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

American Diabetes Association

Diabetes Care. 2003;26(3) 

In This Article


From estimates of per capita health care use and the size of the population, by demographic group, this study estimates total health care use for each demographic group. Applying the etiological fractions for the corresponding demographic groups results in estimates of health resource use attributable to diabetes.

Table 3 shows estimated health care use by type of service aggregated into three broad age-groups. The attributable health care use due to diabetes is greatest for the population aged ≥65 years, despite this population having slightly fewer people with diabetes than the population aged 45-64 years. For instance, office-based physician encounters attributable to diabetes for people over age 65 years is more than double the office-based physician encounters for people between 45 and 64 years. Use of emergency department, home health, and hospice care services is also substantially higher for the population over age 65 years compared with the population between age 45 and 64 years and the population under age 45 years.

Tables 4-6 provide information on health care use attributable to diabetes by medical condition and type of service. Table 4 shows total use of services by type of medical condition attributable to diabetes, Table 5 shows each medical condition's proportion of total use attributable to diabetes, and Table 6 shows the proportion of total U.S. use attributable to diabetes. Examination of these three tables reveals the following trends:

  • Most of the health care use attributable to diabetes is for the treatment of general medical conditions, i.e., visits or inpatient days where the primary diagnosis is neither diabetes nor one of the seven chronic complications analyzed. For example, 63% of hospital inpatient days attributable to diabetes fall under the category of general medical conditions.

  • Of the seven chronic complications analyzed, cardiovascular disease accounts for the largest proportion of health care use attributable to diabetes. For example, in 2002, an estimated 4 million hospital inpatient days were attributable to diabetes where the primary diagnosis is related to cardiovascular disease. This constitutes 24% of total hospital days attributable to diabetes and 19% of total U.S. inpatient days when the primary diagnosis was related to cardiovascular disease.

  • Diabetes accounts for a sizable increase in the use of health care services. An estimated 18% of home health visits in the U.S. are attributable to diabetes. Approximately 15% of nursing home services and 14% of hospice care services in the U.S. are attributable to diabetes.

Health care expenditures attributable to diabetes are those costs incurred by the population with diabetes above what would be expected if this population did not have diabetes. Of the estimated $91.8 billion in health care expenditures attributable to diabetes, $47.6 billion (52%) is for services provided to people ≥65 years of age. An estimated $31.6 billion (34%) is for services provided to people age 45-64 years, whereas the remaining $12.6 billion (14%) is for services provided to people under age 45 years ( Table 7 ). Home and hospice care expenditures attributable to diabetes are incurred primarily by the population ≥65 years of age.

Table 8 shows estimates of attributable health care expenditures by medical condition and type of service. Expenditures for health care events with a primary diagnosis of uncomplicated diabetes and diabetes-related supplies are estimated to be $23.2 billion for 2002, which accounts for 25% of all health care attributable expenditures. At over $44 billion (or 48% of total attributable expenditures), general medical conditions comprise the largest component of expenditures attributable to diabetes. Together, the seven chronic conditions associated with diabetes account for the remaining 27% of attributable expenditures, with cardiovascular disease being the single largest contributor.

Total U.S. expenditures for health care services analyzed in this study are estimated at $865 billion ( Table 9 ), which is 58% of the total U.S. health care expenditures of approximately $1.5 trillion in 2002[25]. (Centers for Medicare and Medicaid Services [CMMS] estimated national health care expenditures of $1.3 trillion in the year 2000, which is adjusted to 2002 using CMMS's projection of an 8% increase in annual cost of health care services in the U.S. resulting from rising medical costs and an increased use of services.) Cost components not included in this analysis include such things as school-based and public health clinics, dental care, podiatric care, optometry care and vision products (with the exception of ophthalmology services, which are included), research, over-the-counter medicines, and other areas. CMMS estimates expenditures in 2000 to be $60 billion for dental care, $44 billion for government public health activities, and $44 billion for investment (i.e., research and construction). Martin et al.[26] estimate expenditures in 1998 to be $16 billion for vision products and other medical durables (e.g., hearing aides, medical equipment rentals, etc.) and $122 billion for over-the-counter medicine and sundries.

This analysis focuses on those areas where health care use patterns have been shown to differ by diabetes status. Therefore, it is unknown what portion of the remaining 42% of U.S. health care costs can be attributed to diabetes. Components of the health care system not analyzed in this study, but where health care use patterns might differ by diabetes status include dentistry, podiatry, optometry, and licensed dietitians. It is known, for example, that people with diabetes are at higher risk for periodontal disease than the general population, but these data are not incorporated here. Thus, it is likely that this estimate of health care costs attributable to diabetes underestimates the true amount.

Of the health care components analyzed, more than $1 in $10 spent on health care services in the U.S. is attributable to diabetes. Expenditures attributable to diabetes are greatest for hospital inpatient stays ($40.3 billion), followed by nursing home care ($13.9 billion) and visits to physician offices ($10 billion). The cost of oral agents to lower blood glucose, insulin, and insulin-related supplies totaled approximately $12 billion. Diabetes is responsible for a substantial proportion of total U.S. expenditures for certain health care services, e.g., 18% of home health expenditures, 15% of nursing home expenditures, and 14% of hospice care expenditures.

The estimated cost to provide health care services to people with diabetes exceeded $160 billion in 2002 (for those components of the health care system included in this study). This includes costs attributable to diabetes as well as non-diabetes-related costs. Although people with diagnosed diabetes comprise only slightly more than 4% of the U.S. population, of the components of the health care system included in this study, almost $1 of every $5 spent on health care in the U.S. is for a person with diabetes.

Because the prevalence of type 2 diabetes increases with age, the population with diabetes tends to be older compared with the population without diabetes. Consequently, people with diabetes incur a substantial proportion of long-term care services. For example, more than $1 in $4 spent for nursing home, home health, and hospice care is spent to provide services to someone with diabetes.

Dividing health care expenditures by the size of the population with and without diabetes creates estimates of per capita expenditures ( Table 10 ). On average, people with diabetes incurred approximately $13,243 in health care expenditures in 2002 across the health care components included in this study. People without diabetes incurred approximately $2,560 in expenditures, for a ratio of ~5 to 1. This comparison is slightly higher than ratios estimated by ADA[2] and Rubin et al.[27], who found a fourfold difference in average annual health care expenditures for people with diabetes compared with others. However, this ratio somewhat overstates the impact of diabetes on per capita costs because the demographic composition of the population with diabetes differs substantially from the demographic composition of the population without diabetes. The population with diabetes tends to be older, on average, than the population without diabetes.

We derived an age-adjusted annual per capita expenditure of $5,642 to control for differences in demographic characteristics of the population with diabetes compared with the nondiabetic population, yielding a ratio of ~2.4-to-1 for health care expenditures among people with and without diabetes. This ratio prevails, roughly, across cost components, ranging from a high of 2.7 to 1 for home health services to a low of 2 to 1 for emergency services.

At an annual cost of $7.5 billion, more than 176,000 cases of permanent disability in 2002 are attributable to diabetes ( Table 11 ). This cost estimate represents a sizeable decrease from the cost of disability in the 1998 report[2], which used the present value of lost lifetime earnings to estimate the cost of disability. We use average annual lost earnings, estimated at $42,462 per case, to represent the productivity loss associated with the disability. Disability cases where diabetes is listed as the primary cause accounts for more than two-thirds of total cases attributed to diabetes. Cases where cardiovascular disease is listed as the primary cause of disability accounts for 7% of all cases attributed to diabetes.

The estimated number of deaths attributable to diabetes is derived from instances where the primary cause of death is diabetes, renal disease, cerebrovascular disease, or cardiovascular disease. The etiological fractions used to estimate health care use attributable to diabetes are applied to the estimates of the number of deaths -- by age, sex, race/ethnicity, and primary cause of death -- to estimate deaths attributable to diabetes. Estimated lost years of life are based on comparing timing of premature death to life expectancy[28].

In 2002, an estimated 186,000 deaths were attributable to diabetes ( Table 12 ). An estimated 19% of all deaths for which cardiovascular disease is listed as the primary cause of death are attributed to diabetes, and this accounts for 108,000 (58%) of all deaths attributable to diabetes.

This finding is consistent with the major findings of a study by DeStefano and Newman[29], which finds that coronary heart disease is the leading cause of mortality among people with diabetes. DeStefano and Newman find that for younger people (i.e., men under age 45 years and women under age 55 years), people with diabetes had a 13-fold greater risk of coronary heat disease mortality than people without diabetes when controlling for other coronary heart disease risk factors. The Centers for Disease Control and Prevention reports that adults with diabetes have heart death rates that are two to four times higher than those of adults without diabetes[30]. An estimated 2,000 deaths with renal disease as the primary cause are attributed to diabetes. Geiss et al.[31] found that age-adjusted renal mortality rates for people with diabetes are more than 2.5 times the rates for people without diabetes.

Combining estimates of health care expenditures and productivity losses attributable to diabetes yields an estimate of the national cost of diabetes ( Table 13 ). In 2002, the estimated cost of diabetes was approximately $132 billion, of which approximately $92 billion (70%) was additional health care expenditures and $40 billion (30%) was lost productivity due to disability and early mortality. Institutional care (i.e., hospital inpatient care and nursing home care) was the largest component of health care costs and comprised 41% of the national cost of diabetes. Outpatient care, at $20 billion in 2002, comprised 15% of the national cost of diabetes. At $17.5 billion, the cost of outpatient medication and supplies comprised 13% of the national cost of diabetes.

As the U.S. population grows in size, ages, and becomes more racially and ethnically diverse, the size of the population diagnosed with diabetes will grow, even if current patterns in diabetes prevalence remain unchanged. Using current diabetes prevalence rates applied to Census Bureau population projections, the national cost of diabetes could grow to $156 billion by 2010 (in 2002 dollars) and to $192 billion by 2020 (Fig. 4). Direct medical costs could increase from $92 billion in 2002 to $138 billion in 2020, whereas indirect costs from lost productivity could increase from $40 billion in 2002 to $54 billion in 2020. The actual future cost of diabetes is likely to be substantially higher than these projected amounts if the prevalence of diabetes continues to grow -- especially for type 2 diabetes, which is correlated with the growing problem of obesity in the U.S. -- even after controlling for changing demographic characteristics.

Projected impact of changing demographic characteristics on the national cost of diabetes: 2002-2020 (in 2002 billions of dollars).