Article
Correlations in the target system in regional target agreements between Statutory Health Insurance and the Association of Statutory Health Insurance Physicians
Search Medline for
Authors
Published: | November 22, 2010 |
---|
Outline
Text
Introduction: The yearly rising drug budget in connection with the target of stable statutory health insurance contributions implies the urgency of the realization of cost reduction potentials. The regional target agreements between Statutory Health Insurances and the National Association of Statutory Health Insurance Accredited Physicians already have realized large potentials. A periodic controlling and further development of targets and parameters is necessary.
Methods: In Schleswig-Holstein two types of targets are used. Type A uses a proportion of lead drugs (“Leitsubstanzen”) within target groups, e.g. Simvastatin within the Statines. Type B uses daily therapy costs based on the international ATC (Anatomical Therapeutic Chemical) and DDD (defined daily dose) classification system. Thereby a proportion of economic drugs with respect to daily costs should be above a threshold value. We will consider the correlation of target type A and B values of the Physicians in 12 target groups. Further on we consider a parameter C which measures the average active agent doses prescribed by the Physicians. Significantly higher average doses may lead to lower daily costs (theoretically calculated by the ATC-DDD concept) for the Physician, but higher costs for the Health Insurance. So the parameter C may indicate inconsistencies in the ratio system of target values. Therefore we consider the partial correlation of A und B with respect to C in order to see, if we get a better correlation between A and B with an elimination of C.
Results: The maximal value of correlation (r^2=0.871) is reached for the Bisphosphonates with lead drug Alendronic acid and the minimal correlation (r^2=0.001) is given for Beta blocking agents with lead drugs Metoprolol, Bisoprolol and Atenolol. Parameter C does not result in a general increase or decrease from correlation to partial correlation, half of the goals change in each direction.
Discussion and conclusion: One could use a parameter D which describes the average price for one packing standardized by the norm packing value which avoids the paradox effect seen using C values as an alternative to he DDD-based considerations. To do so one can consider correlations and partial correlations for every 3 of the four parameters A–D in order to get further improvements in the target agreement system.