Determinants of health care utilization by German sickness fund members--with application to risk adjustment

In many countries, social health insurance systems are being reformed in favor of more competition among insurers, while premiums are community rated by regulation. The implicit incentives for insurers to engage in risk selection can only be curtailed using appropriate systems of risk-adjusted equalization payments among insurers. To develop these systems, predictors of individual utilization patterns have to be identified, e.g. via regression analysis using previous utilization data. In some countries such as Germany, such data are hardly ever available. In the early nineties, a number of sickness funds participated in an experiment in which individual utilization data were collected. Our data set covers more than 70,000 members of company sickness funds over a 5-year period. We analyze socio-demographic determinants of utilization which could be used as risk adjusters in a risk equalization scheme. Our results suggest that besides age and sex, the set of risk adjusters should include income, family status and a dummy for the last year of life.