A Stratification Method for Identifying Subgroups at High-Risk for Type 2 Diabetes in sub-Saharan Africa

Area of Science:

• Global Health
• Epidemiology
• Biostatistics

Background:

• Type 2 diabetes prevalence is increasing globally, particularly in sub-Saharan Africa.
• Low-resource settings face challenges with undiagnosed diabetes due to limited screening.
• Existing risk models may not capture complex factor interactions relevant to African populations.

Purpose of the Study:

• To identify accurate, region-specific risk profiles for Type 2 diabetes in African populations.
• To evaluate the effectiveness of multidimensional subgroup discovery for diabetes risk assessment.
• To improve screening strategies for undiagnosed Type 2 diabetes in low-resource settings.

Main Methods:

• Applied a multidimensional subgroup discovery algorithm to cross-sectional data from three African populations.
• Analyzed combinations of demographic, anthropometric, and lifestyle risk factors.
• Compared identified high-risk subgroups with guideline-based definitions.

Main Results:

• Identified specific combinations of risk factor cutoffs defining high-risk subgroups for Type 2 diabetes.
• A consistent profile (waist-to-hip ratio > 0.9, low physical activity, family history) showed elevated risk across regions.
• This multidimensional approach demonstrated superior predictive performance compared to traditional methods.

Conclusions:

• Interactions between multiple risk factors provide more accurate diabetes risk characterizations in African contexts.
• Core anthropometric and familial factors are broadly applicable, while other factors are region-specific.
• This refined risk stratification enhances predictive performance for targeted diabetes screening.