A Decision Support Tool for Domain-Driven Mobile Clinic Routing: A Case Study in Kenya

Mobile clinics are often employed by organizations to provide affordable and flexible healthcare services to communities with limited healthcare access. The dynamic and unstable situations for mobile clinic deployment require decision-making tools that can periodically aid practitioners in operational and tactical planning to reach the locations in need of healthcare. We propose a routing algorithm that can take care of the uncertain and dynamic nature of healthcare demand by embedding the machine learning model for demand prediction into the optimization framework. We also developed a spreadsheet-based decision support tool that integrates the proposed model to aid practitioners in making data-driven decisions for mobile clinic routing. In association with AMREF, we deploy the tool in the context of COVID-19 vaccination in Kenya. We adopt a Design Science Research approach where the artifact (the proposed model and the tool) of our design is evaluated (quantitatively and qualitatively) and refined iteratively. Future work will include numerical experiments to evaluate the proposed model based on various metrics. The tool will be tested in the field by AMREF, and their user feedback and suggestions will be used to refine the tool. Although the tool is demonstrated for COVID-19 vaccinations, it can be generalized for various scenarios that require mobile clinic deployments and will be made freely available.