Immunisation stands as one of the most powerful public health interventions, yet global vaccination coverage plateaued for nearly a decade before COVID-19, according to the WHO. In 2023, an estimated 14.5 million children worldwide missed out on any vaccination, highlighting a crucial gap in healthcare. Against this backdrop, we were very pleased to present at the Immunity for Health conference in October, hosted by Flanders Vaccine in Ghent. The event gathered leading experts, researchers, industry professionals, and government bodies to explore advancements, share insights, and foster collaborations in the vaccine field.
Leveraging AI to address inequities in vaccination
Dr Sumbul Hashmi, our Global Public Health manager, delivered a very well-received presentation on our innovative use of high resolution data to analyse disparities in vaccination coverage among children in the Philippines, a country where approximately 637,000 zero dose children reside. Our presentation formed part of a broader collaboration with the WHO’s Western Pacific Region Office (WPRO) to map health status across ten countries in the region.
Locating critical vaccination gaps
Our presentation highlighted how high-resolution subnational data can be used to unmask gaps in vaccination coverage. Leveraging the Epi-control platform, which integrates AI and diverse open-source datasets, we uncovered stark regional variations. Areas like Davao showed higher vaccination rates, while regions such as the Zamboanga peninsula and Cagayan valley exhibited significantly lower coverage.
We spoke about the importance of high-resolution mapping in identifying and addressing geographical barriers to immunisation. By integrating a wide range of open-source data—including socioeconomic indicators, wealth indices, travel time to healthcare facilities, and access to improved sanitation—we can accurately pinpoint underserved regions. This enables stakeholders to deploy targeted interventions.
Finding ‘Positive Deviants’: learning from outliers
A key element of our strategy was the idea of the Positive Deviance (PD) approach promoted by the country partners. This methodology identifies communities or individuals – known as Positive Deviants – who achieve better-than-expected results despite facing similar challenges and resource limitations. To map the geographical distribution of PDs, we again use open source data.
By isolating these positive outliers, EPCON supports the country partners to uncover successful local practices or interventions that could be replicated elsewhere. These could include targeted vaccination campaigns, incentivised completion strategies, innovative vaccine delivery models, or unique local policies. More research is definitely needed to fully understand successful practices.
We concluded our presentation with a call for further research and highlighted the potential of AI and high-resolution data in revolutionising public health strategies. Future studies could use AI to develop even more refined models to enhance healthcare delivery in low- and middle-income countries, ultimately reducing inequities in worldwide vaccination coverage.