Patient Satisfaction with an Australian Digital Weight-Loss Service: A Comparative Retrospective Analysis

Abstract

Objective:

This study compared the experiences of patients who received proactive coaching through a digital weight-loss service (DWLS) with those who received reactive coaching.

Methods:

Patients (n = 100) were emailed a mixed-methods survey after 8 weeks of participation in the Juniper DWLS. Reactive and proactive group responses to the survey’s five Likert-scale questions were compared through Mann–Whitney U tests. A Braun and Clarke thematic analysis was conducted on responses to the survey’s open-ended question.

Results:

Proactive group patients were statistically significantly more likely than reactive coaching patients to find their health coach helpful and less likely to feel that their health coach did not communicate with them enough. Generally, patients felt that Juniper health coaching needed to be more personalized and proactive.

Discussion:

This research adds vital foundational knowledge to the scarce literature on real-world comprehensive DWLSs, highlighting the importance of patient preferences around care continuity.

Background

Obesity is a chronic disease affecting roughly 900 million adults worldwide.¹ Among the significant drivers of its rising incidence is the difficulty of accessing quality multidisciplinary care on an ongoing basis, which the disease necessitates.^2,3 To address this problem, many digital weight-loss services (DWLSs) have emerged in recent times.⁴ While it may be generally accepted that such services, in theory, lower access barriers to obesity care,⁵ stakeholders remain skeptical of their quality and safety.^6,7 A key concern is that many DWLSs simply increase access to the much-hyped glucose-like peptide-1 receptor agonist (GLP-1 RA) medications, without providing any follow-up support or lifestyle advice.⁸ This concern aligns with obesity management recommendations from the World Health Organization (WHO) and other major health institutions.^9,10 Some DWLSs, such as Juniper—Australia’s largest DWLS that also operates in Germany, Japan, and the UK,¹¹ claim to offer comprehensive care by combining GLP-1 RA therapy with continuous health coaching and multidisciplinary support. However, lifestyle therapy and care continuity standards are yet to be established for digital GLP-1 RA-supported obesity programs.

Large-scale randomized controlled trials of GLP-1 RAs in weight-loss populations, such as the Semaglutide Treatment Effect in People with Obesity (STEP) and Study of Tirzepatide in Participants With Obesity or Overweight (SURMOUNT) series, have consistently demonstrated the vastly superior weight-loss effect of the medication relative to lifestyle interventions.^12,13 This conclusion appears to persist in the trials that included an intensive lifestyle component as opposed to more standardized diet and exercise counseling.^14,15 Nevertheless, major health stakeholders across industry and academia maintain that lifestyle interventions should underpin any weight management intervention given the uncertainties around GLP-1 RA therapy’s sustainability and the emerging evidence of the post-discontinuation “rebound effect.”^9,16 It is yet to be seen whether real-world obesity programs of any modality adhere to this advice to good effect, let alone those that utilize telehealth technology.

An increasing number of people with complex chronic conditions such as obesity are turning to telehealth solutions. Despite the surge in DWLS uptake over the past few years, research on the modality’s quality and safety remains scarce. To our knowledge, only a handful of studies have been published on real-world GLP-1 RA-supported DWLSs, 3 of which focus on the Juniper program.^17–19 While one of these publications contributed valuable preliminary knowledge on some of the program’s care continuity markers, it contained several limitations, including its inability to compare the markers with other standards.¹⁹ To facilitate the creation of future care continuity standards for DWLSs, deeper and wider insights are required. In accordance with policies and telehealth literature around patient-centered care,^20–22 considerable input needs to come from DWLS users. This study aims to generate foundational patient experience knowledge on the health coaching component of a real-world GLP-1 RA-supported DWLS, Juniper Australia. The Juniper DWLS is a program for women that has been in operation since 2021 and has served over 55,000 patients across Australia, Japan, Germany, and the UK. The study will compare the experiences of Australian Juniper patients who received proactive coaching with those who received reactive coaching.

Methods

Investigators adopted a mixed-methods survey analysis to achieve the aims of the study. During week 8 of their participation in the Juniper DWLS, patients were emailed a 6-question survey containing both open- and closed-ended questions. The analysis compared responses from patients of the two health coaching groups. Bellberry Limited approved the ethics of the study on 22 November, 2023. All participants consented to their data being used in studies submitted for publication.

Doctors follow Ozempic and Saxenda prescribing information guidelines in determining patient eligibility for the Juniper DWLS. Once eligible patients pay their first monthly subscription fee, they are allocated a multidisciplinary team (MDT) consisting of a doctor, registered nurse, and university-qualified health coach (dietetics or nutrition). Under the normal Juniper DWLS, MDTs develop diet and exercise plans based on a patient’s initial health data, which are obtained through medical questionnaires and requested attachments such as photos and blood tests. At the time of writing, the Juniper DWLS has only ever offered GLP-1 RA-supported therapy, i.e., health coaching combined with GLP-1 RA treatment, and thus neither component as standalone therapy.

During the week commencing on 29 January 2024, the Juniper Australia DWLS began to test the potential benefits of an advanced health coaching model. Fifty patients who signed up for the standard weight-loss program that week received proactive health coaching for no additional cost. These patients were selected on a 1:1 allocation ratio, with other new starters receiving the program’s standard reactive health coaching component. All patients were blinded to their coaching group allocation and received the same GLP-1 RA dosing schedule—Ozempic 0.25 mg once a week for 4 weeks, 0.5 mg once a week from weeks 5 to 8, and weekly 1 mg thereafter¹. The reactive coaching group were given a diet and exercise plan based on their initial health data and only received further coaching if they asked questions via the program app. The proactive coaching group received a diet and exercise plan based on their health data and responses to a lifestyle questionnaire, and were engaged by their health coaches at a minimum of every 3 days with personalized advice. All coaching was delivered asynchronously through the Juniper app. No exclusion criteria were applied to the study beyond the those used by doctors to determine program eligibility.

On March 20, 2024, patients from both groups (n = 50 per group) were emailed a 6-question survey specific to their health coaching experience. The first 5 of these questions were 5-point Likert-scale questions, followed by an open-ended question soliciting views on key ways to improve the Juniper health coaching component. The Braun and Clarke thematic analysis method was used to recode and categories responses to the final question.²³ Primary endpoints included percentage distributions across responses to all Likert-scale questions in the two groups and statistical associations between these data and the health coaching method, calculated through Mann–Whitney U tests, and the Vargha and Delaney A (VDA) effect size measure. These statistical tests were selected on account of the evidence that their findings are more accurate than those generated by parametric alternatives for analyses of Likert-type data among samples smaller than 200.²⁴ A thematic analysis of qualitative responses represented the study’s exploratory endpoints.

Results

Responses were obtained from 82 of the 100 patients who were emailed a survey, including 48 (96%) from the proactive health coaching group, and 34 (68%) from patients who received reactive health coaching (Table 1). 79.4% of patients were of Caucasian ethnicity, with mean age = 42.1(±10.6) years and mean body mass index (BMI) = 33.7 kg/m²(±6.1). Nearly three-quarters (71%) of these proactive group patients found their health coach to be helpful to some extent, compared with 42% of patients from the reactive group (Table 2). Consistent with this, a higher percentage of proactive group patients (19%) than reactive group patients (12%) considered the health coaching component to be worth 25% or more of the value of the full Juniper DWLS offering, with the same trend observed in patient willingness to trust their coach’s product recommendations (57% vs. 48%). On the contrary, reactive group patients felt more comfortable sharing vulnerable information with their coach than proactive group patients (33% vs. 31%). Reactive coaching patients were also more likely to express the view that their coach did not communicate with them frequently enough (33% vs. 17%). Two-sample Mann–Whitney U tests revealed a statistically significant difference between the two coaching groups across all five Likert-scale questions (p < 0.001). However, only two of these differences had meaningful effect sizes, albeit small ones: patient views on their coach’s helpfulness (VDA = 0.35) and their satisfaction with their coach’s messaging frequency (VDA = 0.36). Effect sizes across the other three views were all negligible (patient trust of their coach’s product recommendations—VDA = 0.54; their comfort in sharing vulnerable information—VDA = 0.48; and their perceived value of their coaching relative to the whole program—VDA = 0.46).

Table 1.

Baseline Data

	Reactive coaching group	Proactive coaching group	Total
Patients emailed survey—no.	50	50	100
Patients completed survey—no. (% of recipients)	34 (68)	48 (96)	82 (82)
Age in years—mean (SD)	42.6 (±10.4)	41.9 (±10.7)	42.1 (±10.6)
BMI—mean (SD)	33.7 (±6.7)	33.7 (±5.7)	33.7 (±6.1)
Caucasian ethnicity—no. (% of cohort)	26 (76)	39 (81)	65 (79)

SD, standard deviation.

Table 2.

Likert-Scale Responses by Group

Survey question—no. (% of cohort)	Reactive group	Proactive group	Total
How helpful has your coach been for you?
Very helpful	5 (15)	12 (25)	17 (21)
Helpful	4 (27)	19 (46)	23 (28)
Neutral	17 (50)	11 (23)	28 (34)
Useless	3 (9)	1 (2)	4 (5)
Totally useless	0	2 (4)	2 (2)
How comfortable do you feel sharing vulnerable information with your coach?
Very comfortable	6 (18)	4 (8)	10 (12)
Comfortable	10 (29)	23 (48)	33 (40)
Neutral	10 (29)	13 (27)	23 (28)
Uncomfortable	7 (21)	6 (13)	13 (15)
Very uncomfortable	1 (3)	2 (4)	3 (4)
If your coach recommended a product to help you achieve your goals, how likely are you to trust that recommendation?
Highly likely	5 (15)	8 (17)	13 (16)
Likely	11 (32)	19 (40)	30 (37)
Neutral	16 (47)	17 (35)	33 (40)
Unlikely	1 (3)	3 (6)	4 (5)
Highly unlikely	1 (3)	1 (2)	2 (2)
How do you feel about the frequency with which your health coach communicates with you in the Juniper app?
They don’t message me nearly often enough	7 (21)	2 (4)	9 (10)
They don’t message me often enough	4 (12)	6 (13)	10 (12)
They message me at an appropriate frequency	21 (62)	29 (60)	50 (61)
They message me a bit too often	1 (3)	9 (19)	10 (12)
They message me far too often	1 (3)	2 (4)	3 (4)
Out of your entire Juniper experience, how much value does coaching make up?
<5%—I care more about other components of the Juniper program	8 (24)	9 (19)	17 (21)
5–10%—It’s nice to have	9 (26)	13 (27)	22 (27)
11–25%—I value it as part of the whole package	13 (38)	17 (35)	30 (37)
26–50%—It’s a crucial part of my Juniper experience	4 (12)	8 (17)	12 (15)
>50%—I am here mainly for the coaching	0	1 (2)	1 (1)

A thematic analysis of patient views on the health coaching service recoded qualitative data into nine categories (Table 3). A third (33.3%) of proactive group patients felt their health coaching could have been more personalized, with 14.6% and 12.5% desiring more accountability-related contact and more specific dietary advice, respectively. Greater accountability (20.6%) and personalization (14.7%), and more specific dietary advice (8.8%) were also the three most common responses among reactive coaching patients. A Mann–Whtiney U test revealed the difference between the two groups was statistically insignificant (W = 5899, p > 0.05).

Table 3.

Qualitative Views on Improving Juniper Health Coaching by Post-Analysis Categories

How could we improve the health coaching component of the Juniper program?—no. (% of cohort)	Reactive group	Proactive group	Total
More personalization	5 (15)	16 (33)	21 (26)
I’m satisfied with the current program	6 (18)	10 (21)	16 (20)
More accountability encouragement	7 (21)	7 (15)	14 (17)
More specific dietary advice	3 (9)	6 (13)	9 (11)
Unsure/NA	6 (18)	3 (6)	9 (11)
Addition of synchronous consults (F2F/Video/Phone calls)	1 (3)	6 (12)	7 (8)
Health coach not required/desired	2 (6)	3 (6)	4 (5)
Less frequent contact	2 (6)	2 (4)	3 (4)
More frequent contact	1 (3)	2 (4)	3 (4)
More general health guidance	2 (6)	0	2 (2)
Other	1 (3)	1 (2)	2 (2)

Discussion

This study aimed to compare patient satisfaction with different health coaching approaches in a real-world GLP-1 RA-supported DWLS. It found that Juniper DWLS patients who receive more personalized and more frequent lifestyle counseling were more likely to be satisfied with their help they received from health coach than patients whose counseling was more standardized and reactive. The other statistically meaningful discovery was that proactive group patients were less likely than reactive group patients to desire more frequent communication from their health coach. Although the inverse of this scenario is also true, i.e., that proactive group patients were more likely to express discontent with too much messaging from their health coach, this can be reasonably interpreted as a lesser problem than inadequate communication frequency. It is also possible that program satisfaction contributed to the lower survey completion rate among the reactive group. Findings from the thematic analysis indicated that a significant proportion of Juniper DWLS patients, including those who received proactive coaching, desire a higher level of personalization in their coach’s communication and more accountability encouragement.

The study’s limitations included its short assessment period (8 weeks), its predominantly Caucasian female cohort, and its inability to explain the between-group survey completion discrepancy. Despite these limitations, this research adds vital foundational knowledge to the scarce literature on real-world comprehensive DWLSs. As demand for GLP-1 RA-supported DWLSs continues to swell throughout the world, services need to adhere to the WHO recommendations around lifestyle therapy and care continuity. However, in doing so, they need to give strong consideration to patient preferences around the delivery of such program features. Previous telehealth literature has emphasized the importance of soliciting patient preferences in the design and development of novel care models from both a quality of care and patient satisfaction perspective.^22,25 Three earlier studies made various quantitative discoveries around the effectiveness of and adherence to the Juniper DWLS, however, none of these discoveries gave any indication as to the concrete measures the service should take towards improvement.^17–19 This study’s findings suggest that the health coaching component of the Juniper DWLS needs to become more personalized and proactive. Future research should consider similar investigations of other GLP-1 RA-supported DWLSs, along with prospective studies that compare the effectiveness, adherence, and patient satisfaction outcomes of reactive and proactive DWLS coaching over a longer period.

Footnotes

Acknowledgments

The authors would like to thank all patients and clinicians involved in the Juniper weight-loss program over the study period.

Authors’ Contributions

L.T.: Conceptualization, methodology, validation, formal analysis, investigation, data curation, writing—original draft, writing—review and editing. M.V.: Conceptualization, writing—review and editing, validation, project administration. S.W.: Investigation, resources, formal analysis.

Disclosure Statement

L.T., M.V., and S.W. are all paid a salary by Eucalyptus (Juniper parent company).

Funding Information

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Abbreviation Used

References

World Health Organization. Obesity and Overweight. March 1, 2024. Available from: https://www.who.int/news-room/fact-sheets/detail/obesity-and-overweight#:∼:text=In%202022%2C%2043%25%20of%20adults,who%20were%20living%20with%20obesity

Washington

, Johnson

, Kendrick

, et al. Disparities in access and quality of obesity care. Gastroenterol Clin North Am, 2023; 52(2):429–441.

Sarma

, Sockalingam

, Dash

. Obesity as a multisystem disease: Trends in obesity rates and obesity-related complications. Diabetes Obes Metab, 2021; 23 (Suppl 1):3–16.

Irvin

, Madden

, Marshall

, et al. Digital health solutions for weight loss and obesity: A narrative review. Nutrients, 2023; 15(8):1858.

Lei

, Medina Inojosa

, Kumar

, et al. Effectiveness of a weight loss program using digital health in adolescents and preadolescents. Child Obes, 2020; 17(5):311–321.

Fallows

, Ells

, Anand

. Semaglutide and the future of obesity care in the UK. Lancet, 2023; 401(10394):2093–2096.

Johnston

. The role of commercial weight loss programmes. In: Bariatric Surgery in Clinical Practice. In Clinical Practice. ( Haslam

, Malhotra

, Capehorn

. eds.) Springer; 2022.

Tsirtsakis

. Debate over ‘Doctor-Approved’ Weight Loss Drugs Obtained Online. Royal Australian College of General Practitioners. March 30, 2023. Available from: https://www1.racgp.org.au/newsgp/clinical/debate-over-doctor-approved-weight-loss-drugs-obta

World Health Organization. Health Service Framework for Prevention and Management of Obesity. Geneva, 2023.

10.

National Institute for Health and Care Excellence. Semaglutide for managing overweight and obesity; Sep 2023.

11.

Swan

. Australia’s largest online weight-loss business at loggerheads with regulators. Sydney Morning Herald. Mar 18, 2024. Available from: https://www.smh.com.au/technology/australia-s-largest-online-weight-loss-business-at-loggerheads-with-regulators-20240314-p5fcfz.html [Last accessed: April 25, 2023].

12.

Wilding

, Batterham

, Calanna

, et al. STEP 1 Study Group. Once-weekly semaglutide in adults with overweight or obesity. N Engl J Med, 2021; 384(11):989–1002.

13.

Jastreboff

, Aronne

, Ahmad

, et al. SURMOUNT-1 Investigators. Tirzepatide once weekly for the treatment of obesity. N Engl J Med, 2022; 387(3):205–216.

14.

Wadden

, Bailey

, Billings

, et al. STEP 3 Investigators. Effect of subcutaneous Semaglutide vs Placebo as an adjunct to intensive behavioral therapy on body weight in adults with overweight or obesity: The STEP 3 randomized clinical trial. Jama, 2021; 325(14):1403–1413.

15.

Wadden

, Chao

, Machineni

, et al. Tirzepatide after intensive lifestyle intervention in adults with overweight or obesity: The SURMOUNT-3 phase 3 trial. Nat Med, 2023; 29:2909–2918.

16.

Wilding

, Batterham

, Davies

, et al. Weight regain and cardiometabolic effects after withdrawal of semglutide: The STEP 1 trial extension. Diabetes Obes Metab, 2022; 24(8):1553–1564.

17.

Talay

, Alvi

. Digital healthcare solutions to better achieve the weight-loss outcomes expected by payors and patients. Diabetes Obes Metab, 2024; 26(6):2521–2523.

18.

Talay

, Vickers

. Patient adherence to a digital real-world GLP-1 RA-supported weight-loss program in the UK: A retrospective cohort study. J Community Med Public Health, 2024.

19.

Talay

, Vickers

. Effectiveness and care continuity in an app-based, GLP-1 RA-supported weight-loss service for women with overweight and obesity in the UK: A real-world retrospective cohort analysis. Diabetes Obes Metab, 2024; 26(7):2984–2987.

20.

World Health Organization. People-Centered Health Care: A Policy Framework. World Health Organization; 2007.

21.

Australian Commission on Safety and Quality in Health Care. National Safety and Quality Primary and Community Healthcare Standards. Sydney; 2021.

22.

Mason

. The most important telemedicine patient satisfaction dimension: Patient-centered care. Telemedicine and e-Health, 2022; 28(8):1206–1214.

23.

Braun

, Clarke

. Thematic Analysis: A practical guide. Sage Publications; 2021.

24.

Şimşek

. The power and type 1 error of Wilcoxon-Mann–Whitney, Welch’s t, and student’s t tests for Likert-type data. International Journal of Assessment Tools in Education, 2023; 10(1):114–128.

25.

Cranen

, Groothuis-Oudshoorn

, Vollenbroek-Hutten

, et al. Toward patient-centered telerehabilitation design: Understanding chronic pain patients’ preferences for web-based exercise telerehabilitation using a discrete choice experiment. J Med Internet Res, 2024; 19(1):e26.