Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Co-Pilot for Health: Personalized Algorithmic AI Nudging to Improve Health Outcomes

Jodi Chiam, Aloysius Lim, Cheryl Nott, Nicholas Mark, Ankur Teredesai, Sunil Shinde

TL;DR

This study addresses scalable, personalized health nudging by integrating a Graph-Neural Network–based recommendation system with Singapore's national health programs and wearable data. The NudgeRank platform delivers daily, context-aware nudges, achieving statistically significant increases in daily steps and MVPA over a 12-week period among 84,764 participants, with larger effects in those not already highly active. Engagement metrics show moderate open rates and a dose-dependent relationship between nudges opened and activity gains, supporting the approach's feasibility and potential population impact. Limitations include reliance on app data uploads and a relatively short duration; longer-term trials are needed to link nudging to measurable health outcomes.

Abstract

The ability to shape health behaviors of large populations automatically, across wearable types and disease conditions at scale has tremendous potential to improve global health outcomes. We designed and implemented an AI driven platform for digital algorithmic nudging, enabled by a Graph-Neural Network (GNN) based Recommendation System, and granular health behavior data from wearable fitness devices. Here we describe the efficacy results of this platform with its capabilities of personalized and contextual nudging to $n=84,764$ individuals over a 12-week period in Singapore. We statistically validated that participants in the target group who received such AI optimized daily nudges increased daily physical activity like step count by 6.17% ($p = 3.09\times10^{-4}$) and weekly minutes of Moderate to Vigorous Physical Activity (MVPA) by 7.61% ($p = 1.16\times10^{-2}$), compared to matched participants in control group who did not receive any nudges. Further, such nudges were very well received, with a 13.1% of nudges sent being opened (open rate), and 11.7% of the opened nudges rated useful compared to 1.9% rated as not useful thereby demonstrating significant improvement in population level engagement metrics.

Co-Pilot for Health: Personalized Algorithmic AI Nudging to Improve Health Outcomes

TL;DR

This study addresses scalable, personalized health nudging by integrating a Graph-Neural Network–based recommendation system with Singapore's national health programs and wearable data. The NudgeRank platform delivers daily, context-aware nudges, achieving statistically significant increases in daily steps and MVPA over a 12-week period among 84,764 participants, with larger effects in those not already highly active. Engagement metrics show moderate open rates and a dose-dependent relationship between nudges opened and activity gains, supporting the approach's feasibility and potential population impact. Limitations include reliance on app data uploads and a relatively short duration; longer-term trials are needed to link nudging to measurable health outcomes.

Abstract

The ability to shape health behaviors of large populations automatically, across wearable types and disease conditions at scale has tremendous potential to improve global health outcomes. We designed and implemented an AI driven platform for digital algorithmic nudging, enabled by a Graph-Neural Network (GNN) based Recommendation System, and granular health behavior data from wearable fitness devices. Here we describe the efficacy results of this platform with its capabilities of personalized and contextual nudging to individuals over a 12-week period in Singapore. We statistically validated that participants in the target group who received such AI optimized daily nudges increased daily physical activity like step count by 6.17% () and weekly minutes of Moderate to Vigorous Physical Activity (MVPA) by 7.61% (), compared to matched participants in control group who did not receive any nudges. Further, such nudges were very well received, with a 13.1% of nudges sent being opened (open rate), and 11.7% of the opened nudges rated useful compared to 1.9% rated as not useful thereby demonstrating significant improvement in population level engagement metrics.
Paper Structure (13 sections, 2 figures, 14 tables)

This paper contains 13 sections, 2 figures, 14 tables.

Table of Contents

  1. Introduction
  2. Results
  3. Effect of Nudges on Physical Activity Behavior
  4. Nudge Engagement
  5. Discussion
  6. Methods
  7. Algorithmic AI Nudging System
  8. Study Design
  9. Code Availability
  10. Data Availability
  11. Competing Interests
  12. Author Contributions
  13. Extended Data

Figures (2)

  • Figure 1: Effect of daily personalized nudges on steps and MVPA, nudge open rates and ratings, and dose-response relationship between nudges opened and physical activity.
  • Figure 2: A. Schema of the NudgeRank™ system to deliver personalized algorithmic nudges to participants. B. Example of the knowledge graph used to select nudges from the Nudge Library. See Methods section for more details.