Authors:

Thomas Curtis¹, Laura Falvey¹, Adrian Brown², Jonathan Nguyen Van Tam³, Hywel Room⁴, Siri Steinmo⁵, Claudia Ashton¹, Dipesh Patel⁶, Jonathan Kwan⁷, Barbara McGowan⁸‍

Affiliations:

‍¹Reset Health Ltd, London, UK ²University College London, London, UK ³University of Nottingham School of Medicine, Nottingham, UK ⁴Colchester Hospital, East Suffolk and North Essex NHS Trust, UK ⁵UCLH, London, UK ⁶Royal Free Hospital, London, UK ⁷Darent Valley Hospital, Kent, UK ⁸Guys & St Thomas Hospital, London, UK‍

Background:

Real-world adherence to incretin-based therapies for obesity typically lags behind clinical trials data, with typical attrition of 30–40% at six months. Despite established efficacy, adverse events (AEs) are considered problematic and service models required to achieve retention in medically-complex patient groups are inadequately specified. We evaluated the impact of a high-engagement digital NHS service (Roczen) on retention, weight, AEs and clinical touchpoints.

Method:

A retrospective analysis was performed on 105 patients enrolled in the digital arm (Roczen) of a specialist obesity service in the NHS (SNEE ICB – ESNEFT Complex Obesity Service). Patients received ongoing clinical support and proactive medication management. Cumulative documented AEs and adherence (confirmed by real-time clinical contact) were assessed for 24 weeks. Patients were stratified into AE groups: low (0–2 events; n=22), medium (3–5 events; n=60) and high (≥6 events; n=23). 24-week retention and weight loss (168±28 days) were compared, overall, and by AE group. Weight-loss and dosing decision analysis was restricted to patients initiating incretin-based therapy with complete follow up data (n=90); retention and digital touchpoints were counted in the intention-to-treat population (n=105).

Results:

The cohort (79.0% female, age 54.4±12.8 years) was characterised by medical complexity (mean body mass index 47.0±10.3 kg/m²; mean comorbidities per patient 7.0±3.4). Treatment consisted of semaglutide (n=41; 39.1%), tirzepatide (n=38; 36.2%), or sequentially switched therapy (n=26; 24.8%). Two patients were excluded for planned surgery.

Overall, the service achieved 96.1% adherence and 10.3±5.3% mean weight loss at six months. Neither were affected by AE frequency. Patients in the high, medium and low AE groups achieved retention rates of 100% (22/22), 94.9% (56/59) and 95.5% (21/22) respectively (p=0.81, Fisher’s exact test [Monte Carlo]). The high AE group achieved a mean weight loss of 10.2±4.5% (n=21), the medium 10.0±5.4% (n=49) and low AE group 11.3±6.2% ([n=20]; p=0.66 [one-way ANOVA]).

Of 612 dosing decisions, 31.1% were personalised deviations from standard escalation or maximal dosing. The mean level of digital touchpoints required in the high AE group was higher (5.7±2.9/week) than in medium or low groups (5.2±3.4, 4.8±3.6, respectively), but did not reach significance (p=0.62 [one-way ANOVA]).

Switching medication on clinical grounds (n=26, 25.2%) did not adversely affect retention (100%).

Conclusion:

‍In a medically-complex state-funded NHS cohort, Roczen was associated with strong adherence to incretin-based therapies across AE groups. We did not observe an expected relationship between AE frequency and attrition. We postulate that high attrition rates often observed in the real-world are not an inevitable consequence of the therapies themselves, but rather related to the level of clinical oversight and cost-based withdrawal. Roczen’s model provides sufficient support to help state-funded patients continue on medication for six months, and has the potential to increase long-term adherence.