PROJECT ABSTRACT Diabetes affects 10% of the U.S. population and is responsible for $237 billion in direct medical costs annually. Primary care and specialist visits, medications and testing supplies, durable medical equipment, lab testing, and potential emergency and hospital visits can be costly. Patients with diabetes report a high level of financial burden relative to patients without diabetes, struggling to pay for necessary health care, food, housing, and retirement. Cost-related non-adherence (delaying filling a prescription or taking less medicine) for patients with diabetes is high, and subsequently, patients in health insurance plans with poor coverage for needed services and resources experience adverse outcomes like hyper- or hypoglycemia, acute cerebrovascular disease, and ischemic heart disease due to delayed or forgone care. Therefore, choosing a health insurance plan that meets your needs is critical to improve health, especially for patients with low income socioeconomic status. Unfortunately, many consumers struggle to choose a health plan from among the many offered, focusing on a single factor like costs, using heuristics to make decisions, falling subject to biases that challenge rational decision-making theory, and failing to weigh trade-offs between plan features. We hypothesize that health economic stated preference methods like discrete choice experiments and multi- criteria decision analyses can be used to elicit health plan preferences and values. These tools specifically ask patients to weigh trade-offs between multiple plan features and can assess which features influence patient decisions the most. In this observational natural experiment study, we will use stated preference surveys to systematically assess patient preferences for health plans that adequately cover diabetes care. We will leverage our access to health plan and claims data from Harvard Pilgrim Health Care and the Tufts Health Plan, two New England-based insurers with a combined 2.2 million covered lives. We will analyze the concordance between patient preferences and their real-world health plan. Finally, we will assess the relationship between preference-aligned health plan enrollment and adverse diabetes health outcomes, cost- related treatment non-adherence, receipt of appropriate high-quality diabetes care, health care costs and utilization, patient financial burden, and plan satisfaction. If preference-aligned health plans are associated with better health and health care outcomes, future work will develop a decision aid that allows patients to clarify their values and preferences for health care and choose a plan that will enable them to use preventive care that reduces adverse health events and lowers their financial burden. If not, future research can use study results to design a decision aid that directs consumers to plans with features associated with improved health, health care utilization, health care costs, and consumer satisfaction.