# Determinants of old age disability in Botswana: an empirical investigation using generalized linear models

## TL;DR

This study identifies personal, household, and community factors that influence disability among older adults in Botswana, highlighting the need for inclusive policies.

## Contribution

The study provides a comprehensive multilevel analysis of disability determinants in Botswana using nationally representative data and generalized linear models.

## Key findings

- Older adults aged 65–69 and 70–74 had significantly lower disability rates compared to those aged 80+.
- Males had lower disability counts than females, and higher education was associated with lower disability rates.
- Urban residence and access to infrastructure were linked to reduced disability in older populations.

## Abstract

As Botswana experiences a demographic transition marked by an expanding population of older adults, understanding the determinants of disability among older adults becomes critical for shaping inclusive health and social policies. Disability in later life often stems not only from biological ageing but also from intersecting personal, household, and community-level conditions. Despite increasing attention to ageing in sub-Saharan Africa, few studies have comprehensively assessed the multilevel factors influencing functional limitations in older populations.

Drawing on nationally representative microdata from the 2022 Botswana Population and Housing Census, this study employed Generalized Linear Models (specifically Poisson regression) to examine the severity of disability among individuals aged 65 years and above. Disability was measured using a composite count variable derived from three functional domains: mobility, self-care, and cognition. The final analytical sample comprised 47,309 older adults. The model integrated a wide range of covariates across individual, household, and community levels. Model selection was based on goodness-of-fit statistics, including AIC, BIC, and deviance diagnostics.

The multilevel Poisson model revealed that age, gender, education, marital status, and employment status were significant individual-level predictors of disability. Older adults aged 65–69 and 70–74 were significantly less likely to experience multiple disabilities compared to those aged 80+, with IRRs of 0.434 and 0.576 respectively. Males had lower disability counts than females (IRR = 0.648), and those with only primary or less education had higher disability rates than those with tertiary education (IRR = 1.207). At the household level, individuals in smaller households (number of rooms) reported significantly higher disability levels. At the community level, urban residence, access to electricity, internet, and transportation services were all associated with reduced disability. Although interaction terms were not explicitly specified, the integrated model structure supported inferences about cross-level interactions between environmental infrastructure and individual vulnerabilities.

This study provides a comprehensive analysis of the multifactorial determinants of old-age disability in Botswana. The findings underscore the need for integrated, multisectoral strategies that go beyond healthcare access to include educational equity, age-friendly infrastructure, digital inclusion, and gender-sensitive social protection. Policies must address not only individual risk but also household and community conditions that jointly shape disability outcomes. These insights provide a critical roadmap for building age-inclusive societies as Botswana continues to age.

The online version contains supplementary material available at 10.1186/s12877-025-06534-z.