A spatial decision support system for optimizing expansion of COVID immunization services

Abstract Background Access to COVID 19 continues to be a significant challenge in most low- and middle-income countries. There have been longstanding health system barriers towards accessing services in these countries over time, with geographic barriers being significant particularly given majority of people live in rural areas with poor facility network coverage. There has been no systematic assessment of existing networks in addition to how they can be optimised. Methods Using Kenya as an example, we mapped facilities providing COVID immunization facilities in the country. We then used a cost distance algorithm to assess geographic access to these facilities using a model that adjusts for differences in travel speeds across differences in elevation, land use and travel roads. We then extracted populations living within 1 hour of these facilities as a proxy of access for the 47 counties in the country. Finally, we used the same cost distance parameters to select facilities that would be used to optimize selection of additional facilities that would move access to 80%. Results We identified 621 facilities that were administering COVID vaccines as of August 2021. The accessibility model highlighted that only 60% of the country’s population was living within 1 hour of these facilities. The scale up activity using the cost distance algorithm identified an additional 560 facilities that can be prioritised to move theoretical access to 80%. These were not evenly distributed in the country, with counties such as Kakamega, Mandera, Nakuru and Nandi having additional 30 facilities selected. Discussion We highlight that 2 in five adults in Kenya live more than one hour from the nearest COVID immunization facility and this may be a barrier towards achieving universal access. We also highlight where efforts to improve coverage by increasing cold chain services can be concentrated to improve access. Furthermore, results can be used to improve decision making on where other interventions like community campaigns can be actively pursued for universal coverage.