Diabetes Mellitus Health Article

COSTS OF DIABETES

Health care costs incurred by people with diabetes include non-diabetes-related and diabetes-related costs. In the United States, in 1992, the direct cost of non-diabetes-related and diabetes-related medical care incurred by people with diabetes was estimated to be $105.2 billion. The direct cost of medical care attributable to diabetes was estimated to be $45.2 billion and the indirect cost of diabetes was estimated to be $46.6 million (see Table 5).

In 1992, per capita health care expenditures for people with diabetes averaged $9,493, compared to $2,604 for people without diabetes. When adjusted for age, per capita health care expenditures for people with diabetes were approximately

Table 5

$3,800 higher for people with diabetes than for people without diabetes ($6,425 versus $2,604).

The fact that 62 percent of direct health care costs among people with diabetes and 82 percent of costs directly attributable to diabetes are incurred in the hospital setting suggests that the majority of costs are associated with the treatment of the late, chronic complications of diabetes.

SCREENING FOR TYPE 2 DIABETES

One-third of diabetes in the United States is undiagnosed, and one-third to one-half of all diabetes worldwide is undiagnosed. This finding, combined with the fact that glycemic management can prevent or delay the development of complications, and the fact that diabetic patients may already have complications at clinical diagnosis, have lead some to call for public health screening for type 2 diabetes. In general, screening is appropriate in asymptomatic populations when six specific conditions are met (see Table 6).

Table 6

Diabetes imposes substantial morbidity and mortality on the population. The natural history of type 2 diabetes is well understood, and with systematic testing, diabetes can be diagnosed in asymptomatic, preclinical, subjects. Unfortunately, although it is clear that intensified management can improve outcomes, no studies have demonstrated the effectiveness or safety of early treatment. Likewise, there is no consensus as to the optimal approach to screening for type 2 diabetes. Ideally, a screening test should be both sensitive and specific. Generally, however, trade-offs must be made between sensitivity and specificity (increasing sensitivity reduces specificity, and increasing specificity reduces sensitivity). In some health systems, the costs of screening and treatment are reasonable, but in others they are simply unaffordable. Finally, although it is recognized that screening must be an ongoing process, no empirical data exist to indicate the optimal screening frequency.

Questionnaires that use self-reported demographic, behavioral, and past medical history to assign a person to a higher or lower risk group; fasting, random, and postprandial urine glucose tests; fasting, random, and postprandial capillary whole blood and capillary plasma glucose tests; fasting, random, and postprandial venous whole blood and plasma glucose tests; and hemoglobin A1c have all been evaluated as screening tests for diabetes. In general, questionnaires perform rather poorly as screening tests for diabetes. Measurement of glycosuria using a cut-off point greater than or equal to a trace value generally has a low sensitivity and a high specificity. Capillary or venous whole blood or plasma glucose determinations have generally performed better than urine glucose testing. With both urine and blood testing, random, postprandial, and glucose-loaded tests perform better than fasting tests. There is little consensus, however, as to optimal cut-points for defining positive tests. Screening with hemoglobin A1c has suffered from lack of standardization of the assay. Even as this problem has been addressed, the test has generally been found to be specific but less sensitive than glucose measurements.

Accordingly, the American Diabetes Association has recommended that clinicians should be vigilant and recognize clinical histories and signs suggestive of diabetes that warrant testing. Generally, screening of high-risk individuals for type 2 diabetes should be performed only as part of ongoing medical care, understanding that the evidence is incomplete and questions remain as to the benefits and risks of early treatment, the optimal screening methods and cut-points, and screening frequency. Community-based screening for diabetes is generally associated with a low yield and poor follow-up, and it probably does not represent a good use of resources.

WILLIAM H. HERMAN

LIZA L. ILAG

(SEE ALSO: Cardiovascular Diseases; Glycosylated Hemoglobin; Noncommunicable Disease Control; Nutrition; Screening)