A simple screening score for diabetes for the Korean population: development, validation, and comparison with other scores

Abstract

OBJECTIVE: We developed and validated a self-assessment score for diabetes risk in Korean adults and compared it with other established screening models.

RESEARCH DESIGN AND METHODS: The Korea National Health and Nutrition Examination Survey (KNHANES) 2001 and 2005 data were used to develop a diabetes screening score. After excluding patients with known diabetes, 9,602 participants aged ≥ 20 years were selected. Undiagnosed diabetes was defined as a fasting plasma glucose ≥ 126 mg/dL and/or nonfasting plasma glucose ≥ 200 mg/dL. The SAS Survey Logistic Regression analysis was used to determine predictors of undiagnosed diabetes (n = 341). We validated our model and compared it with other existing methods using the KNHANES 2007-2008 data (n = 8,391).

RESULTS: Age, family history of diabetes, hypertension, waist circumference, smoking, and alcohol intake were independently associated with undiagnosed diabetes. We calculated a diabetes screening score (range 0-11), and a cut point of ≥ 5 defined 47% of adults as being at high risk for diabetes and yielded a sensitivity of 81%, specificity of 54%, positive predictive value of 6%, and positive likelihood ratio of 1.8 (area under the curve [AUC] = 0.73). Comparable results were obtained in validation datasets (sensitivity 80%, specificity 53%, and AUC = 0.73), showing better performance than other non-Asian models from the U.S. or European population.

CONCLUSIONS: This self-assessment score may be useful for identifying Korean adults at high risk for diabetes. Additional studies are needed to evaluate the utility and feasibility of this score in various settings.