Multimorbidity patterns, sociodemographic characteristics, and mortality: Data science insights from low-resource...

Area of Science:

• Epidemiology
• Public Health
• Biostatistics

Background:

• Multimorbidity analysis often uses simple disease counts, missing complex interrelationships.
• Understanding these relationships is vital for accurate mortality risk assessment.

Purpose of the Study:

• To identify clusters of co-occurring chronic conditions.
• To examine associations between these clusters and all-cause mortality.
• To detect distinct subpopulations with varying mortality risks based on multimorbidity patterns.

Main Methods:

• Utilized Partitioning Around Medoids and multidimensional scaling for cluster identification.
• Employed finite mixture modelling to identify latent subpopulations.
• Analyzed 8-year follow-up data from the CRONICAS Cohort Study in Peru.

Main Results:

• Identified three distinct multimorbidity clusters: respiratory, lifestyle/metabolic, and circulatory.
• Circulatory disease cluster significantly associated with mortality overall.
• Two subpopulations identified with differential mortality risks linked to specific clusters.

Conclusions:

• Multimorbidity clustering and subpopulation analysis provide deeper insights into mortality risks than simple disease counts.
• Sociodemographic factors interact with multimorbidity clusters to influence mortality.
• Personalized interventions targeting specific multimorbidity profiles within subpopulations are needed to effectively reduce mortality.