The purpose of the study is to evaluate hospital costs in Kaunas Medical University Hospital (KMUH). KMUH is the largest hospital in Lithuania, having 1995 in-patient beds, 26 specialised in-patient departments, 5130 employees, and providing wide range of in-patient services.

Methods. Methods, used in the study include assessment of inputs and outputs, evaluation of average cost per case, estimation of cost structure, estimation of case-mix dimensions in in-patient departments and clinical categories and assessment of impact of case-mix dimensions to cost per case, using multiple regressionanalysis. Cross-sectional study designwas used in the study, evaluating mainly cases and expenses of all 26 specialised in-patient departments of KMUH per year 2002. Five cost groups have been used and defined inmonetary terms in each in-patient department: labour costs; medication costs; laboratory, radiology and anaesthesiology costs; running costs of medical equipment supply andother costs (including in-patients’ mealcosts, transportation, laundry, communication, etc. costs). Case was defined as one treatment episode in particular in-patient department. Cases were analysed using following case-mix dimensions: sex, age, absenceor presence of surgical operation, patient separation status and in-patientservice group.

Results. Average costs per case vary widely among in-patient departments, ranging from 126.01 Litas (36.52 Euro) to 3451.68 Litas (999.73 Euro) per case.During the study average cost per case were also estimated in clinical profiles – surgery – 1161.0 Litas (336.24 Euro), therapy – 1312.15 Litas (380.02 Euro),obstetrics and gynaecology –685.82 Litas (198.62 Euro), newborn and child care – 893.54 Litas (258.78 Euro) and intensive care – 1292.92 Litas (374.45 Euro). Using multiple regression analysis method, costper case ineach in-patient department and clinical category according case-mix dimensions were predicted. In all in-patient departments predicted values of average costs per case according case-mix dimensions, comparing with actual values, did not differ so much. Positive contributions to predictedvalue of cost per case, shows only one variable – IA in-patient service group. In any predicted case contributions of independent variables have notbeen observedas significant (p>0.05).

Conclusions. Inputs (measured in the number of beds) and outputs (measured in the number of in-patientcases and the number of bed-days) are different across in-patient departments, as well as outputs (measured inthe number of treatment episodes according to case-mix dimensions). The average costs per case vary widely across in-patient departments and clinical categories. The analysis of the structure of average costs per case demonstrated striking differences in in-patient departments. In all in-patient departments the predicted values of the average costs per case according to case-mix dimensions, do not differ so much comparing with theactual observed costs per case. Positive contributions to the predicted value of the cost per case, shows only onevariable – IA in-patient service group. The results of the study have proved the evidence that clinical casestreated within the same in-patient department of the hospital are not similar. The results of studyhave showedthe failure of use of “in-patient service groups” as proxy of International Disease Classification due to numberof reasons