A risk scores for predicting prevalence of diabetes and pre-diabetes in the LAO population

Abstract

Objective: To develop a risk scores for predicting the prevalent diabetes and pre-diabetes in Lao population.

Research design and methods: a cross-sectional investigation was conducted with 1098 subjects aged between 30 to 70 years of Lao population in Vientiane. Multiple logistic regressions with backward stepwise selection were utilized in the statistical modeling, and the diabetes and pre-diabetes risk score values were derived from the relevant β-coefficients. Performances of the scores were determined by the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity for the optimal cut-off values.

Result: the prevalence of type 2 diabetes and pre-diabetes was 7% and 15.5 respectively. Factors included in the predictive model were 17 (> 40 of age) + 14 higher WC) + 11 (hypertension) + 7 (family history of diabetes) for “DM risk score” and 5 (> 40 of age) + 5 (hypertension) + 1 BMI for “Pre-DM risk score”. The cutoff point of “DM” and “Pre-DM” risk scores of 29.5 out of 49 and 5.5 out of 12 produced the optimal sum of sensitivity .75 and .76 , specificity .55 and .54, the AUC was .698 (p < .002) and .682 (p < .0001) in validation group respectively.

Conclusion: the researchers have developed a simple risk scores to be use in the screening in Lao population at high risk of diabetes and pre-diabetes. Their generalizability and validity for other Lao population, however, need further investigation.