A spatiotemporal recommendation engine for malaria control.

Area of Science:

• Epidemiology
• Public Health
• Health Economics

Background:

• Malaria remains a significant global health burden, necessitating efficient resource allocation for effective control.
• Current malaria interventions require optimized strategies to manage limited resources in real-time.
• Policy-making for malaria control needs to balance disease suppression with interpretability and equity.

Purpose of the Study:

• To develop a framework for optimizing real-time resource allocation for malaria control.
• To formalize a policy for sequential decision-making in resource allocation based on disease data.
• To construct an interpretable class of resource allocation policies.

Main Methods:

• Developed a framework combining a hierarchical Bayesian spatiotemporal model for disease transmission.
• Integrated a policy-search algorithm to estimate an optimal resource allocation policy.
• Constructed an interpretable policy class accommodating continuous resource domains.

Main Results:

• The estimated optimal policy significantly improved cumulative long-term outcomes in simulations.
• The framework demonstrated improved malaria intervention outcomes in the Democratic Republic of the Congo.
• The proposed policy framework outperformed naive resource allocation approaches.

Conclusions:

• The developed framework provides an effective approach for real-time, interpretable, and equitable resource allocation in malaria control.
• Optimized resource allocation strategies are crucial for reducing the global burden of malaria.
• This approach offers a valuable tool for policymakers in managing infectious diseases with limited resources.