Abstract
OBJECTIVE: To estimate the impact of state-level insulin out-of-pocket caps on changes in out-of-pocket and total costs of insulin and healthcare for insulin users with employer-sponsored insurance. STUDY SETTING AND DESIGN: We evaluated changes in costs using a quasi-experimental (triple difference-in-differences; "DDD") design to analyze multi-carrier claims from insulin users enrolled in fully insured (state-regulated) and self-funded (generally exempt) employer-sponsored plans in 10 states with caps by January 2021 compared to no-cap states pre-/post-cap implementation. Primary outcomes were changes in insulin out-of-pocket spending, total (plan + member) paid for insulin, and total healthcare costs. Secondary outcomes were intermediary (e.g., pharmaceutical) changes in out-of-pocket and total costs. DATA SOURCES AND ANALYTIC SAMPLE: In the policy year (no-cap states: 2021), we identified 218,441 insulin-users in the Health Care Cost Institute 2.0 Dataset (cap states: 27,834 in fully insured and 22,131 in self-funded plans; no-cap states: 97,239 in fully insured and 71,237 in self-funded plans) and 215,635 in the year prior. PRINCIPAL FINDINGS: We found evidence of modest decreases in 30-day standardized (DDD: -$5 [95% CI: -$6 to -$4]; p < 0.001) and annual (DDD: -$67 [95% CI: -$82 to -$51]; p < 0.001) insulin out-of-pocket spending. Savings increased by spending quantile (e.g., 95th-percentile change:-$347 [95% CI: -$460 to $233]). Difference-in-differences (DiD) comparing fully insured to self-funded plans within cap-states showed larger changes (e.g., 95th-percentile annual insulin out-of-pocket:-$484 [95% CI: -$651 to -$318]), likely due to policy spillover effects (i.e., fully insured plans decreased out-of-pocket in no-cap states). Change in annual total paid for healthcare was not statistically significant (DDD:-$1082 [95% CI: -$2918 to $755]; p < 0.25). We saw no evidence of caps increasing out-of-pocket or total spending on insulin, prescriptions, or healthcare. CONCLUSIONS: Our findings suggest early caps had modest effects on out-of-pocket spending among fully insured insulin users, with larger savings for those at the top of the spending distribution and no total cost increases. Policy effects may be greater than observed; they likely lag implementation and develop over time.