Purpose: Cost-effectiveness analyses (CEA) of diabetes treatments have found that the components of diabetes care differ in their economic value. These analyses have assumed no quality of life effects of treatments and have not accounted for the preferences/utilities of individual patients. We set out to evaluate how patient utilities might alter CEA results and prioritization decisions.

Method: We surveyed 701 adult diabetes patients, from 5/03-5/06. We elicited patients' utilities using time trade-off questions for complications and treatments related to diabetes. Individual patient demographics (age, gender, race) were incorporated into the CDC Diabetes Model which provides comparisons of 1) intensive vs. conventional glucose control, 2) intensive vs. conventional blood pressure control, and 3) cholesterol control with a statin versus diet and exercise. For each individual patient, the model was first run with the utilities originally used in the CDC model. Simulations of the model were then repeated using patient-specific utilities that we collected. The average difference in lifetime costs, effectiveness (quality-adjusted life years, QALYs), and incremental-cost-effectiveness ratios (ICERs) were compared. The relative rankings of expected benefits for the three components of care were also compared.

Results: When using original utility assumptions, for the study population as a whole, intensive blood pressure control was found to be beneficial (Δ QALYs 0.40) and cost saving (-$898), while intensive glucose (ICER $29,473/QALY) and cholesterol control (ICER $53,708/QALY) were cost effective. When using patient utilities, intensive glucose control (Δ QALYs 0.27→ -0.56) and cholesterol control (Δ QALYs 0.33→ -0.57) became harmful. Intensive blood pressure control remained beneficial, although the size of average benefit declined (Δ QALYs 0.40→ 0.02). Incorporating patient utilities also shifted the priorities of diabetes care for individual patients. With the original utilities, 44% of patients were expected to have the largest health benefits from intensive blood pressure control, followed by cholesterol control and then glucose control. With patient utilities, 17% of patients were expected to have this order of benefits.

Conclusions: Overall, incorporating patient utilities lowered the expected benefits of each of the three components of diabetes care, but blood pressure control remained beneficial and cost saving. Incorporating patient utilities led to greater variation in health care priorities for individual patients. In clinical practice, the actual priorities and goals of diabetes care should acknowledge patient treatment preferences.