A multi-stage stochastic programming approach to epidemic resource allocation with equity considerations

Area of Science:

• Epidemiology
• Operations Research
• Public Health

Background:

• Existing epidemiological models struggle with optimizing resource allocation under disease growth uncertainty.
• Effective epidemic control requires dynamic strategies that integrate disease progression and resource management.

Purpose of the Study:

• To develop a multi-stage stochastic programming compartmental model for optimizing resource allocation in infectious disease outbreaks.
• To integrate uncertain disease progression with resource distribution for improved epidemic control.
• To introduce and analyze novel equity metrics for fair resource allocation.

Main Methods:

• Formulated a multi-stage stochastic program incorporating disease growth scenarios.
• Optimized distribution of treatment centers and resources to minimize infections and funerals.
• Defined and applied infection and capacity equity metrics for resource allocation.
• Calculated the multi-stage value of the stochastic solution (VSS) to assess model superiority.

Main Results:

• The stochastic model effectively balances infection proportions across regions, even without explicit equity constraints.
• Allocating resources proportionally to population size was found to be sub-optimal and potentially detrimental.
• The model demonstrated the superiority of stochastic programming over deterministic approaches for epidemic control.
• The proposed model was applied to control Ebola Virus Disease (EVD) in West Africa.

Conclusions:

• The developed multi-stage stochastic epidemic-logistics model offers a practical approach for optimizing resource allocation in infectious disease outbreaks.
• The model provides a framework for dynamic and equitable resource distribution, adaptable to various diseases and evolving situations.
• Fairness considerations in resource allocation can significantly impact overall epidemic control outcomes and costs.