[How does the German DRG system differentiate and reimburse vitreoretinal surgery in diabetic patients?]

The German DRG system (G-DRG system) is required to assign medical cases with similar costs correctly into a particular group, each case within the group receiving the same amount of reimbursement. At the same time the system should allow all-inclusive reimbursement, not necessarily reflecting the exact costs of each case. These opposite goals and the so far limited calculation basis raise the question of how the G-DRG system actually processes and reimburses empirically collected in-hospital treatment data.

In 2005, 112 patients were admitted to the University Eye Hospital, University of the Saarland. All patients had diabetic retinopathy and required at least one vitreoretinal procedure. Demographic and clinical data were collected by using the hospital information system and the coding software KODIP. For statistic evaluation, principal diagnoses, ancillary diagnoses and procedures were each reassigned to particular groups. Reimbursement was calculated based on the case data of the year 2005. Also, the case data were reassigned with respect to calculation of reimbursement for the years 2006 and 2007. The results were compared with federal G-DRG calculation data.

Mean age of the patients was 65.8 +/- 11.1 years, length of stay in-hospital was 9.3 +/- 3.2 days. In the 66 patients requiring general anaesthesia the cumulative length of stay in the operation room was 148.4 +/- 39.5 minutes, the cumulative duration of surgery was 86.3 +/- 34.1 minutes. In the 50 patients requiring local anaesthesia the cumulative length of stay in the operation room was 137.8 +/- 51.8 minutes, the cumulative duration of surgery was 81.6 +/- 43.6 minutes. The patients had 1.9 +/- 0.8 principal diagnoses, 14.4 +/- 5.8 ancillary diagnoses and 3.4 +/- 1.6 procedures. Twenty-five of 112 patients (22.3 %) were assigned to DRG C 03Z (1), 82 of 112 patients (73.2 %) were assigned to DRG C 17Z (2). Five patients were assigned to other DRG. Compared with the federal calculation data, our own data for 2005, 2006 and 2007 showed more high primary clinical complexity levels and a longer duration of in-hospital stay. For each of the three years the amount of reimbursement was equal in about two thirds of the own patients. Reimbursement was only differentiated for outliers beyond the trim point of the duration of in-hospital stay.

The demographic and clinical G-DRG data of the included patients showed substantial cost-effective inhomogeneities. These inhomogeneities were not sufficiently considered for reimbursement based upon Z-DRG. Specialised departments with higher numbers of difficult cases may be discriminated. Wrong incentives may result in the selection of "low-risk cases".