Diabetes risk and provision of diabetes prevention activities in 44 low-income and middle-income countries: a cross-sectional analysis of nationally representative, individual-level survey data

Abstract

Background: The global burden of diabetes is rising rapidly, yet there is little evidence on individual-level diabetes prevention activities undertaken by health systems in low-income and middle-income countries (LMICs). Here we describe the population at high risk of developing diabetes, estimate diabetes prevention activities, and explore sociodemographic variation in these activities across LMICs.

Methods: We performed a pooled, cross-sectional analysis of individual-level data from nationally representative, population-based surveys conducted in 44 LMICs between October, 2009, and May, 2019. Our sample included all participants older than 25 years who did not have diabetes and were not pregnant. We defined the population at high risk of diabetes on the basis of either the presence of impaired fasting glucose (or prediabetes in countries with a haemoglobin A_1c available) or overweight or obesity, consistent with the WHO Package of Essential Noncommunicable Disease Guidelines for type 2 diabetes management. We estimated the proportion of survey participants that were at high risk of developing diabetes based on this definition. We also estimated the proportion of the population at high risk that reported each of four fundamental diabetes prevention activities: physical activity counselling, weight loss counselling, dietary counselling, and blood glucose screening, overall and stratified by World Bank income group. Finally, we used multivariable Poisson regression models to evaluate associations between sociodemographic characteristics and these activities.

Findings: The final pooled sample included 145 739 adults (86 269 [59·2%] of whom were female and 59 468 [40·4%] of whom were male) across 44 LMICs, of whom 59 308 (40·6% [95% CI 38·5-42·8]) were considered at high risk of diabetes (20·6% [19·8-21·5] in low-income countries, 38·0% [37·2-38·9] in lower-middle-income countries, and 57·5% [54·3-60·6] in upper-middle-income countries). Overall, the reach of diabetes prevention activities was low at 40·0% (38·6-41·4) for physical activity counselling, 37·1% (35·9-38·4) for weight loss counselling, 42·7% (41·6-43·7) for dietary counselling, and 37·1% (34·7-39·6) for blood glucose screening. Diabetes prevention varied widely by national-level wealth: 68·1% (64·6-71·4) of people at high risk of diabetes in low-income countries reported none of these activities, whereas 49·0% (47·4-50·7) at high risk in upper-middle-income countries reported at least three activities. Educational attainment was associated with diabetes prevention, with estimated increases in the predicted probability of receipt ranging between 6·5 (3·6-9·4) percentage points for dietary fruit and vegetable counselling and 21·3 (19·5-23·2) percentage points for blood glucose screening, among people with some secondary schooling compared with people with no formal education.

Interpretation: A large proportion of individuals across LMICs are at high risk of diabetes but less than half reported receiving fundamental prevention activities overall, with the lowest receipt of these activities among people in low-income countries and with no formal education. These findings offer foundational evidence to inform future global targets for diabetes prevention and to strengthen policies and programmes to prevent continued increases in diabetes worldwide.

Funding: Harvard T H Chan School of Public Health McLennan Fund: Dean's Challenge Grant Program and the EU's Research and Innovation programme Horizon 2020.

Figure 1:. Proportion of the population at high risk of developing diabetes by country

Figure shows the weighted estimated proportion of each country’s population deemed at high risk of developing diabetes based on the WHO Package of Essential Noncommunicable Disease Guidelines. Division of income is according to World Bank income group. Sample excludes those who meet criteria for a diagnosis of diabetes or are pregnant. Error bars represent 95% CIs.

Figure 2:. Self-reported number of diabetes preventive services received

Number of countries in each World Bank income group listed below each label. Estimates of prevalence were calculated with sampling weights readjusted to weigh each country equally. Error bars represent 95% CIs.

Figure 3:. Proportion of the population at high risk who self-reported having received each diabetes preventive intervention by World Bank income group

Sampling weights were adjusted to represent each country equally. Error bars represent 95% CIs.

Figure 4:. Association between individual characteristics and having received diabetes preventive services

The marginal differences and 95% CIs in predicted probability of reporting having received each preventive service are depicted, with the base category for each sociodemographic characteristic serving as the reference point. Units for the estimated average marginal effect are percentage point differences in predicted probability. Error bars represent 95% CIs. Models were estimated with sampling weights readjusted to represent each country equally, and accounted for country-level fixed effects.