Comparison of near-patient capillary glucose measurement and a risk assessment questionnaire in screening for type 2 diabetes in a high-risk population in rural India.

Abstract

To assess the utility of a point-of-care (POC) capillary blood glucose measurement as compared with routine clinical parameters in predicting undiagnosed diabetes in a low-resource rural India setting. Nine hundred and ninety-four participants aged >30 years and stratified by age and sex were randomly selected from 20 villages in India. A clinical questionnaire, sampling for laboratory venous fasting plasma glucose (FPG), and POC capillary blood glucose assay were performed simultaneously. Diabetes diagnosis was based on the World Health Organization (WHO) definition using FPG. The capacity of the POC glucose to predict the presence of diabetes was assessed and compared with the questionnaire using area under the receiver operating characteristic curves (AUCs). The AUC for POC glucose alone in predicting diabetes was 0.869 (95% CI 0.810-0.929). This was significantly better (P < 0.001 for AUC comparison) than the models based upon clinical variables alone (AUC for the best clinical model including age, BMI, hypertension, waist circumference: 0.694 [95% CI 0.621-0.766]). POC glucose appropriately reclassified the risk of up to one-third of participants ranked according to the clinical models. Adding the clinical variables to the POC glucose assay did not significantly improve the discriminatory capability beyond that achieved with the POC glucose measurement alone (all P > 0.37). POC glucose testing appears to be a simple and reliable tool for identifying undiagnosed diabetes in a high-risk, resource-poor rural population. However, studies evaluating the cost effectiveness of introducing POC glucose testing are needed prior to widespread implementation.