The objective of this study was to estimate the national medical costs associated with gestational diabetes mellitus (GDM) in 2007. We analyzed the National Hospital Discharge Survey to estimate the national prevalence of GDM. Using Poisson regression analysis with medical claims for about 27,000 newborns and their mothers, we estimated rate ratios that reflect the increase in use of health care services associated with GDM. Combining GDM prevalence rates with these rate ratios, we calculated etiological fractions that reflect the proportion of national health care resource use associated with GDM. We then multiplied these fractions by estimates of national health care use and costs in 2007. GDM prevalence increases with age, rising from 1.3% of pregnancies of women younger than age 21 to 8.7% of pregnancies of women older than age 35. For the estimated 180,000 GDM pregnancies resulting in delivery, average expenditures increased $3,305 per pregnancy plus $209 in the newborn's first year of life. GDM increased national medical costs by $636 million in 2007-$596 million for maternal costs and $40 million for neonatal costs. Approximately $230 million (36%) of GDM-related medical costs are covered by government programs (primarily Medicaid), $355 million (56%) are covered by private insurers, and $51 million (8%) are covered by self-pay and charity care. GDM imposes a significant economic burden. These estimates of the economic burden of GDM are likely conservative because we focus on near-term medical costs, omitting the increased risk for long-term sequelae.

The objective is to estimate the national economic costs associated with undiagnosed diabetes mellitus (UDM). UDM is defined as unknowingly having an elevated glucose level that meets the definition of diabetes. National Health and Nutrition Examination Survey (NHANES) data are used to estimate the prevalence of UDM. Because UDM cannot be directly observed in medical claims for analyzing per capita patterns of health care use, we analyze annual medical claims from a proxy population--people within 2 years of first diagnosis of diabetes. For a commercially insured population first diagnosed with diabetes in 2006 (n = 29,770), we compare their annual health care use in 2004 and 2005 to that of patients with no history of diabetes between 2004 and 2006 (n = 3.2 million). We combine estimates of UDM prevalence from NHANES with health care use patterns from the proxy population to estimate etiological fractions that reflect the portion of national health care use associated with UDM. Approximately 6.3 million adults in the United States have UDM in 2007. Annual per capita use of health care services for the UDM proxy population is higher than for a comparable group with no history of diabetes, but lower than for a comparable group with a history of diabetes. The estimated economic costs of UDM in 2007 is $18 billion ($2864 per person with UDM), including medical costs of $11 billion and indirect costs of $7 billion. Although the high prevalence of UDM makes it an important health issue to be studied, data limitations have contributed to a dearth of information on the health care use patterns and economic costs of UDM. By omitting UDM, estimates of the total national cost of diabetes are underestimated.

In this article, we estimate national health care resource use and medical costs in 2007 associated with prediabetes (PD), defined as either fasting plasma glucose between 100 and 125 or oral glucose tolerance test between 140 and 200. We use Poisson regression with medical claims for an adult population continuously insured between 2004 and 2006 to analyze patterns of health care resource use by PD status. Combining rate ratios that reflect health care use patterns with national PD prevalence rates from the National Health and Nutrition Examination Survey, we calculate etiological fractions to estimate the portion of national health resource use associated with PD. The findings suggest that PD is associated with statistically higher rates of ambulatory visits for hypertension; endocrine, metabolic, and renal complications; and general medical conditions. PD is associated with a slight increase in visit rates for neurological symptoms, peripheral vascular disease, and cardiovascular disease, but the increase is not statistically significant. There is no indication that PD is associated with an increase in emergency visits and inpatient days. Extrapolating these patterns to the 57 million adults with PD in 2007 suggests that national annual medical costs of PD exceed $25 billion, or an additional $443 for each adult with PD. PD is associated with excessive use of ambulatory services for comorbidities known to be related to diabetes. Our findings strengthen the business case for lifestyle interventions to prevent diabetes by adding additional economic benefits that potentially can be achieved by preventing or delaying PD.

No abstract