Recent changes in health coverage benefits and increases in diabetes medication prices may have affected out-of-pocket (OOP) expenditures for people with diabetes. High OOP expenditures is a barrier to get adequate medical care. We assessed national trends in OOP expenditures overall and by OOP-component among adults aged 18-64 years with diabetes from 2001 to 2017 using data from the U.S. Medical Expenditure Panel Survey. OOP expenditures were defined as the amount paid by a family with one or more members with diabetes, including insurance premiums, deductions, copays, coinsurance and payments not covered by insurance. High OOP burden was defined as spending >10% of the family’s pretax income on OOP expenditures. We examined trends overall as well as by insurance type, income (as a ratio of federal poverty level, FPL), type of service, and whether patients used insulin. We tested for changes in trends using joinpoint analysis. OOP expenditures were adjusted to 2017 U.S. dollars. From 2001 to 2017, OOP expenditures decreased 4.3%, from $4328 to $4139, and the percentage of persons with high OOP burden fell 32%, from 28% to 19% (P<0.001). No significant changes in trends were identified for either measure. Both OOP expenditures and percentages with high OOP burden declined for those with public or no insurance and those with family income < 200% of FPL (P<0.001). There were increases in OOP expenditures (P<0.001) but no significant changes in percentages with high OOP burden for those with private insurance and those with income ≥ 400% of FPL. OOP expenditures on prescriptions decreased for all insurance types and income subgroups (P<0.001). Decreases in total (-$58 vs. -$37, P<0.001) and prescription (-$79 vs. -$68, P<0.001) OOP expenditures were larger for insulin users than non-users. These most recent data show that OOP expenditures for U.S. adults with diabetes are steadily declining, especially for those least able to afford them. Future studies are needed to explain these patterns.


Y. Wang: None. J. Park: None. R. Li: None. E. Luman: None. P. Zhang: None.

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at