Skip to main content
Log in

What accounts for the regional differences in the utilisation of hospitals in Germany?

  • Original Paper
  • Published:
The European Journal of Health Economics Aims and scope Submit manuscript

Abstract

There are huge regional variations in the utilisation of hospital services in Germany. In 2007 and 2008 the states of Hamburg and Baden-Württemberg had on average just under 38 % fewer hospitalisations per capita than Saxony-Anhalt. We use data from the DRG statistics aggregated at the county level in combination with numerous other data sources (e.g. INKAR Database, accounting data from the National Association of Statutory Health Insurance Physicians (KBV), Federal Medical Registry, Germany Hospital Directory, population structure per county) to establish the proportion of the observed regional differences that can be explained at county and state levels. Overall we are able to account for 73 % of the variation at state level in terms of observable factors. By far the most important reason for the regional variation in the utilisation of in-patient services is differences in medical needs. Differences in the supply of medical services and the substitutability of outpatient and inpatient treatment are also relevant, but to a lesser extent.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6

Similar content being viewed by others

Notes

  1. Here we conflate the years 2007 and 2008, as the analysis that follows refers to these two years. However, the values barely change if either only the year 2007 or only the year 2008 is taken.

  2. This is particularly the case for regional differences in the healthcare expenditures in the different American states, c.f. e.g. [27] are among those who also examine regional differences in the utilisation of medical services. Recent European studies on regional variation in health expenditure—though not health care utilisation—are, e.g. [8] for Switzerland and [9] for Spain.

  3. The Federal Medical Registry maintained by the National Association of Statutory Health Insurance Physicians (KBV) lists all medical doctors on contract to the statutory health insurance funds.

  4. The county boundaries are as defined in the year 2008 after the boundary reforms in Saxony-Anhalt and Saxony.

  5. The data for this come from the Federal Statistical Office Destatis.

  6. As is well-known, by definition the R 2 does not fall when additional variables are included in the regression, even when the latter have no explanatory content. We thus additionally show the corrected coefficient of determination (adjusted R 2), which does not have this problem. This in turn is by definition smaller than the R 2. However, the differences between the two are slight.

  7. The variation of the residuals is in fact exactly the same as the variation of the number of cases adjusted by the influencing factors given in the left-hand column.

  8. This procedure is itself analogous to the procedure described above and identical to simply taking the average of the residuals of the regressions per state and deducting the respective differences as a percentage.

References

  1. Statistisches Bundeamt.: Gesundheit: Ausgaben 1995 bis 2009. Wiesbaden (2011)

  2. Cutler, D.M., Sheiner, L.: The geography of medicare. Am. Econ. Rev. 89, 228–233 (1999)

    Article  Google Scholar 

  3. Rettenmaier, A.J., Wang, Z.: Regional variations in medical spending and utilization: a longitudinal analysis of US Medicare population. Health Econ. 21(2), 67–82 (2012)

    Google Scholar 

  4. Doyle, J.J.: Returns to local-area healthcare spending: evidence from health shocks to patients far from home. Am. Econ. J. Appl. Econ. 3(3), 221–243 (2011)

    Article  Google Scholar 

  5. O’Connor, G.T., Quinton, H.B., Traven, N.D., Dodds, L.D., Ramunnoand, T. Andrew, Marciniak, T.A., Wennberg, J.E.: Geographic variation in the treatment of acute myocardial infarction: the cooperative cardiovascular project. J. Am. Med. Assoc. 281, 627–633 (1999)

    Article  Google Scholar 

  6. Fuchs, V.R.: More variation in use of care, more flat-of-the-curve medicine. Health Aff. 23(6), 104–107 (2004)

    Google Scholar 

  7. Fuchs, V.R., McClellan, M., Skinner, J.: Area differences in utilization of medical care and mortality among US elderly. In: Wise, D.A. (eds.) In Perspectives on the Economics of Aging. NBER Conference Report Series, pp. 367–406. University of Chicago Press, Chicago (2004)

  8. Reich, O., Weins, C., Schusterschitz, C., Thöni, M.: Exploring the disparities of regional health care expenditures in Switzerland: some empirical evidence. Eur. J. Health Econ. 13, 193–202 (2012)

    Article  PubMed  Google Scholar 

  9. Prieto, D.C., Lago-Penas, S.: Decomposing the determinants of health care expenditure: the case of Spain. Eur. J. Health Econ. 13, 19–27 (2012)

    Article  PubMed  Google Scholar 

  10. Felder S., Tauchmann, H.: Federal state differentials in the efficiency of health production in Germany: an artifact of spatial dependence? Eur. J. Health Econ. (2011) (forthcoming)

  11. Kopetsch, T.: Der Zusammenhang zwischen dem Leistungsgeschehen im ambulanten und stationären Sektor des deutschen Gesundheitswesens: Eine empirische Untersuchung. Jahrbücher für Nationalökonomie und Statistik 227(1), 49–64 (2007)

    Google Scholar 

  12. Krämer, W.: Eine ökonometrische Untersuchung des Marktes fr ambulante kassenärztliche Leistungen. Zeitschrift für die gesamte Staatswissenschaft 1, 45–61 (1981)

    Google Scholar 

  13. Breyer F.: Die Nachfrage nach medizinischen Leistungen: Eine empirische Analyse von Daten aus der Gesetzlichen Krankenversicherung. Berlin (1984)

  14. Kopetsch, T.: Arztdichte und Inanspruchnahme ärztlicher Leistungenin Deutschland: Eine empirische Untersuchung der These von der angebotsinduzierten Nachfrage nach ambulanten Arztleistungen. Schmollers Jahrbuch 127, 373–405 (2007)

    Google Scholar 

  15. Hendrik J.: Health insurance status and physician behavior in Germany. J. Appl. Soc. Sci. Stud. (Schmollers Jahrbuch) 129(2), 297–308 (2009)

    Google Scholar 

  16. Kopetsch, T.: Gilt Roemer’s Law auch in Deutschland? Eine empirische Untersuchung zur Überprufung der These der Angebotsinduzierung im stationären Sektor des deutschen Gesundheitswesens. Jahrbücher für Nationalökonomie und Statistik 226(6), 646–669 (2006)

    Google Scholar 

  17. Kemptner, D., Jürges, H., Reinhold, S.: Changes in compulsory schooling and the causal effect of education on health: evidence from Germany. J. Health Econ. 30(2):340–354 (2011)

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Hendrik Schmitz.

Appendix

Appendix

RSA risk factor

The average morbidity in a county can be determined by the RSA risk factor. The German risk adjustment scheme (Risikostrukturausgleich, RSA) redistributes money between all sickness funds according to differences in demographic factors and morbidity. A fund receives an assignment for each insured i depending on her age, sex and health status, where the health status (morbidity) of each insured is measured by 80 diseases. Assignments are higher for individuals that are more likely to incur higher costs and lower for healthy individuals. The RSA risk factor (RF RSA) assesses to what extent the morbidity of insured i in group G (say, a sickness fund or a county G) – as measured by the criteria of the risk adjustment scheme – deviates from the average morbidity in the population:

$${\text {RF}}_{\rm RSA} = \frac{\sum_{i \in G} {\text {Assignment}}_i}{\sum_{i \in G}{\text {Insured}}_i}\cdot \frac{\sum_{i} {\text {Insured}}_i}{\sum_{i}{\text {Assignment}}_i} $$

An RSA risk factor of 1.0 reflects average morbidity in the population whereas a value of 1.1 expresses that the morbidity level in the respective group is such that one may expect by 10 % higher expenditures than average expenditures. An RSA risk factor below 1.0 reflects a morbidity level better than average. Typically, the RSA risk factor is determined for sickness funds to identify their average mortality but can analogously be determined for other units such as citizens of a county (as is done here) (Tables 6, 7).

Table 6 Full regression results
Table 7 Reduction of the variation with alternative sequences

Rights and permissions

Reprints and permissions

About this article

Cite this article

Augurzky, B., Kopetsch, T. & Schmitz, H. What accounts for the regional differences in the utilisation of hospitals in Germany?. Eur J Health Econ 14, 615–627 (2013). https://doi.org/10.1007/s10198-012-0407-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10198-012-0407-6

Keywords

JEL Classification

Navigation