Transforming User Engagement: Applying Behavioral Science to Improve Adherence in Chronic Disease Management
Applying behavioral science and user experience research to design and test effective strategies for improving clinical adherence and achieving meaningful health outcomes
Background & The Problem
Ping’an Insurance, an $800 billion company, tasked its digital health department with developing a chronic disorder management app to improve health outcomes, reduce costs, and enhance user engagement. The app was designed to help users track their diet, exercise, medication intake, and overall health, with clinical support integrated into the experience.
However, the app faced a major challenge—low patient adherence and engagement; specifically, the dropout rates were high and daily usage was low.
The Approach
Uncovering the Root Causes of Low User Adherence
Identifying the Dropout Points
To diagnose the problem, I mapped out the full user journey and collaborated with data scientists to analyze conversion rates at each step:
The highest dropout rate occurred at the program's daily check-in stage.
Understanding the Why: User Interviews and Behavioral Insights
To investigate why users dropped out, I conducted in-depth interviews with 20+ users who had discontinued use of the app. Patterns began to emerge:
For Users Who Dropped Out During the Daily Check-In
-
Tracking Diet Was Too Difficult: Many users didn’t understand what "daily salt intake (less than 5g)" meant practically. They lacked tools to measure salt intake and found it impossible to estimate nutrients in shared Chinese dishes.
-
Complexity and Friction: Manually entering meal details, exercise, and medications felt like a chore, discouraging consistent engagement.
-
Guilt and Avoidance: Users who ate unhealthy meals often avoided logging their food intake to escape feelings of guilt.
For Users Who Dropped Out at Earlier Stages
-
App Navigation Issues: Some older users didn't know how to use the App and navigate to the program page
-
Lack of Perceived Urgency: Some users, particularly pre-clinical individuals, didn’t believe their condition was serious enough to require daily monitoring.
Designing and Testing Solutions: Applying Behavioral Science
Drawing on behavioral change models and cognitive behavioral therapy (CBT) techniques, I researched their application in chronic disease management. While most interventions were designed for in-person settings, I explored ways to adapt these methods for a digital platform.
After presenting my findings to leadership, I got the opportunity to lead a cross-functional team (1 product manager, 1 product operations specialist, 2 clinicians, 1 designer, and 1 engineer) to run a pilot to identify and integrate effective behavioral change techniques to improve app adherence.
The Targeted Solutions Included:
We followed an agile model, testing one feature per week before moving on to the next. After four weeks of testing, we finalized the solution: in addition to health education and AI-powered meal recognition, we implemented a community feature and retained incentives through rewards (e.g., free weight scales, free physical examination packages, etc.).
The Impact
1. Improved User Adherence: After implementing AI meal recognition, community features, and gamification, user adherence rose by 40%, with more users consistently tracking their health.
2. Secured Leadership Buy-In and Funding: I was invited to present the research findings, pilot design, and user insights to department leadership and investors, which led to securing $1M in funding for app expansion.
3. ​​Institutionalizing UX Research in Product Development: This project demonstrated the value of user research, leading to its integration into the department’s product development process.
Lessons Learned
1. Iterate, Test, and Adapt: The pilot reinforced the importance of agile experimentation—testing one solution at a time allowed us to refine interventions based on real user behavior.
2. Execution Matters as Much as Insights: A well-researched idea only works if it’s implemented effectively. Balancing strategy, execution, and continuous iteration was key to making a real impact.