Introduction
Adherence to obesity pharmacotherapy is essential for achieving meaningful weight and health benefits, yet remains a challenge in the real world. Discontinuation rates for GLP-1 receptor agonist (GLP-1) therapy range from 40–75% within the first 3–6 months, and were approximately 50% in a previous prospective review of two outpatient obesity clinics in Ireland between 2021 and 2023. Discontinuation reflects a combination of patient, medication-related and clinician or system factors, occurring across different phases of adherence. In the prior study, side effects (36%), limited access to follow-up (24%), and cost (23%) were the most commonly reported reasons for discontinuation.
This study evaluates an AI-based behavioural companion in an Irish obesity clinic to assess early signals of improved medication adherence and platform engagement.
Methods
A total of 129 adults receiving GLP-1 receptor agonist therapy enrolled in a messaging-based digital platform at an Irish obesity clinic between March and December 2025 and used it for ≥3 months. Active users were defined as those who consented and did not pause or discontinue the program. The platform delivered proactive behavioral support and addressed patient questions on medication use, side effects, appetite, nutrition, activity, and expectations. Engagement was defined as ≥2 bidirectional messages per month and categorized as minimal (1), light (<5), regular (6–20), or power user (≥21). Outcomes included three-month self-reported medication adherence, user engagement, treatment discontinuation, message volume, and sentiment, derived using an algorithmic analysis of naturalistic conversation data.
Results
104 out of 129 enrolled patients met criteria for active use. At 3 months, 83% of active users remained adherent to GLP-1 RA therapy. 85% of users were engaged, with over 73% of users classified as regular or power users. Discontinuation of the program was low (1%).
Users exchanged an average of 18–24 messages per active user (from 1 to 163 messages, median 16), with interactions clustering around clinically meaningful micro-events such as dose escalation, side effects, appetite changes, weight plateaus, travel, and stress. Many users re-engaged after 3–6 weeks of inactivity, indicating more frequent touchpoints than standard care. Conversations captured symptom- and appetite-related concerns within 24–48 hours of treatment initiation, highlighting high-risk moments for discontinuation not typically addressed during scheduled visits. Sentiment analysis showed predominantly positive experiences, including improved confidence, motivation, and perceived support.
Conclusion
Use of a WhatsApp-based behavioral support program in an outpatient obesity clinic was associated with early signals of improved short-term adherence, sustained engagement, and low program discontinuation among adults receiving GLP-1 receptor agonist therapy. By addressing event-driven needs and delivering timely, patient-centred support between visits, digital messaging may help close gaps in obesity care; longer-term data is needed to assess durability and long-term outcomes.
Interested in collaborating with us on research? We welcome partnerships with academic institutions and clinical researchers. Reach out to us at [email protected].