Evaluation of Madras Diabetes Research Foundation-Indian Diabetes Risk Score in detecting undiagnosed diabetes in the Indian population: Results from the Indian Council of Medical Research-INdia DIABetes population-based study (INDIAB-15)

Background & objectives

Screening of individuals for early detection and identification of undiagnosed diabetes can help in reducing the burden of diabetic complications. This study aimed to evaluate the performance of Madras Diabetes Research Foundation (MDRF)-Indian Diabetes Risk Score (IDRS) to screen for undiagnosed type 2 diabetes in a large representative population in India.

Methods

Data were acquired from the Indian Council of Medical Research-INdia DIABetes (ICMR-INDIAB) study, a large national survey that included both urban and rural populations from 30 states/union territories in India. Stratified multistage design was followed to obtain a sample of 113,043 individuals (94.2% response rate). MDRF-IDRS used four simple parameters, viz. age, waist circumference, family history of diabetes and physical activity to detect undiagnosed diabetes. Receiver operating characteristic (ROC) with area under the curve (AUC) was used to assess the performance of MDRF-IDRS.

Results

We identified that 32.4, 52.7 and 14.9 per cent of the general population were under high-, moderate- and low-risk category of diabetes. Among the newly diagnosed individuals with diabetes [diagnosed by oral glucose tolerance test (OGTT)], 60.2, 35.9 and 3.9 per cent were identified under high-, moderate- and low-risk categories of IDRS. The ROC-AUC for the identification of diabetes was 0.697 (95% confidence interval: 0.684-0.709) for urban population and 0.694 (0.684-0.704) for rural, as well as 0.693 (0.682-0.705) for males and 0.707 (0.697-0.718) for females. MDRF-IDRS performed well when the population were sub-categorized by state or by regions.

Interpretation & conclusions

Performance of MDRF-IDRS is evaluated across the nation and is found to be suitable for easy and effective screening of diabetes in Asian Indians.

Prevalence of diabetes and prediabetes in 15 states of India: results from the ICMR-INDIAB population-based cross-sectional study

BACKGROUND

Previous studies have not adequately captured the heterogeneous nature of the diabetes epidemic in India. The aim of the ongoing national Indian Council of Medical Research-INdia DIABetes study is to estimate the national prevalence of diabetes and prediabetes in India by estimating the prevalence by state.

METHODS

We used a stratified multistage design to obtain a community-based sample of 57 117 individuals aged 20 years or older. The sample population represented 14 of India's 28 states (eight from the mainland and six from the northeast of the country) and one union territory. States were sampled in a phased manner: phase I included Tamil Nadu, Chandigarh, Jharkhand, and Maharashtra, sampled between Nov 17, 2008, and April 16, 2010; phase II included Andhra Pradesh, Bihar, Gujarat, Karnataka, and Punjab, sampled between Sept 24, 2012, and July 26, 2013; and the northeastern phase included Assam, Mizoram, Arunachal Pradesh, Tripura, Manipur, and Meghalaya, with sampling done between Jan 5, 2012, and July 3, 2015. Capillary oral glucose tolerance tests were used to diagnose diabetes and prediabetes in accordance with WHO criteria. Our methods did not allow us to differentiate between type 1 and type 2 diabetes. The prevalence of diabetes in different states was assessed in relation to socioeconomic status (SES) of individuals and the per-capita gross domestic product (GDP) of each state. We used multiple logistic regression analysis to examine the association of various factors with the prevalence of diabetes and prediabetes.

FINDINGS

The overall prevalence of diabetes in all 15 states of India was 7·3% (95% CI 7·0-7·5). The prevalence of diabetes varied from 4·3% in Bihar (95% CI 3·7-5·0) to 10·0% (8·7-11·2) in Punjab and was higher in urban areas (11·2%, 10·6-11·8) than in rural areas (5·2%, 4·9-5·4; p<0·0001) and higher in mainland states (8·3%, 7·9-8·7) than in the northeast (5·9%, 5·5-6·2; p<0·0001). Overall, 1862 (47·3%) of 3938 individuals identified as having diabetes had not been diagnosed previously. States with higher per-capita GDP seemed to have a higher prevalence of diabetes (eg, Chandigarh, which had the highest GDP of US$ 3433, had the highest prevalence of 13·6%, 12.8-15·2). In rural areas of all states, diabetes was more prevalent in individuals of higher SES. However, in urban areas of some of the more affluent states (Chandigarh, Maharashtra, and Tamil Nadu), diabetes prevalence was higher in people with lower SES. The overall prevalence of prediabetes in all 15 states was 10·3% (10·0-10·6). The prevalence of prediabetes varied from 6·0% (5·1-6·8) in Mizoram to 14·7% (13·6-15·9) in Tripura, and the prevalence of impaired fasting glucose was generally higher than the prevalence of impaired glucose tolerance. Age, male sex, obesity, hypertension, and family history of diabetes were independent risk factors for diabetes in both urban and rural areas.

INTERPRETATION

There are large differences in diabetes prevalence between states in India. Our results show evidence of an epidemiological transition, with a higher prevalence of diabetes in low SES groups in the urban areas of the more economically developed states. The spread of diabetes to economically disadvantaged sections of society is a matter of great concern, warranting urgent preventive measures.

FUNDING

Indian Council of Medical Research and Department of Health Research, Ministry of Health and Family Welfare, Government of India.