Why do so many diabetics remain undiagnosed? results from a longitudinal investigation

Undiagnosed diabetes remains a major public health challenge. Approximately 13% of the adult US population has diabetes, of which 40% is undiagnosed. Many diabetics have the disease for 7-10 years prior to diagnosis. The objective of this research was to identify and quantify factors that differentiate those diagnosed with diabetes versus those who remain undiagnosed. The Boston Area Community Health (BACH) Survey is a longitudinal (2002-2012), racially/ethnically diverse (Black, Hispanic, White) cohort study. Using a validated screening algorithm (74% sensitivity, 41% specificity), we identified 231 likely undiagnosed diabetics who were confirmed at follow-up to have either diagnosed or undiagnosed (fasting glucose > 125 mg/dL) diabetes. After an average of 7.2 years of follow-up, 12.4% had been diagnosed, while 13.2% remained undiagnosed. Contrasting with previous studies we found neither sociodemographic factors (race/ethnicity, age, employment status, occupation, education, nativity, marital status), medical care access measures (insurance, regular care, visits to the doctor), nor most health measures (comorbid conditions, family history of diabetes, medication use at baseline) distinguished the diagnosed from the undiagnosed. Multivariate logistic regression models revealed diabetes diagnosis was related to starting use of statins (OR=6.92, CI 2.35-20.37,p=<0.001) or antihypertensive medications (OR 0.30, CI 0.09-0.93, p=0.037). This finding may indicate diabetes is diagnosed serendipitously (an unexpected finding obtained from a panel while checking cholesterol levels), causing a physician to contemplate diabetes and test appropriately. More systematic early and cost efficient detection, and less haphazard accidental or serendipitous diagnosis, is required if we are to successfully manage the burgeoning epidemic of diabetes.