Evaluation of a scalable behavioural weight loss programme in adults at risk of developing type 2 diabetes
This WP evaluated the 5-year effectiveness and cost-effectiveness of an existing scalable behavioural weight loss programme in adults who are at risk of T2D because of their high BMI. It also qualitatively explored experiences of weight management after the intervention ended, in order to inform the development of a new weight loss maintenance programme (WP2.1).
Rationale
In the UK NHS, one of the most commonly used weight loss interventions is referral to a commercial open-group BWMP for 12 weeks. In the Weight loss Referrals for Adults in Primary care (WRAP) RCT,21,22 we previously demonstrated that primary care referral to 12 or 52 weeks of a scalable (inter)nationally available weight loss programme (WW) achieved greater weight loss than a brief intervention (a booklet of self-help weight management strategies) over 2 years. Participants allocated to the 52-week WW programme lost more weight than those allocated to the 12-week WW programme. Participants in the 52-week group also had the greatest reductions in glycaemia at 1-year follow-up. Modelled lifetime cost-effectiveness suggested that offering the 52-week programme (rather than the standard 12-week programme) would be cost-effective because of greater reductions in health-related diseases and associated healthcare costs.
The long-term health impact and cost-effectiveness of both the 12- and 52-week programmes depend on assumptions about weight regain.38 This WP evaluated the 5-year outcomes of WRAP, enabling precise measurement of the impact of this scalable (inter)nationally available behavioural weight management intervention on body weight, glycaemia, and incidence of T2D, and facilitating more realistic evaluation of cost-effectiveness, both within trial and over the longer term (WP1.3).
Long-term follow-up of this trial cohort also provided a unique opportunity to qualitatively explore and compare experiences of post-treatment weight management for people who had lost weight during the intervention. Such a comparison was necessary to inform the development of an intervention to support better weight loss maintenance (WP2.1). Prior evidence synthesis of qualitative research on weight loss maintenance had identified that few studies had compared behavioural and cognitive strategies used by people with different weight trajectories.50 Previous studies relied on self-reported weight history to categorise participants and included samples which were less representative of the target population. In this WP, we compared the cognitive and behavioural strategies for weight management used by participants who lost weight as part of the WRAP trial and either maintained or regained their weight loss at 5-year follow-up.
Objectives
- To evaluate the 5-year impact of a scalable behavioural weight loss programme on body weight, glycaemia, diabetes status, modelled cardiovascular risk and resource use in adults at high risk of developing T2D because of overweight or obesity.
- To evaluate the 5-year (within trial) cost-effectiveness of this programme in this population.
- To provide longitudinal data on 5-year weight and related risk factors to inform improved modelling of longer-term outcomes (WP1.3)
- To qualitatively explore and compare experiences of weight loss ‘maintainers’ and ‘regainers’ in order to inform the development of a weight loss maintenance programme (WP2.1).
Five-year effectiveness and cost-effectiveness of a scalable weight loss programme in adults at high risk of type 2 diabetes
The full report addressing objectives 1–3 of this WP was published in Ahern et al.51 and was awarded the RCGP Paper of the Year 2023 in the Category of Health Services Research (including Public Health and Implementation).
Methods
The Weight loss Referrals for Adults in Primary care (WRAP) trial is a parallel-group non-blinded RCT in which adults living with overweight or obesity (BMI ≥ 28 kg/m2) were randomised to one of three interventions (brief intervention, 12-week WW, 52-week WW) in an uneven (2 : 5 : 5) allocation and followed for 2 years, with assessments at 3 months, 1 year and 2 years. Participants (N = 1267, mean BMI 34.5 kg/m², mean age 53.2 years, 68% female, 90% White) were recruited between October 2012 and February 2014 from 23 primary care practices across England, and over half of these were in areas with an Index of Multiple Deprivation (IMD) above the national median. Primary care providers identified eligible patients by searching medical records and invited them to participate via letter. The full protocol and 2-year outcomes of the WRAP trial were published previously.21,22
Between 5 February 2018 and 3 August 2019, we collected 5-year outcome data for WRAP participants via measurement visits at local primary care practices or a research centre, self-report questionnaires, and review of primary care electronic medical notes. Five-year change in weight was the primary outcome and secondary outcomes included changes in HbA1c, lipid profile, blood pressure and diabetes status. Self-reported health-related quality of life and healthcare resource use were measured to assess cost-effectiveness.
Full details of the protocol and statistical analysis plan were pre-registered (ISRCTN64986150). We used analysis of covariance (adjusted for baseline value and randomisation stratifiers) to estimate between-group differences in continuous clinical outcomes (52-week programme vs. 12-week programme, 12-week programme vs. brief intervention and 52-week programme vs. brief intervention). For the primary outcome, we examined potential interactions between intervention and gender, education and diabetes status.
We undertook a within-trial cost-effectiveness analysis that compared health outcomes and NHS costs incurred in each intervention group between baseline and 5-year follow-up. Full details of these methods are described on pages 5–8 of the supplementary material for the main trial paper.22
Key findings
We ascertained the primary outcome (weight) for 871 (69%) of 1267 participants, with a mean duration of follow-up of 5.1 years [standard deviation (SD) 0.3]. Participants with data at follow-up were broadly similar to those with missing data, but had a slightly higher mean age and lower mean weight.
Figure 2 shows mean weight change over 5 years by intervention group. At 5 years, participants in all three groups had regained some of the weight they had lost during the initial 2-year trial. On average, participants randomised to the two WW programmes were still around 2 kg below baseline weight (5-year weight change was−1.95 (9.55) kg in the 12-week programme group; −2.67 (9.81) kg in the 52-week programme], whereas those in the brief intervention group were < 0.5 kg below baseline [−0.46 (SD 8.31) kg]. However, although pairwise comparisons favoured the WW programmes, confidence intervals (CIs) were wide and crossed zero. The adjusted difference between groups was −1.76 (95% CI −3.68 to 0.17) kg between the 52-week programme and the brief intervention; –0.96 (−2.90 to 0.97) kg between the 12-week programme and the brief intervention, and −0·80 (−2.13 to 0.54) kg between the 52-week and the 12-week programme. While the consistency of effects over time and the dose-response effect suggest that differences between groups are not due to chance, we could not rule this out. We found no evidence that the effect of the intervention on weight was moderated by gender, education or diabetes status.
FIGURE 2 Weight change over 5 years in the WRAP trial. Data are mean of all measured weights at each time point (standard error). Reproduced with permission from Ahern et al.51 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) licence, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure above includes minor additions and formatting changes to the original text. (Figure not reproduced here; see the original article.)
Participants who attended the 52-week WW programme had greater odds of clinically significant weight loss (≥ 5%) at 5 years than brief intervention participants [odds ratio (OR) 1.79 (95% CI 1.13 to 2.83)] but other comparisons between groups had wide confidence intervals that crossed 0. We did not find evidence of between group differences in HbA1c, fat mass, lipid profile, blood pressure or development of T2D.
During the trial, per person NHS and social care costs and changes in quality-adjusted life-years (QALYs) were very similar across groups, but point estimates suggested that the 12-week programme incurred the lowest cost and produced the highest QALYs (Table 1). The uncertainty around this is explored in the appendix to the published paper (pp. 9–24).
| Total discounted NHS and social care costs per person (£) | Total discounted QALYs per person | Incremental costs (£) | Incremental QALYs | ICER vs. next most effective option (£) | Expected net monetary benefit (£)a | Incremental expected net monetary benefit (£) | Rank on expected net monetary benefit ¥ | |
|---|---|---|---|---|---|---|---|---|
| Brief intervention | 4118 (3654–4622) | 3.82 (3.74–3.91) | – | – | Dominated | 72,332 (70,410–74,285) | – | 2 |
| 12-week programme | 3987 (3661–4322) | 3.82 (3.77–3.87) | −130.74 (−754.85 to 427.92) | −0.00 (−0.11 to 0.1) | Estimated as most cost-effective | 72,415 (71,132–73,616) | 83.16 | 1 |
| 52-week programme | 4159 (3816–4549) | 3.79 (3.73–3.85) | 40.64 (−572.23 to 637.55) | −0.03 (0.14 to 0.07) | Dominated | 71,675 (70,268–72,967 | −656.51 | 3 |
Qualitative comparison of experiences of weight loss maintainers and regainers following a scalable behavioural weight loss programme
A paper addressing the aim to qualitatively explore and compare experiences of weight loss ‘maintainers’ and ‘regainers’ in order to inform the development of a weight loss maintenance programme, was reported in detail in Lawlor et al.52
Methods
Following the 5-year follow-up assessment, we conducted in-person interviews with 26 participants (58% female) who had achieved ‘clinically significant’ (≥ 5%) weight loss in year 1 and either maintained their lower weight at 5 years (N = 10) or regained weight (N = 16). Participants were purposively sampled for an even split of ‘maintainers’ and ‘regainers’ and a range of demographic characteristics. Interviews explored participants’ experiences of weight management after the weight loss programme ended. Reflexive thematic analysis explored differences between maintainers and regainers in the cognitive and behavioural strategies used to manage their weight and to overcome lapses and prevent relapse. Analysis was supported by a patient and public involvement (PPI) coauthor and early findings were discussed with our patient user group to support interpretation and theme development.
Key findings
We identified that ‘maintainers’ used more self-regulation techniques to manage their weight than ‘regainers’. They continued to pay close attention to their diet, were more likely to anticipate lapses (e.g. indulging during social occasions), and made conscious plans to avoid or compensate for theses lapses and continue their weight management efforts. Although regainers also tried to self-regulate their behaviours, they were more relaxed about self-monitoring and did not anticipate or plan for lapses. Regainers found distraction techniques to be ineffective and struggled to navigate the role that food plays within interpersonal relationships.
This study identified that an intervention to support weight loss maintenance should support development of skills in self-monitoring, planning for high-risk situations and managing interpersonal relationships in relation to food. These findings influenced the design of our intervention to Support Weight loss Maintenance (SWiM; WP2.1)
Strengths and limitations
Weight loss Referrals for Adults in Primary care is the first RCT of a commonly commissioned scalable behavioural weight loss programme to collect objectively measured 5-year outcome data, including weight, glycaemia and diabetes status. It informs decision-makers about the value for money of such interventions and offers a valuable data set for future research. Retention was good relative to similar studies, with 69% follow-up, but some outcomes were self-reported. The trial was designed to detect moderate differences in weight at 12 months and was not powered to identify small differences at 5 years, so results should be interpreted cautiously. This study did not identify moderation of intervention effects by demographic characteristics, but our subsequent individual participant data meta-analysis of UK behavioural weight loss trials53 (which have greater power to detect intervention moderation effects) found a larger effect of behavioural interventions on weight in men (relative to women) and in those of White ethnicity (relative to people from minority ethnic groups), so this finding should be interpreted with caution.
This is the first study to qualitatively compare the experiences of maintainers and regainers 5 years after starting a BWMP. We purposively sampled a demographically diverse sample, including a high proportion of men, people from minority ethnic groups and a spread of socioeconomic and geographic backgrounds. While duration of follow-up is a key strength, it increases the likelihood of recall bias. Weight measurements during the trial facilitated objective categorisation of maintainers and regainers; however, binary categorisation means some participants with similar weight trajectories might be in different groups.
Conclusions
Weight regain following a commercial BWMP has previously been overestimated. While differences between groups in weight loss, glycaemia and cardiovascular risk factors were not significant at 5 years, there is evidence that some weight loss is sustained. Both interventions were cost-effective within 5 years, with some suggestive evidence (but considerable uncertainty) that the 12-week programme has better return on investment within the 5-year time horizon.
Many people do regain weight following a behavioural programme and could benefit from additional support to maintain weight and maximise the benefits of weight loss. Psychological challenges of maintaining weight were recognised as a key target for intervention. We identified some cognitive and behavioural strategies that support better long-term weight management and should be included in an intervention to support weight loss maintenance following a behavioural programme.
Interrelation with other work packages
The data on 5-year metabolic trajectories (BMI, systolic blood pressure, total cholesterol and HbA1c profile) were used to improve the estimation of long-term weight trajectories used in the lifetime perspective health economic modelling of the long-term impact on incidence and complications of T2D (WP1.3).
Key cognitive and behavioural skills and strategies identified as supporting post-treatment weight loss maintenance informed the development of our weight loss maintenance programme (WP2.1).