Effectiveness and care continuity in an app-based, glucagon-like peptide-1 receptor agonist-supported weight-loss service for women with overweight and obesity in the UK: A real-world retrospective cohort analysis
1 BACKGROUND
Many digital weight-loss services (DWLSs) have emerged throughout the world in recent years to combat rising obesity rates.1 A key rationale behind digital modalities is that they increase care access, which is a considerable barrier to the treatment of chronic conditions such as obesity, for which patients need to consult with a coordinated multidisciplinary care team (MDT) on an ongoing basis.2, 3 However, influential stakeholders remain sceptical about the quality of many DWLSs, raising regular concerns about providers who prescribe glucagon-like peptide-1 receptor agonists (GLP-1 RAs) without providing any follow-up therapy.4-6
While it may be true that certain DWLSs offer substandard care continuity, many others appear to deliver more comprehensive obesity care. Programs of this description, such as the Juniper women's weight-loss service, typically combine GLP-1 RA therapy with personalized diet and exercise coaching. A 2024 study7 found weight-loss outcomes in a Juniper Australia cohort to be superior to those observed in the SCALE clinical trial8 and another real-world intervention (ADDRESS),9 in which patients followed the same GLP-1 RA schedule, yet with standardized and unspecified health counselling, respectively. However, the Juniper cohort's high drop-out rate was a major caveat of this comparison, as were the ADDRESS (non-reimbursed) cohort's small sample size (n = 63) and the SCALE study's fully subsidized clinical trial setting. Because of the scarcity of literature on real-world GLP-1 RA-supported programmes and the methodological variation within it, comparisons between comprehensive obesity services and those centred around GLP-1 RA therapy remain unreliable. Nevertheless, unless a substantial body of counterevidence emerges, DWLSs that utilize GLP-1 RAs should follow the National Institute for Health and Care Excellence (NICE) and World Health Organization (WHO) obesity management recommendations of maintaining lifestyle therapy as the centrepiece of any weight-loss intervention.10, 11
An additional benefit of digital modalities is that they facilitate the integration of data automation and analytics tools to enhance the precision and efficiency of their care models.12-14 These data coordination features could feasibly improve communication, clinician tracking and patient adherence, which, when combined with comprehensive treatment, should generate meaningful weight-loss outcomes. However, despite the widespread acknowledgement of the importance of care continuity in the treatment of obesity, there does not appear to be evidence available on the extent to which its principles are implemented in real-world settings, let alone those that ostensibly benefit from digitization. This study aims to initiate the exploration of this knowledge gap by analysing several care continuity markers in the Juniper UK weight-loss programme, along with the programme's effectiveness. Although the study cannot reliably compare these outcomes with those of face-to-face programmes or GLP-1 RA-focused services, its foundational care continuity data and rare effectiveness data will be of significant value to the fledgling field of real-world DWLS research.
2 METHODS
The study analysed a retrospective cohort of 8276 UK-based patients who commenced Juniper weight-management treatment between 28 April 2022 and 1 April 2023. Juniper doctors follow NICE guidelines on ‘Semaglutide for managing overweight and obesity’ in determining programme eligibility.10 To be included in the final analysis, patients were required to complete the first follow-up questionnaire between 140 and 170 days after programme initiation and to have received a minimum of six semaglutide orders by the completion of this questionnaire. Exclusion criteria included weight loss or gain of >30% from baseline and doses >1 mg of weekly semaglutide (Ozempic). Across the study period, the cost of the Juniper weight-loss programme ranged between 195 and 249 British pounds per month, with the higher costs occurring during periods of Ozempic supply constraints. The study was approved by a Human Ethics Research Committee (Bellberry) in November 2023.
All Juniper UK patients were allocated an MDT consisting of a prescribing doctor, a health coach and a medical support officer. MDTs guided patients through personalized diet and exercise programmes and semaglutide therapy, communicating via the Juniper in-app chat feature or email. Patients were able to modify their diet and exercise plan in consultation with their health coach at any stage of their care journey. MDTs are required to check in with patients once a month and use their professional discretion to determine the frequency of additional communication. The follow-up questionnaire contains 15 standard questions pertaining to patient experience, side effects and weight and often includes additional questions if MDTs seek further information. Patients measure weight with a standardized set of Juniper scales and report weight data themselves. Follow-up questionnaires are not considered complete until answers to all questions are provided.
The study's coprimary endpoints were mean weight loss from baseline and mean monthly patient-MDT communications. Secondary endpoints included the proportion of patients who achieved 5%, 10% and 15% weight loss milestones from baseline, mean body mass index (BMI) change from baseline, mean maximum days without contact from patient MDT and mean response speed to patient-reported adverse events. Study data were extracted from Juniper's central data repository on Metabase, which houses all patient-clinician interactions and patient-reported weight data from the completion of an initial questionnaire onwards.
3 RESULTS
Of the 8276 patients who were eligible for this study, 6361 failed to complete the follow-up questionnaire, and a further 33 provided erroneous data (over 30% weight loss), leaving 1882 (22.7%) patients for inclusion in the final analysis (see Figure 1). These patients were predominantly White (85%) and had a mean baseline BMI of 34.6 kg/m2 (see Table S1 for all baseline data).
In the final analysis cohort, follow-up questionnaires were completed at an average of 153.84 (±6.66) days after programme initiation (Table 1). The mean weight loss at this point was 10.73%. Regarding milestones, 82.36% of the programme's adherers lost ≥5% of their baseline weight, 52.07% lost ≥10%, and 23.22% lost at least 15%. The median BMI loss was 3.18 kg/m2.
| Coprimary endpoints, mean (SD) | |
| % Change in body weight | 10.73 (±6.62) |
| Messages per month | 5 (±2.68) |
| Secondary endpoints, mean (SD)/median | |
| Maximum period without MDT contact | 33.87 (±28.34) days/21 days |
| MDT response time to normal AEs | 19.05 (±25.63) hours/8.17 h |
| MDT response time to urgent AEs | 22.27 (±31.59) hours/14.3 h |
| Side effect incidents reported at follow-up, na | |
| Gastrointestinal issues | 1265 |
| Fatigue or dizziness | 426 |
| Headaches | 227 |
| Other | 193 |
| Mood changes | 68 |
| Increased heart rate | 59 |
| Total patients with any side effects (% of cohort) | 840 (44.5) |
| Other data | |
| Days from baseline to follow-up, mean (SD) | 153.84 (±6.66) |
| Total number of normal AEsb | 227 |
| Total number of urgent AEs | 116 |
- a Side effects were reported retrospectively by patients during their follow-up consultation.
- b AEs reported by patients at the time of the event, along with their severity ratings.
- Abbreviations: AEs, adverse events; MDT, multidisciplinary care team; SD, standard deviation.
For the study's other primary endpoint, patients were found to have received a mean of 5.0 messages per month from their MDTs. Corresponding secondary endpoints observed a mean maximum period without MDT contact of 33.87 (±28.34) days and a mean response time to patient-reported adverse events of normal severity of 19.05 h and urgent severity of 22.27 h. It should be mentioned that the distributions of the latter three data points were highly skewed, and therefore median scores are probably more informative. The median maximum period without MDT contact was 21 days, while the median response to normal adverse events was 8.17 h and 14.30 h to urgent adverse events. In total, patients reported 227 adverse events of normal severity and 116 of urgent severity over the study period. A Pearson test revealed that weight loss was not significantly associated with monthly message volume [r (1880) = 0.04, p > .1], and a Spearman test found that the relationship between weight loss and maximum period without MDT contact was not statistically significant [r (1872) = 0.03, p > .1].
4 CONCLUSIONS
To the knowledge of the authors, this is the first study to investigate both the effectiveness and care continuity levels of a real-world DWLS. Our analysis found that over three-quarters (77.3%) of patients failed to meet study inclusion criteria, indicating a low programme adherence rate. Although mean 5-month weight loss (10.73% from baseline) and the proportion of patients who lost a meaningful amount of weight (5%) at this point (82.36%) were substantial for the 22.7% of patients who were included in the final analysis, the high attrition rate is the most salient conclusion that can be drawn about the effectiveness of the Juniper UK service. These findings are relatively consistent with those from a study on a Juniper Australia cohort (using liraglutide rather than semaglutide) earlier this year, which observed an attrition rate of 94% and mean weight loss of 11.6% at 32 weeks (approximately 7.4 months) among the 6% of patients who satisfied study criteria.
Weight loss in the Juniper UK cohort was not found to have been significantly impacted by mean monthly messages or the mean maximum period without MDT contact. Although the observed mean of 5 messages per month over 5 months appears to adhere to an acceptable standard of care continuity, its significance is difficult to assess given the scarcity of comparable data across the obesity management sector. Moreover, the figure is no doubt inflated by the higher volume of messages Juniper UK patients receive during their onboarding process, as indicated by the maximum period without MDT contact results. It could be argued that the observed median score for this measure (21 days) appears to adhere to a high standard of care continuity; however, the fact the data were so skewed (mean = 33.87 days) suggests a problematic inconsistency. An even greater concern is that the mean and median response times to normal adverse events (19.05 h; 8 h) were nearly half those recorded for urgent adverse events (22.27 h; 14.3 h).
The study had several limitations. First, the cohort was predominantly White and female, and thus largely unrepresentative of the British population. Secondly, the study's care continuity data would have been enriched with a sub-analysis of communication type, for example, a distinction between automated reminders and personalized questions. Thirdly, weight data were patient-reported. Finally, and most significantly, only 22.7% of the initial cohort satisfied the study criteria to be included in the final analysis, which could not be compared with a full analysis set as weight-loss data from earlier points were unavailable. As such, the final results only reflect a cohort of patients who adhered to the programme treatment as intended for 5 months. The limitation was arguably compounded by the study's inability to explain the cohort's high drop-out rate. Some probable factors include programme cost, service dissatisfaction and side effect intolerance. It is also feasible that a significant percentage of patients paused their subscription to accommodate other commitments or rescheduled their follow-up consultation outside the study window inclusion criterion. Some patients may have even achieved their weight-loss goals in under 5 months and decided (against common advice) to end treatment at this point. It is worth mentioning that high drop-out rates are common in the scarce literature on real-world GLP-1 RA-supported weight-loss interventions15, 16 and have even been observed to a lesser degree in subsidized, highly controlled clinical trials.17 Nevertheless, an informed explanation of patient discontinuation in real-world, comprehensive DWLSs such as the Juniper UK programme can only come from a dedicated programme adherence analysis, which should become a priority in future DWLS research initiatives.
Despite its shortcomings, this study added to the scarce literature on DWLS effectiveness and laid an important foundation for further research on care continuity across all obesity care models. It found that most patients in a real-world DWLS seemed to be unable to adhere to the service for long enough to derive its potential benefits. Although adherers experienced good weight-loss outcomes and a somewhat reasonable standard of MDT engagement, it is possible that the 77.3% of patients who dropped out of the study neither lost a significant amount of weight nor received enough care from their MDT. Moreover, as no statistical association was detected between weight loss and the various care continuity markers among adherers, none of the weight loss observed in these patients can be attributed to anything in the Juniper DWLS beyond its GLP-1 RA component. The study also discovered that the service's MDTs typically took longer to respond to urgent adverse events [22.27 (±31.59) h] than those of ‘normal’ severity [19.05 (±25.63) h] and that all care continuity data were asymmetrical, suggesting tighter clinical governance controls are needed. Ultimately, if GLP-1 RA-supported DWLSs seek to be endorsed by reputable health institutions as reliable obesity care modalities, they need to deliver substantially better adherence and care continuity outcomes than those observed in this study.
Open Research
PEER REVIEW
The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/dom.15607.
The authors would like to thank all patients and clinicians involved in the Juniper weight-loss program over the study period. Open access publishing facilitated by The University of Sydney, as part of the Wiley - The University of Sydney agreement via the Council of Australian University Librarians.
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
LT and MV are paid a salary by Eucalyptus (Juniper parent company).
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.
