The Re:Mission study provides one of the first and largest real-world evaluations of a T2D treatment programme delivered through TDR and behaviour change support using three different commercially provided delivery models. While each of the threaded publications provides a detailed discussion and contextualisation of the reported findings within the current evidence base, here we describe the key overarching findings from the Re:Mission study within the context of the wider literature.
Currently the implementation and delivery of specific TDR remission-focused interventions as part of routine care remain largely under-researched. Therefore, our findings on the effective mobilisation and delivery of the LCD programme provide important new insights. We found that barriers and facilitators to the programme mobilisation aligned across NHS LHS leads and primary care referral staff, who acknowledged the significant impact of the COVID-19 pandemic on resources and staff turnover, a finding which is supported by wider evidence on the impact of COVID-19-related burnout within NHS staff.22,23 Work with NHS locality staff across WP2 and WP5 also found that local adaptations to referral strategies (such as utilising local population sociodemographic data) were necessary to ensure equitable uptake, and referrals that reflect the local target population, although we were unable to identify a recognised approach or strategy across areas. WP2 also identified a need for ongoing surveillance of referrals, population uptake and outcomes beyond the mobilisation stage to support the continued development and adaptation of the programme and response to local needs. In some areas, a diabetes strategy group of multiple stakeholders existed to support this. Despite an ongoing debate about the use of targeted and universal strategies to address health inequities,24 proportionate universalism is an example of a policy approach or strategy that would be appropriate for tackling the social gradient in health.
When assessing programme delivery, we identified variability in both the theoretical underpinning of service delivery and a drift in the implementation of fidelity. Interestingly, the provider with the strongest theoretical underpinnings was also found to have the strongest fidelity in their BCT content, while the two providers with the weakest theoretical underpinnings also had the weakest fidelity in their BCT content. This supported the notion that unclear theoretical underpinnings might result in a drift in programme fidelity.25–27 However, far less drift in the fidelity of the behaviour change was observed for the digital delivery. These findings reflect a similar pattern to the behaviour change fidelity assessment undertaken for the National Diabetes Prevention Programme group and digital delivery (the DIPLOMA study),27 which is perhaps not surprising given the overlap in SPs across both programmes. This is important as loss of fidelity might impact participants’ ability to initiate and sustain behavioural changes and improve their T2D. A recent meta-analysis conducted by our research group identified eight BCTs individually and significantly associated with weight loss following a LCD (RA5). Although the majority of these BCTs were included in the NHS-LCD programme designs, they were not always observed during programme delivery. If BCTs are associated with weight outcomes, it is important that they are delivered with fidelity in order to optimise intervention effects.
When challenges to programme delivery were examined across WP2 and WP3, there was alignment in the findings between SPs and users in the identification of barriers to the programme uptake and engagement. Key learning aligns with the evidence base in demonstrating the importance of patient (service user)-reported outcomes,28 person-centred care29 and effective peer support.30 Ongoing support needs for psychological challenges were also documented. Although baseline assessment (RA12) of well-being scores in service users who completed the evaluation survey reflected the UK population norms,31 quality-of-life scores were slightly lower than UK population norms.32 This is an important finding given the DiRECT study33 found that better baseline quality of life was a predictor of remission at 12 and 24 months, and re-emphasised the need for robust post-intervention follow-up. Unfortunately, a lack of longitudinal survey data prohibited a robust analysis of changes in service users’ mental health status over time; however, a recent systematic review with meta-analysis found no adverse effects of TDR programmes on any mental well-being subdomain.34 Emotional and disordered eating were also highlighted as an area of concern across the interview and survey data collected. Although we were unable to assess changes in emotional or disordered eating behaviours over time due to insufficient longitudinal data, the baseline prevalence of potential BED diagnosis observed in our survey sample (24%) was over double the rate (11%) reported in the Look AHEAD trial.35 However, systematic review evidence suggests that severe dietary energy restriction (such as TDR) appears safe and beneficial for people with pre-existing BED under close clinical supervision.36 This is an important consideration for referral and service delivery staff, as although patients with a diagnosed eating disorder are excluded from the programme, the prevalence of BED is higher in people living with T2D but remains largely under diagnosed.37
It is important to acknowledge the impact of the programme in achieving weight loss, and improving glycaemic control for many service users, although the T2D remission rates were lower than those reported in the DiRECT trial.38 This may be reflective of the anticipated dilution effect when trial evidence is translated into real-world practice,39 but may also have been impacted by the service users who started and had time to complete the programme with two HbA1c measurements < 48 mmol/mol but were not considered to have achieved remission as they remained on glucose-lowering medication. Nonetheless, the ability to achieve remission has significant benefits, as demonstrated in the recent 12-year data from the Look AHEAD study, which reported a 33% reduction in chronic kidney disease and a 40% lower rate of composite cardiovascular disease measures in patients with evidence of remission.40 This is in addition to 5-year data from the DiRECT trial which reported 47% fewer serious adverse events in the intervention group compared to the usual care control, driven by reductions in infections and new cancer diagnoses.20 However, the fact that only 36% of those who were referred and 55% who started the programme completed it suggests this approach may not be acceptable to everyone. MDS data analysed by NHSE also suggest some sociodemographic inequities in uptake and outcomes, which correlates with similar patterns observed in the first cohort analysis of the NHS National Diabetes Prevention Programme.41
Although the DiRECT trial was predominantly based on White British adults, recent trial data from Qatar demonstrated 61% remission rates in a cohort of Middle Eastern and North African participants11 following a similar LCD programme. A further real-world service evaluation42 demonstrated 78% completion rates, and 66% remission rates within an ethnically diverse UK population. The recent SouTh AsiaN DiaBetes remission feasibility studY (STANDby) trial8 also assessed the impact and feasibility of TDR-led weight loss in UK-based South Asian adults and demonstrated acceptability of this approach. These findings suggest a need for improved cultural competency of the existing NHS programme, which was a theme drawn out from this evaluation study, with recommendations including culturally tailored resources and language-specific groups, enhanced coach training and a wider range of more culturally tailored TDR products. These findings align with a recent participatory research study among South Asian adults with a lived experience of T2D in Manchester UK, which concluded that TDR in the form of soups and shakes was not acceptable, with a preference for a more culturally tailored low-energy food-based diet.43 They also align with a systematic review undertaken by the research team (RA18), which concluded that weight management programmes for T2D need to be sensitive to the needs of the population group, and implementable within the community. Recommendations from the review also included consideration of social, habitual, economic and conceptual factors, which include consideration of local ethnic and cultural norms, and community relationships that facilitate the creation of culturally tailored programmes co-designed with target communities.
The cost-effectiveness analyses using short-term follow-up data indicated there is potential for programme to be cost-effective, dependent on assumptions about long-term weight and HbA1c regain trajectories. In contrast to the DiRECT economic analysis,19 we conclude the programme is unlikely to be cost saving. This difference is likely to be partly due to a smaller observed mean reduction in HbA1c at 12 months, and to our use of a microsimulation model to project long-term outcomes, which differs from the three-state Markov model used in DiRECT. Future economic evaluations would benefit from longer-term follow-up data to determine how long observed weight and HbA1c reductions are maintained among people who have achieved remission at 12 months.
Overall, this study, along with the data on effectiveness from clinical trials, indicates that while LCD can enable some participants to achieve T2D remission, the high incidence of T2D among eligible adults and the low likelihood of LCD being cost saving has the potential to significantly impact NHSE budgets. This is not an unusual conclusion, since it is rare for a new medical treatment for any major chronic disease to be both health-improving and cost saving. Indeed, systematic reviews of novel healthcare interventions delivered in primary care and the community to promote diet or physical activity44,45 and other studies that have used UKPDS-OM2 to simulate the cost-effectiveness of interventions,46–48 have also generally made the same conclusion of cost-effectiveness but not cost saving. Depending on the level of uptake, the LCD budget impact could approach NHSE’s budget impact threshold.49 A useful comparative example is the economic evaluation conducted alongside the DROPLET trial50 which estimated the long-term outcomes associated with a similar commercially provided TDR-based LCD programme compared with nurse-led behavioural support among eligible participants with or without T2D and a BMI > 30 kg/m2. The trial had a comparable sample size to DiRECT and used a pre-built simulation model suited to populations without diabetes (PRIMEtime-CE). While there are key differences with our study in terms of the intervention and the comparison group, the study estimated an incremental cost-effectiveness ratio of £12,955 (95% CI: £8082 to £17,827), which is comparable to our study in terms of demonstrating that the intervention improves health but is not cost saving.
Strengths and limitations of the study
While the strengths and limitations of each part of this study are clearly documented within each of the threaded publications, overarching strengths and limitations include the following:
Strengths
- The incorporation of PPI activity at every stage of this study has ensured a person-centred focus throughout.
- Collaboration with Diabetes UK enabled the team to support and incorporate learning from this study, to develop the user experience surveys which are now being used in the national programme to ensure continued monitoring of SP delivery.
- While the independence of this evaluation was paramount, effective coproduction with NHSE was also critical in enabling real-time translation of evidence into practice, which informed the development of the national roll-out prior to the official study end date.
- Effective cross-institutional working has allowed us to pool the breadth of expertise required for this complex real-world evaluation and deliver within a tight time frame and budget.
- The use of real-world data provides evidence of how the intervention works in practice and provides complementary data to the trial evidence but is delivered across the breadth of eligible populations, including the populations who are often under-represented in clinical trials.
- Use of a realist approach pulls together findings from across a complex programme of threaded research to provide a comprehensive insight into the underpinning mechanisms and context that drive the programme outcomes.
Limitations
- Starting the evaluation after the programme had been commissioned led to some limitations such as not being able to observe live staff training which had largely been completed prior to the study commencement.
- There were insufficient data to correlate the fidelity assessment with the clinical outcomes, which would be helpful in informing future service development.
- Given the time taken to acquire and process the NHS data, we are reliant on learning from data from the first cohort of service users who had time to complete the programme. As the data set grows, so will opportunities to expand the analyses.
- Developing relationships with some providers was a challenge, as interest and commitment to the evaluation process varied significantly between providers, despite monthly feedback where data were available.
- Varying levels of provider engagement impacted uptake and engagement in the service user survey (which was reliant on the provider as a gatekeeper for governance and logistical reasons). The longitudinal completions were most significantly impacted which limited the longitudinal analyses we could undertake.
- One provider was unable to take part in the session observations which limited the completeness of our data and the learning from this workstream.
- Data available from this study focused on those referred to the programme; therefore, more data are needed to explore the needs and characteristics of patients who were offered but declined a referral.
- It was challenging to engage NHS staff to take part in the evaluation which took place during the COVID-19 pandemic, which may have impacted the range of staff available to participate.
- As a result of the pandemic, the group and 1 : 1 services had to rapidly convert to remote delivery which may have impacted their delivery plans, and our ability to gain an accurate assessment of differences in costs and effectiveness by delivery model.
- The small number of people completing MyFood24 substantially limited the value of the dietary data intake data.
Other reflections and study learning
Our effective cross-institutional working provided a fantastic opportunity to share expertise, foster cross-university partnerships which have led to other successful funding applications (NHIR Health and Social Care Delivery Research funded complications in excess weight clinic evaluation, and a multiple long-term condition study funded by the Nuffield foundation), and others under development. The study has helped support the career development of our two project PhD students (Evans and Dhir) and Early Career Researchers (Drew and Marwood), and the promotion of three of our senior research staff to reader/associate professor (Matu, Brown and Homer). The study also allowed the development of a close and effective working relationship with the NHSE team, and opportunities such as the involvement of the team in the national procurement process, and the co-development of ongoing research.