Multimodal Machine Learning for Multifactor Geospatial Prioritization of Suitable Healthcare Facility Sites

This study introduces a comprehensive multifactor geospatial methodology to identify and prioritize optimal locations for the construction of new healthcare facilities. Building on a previously developed geospatial machine learning pipeline designed to predict suitable sites, we introduce two complementary machine learning-based methodologies to enhance the decision-making process. The first methodology employs a population-density-based approach with the objective of maximizing the impact of new healthcare facilities by prioritizing areas with the highest population densities. This strategy ensures that the greatest number of individuals can benefit from new healthcare services. The second methodology is a geographical-distribution-based approach, which focuses on promoting equitable access to healthcare by targeting underserved and underrepresented areas, regardless of administrative boundaries. By prioritizing locations that currently lack sufficient healthcare resources, this approach aims to reduce disparities in healthcare accessibility and improve overall equity in resource allocation. The integration of these two methodologies provides a balanced framework that addresses both the need for widespread impact and the imperative for equitable access. The results of this study are expected to help improve healthcare access, ensuring that new facilities are strategically placed to serve densely populated regions and underserved communities, ultimately leading to more equitable and effective healthcare delivery.