Economic Costs of Diabetes in the U.S. in 2002

American Diabetes Association

Diabetes Care. 2003;26(3) 

In This Article


Health care spending in 2002 for people with diabetes is more than double what spending would be without diabetes. This costs the U.S. economy an estimated $92 billion in higher health care expenditures. Lost productivity attributed to diabetes resulting from lost workdays, lost home services, permanent disability, and premature mortality is estimated at $40 billion. Compared to people without diabetes, people with diabetes and their families bear a disproportionate share of health care expenditures.

This cost estimate is conservative and likely understates the true burden of diabetes for the following reasons:

  • This estimate omits the cost of intangibles such as pain and suffering, the cost of care provided by informal caregivers, and administrative costs of insurers.

  • The cost components included in this analysis account for only 58% of the estimated $1.5 trillion in U.S. health care expenditures in 2002. For example, over-the-counter medications and sundries, which Martin et al.[26] estimate at $122 billion in 1998, are omitted from the cost estimate. Whereas the areas of health care expenditures analyzed are those where health care use patterns have been shown to differ by diabetes status, there are several areas omitted from the analysis where people with diabetes probably use services at higher rates than people without diabetes, e.g., dental care, optometry care, and the use of licensed dietitians.

  • The average price per health service used could differ by diabetes status. If health care conditions classified as "general medical conditions" (e.g., pneumonia) are more severe for people with diabetes than without diabetes, then the cost estimate would be too low. The study controls for differences in health care use attributable to diabetes, e.g., the number of hospital inpatient days, but does not control for differences in mix of health care professionals seen (e.g., if people with diabetes are more likely to see a specialist instead of a primary care physician).

  • In this study, people with undiagnosed diabetes are categorized with the nondiabetic population. If per capita use of health care services is greater for people with undiagnosed diabetes than for people without diabetes, the health care costs attributable to diabetes will be underestimated.

Future research might investigate the cost of diabetes in these areas omitted from the present analysis.

The estimated national cost of diabetes was calculated using prevalence-based cost-of-illness methods with data from 1998 through 2002. For some components of the cost estimate (e.g., the cost of supplies), multiple data sources were analyzed and the results were compared to ensure robust results. One change from the approach used in ADA's 1998 study was to combine multiple years of national health use databases to increase sample size and allow for finer disaggregation of the U.S. population -- both of which would improve the accuracy of the findings, because the prevalence of diabetes and the use of health care services varies substantially by age-group, sex, and race/ethnicity. Greater disaggregation also allows for more accurate projections of the national cost of diabetes in future years as the U.S. population grows, ages, and becomes more racially and ethnically diverse. However, if lifestyle trends in the U.S. (such as the growing problem of obesity) increase diabetes prevalence rates, future costs could grow in excess of those extrapolated based on current prevalence rates.

Although this study includes the same cost components of ADA's 1998 study[2], the change in estimated diabetes-attributed costs between 1997 and 2002 for some cost components reflects a refinement in the cost estimates as opposed to an actual change in true costs. As discussed previously, the 1998 study estimated disability-related costs at $32.5 billion in 1997, compared with the current study, which estimates disability-related costs at $7.5 billion. Much of the decrease in attributed costs is the result of using foregone expected annual expenditures instead of foregone expected lifetime earnings to estimate the pecuniary cost of lost productivity, which may have been an inadvertent overstatement in the previous report. This large decrease in attributed costs is offset by substantially higher cost estimates for certain health care components such as nursing home care, home health care, and physician office-based care.

One factor contributing to the large increase in attributed cost for nursing home care is the higher estimated cost per day in nursing homes ($169 per day used in this study vs. $79 per day [$97 per day in 2002 dollars] used in the 1998 study). This study estimates a much higher cost of home health care services, with an estimated 18% of total U.S. home health care services costs attributed to diabetes compared with an estimated 0.2% of the total U.S. cost of home health care services attributed to diabetes in the 1998 study. Martin et al.[26] estimated national expenditures of approximately $30 billion for home health care in 1997, compared with the estimate of $19 billion in the 1998 report[2].

Eliminating or reducing the health problems caused by diabetes through factors such as better access to preventive care, more widespread diagnosis, more intensive disease management, and the advent of new medical technologies could significantly improve the quality of life for people with diabetes and their families, while at the same time potentially reducing national expenditures for health care services and increasing productivity in the U.S. economy.

In conclusion, the cost of diabetes, both direct medical expenditures and the costs of foregone productivity, is estimated to have been $132 billion in 2002. This represents a substantial cost burden to society and, in particular, to those in-dividuals with diabetes and their families. Nevertheless, this estimate is conservative and probably underestimates the true cost of the disease.