Social network fragmentation and community health

Area of Science:

• Public Health
• Network Science
• Epidemiology

Background:

• Community health interventions aim to prevent disease spread by targeting individuals.
• Conventional methods often rely on established community roles, but their effectiveness in network fragmentation is unclear.
• Network-based strategies offer a potential alternative for optimizing intervention targeting in low-income settings.

Purpose of the Study:

• To compare the efficiency of network-based acquaintance algorithms versus traditional community roles in fragmenting social and health advice networks.
• To assess the effectiveness of acquaintance algorithms in improving treatment compliance during mass drug administration (MDA).

Main Methods:

• Collected complete friendship and health advice networks in 17 rural Ugandan villages.
• Applied acquaintance algorithms (targeting neighbors of random nodes) and role-based targeting (health workers, leaders, teachers).
• Simulated the impact of network fragmentation on achieving target treatment compliance (≥75%) for deworming MDA.

Main Results:

• Acquaintance algorithms were significantly more efficient in fragmenting both friendship and health advice networks compared to targeting community roles.
• Community roles were poor indicators of household proximity and connections to sick individuals.
• Targeting health advice networks with acquaintance algorithms required removing only 32% of nodes to reduce non-compliance risk below 25%.

Conclusions:

• Network-based acquaintance algorithms represent a more efficient strategy for targeting individuals in rural health interventions than traditional role-based approaches.
• This approach shows promise for enhancing treatment compliance in mass drug administration (MDA) programs.
• Findings support the adoption of network analysis for optimizing public health strategies in resource-limited settings.