A simple nomogram for identifying individuals at high risk of undiagnosed diabetes in rural population.

AIMS: To sought for an easily applicable nomogram for detecting individuals at high risk of undiagnosed type 2 diabetes. METHODS: The development cohort included 2542 participants recruited randomly from a rural population in 2011.The glycemic status of subjects was determined using the fasting plasma glucose test and the oral glucose tolerance test. The Bayesian Model Average approach was used to search for a parsimonious model with minimum number of predictor and maximum discriminatory power. The corresponding prediction nomograms were constructed and checked for discrimination, calibration, clinical usefulness, and generalizability in nationwide population in 2012. RESULTS: The non-lab nomogram including waist circumference and systolic blood pressure was the most parsimonious with the area under receiver operating characteristic curve (AUC) of 0.71 (95 %CI = 0.64-0.76). Adding low-density lipoprotein cholesterol in the non-lab nomogram generated the lab-based nomogram with significantly improved AUC of 0.83 (0.78-0.87, P < 0.001). The nomograms had a positive net benefit at threshold probability between 0.01 and 0.15. Applying the non-lab nomogram to the national population yielded the AUC of 0.66 (0.63-0.70) and 0.68 (0.65-0.71) in the cohorts aged 40-64 and 30-69 years, respectively. CONCLUSIONS: The novel nomograms could help promote the early detection of undiagnosed diabetes in rural Vietnamese population.

Prevalence and correlates of hyperglycemia in a rural population, Vietnam: implications from a cross-sectional study.

BACKGROUND: Despite the increasing prevalence of type 2 diabetes in urban areas, relatively little has been known about its actual prevalence and its associations in rural areas, Vietnam. The purpose of this study was to evaluate the prevalence of impaired fasting glucose (IFG), impaired glucose tolerance (IGT), diabetes and their risk factors in a rural province, Vietnam. METHODS: A cross-sectional study with a representative sample was designed to estimate the hyperglycemia prevalence, using 75-g oral glucose tolerance test. Potential risk factors for hyperglycemia were analyzed using multinomial logistic regression, taken into account influences of socio-economic status, anthropometric measures, and lifestyle-related factors. RESULTS: The age and sex-adjusted prevalence rates (95% CI) of isolated IFG, isolated IGT, combined IFG-IGT, and diabetes were 8.7 (7.0-10.5), 4.3 (3.2-5.4), 1.6 (0.9-2.3), and 3.7% (2.7-4.7%), respectively. There were still 73% of diabetic subjects without knowing the condition. Blood pressure, family history of diabetes, obesity-related measures (waist circumference, waist-hip ratio, body fat percentage, and abdominal obesity) were the independent risk factors for hyperglycemia (IFG, IGT, and diabetes). CONCLUSIONS: The prevalence of hyperglycemia in rural areas has not been as sharply increased as that reported in urban cities, Vietnam. Blood pressure and obesity-related measures were the most significant predictors for hyperglycemia level and they can be taken into account in building prognosis models to early detection of diabetes in rural Vietnamese populations.