Context: Bronx, New York, an urban county with a large low-income, immigrant and minority population, has a prevalence of diabetes that is among the highest in the United States.
Objective: The aim of the study was to evaluate the utility of hemoglobin A1c (HbA1c) in identifying patients at risk for diabetes on an in-patient medical service of a hospital serving a high prevalence community.
Design and setting: We conducted a prospective cohort study at an urban public hospital.
Patients: The study included 971 patients (1132 admissions) admitted to the general medicine service over 4 months.
Main outcome measures: HbA1c was measured on all patients. Records were checked for prior diagnosis of diabetes and other clinical data. Follow-up data were obtained for those with repeat HbA1c testing or glucose within 1 yr after admission.
Results: We found that 35.2% of the patients (n = 342) had an established diagnosis of diabetes. The remaining 629 patients defined the study cohort of patients without known diabetes. Mean HbA1c was 6.05 +/- 0.87%. A total of 152 patients (24%) had admission HbA1c of at least 6.5% and 62 (9.9%) had HbA1c of at least 7.0%. Fifty-five patients with HbA1c of at least 6.5% had follow-up HbA1c within 1 yr. Of those, 44 (80.0%) met the criteria for diabetes as proposed by The International Expert Committee using repeated HbA1c testing.
Conclusion: In communities with high prevalence of diabetes, a large percentage of patients without a diagnosis of diabetes who are admitted as in-patients have HbA1c of at least 6.5% and 7.0%. Hospital-based HbA1c testing might identify patients for whom further testing is indicated to make the diagnosis of diabetes.