Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Optimizing HIV Patient Engagement with Reinforcement Learning in Resource-Limited Settings

África Periáñez, Kathrin Schmitz, Lazola Makhupula, Moiz Hassan, Moeti Moleko, Ana Fernández del Río, Ivan Nazarov, Aditya Rastogi, Dexian Tang

TL;DR

The paper addresses improving HIV care in resource-limited settings by empowering Community Health Workers with AI-driven, adaptive interventions. It introduces CHARM, an Android app and CHARM Central integrated with the CF RL platform to collect time-varying data and deliver personalized support, guided by survival-based risk modeling $P(T \ge t \mid X=x)$ using data $(x_j, t_j, c_j)_{j=1}^m$, and bandit-based decisions with rewards $r_t = x_t^\top \theta_{a_t} + \varepsilon_t$. It combines linear and beyond-linear bandits and restless bandits (RMABs) to optimize intervention allocation under resource constraints, with experimental designs including micro-randomized trials. The work is being piloted through a soft launch in 2024 across seven countries, aiming to quantify onboarding, risk profiling, prioritization, and outcome improvements, and is supported by the Gates Foundation and Google.org, under CC BY 4.0.

Abstract

By providing evidence-based clinical decision support, digital tools and electronic health records can revolutionize patient management, especially in resource-poor settings where fewer health workers are available and often need more training. When these tools are integrated with AI, they can offer personalized support and adaptive interventions, effectively connecting community health workers (CHWs) and healthcare facilities. The CHARM (Community Health Access & Resource Management) app is an AI-native mobile app for CHWs. Developed through a joint partnership of Causal Foundry (CF) and mothers2mothers (m2m), CHARM empowers CHWs, mainly local women, by streamlining case management, enhancing learning, and improving communication. This paper details CHARM's development, integration, and upcoming reinforcement learning-based adaptive interventions, all aimed at enhancing health worker engagement, efficiency, and patient outcomes, thereby enhancing CHWs' capabilities and community health.

Optimizing HIV Patient Engagement with Reinforcement Learning in Resource-Limited Settings

TL;DR

The paper addresses improving HIV care in resource-limited settings by empowering Community Health Workers with AI-driven, adaptive interventions. It introduces CHARM, an Android app and CHARM Central integrated with the CF RL platform to collect time-varying data and deliver personalized support, guided by survival-based risk modeling using data , and bandit-based decisions with rewards . It combines linear and beyond-linear bandits and restless bandits (RMABs) to optimize intervention allocation under resource constraints, with experimental designs including micro-randomized trials. The work is being piloted through a soft launch in 2024 across seven countries, aiming to quantify onboarding, risk profiling, prioritization, and outcome improvements, and is supported by the Gates Foundation and Google.org, under CC BY 4.0.

Abstract

By providing evidence-based clinical decision support, digital tools and electronic health records can revolutionize patient management, especially in resource-poor settings where fewer health workers are available and often need more training. When these tools are integrated with AI, they can offer personalized support and adaptive interventions, effectively connecting community health workers (CHWs) and healthcare facilities. The CHARM (Community Health Access & Resource Management) app is an AI-native mobile app for CHWs. Developed through a joint partnership of Causal Foundry (CF) and mothers2mothers (m2m), CHARM empowers CHWs, mainly local women, by streamlining case management, enhancing learning, and improving communication. This paper details CHARM's development, integration, and upcoming reinforcement learning-based adaptive interventions, all aimed at enhancing health worker engagement, efficiency, and patient outcomes, thereby enhancing CHWs' capabilities and community health.
Paper Structure (13 sections, 2 figures)

This paper contains 13 sections, 2 figures.

Table of Contents

  1. Introduction
  2. The CHARM App and CHARM Central
  3. Community Based Digital Health
  4. An AI Platform for Adaptive Intervention
  5. Current Development Status
  6. Methodology
  7. Predictive Modeling
  8. Bandits for Intervention
  9. Linear Bandits
  10. Beyond Linear Bandits
  11. Restless bandits
  12. Experimentation
  13. Summary and Conclusions

Figures (2)

  • Figure 1: Screenshots of the AI-optimized mobile health application CHARM (Community Health Access and Resource Management) from mothers2mother organization.
  • Figure 2: Reinforcement Learning platform to collect and organize data and deliver personalized, just-in-time adaptive interventions for CHWs and patients, based on contextual and restless bandits, with m2m CHARM application with a focus on diagnosis, prevention, and management of HIV, among other diseases.