Variation and Prevalence of Multimorbidity Clusters across Rural Africa, 2018-2021

Cyril Chironda, University of the Witwatersrand, Johannesburg
Chodziwadziwa Kabudula, University of the Witwatersrand
Farai Mlambo, University of the Witwatersrand
Tobias Chirwa, University of the Witwatersrand
Francesc Xavier Gomez-Olive, University of the Witwatersrand
Rumbidzai Mupfuti, University of the Witwatersrand

Objectives: This study examines the prevalence and patterns of multimorbidity clusters among the deceased across rural Africa in HDSS sites (ARHI, AGINCOURT, DIMAMO, HARAR, KARONGA, KERSA, NAVRONGO, SIAYA) from 2018 to 2021, aiming to identify patterns common and distinct multimorbidity patterns that influence mortality across these sites. Methods: Using verbal autopsy data from the WHO 2016 VA questionnaire, the study involves data cleaning, exploratory analysis, and statistical and machine learning modelling with R and Python. Hierarchical clustering, k-medoids (PAM) clustering, and latent class analysis will investigate multimorbidity cluster patterns. Data Analysis: The study will assess multimorbid conditions’ distribution by age and sex, using clustering methods to identify disease co-occurrence patterns and latent profiles. Upset plots will visualize chronic condition intersections, and multinomial regression will identify demographic and behavioural risk factors. Expected Outcomes: The study aims to reveal significant multimorbidity patterns impacting mortality and variations across HDSS sites, providing insights into multimorbidity’s influence on health outcomes and informing targeted healthcare strategies. Significance: By highlighting multimorbidity clusters and their mortality relationships, this research will offer valuable insights into chronic disease management in resource-limited settings, informing public health policies and interventions to improve health outcomes and reduce mortality.

Keywords: Mortality and Longevity, Health and Morbidity

See extended abstract.