Deploying ADVISER: Impact and Lessons from Using Artificial Intelligence for Child Vaccination Uptake in Nigeria

TL;DR

This paper reports the first real-world deployment of an AI-driven health intervention optimizer, ADVISER, to improve child vaccination uptake in Oyo State, Nigeria. The framework uses an $ILP$-based system to allocate one intervention per beneficiary by learning a heterogeneous success model $P_{s}(i,a)$ via logistic regression and solving a constrained optimization for resource-limited settings. Deployments covered 13,783 mothers across 99 clinics, leveraging a vaccination tracker, cloud infrastructure, and a grid-based travel-time approximation to reduce costs. Results show a substantial uplift in vaccination completion compared with baseline and provide actionable lessons on data, partnerships, and cost-effective deployment in low-resource environments. The work demonstrates the feasibility and impact of AI-augmented public health interventions in Nigeria and potentially other resource-constrained regions.

Abstract

More than 5 million children under five years die from largely preventable or treatable medical conditions every year, with an overwhelmingly large proportion of deaths occurring in underdeveloped countries with low vaccination uptake. One of the United Nations' sustainable development goals (SDG 3) aims to end preventable deaths of newborns and children under five years of age. We focus on Nigeria, where the rate of infant mortality is appalling. In particular, low vaccination uptake in Nigeria is a major driver of more than 2,000 daily deaths of children under the age of five years. In this paper, we describe our collaboration with government partners in Nigeria to deploy ADVISER: AI-Driven Vaccination Intervention Optimiser. The framework, based on an integer linear program that seeks to maximize the cumulative probability of successful vaccination, is the first successful deployment of an AI-enabled toolchain for optimizing the allocation of health interventions in Nigeria. In this paper, we provide a background of the ADVISER framework and present results, lessons, and success stories of deploying ADVISER to more than 13,000 families in the state of Oyo, Nigeria.

Figures (4)

• Figure 1: The figure denotes under-five mortality rate, defined as the probability per 1,000 that a newborn baby will die before reaching age five. Nigeria, denoted by the orange point in the map, currently has the third highest rate of child mortality in the world. (Source: World Bank, shared under CC BY-4.0 License)
• Figure 2: A snapshot from ADVISER's deployment. We envision that the uptake in vaccination through our deployment will potentially pave the way for eliminating infant and child mortality in Nigeria. (Image Courtesy: HelpMum, shared with permission)
• Figure 3: Different stages in the ADVISER framework. 1) We collect data from mothers either through a cell phone application or in-person at health centers and hospitals. 2) The data is cleaned and stored in a database server. 3) We use historical data and surveys to estimate the success of interventions. 4) The resource allocation problem is formulated as an integer linear program, whose solution space is greedily pruned. 5) We do guided local search for generating promising vehicle routes. 6) A significantly smaller ILP, with a reduced solution space, is solved by branch and bound. 7) The matches are then deployed in the field.
• Figure 4: Images from ADVISER's deployment (all images have been shared with permission) (a) Our staff collecting information to be updated to the Vaccination Tracker database. This data serves as the fundamental building block of the ADVISER framework. (b) A nurse vaccinating an infant during a vaccination drive. (c) Nurses prior to a vaccination drive. The cold storage box (kept in front of the nurses) had limited efficacy during the deployment. (d) Two mothers boarding the pickup service bus with their children. (Images Courtesy: HelpMum, shared with permission)