status induces cost savings for the Social Security. In addition, in a lasting cost containment context, the out-ofpocket payments of the elderly will probably increase. This in turn might induce a lack of care. It is obvious that compliance with a strict budget will imply painful choices either between the young and the elderly pharmaceutical care or between the acute or chronic care. Thus, the risk of a lack of care for the elderly becomes true. Neglecting the needs of the elderly would be unethical. The policymakers have to pay attention to this contingency and figure out this potential problem will be an important challenge. It should be pointed out that, besides the 49.9 billion francs which the INAMIRIZIV spent in 1996 on the reimbursement of ambulatory pharmaceuticals, another 19.2 billion was spent on hospital pharmaceuticals. To get an overall picture of developments in the pharmaceutical sector, it would be useful to study the demographic impact on spending on hospital pharmaceuticals as well.

14

R. Van Tielen et al. / Health Policy 45 (1998) 1–14

References [1] Gerdtham UG, Sogaard J, Andersson F, Jo¨nsson B. An econometric analysis of health care expenditures: a cross-section study of the OECD countries. Journal of Health Economics 1992;11:63–84. [2] Leidl R. Health economic issues relevant to countries with aging populations. World Health Statistics Quarterly 1992;45. [3] OCDE. Le vieillissement dans les pays de l’OCDE: un de´fi fondamental pour la politique (Ageing in OECD countries: a fundamental challenge for politics). Etudes de politique sociale 1996;20:126. [4] Mizrahi A, Mizrahi A. Les facteurs de´mographiques et la croissance des consommations me´dicales (Demographic factors and the medical consumptions growth). Consommation 1974;1:73. [5] Sandier S. Le vieillissement de la population en Europe et le cot des soins mdicaux (Ageing in Europe and the medical care costs). CREDES pour le BIT, 1987. [6] Defeyt P. Soins de sante´ et vieillissement de´mographique (Health care and ageing). IRES 1989:130. [7] Gadreau M. Vieillissement et consommation mdicale(Ageing and medical consumption). Revue franc¸aise des Affaires Sociales 1989;2:131 – 41. [8] Hourriez J-M. La consommation me´dicale a` l’horizon 2010 (Medical consumption towards 2010). Economie et Statistiques, INSEE 1993;5:17 – 30. [9] Mizrahi An, Mizrahi Ar. Evolution de la consommation pharmaceutique par personne et par maladie (Evolution of drug consumption per capita and per disease). Mai 1995, Biblio no. 1071. [10] National Statistical Office. Population perspectives 1992 – 2050. Brussels: INS-NIS, 1993, p. 274. [11] Gerdtham UG. The impact of ageing on health care expenditure in Sweden. Health Policy 1993;24:1–8. [12] Kesenne J. De gevolgen van de socio-demografische ontwikkelingen voor de financiering van de ziekteverzekering (Consequences of sociodemographic evolutions for the health insurance financing). Studiedag van het Belgisch Instituut Voor Openbare Financien op 11 mei 1990. [13] Mendelson DN, Schwartz WB. The effects of aging and population growth on health care costs. Health Affairs Spring 1993;12:119–25. [14] Launois R, Pelc A. Le vieillissement de la population et la consommation me´dicale (Ageing and medical consumption). Solidarite´ Sante´, Etudes statistiques, vol. 4, 1988, Juillet-Aouˆt. [15] Pelc A. L’influence de la de´mographie sur le couˆt de la sante´ et son financement (The demographic influence on the health care costs and the financing). Analyse financie`re, 2e trimestre 1991, pp. 14–20.

.