To develop risk scores for predicting the prevalence of diabetes in the Lao population. This was a cross-sectional study of both men and women (age 30 to 70 years) living in rural villages of the Vientiane municipality in the Lao PDR. Multiple logistic regressions with backward stepwise selection were used; the diabetes risk score was derived from the β-coefficient. Performance of the score was determined by the area under the receiver operating characteristic curve (AUC), the sensitivity, the specificity, and the positive predictive value for the specified cut-off value. The prevalence of undiagnosed diabetes was 7%. The factors included in the predictive in model were 17 (40 to 70 years of age) + 14 (high waist circumference) + 11 (hypertension) + 7 (family history of diabetes). A cut-off point of risk scores of 29.5 out of 49 produced the optimal sum, leading to a sensitivity of 0.75, a specificity of 0.55, a positive predictive value of 17.8%, and an AUC of 0.70. Data suggested that the combination of age, waist circumference, hypertension, and family history of diabetes could be utilized to identify Lao individuals at high risk of undiagnosed diabetes. The generalizability for other Lao population needs further investigation.
Read full abstract