Analysis of Health Care Costs in Elderly Patients with Multiple Chronic Conditions Using a Finite Mixture of Generalized Linear Models.

In this paper we analysed healthcare costs in a sample of elderly patients suffering from multimorbidity. On the one hand, multimorbid individuals consume a disproportionally large share of healthcare resources. On the other hand, the patient specific number and combination of co-occurring single diseases result in inhomogeneous data leading to biased estimates when using traditional regression techniques. Therefore, we applied a mixture of regressions in order to control for unobserved heterogeneity focussing on the identification of multimorbidity patterns. We used a subsample of N = 1050 patients from a multicentre prospective cohort study of randomly selected multimorbid primary care patients aged 65 to 85 years in Germany (ISRCTN 89818205) who completed a detailed questionnaire on healthcare utilization during the 6-month period preceding the interview. Disease combinations of 1047 were included. We detected four different groups of patients with regard to total costs. These groups corresponded largely to findings from the epidemiological literature. The effect of the presence of an additional disease on costs differed between groups. Moreover, two diametrically opposed cost trends were detected with respect to the number of co-occurring diseases. While in one group costs increased with the number of co-occurring diseases, in a second group cost tended to decrease.