Development of a model to predict the cost of management of diabetes mellitus and its complications at Groote Schuur Hospital
Diabetes mellitus places a substantial financial burden on the funder for treatment of this disease. This burden is compounded by the development of diabetes mellitus related complications. The cost of management of diabetes mellitus and its complications in South Africa in a tertiary level hospital is unknown. The objectives of this study were to: (1) develop a method to determine the cost of management of diabetes mellitus and its complications at Groote Schuur hospital, (2) quantify the direct medical costs associated with diabetes mellitus and its complications, and (3) develop models which could predict the cost of management of diabetes mellitus and its complications at Groote Schuur hospital. Patients were conveniently selected from Groote Schuur Hospital. Retrospective data were collected for 20 months from the two data sources available at Groote Schuur hospital, i.e., (1) electronic patient records and (2) patient folders. Two methods of costing were developed, i.e., the combined method (using data from electronic database and patient folders) and the electronic method (using data from the electronic database only). The patient folders were used to complete any missing information from the electronic patient records. The combined method allowed better categorisation of costs compared to the electronic method. With the combined method the costs associated with diabetes mellitus (type I or type II) and diabetes mellitus complications could be categorized as the type of diabetes mellitus complications and the reasons for emergency room visits and hospitalisations were known. However, with the electronic method this categorisation could not be done. Both methods provided a total cost and cost per patient associated with diabetes mellitus and its complications. Data from the combined method were arranged in the following main sub-groups for analysis: (1) type I diabetes mellitus, (2) type II diabetes mellitus, (3) diabetes mellitus, (4) diabetes mellitus with complications (5) diabetes mellitus complications only. The electronic method estimated a total cost which was 6.4% more than the combined method .The electronic method did not require the perusal of patient folders for additional information and hence it is a simplified method that could be used. When comparing the total costs and the average cost per patient for diabetes mellitus and diabetes mellitus with complications, there was no statistically significant difference between the combined method and the electronic method (p= 0.41). The average cost per patient per year for diabetes mellitus and diabetes mellitus with complications was R 1 231.54 and R 3 208.71, respectively. These results show that a patient with diabetes mellitus and its complications cost 2.6 times more to treat than a patient with diabetes mellitus only. The complications of diabetes mellitus contributed 60.7% to the total cost of diabetes mellitus and its complications. The cost of treating diabetes mellitus only and the cost of treating diabetes mellitus complications only were separated into 2 distinct categories. In this case, the cost of treating the complications of diabetes mellitus only was 2 times more than treating diabetes mellitus only. Type I diabetes mellitus cost R 3 011.32 and type II diabetes mellitus cost R 2 649.40 per patient per year. The treatment costs for type I diabetic patients was 13.7% more than that for type II diabetic patients. The decision tree model determined a cost which was 0.23% more than the actual cost obtained by the combined method. Hence, the decision tree method could be used to predict the total cost of diabetes mellitus and its complications. Patients who attended Groote Schuur Hospital also received treatment for their diabetes at community health centres. A method was developed which could predict the cost of management of diabetes mellitus at community health centres. This will allow the calculation of the total cost of diabetes mellitus and its complications from the perspective of the government. Four methods (i.e. combined, electronic, decision tree and prediction of costs at cost at community health centres) were developed in this study.