WHO ARE THEY FOOLING?- COST OF DISEASE OR COMPLICATIONS CAN SIGNIFICANTLY BIAS ESTIMATES UNLESS CONTROL (NON-DISEASED) COSTS ARE NOT ALSO ACCOUNTED FOR IN THE ANALYSIS

OBJECTIVES: Costing studies often do not identify the excess costs incurred by the healthcare system for patients with a disease vs. patients without the disease. Using cohort based cost estimates without controlling for costs also incurred by a non-disease population can bias projections, long-term modeling and economic evaluation analyses. The objective of this study was to estimate the prevalence, total and excess costs attributable to diabetes and its complications in Ontario, Canada over 11 years (1995 to 2005). METHODS: Newly diagnosed type 1 and 2 diabetes cases aged 35 and over were identified from the Ontario Diabetes Database and matched 1:2 using propensity scores with controls (non-diabetes cases). The following complications were identified: myocardial infarction, stroke, angina, heart failure, blindness in 1 eye, amputation, nephropathy and cataracts. Excess costs of diabetes were estimated as the difference between costs attributed to patients with diabetes vs. those attributed to patients without diabetes. RESULTS: The prevalence of diabetes rose drastically, from 6.5 to 10.5%. Excess costs were $2930 in the year of diabetes diagnosis and $1240 in subsequent years. In the year of an event, cost differences were greatest for patients with diabetes who had an amputation ($5133), followed closely by nephropathy ($4117) and stroke cases ($3965). Excess costs were apparent for both females and males, and the cost amount was strongly associated with increasing age. CONCLUSIONS: Results demonstrate that relying on costs from a population with only the disease (i.e. diabetes) with no control can overestimate costs of the disease and associated complications. Assessing excess costs of disease is important for costing studies, longer-term modeling and economic evaluations in general. Existing studies which do not account for excess cost may overestimate cost and potentially bias estimates of cost-effectiveness or cost savings due to effective patient management.