Development and validation of a simple risk score for prevalent undiagnosed type 2 diabetes in Southern Chinese population

This study aimed to develop a simple risk score for detecting individuals with undiagnosed type 2 diabetes mellitus in a Southern Chinese population. A sample of participants from the 2006-2007 Guangzhou type 2 diabetes cross-sectional study, aged 20 years and above, was used to develop scores for men and women separately by multiple logistic regression analysis. External validation was performed based on three other type 2 diabetes studies (the Zhuhai rural sample, the 2008-2010 Guangzhou urban sample, and the Tibet sample). Performances of the scores were measured using Hosmer-Lemeshow goodness-of-fit test and receiver operating characteristic (ROC) c-statistic. Age, waist circumference, BMI, and family history of diabetes were included in the risk score for both men and women, and the risk factor of hypertension was included for men. The ROC c-statistics were 0.70 for men and 0.78 for women in the Zhuhai rural sample, and 0.76 for men and 0.69 for women in the 2008-2010 Guangzhou urban sample. At the risk score value of a parts per thousand yen36 for men and a parts per thousand yen28 for women, the scores gave a sensitivity, specificity, positive predictive value, and negative predictive value of 56.3, 73.2, 9.0, and 97.0 % for men and 62.5, 76.7, 11, and 97.0 % for women in the Zhuhai sample, respectively. In the 2008-2010 Guangzhou urban sample, the above measures are 85.7, 60.5, 3, and 100 % for men and 77.3, 53.0, 2, and 99 % for women. However, the scores performed poorly in the Tibet sample. The results indicate that the simple diabetes risk scores can be generalized in the Guangzhou city and the nearby rural regions and thus can help primary health care workers to identify individuals with undiagnosed type 2 diabetes.