Modelling Interventions to Combat Antibacterial Resistance in East Africa Using Causal Bayesian Networks.

Area of Science:

• Public Health
• Infectious Diseases
• Epidemiology

Background:

• Antibacterial resistance (ABR) is a global health threat, complicated by interconnected drivers.
• Multi-drug resistant (MDR) bacteria increasingly impact common infections like urinary tract infections (UTIs).
• A systems-based approach is needed to understand and address ABR drivers.

Purpose of the Study:

• To construct a causal diagram of the drivers influencing the prevalence of MDR UTIs.
• To identify key intervention points for reducing MDR UTI prevalence using a systems perspective.

Main Methods:

• Analysis of 2,007 adult outpatients with UTIs in Kenya, Tanzania, and Uganda (2019-2020).
• Application of structure learning in Bayesian networks combined with expert knowledge.
• Conducting hypothetical interventions to estimate causal effects of identified drivers.

Main Results:

• MDR UTI prevalence was significantly influenced by demographic, socioeconomic, and environmental factors.
• Factors such as education access, water and sanitation quality, and overcrowding showed greater impact than recent antibiotic use.
• Hypothetical interventions indicated that improving living conditions could substantially decrease MDR prevalence.

Conclusions:

• A systems-based approach effectively identifies underlying causal patterns in MDR UTI prevalence.
• Interventions targeting socioeconomic and environmental determinants are crucial for combating MDR infections.
• Complexity-aware strategies are essential for developing effective, targeted interventions against ABR.