Engagement and behaviour modification of patients with atrial fibrillation for improved health outcomes

Start Date

15 May 2020

End Date

31 March 2026

Project Participants

Tags: cardiac care, chronic disease management, conversational agents, hospital services, outpatient clinics, targeted health information, treatment adherence

Opportunity

Atrial fibrillation (AF) is associated with increased risk of stroke, heart failure, reduced quality of life, and increased health service use. This project aimed to implement and evaluate a multitouch digital support program for patients with AF involving the use of interactive voice response technology, text messages, and a web-based education portal. It was envisaged that the digital support program would ensure patient compliance with treatment plans, including medication adherence, attendance at regular office visits, timely completion of assessments and testing, and decreased Emergency Department (ED) utilisation.

Objectives

This project was a proof-of-concept study to investigate the feasibility and effectiveness of applying an automated, multi-touch digital outreach system to support the care of patients diagnosed with atrial fibrillation: the Conversational Health support in Atrial Fibrillation Self-Management (CHAT-AF-S) intervention. It explored whether voice-based conversational artificial intelligence (AI) could provide scalable education, self-management support, and monitoring for people with AF, assessing implementation outcomes including engagement, satisfaction, feasibility, user experience, barriers, and enablers to use. This would provide Australian evidence responding to growing interest in digital tools that can support patients in chronic disease care outside traditional clinical encounters.

Publications

Final Outcomes

CHAT-AF-S delivered a three-month automated intervention comprising of:

six fortnightly AI-powered voice-based outreach calls
educational text messages with AF-related information and links to resources
automated alerts to the research team where follow-up was needed.

The intervention was evaluated in a multi-site randomised controlled trial at two Sydney hospitals. While final analyses of the primary clinical outcome (AF‑related quality of life) are ongoing, the evaluation showed strong uptake and engagement with the technology-enabled health program. Most participants answered the automated calls, and satisfaction ratings were high across the intervention period.

Impact and Benefits

The project has generated evidence that automated, voice-based conversational AI can be an acceptable and engaging way to deliver patient education and support AF self-management in a real-world clinical population. The findings highlight potential for similar approaches in other chronic conditions requiring ongoing education, behavioural support and monitoring. Key learnings can inform future digital health design and implementation, including opportunities to improve conversational flow, multilingual delivery, speech recognition and hybrid digital-human support models.