Introduction

Rationale for research and background

Type 2 diabetes (T2D) is now one of the leading causes of death globally, and it affects millions of people in the UK.1,2 Currently, over a quarter of adults in England live with obesity:3 the most significant risk factor for developing T2D.4 However, obesity and T2D do not affect all populations equally, with higher prevalence in older age groups, those living in area-level deprivation, and people of black and South Asian ethnicity.5 It is estimated that the NHS and wider societal costs associated with obesity and diabetes will continue to accelerate unless urgent action is taken.4 The NHS long-term plan6 therefore pledged to provide targeted support, and access to weight management services in primary care for people with a diagnosis of T2D and a body mass index (BMI) of ≥ 27 (adjusted appropriately for ethnicity), with the aim of significantly improving health, reducing health inequalities and cutting associated NHS costs.

The most recent trial7–11 systematic12,13 and narrative review14 evidence has shown that a low-calorie diet (LCD) achieved by total diet replacements (TDRs) can lead to clinically significant weight loss and support remission of T2D. Based on evidence from two large UK trials [Doctor Referral of Overweight People to Low Energy total diet replacement Treatment (DROPLET) and Diabetes Remission Clinical Trial (DiRECT)],7,10 a commitment was made by the NHS to pilot a LCD programme delivered through TDR with behaviour change support, for people living with overweight/obesity and T2D. This pilot programme was launched in 2020, initially across 10 sociodemographically diverse Integrated Care Board (ICB) areas (wave 1), with a further 11 ICB areas added in 2022 (wave 2). The pilot was delivered by a range of commercial service providers, commissioned through NHS procurement processes to provide a service aligned to a standard service specification. A summary of the programme is provided in Figure 1, but in brief: adults (18–65 years) residing in the pilot ICB localities, with a recent (< 6 years) diagnosis of T2D and a BMI of ≥ 27 (adjusted appropriately for ethnicity) who did not meet the NHS exclusion criteria (which includes use of insulin, active cancer, eating disorder, heart failure, renal impairment, bariatric surgery, or an inability to follow or attend the programme) were eligible to be referred through primary care. Once referred, service users received an initial assessment (IA), then commenced the programme of 12 weeks TDR, followed by 4–6 weeks of food reintroduction (FR), then weight maintenance (WM) support until the end of the programme (52 weeks). Behaviour change support was delivered via one of three delivery models: 1 : 1, group or digital delivery [where each area was allocated one delivery model and one commissioned service provider (SP)]. As the programme was launched during the COVID-19 pandemic, 1 : 1 and group support was initially delivered remotely, with in-person delivery tested from 2022 in wave 2.

FIGURE 1 An overview of the NHS LCD programme. (Figure not reproduced here; see the original article.)

A comprehensive evaluation of the translation and implementation of the LCD intervention into real-world practice was critical, particularly given the lack of diversity in ethnicity (primarily white), SPs (only Counterweight/Cambridge Weight Plan) and behaviour support models (only one-to-one support) within the underpinning trials. The Re:Mission study reported here provides the outcome of this independent evaluation funded by the National Institute for Health and Care Research (NIHR) Health Service and Delivery Research stream.

Aim

To deliver a coproduced, comprehensive qualitative and economic evaluation of the NHS LCD pilot, that will be integrated with the NHS England (NHSE) quantitative analyses, to provide an enhanced understanding of the long-term cost-effectiveness of the programme and its implementation, equity, transferability and normalisation across broad and diverse populations.

Objectives

Each objective is assigned to a designated work package (WP):

1. assess different providers’ experiences of the programme, including any barriers and facilitators to implementation across the different populations (WP2)
2. assess the experiences and attitudes of NHS staff involved in mobilising the programme across each pilot area and referring and supporting patients on the programme, and their opinions on the management of the programme implementation (WP2)
3. assess patients’ experiences of the programme: including patients with a range of sociodemographics (e.g. socioeconomic status, ethnicity, sex, start BMI), and with differing engagement experiences (e.g. adhered to, or dropped out of the programme) within each of the different delivery models, to gain insight into what worked, and what did not, for whom and why, and how the programme could be improved in the future (WP3)
4. estimate the long-term cost-effectiveness of each NHS LCD delivery model [in terms of incremental cost per quality-adjusted life-year (QALY)] when compared to a counterfactual scenario, including how this varies by delivery model, to enable comparisons with other demands on healthcare resources and thus support commissioning decisions (WP4)
5. assess national roll-out of the NHS LCD through a transferability and policy impact assessment (WP5)
6. integrate findings from WP2 to WP5 with the quantitative analyses conducted by NHSE, using a realist-informed approach, to provide a comprehensive understanding of the programme implementation and impact, to inform future service development and commissioning (WP1).

Research questions and methods

A mixed-methods approach was undertaken involving documentary reviews, session observations, semistructured interviews, focus groups and surveys. These data were drawn together with findings from the NHSE clinical data analyses, using a realist-informed approach to understand what works, what does not work, for whom, why and in what context. A comprehensive description of the methods deployed within each WP is provided in the study protocol v6,15 but a brief overview of the WP methods and associated research questions (RQs) is as follows, with a schematic presentation of the research pathway shown in Figure 2. Methods and learning from the associated study within a study programme (SWAP) are provided in Appendix 1.

FIGURE 2 A schematic representation of the research pathway. PPI, patient and public involvement. DUK, Diabetes UK, ICB: Integrated Care Board; PPI, patient and public involvement; SU, service user; SP, service provider. (Figure not reproduced here; see the original article.)

Work package 1 project management, coproduction, patient involvement and dissemination

(1) facilitated liaison with all key stakeholders, NHSE and the LCD advisory and patient groups: ensuring that patient involvement and coproduction underpin every stage of each WP; (2) provided overarching project management: ensuring timely completion, cohesive working and quality assurance; (3) co-ordinated the interim and final evaluation reports: drawing together the evidence from WP2 to WP5 with NHSE quantitative analyses; (4) delivered a comprehensive programme of dissemination and communication.

Work package 2 service delivery and fidelity

Used a combination of documentary review session observations (two sessions observed from start to finish for each delivery model and provider); thematically analysed semistructured interviews with NHS staff (19 referral staff and 10 locality leads – 1 from each of the wave 1 pilot sites interviewed at years 1 and 2); focus groups with providers (n = 13); and additional semistructured interviews with 7 coaches from, or responsible for delivery to, diverse ethnic groups, to answer the following RQs:

RQ1 What are the theoretical principles, behaviour change components, content and mode of delivery of the programme, and how do these vary across sites and providers?

RQ2 To what extent does the staff training delivered by each provider address behaviour change theory and content, and how does this vary across sites and providers?

RQ3 To what extent is the delivery of the NHS Low Calorie Diet delivered with fidelity to the specification as set out by NHSE?

RQ4 What are provider and NHS support staff experiences of the service, and what do they perceive to be the key barriers and facilitators to effective delivery, integration and normalisation into routine care?

Work package 3 service user experience and inequalities

Were underpinned by a pluralistic approach, undertaken using cross-sectional and longitudinal service user surveys (n = 719), semistructured thematically analysed cross-sectional (n = 37) and longitudinal (n = 30) interviews supported by adapted photovoice methodology. These findings will be aligned to, and integrated with quantitative process and outcome data from NHSE, to answer:

RQ5 To what extent is the content of the NHS Low-Calorie Diet understood and applied by service users?

RQ6 Do sociodemographic characteristics (such as sex, socioeconomic status and ethnicity) influence access, uptake, compliance and success on the NHS Low-Calorie Diet, and does this vary across the different (1 : 1, group or digital) behaviour change delivery models?

RQ7 What aspects of the service work and do not work, for whom, in what context and why?

RQ8 If effective, how can the service be improved in the future, to enhance service user experience and ensure any inequities are addressed?

Work package 4 Economic evaluation

Used patient-level simulation modelling to estimate the long-term cost-effectiveness of the NHS LCD (in terms of incremental cost per QALY) when compared to a counterfactual scenario. This enabled comparisons with other demands on healthcare resources and thus supported commissioning decisions, and included a micro-costing exercise for each of the three delivery models, to address:

RQ9 What are the costs of delivering the NHS Low-Calorie Diet programme from an NHSE perspective and how do they (i) differ across the different delivery models and (ii) compare to estimates provided in the DiRECT trial?

RQ10 What are the costs of the NHS Low-Calorie Diet programme to participants, and how do they differ by delivery model and service user-level characteristics?

These cost estimates were then used along with the patient-level demographic and clinical information collected over 12 months by NHSE as inputs in the patient-level simulation model to answer:

RQ11 What is the long-term cost-effectiveness of the NHS Low-Calorie Diet (in terms of incremental cost per QALY) when compared to a counterfactual scenario, and how does this vary by delivery model and patient characteristics?

We also compared the cost and short-term outcome data collected in this study with those from DiRECT, to enable comparison with the cost-effectiveness estimates, to answer:

RQ12 How does the cost and (short-term) outcome data collected in this study affect the estimates of cost-effectiveness in previous trials?

Work package 5 transferability assessment

Employed a theoretical model for the assessment of transferability and normalisation of health interventions to understand the role of the delivery context in influencing the implementation and service user outcomes, and the implications for widespread adoption of the programme. We undertook semistructured interviews with locality leads (n = 7), referrers (n = 9) and providers (n = 9) in sites involved in stage two of the pilot (whereby the programme was delivered in further 11 areas ahead of the national roll-out). Interviews were structured and thematically analysed according to constructs within the Population, Intervention, Environment, Transfer16 conceptual model of transferability with emerging findings from WP2 to WP4 reflected upon during interviews. Similarities and differences between sites with respect to delivery models were also explored to address:

RQ13 What are the core elements of the intervention that are required to achieve impact?

RQ14 What elements can be adapted to suit local context?

RQ15 What are the policy implications for widespread adoption of the programme?

Results summary

A realist-informed framework was used to draw together learning across the study (see Appendix 2) to provide a comprehensive understanding of the programme implementation and impact, as well as inform future service development and commissioning. The clinical outcome data from the first available cohort of service users analysed by NHSE using the LCD minimum data set (MDS)17 are summarised in Box 1. These quantitative NHS findings were contextualised within key study findings from across the threaded publication series [Research Articles (RA) – see Box 2], which are presented under each WP and associated RQs below.

A summary of the clinical and sociodemographic outcome data from the first available cohort of LCD service users – provided by NHSE:17

• 7540 people were referred to the NHS LCD from September 2020 to December 2022.
• At referral:
  • 43% were men.
  • Average age was 50 years.
  • 19% were of Asian, 8% of black, 3% of mixed, 1% of other and 64% of white ethnicities.
  • 26% were from the least deprived areas and 14% from the most affluent.
  • 47% had been diagnosed with diabetes for < 1 year.
  • Average: weight was 109.2 kg, BMI 38 and glycated haemoglobin (HbA1c) was 58.5 mmol/mol.
  • 50% were taking one, 16% were taking more than one and 34% were not taking any glucose-lowering medication.
  • Younger adults, those diagnosed with diabetes 4–6 years before referral, of mixed ethnicity and from more affluent backgrounds, were more likely to start TDR.
  • Men, older adults, those with a BMI between 25 kg/m² and 29 kg/m² or of Asian, black or other ethnicity were less likely to start TDR.
• Of the 1740 people who started TDR and had sufficient time to complete the programme by December 2022:
  • 84% finished TDR.
  • 79% finished FR.
  • 55% completed the programme.
• Average time on the programme was 8 months.
• People who were more affluent, had a lower starting BMI (25–30) and aged 50–65 years were more likely to complete the programme.
• There were differences in completion rates between providers but not delivery models.
• People of Asian, black and mixed ethnicities lost a smaller proportion of their starting weight compared to those of white ethnicity.
• There were differences in weight change between provider and delivery models.
• Service users who had the required two HbA1c measurements to calculate remission rates revealed a remission rate of 32%.

Data extracted with permission from Valabhji et al.17

Higher remission rates were related to higher weight losses, lower starting glycated haemoglobin (HbA1c) and those with a more recent T2D diagnosis (< 1 year).

RA1_WP1: Clare K, Ojo A, Teke J, Willis M, Akhtar G, Clegg B, et al. ‘Valued and listened to’: the collective experience of patient and public involvement in a national evaluation. Perspect Public Health 2022;142:199–201. https://doi.org/10.1177/17579139221103184

RA2_WP2: Evans TS, Hawkes RE, Keyworth C, Newson L, Radley D, Hill AJ, et al. How is the NHS Low-Calorie Diet Programme expected to produce behavioural change to support diabetes remission: an examination of underpinning theory. Br J Diabetes 2022;22:20–29. https://doi.org/10.15277/bjd.2022.341

RA3_WP2: Evans TS, Dhir P, Radley D, Duarte C, Keyworth C, Homer C, et al. Does the design of the NHS Low-Calorie Diet Programme have fidelity to the programme specification? A documentary review of service parameters and behaviour change content in a type 2 diabetes intervention. Diabet Med 2023;40:e15022. https://doi.org/10.1111/dme.15022

RA4_WP2: Jones S, Brown T, Watson P, Homer C, Freeman C, Bakhai C, Ells L. Commercial provider staff experiences of the NHS Low Calorie Diet Programme pilot: a qualitative exploration of key barriers and facilitators. BMC Health Serv Res 2024;24:53. https://doi.org/10.1186/s12913-023-10501-y

RA5_WP2: Evans T, Dhir P, Matu J, Radley D, Hill AJ, Newson L, et al. Behaviour change techniques in low-calorie and very low-calorie diet interventions for weight loss: a systematic review with meta-analysis. Obes Rev 2025:e13896. https://doi.org/10.1111/obr.13896

RA6_WP2: Drew KJ, Homer C, Radley D, Freeman C, Kinsella K, Maynard M, et al. Equity and Local Health Systems – a qualitative evaluation of the experiences of Local Health Service Leads during the first two years of the NHS Low Calorie Diet Programme pilot. Br J Diabetes 2023;23:77–85. https://doi.org/10.15277/bjd.2023.416

RA7_WP2: Marwood J, Kinsella K, Homer C, Drew KJ, Brown T, Evans TS, et al. Is the NHS low calorie diet programme delivered as planned? An observational study examining adherence of intervention delivery to service specification. Clin Obes 2024;14:e12652. https://doi.org/10.1111/cob.12652

RA8_WP2: Evans TS, Drew KJ, McKenna J, Dhir P, Marwood J, Freeman C, et al. Can the delivery of behavioural support be improved in the NHS England Low-Calorie Diet Programme? An observational study of behaviour change techniques. Diabet Med 2024;41:e15245. https://doi.org/10.1111/dme.15245

RA9_WP2: Radley D, Evans TS, Marwood J, Keyworth C, Homer C, Ells LJ. The NHS Low-Calorie Diet Digital Programme: fidelity of behaviour change technique delivery. Diabet Med 2024;41:e15350. https://doi.org/10.1111/dme.15350

RA10_WP3: Homer C, Kinsella K, Brown T, Marwood J, Drew K, Radley D, et al. ‘A Fresh Start’: qualitative experiences of the Total Diet Replacement phase of the NHS Low Calorie Diet Programme Pilot. Br J Diabetes 2024;24:60–66. https://doi.org/10.15277/bjd.2024.435

RA11_WP3: Dhir P, Maynard M, Drew KJ, Homer C, Bakhai C, Ells L. South Asian individuals’ experiences on the NHS low calorie diet programme: a qualitative study in community settings in England. BMJ Open 2023;13:e079939. https://doi.org/10.1136/bmjopen-2023-079939

RA12_WP3: Marwood J, Radley D, Evans T, Matu J, Clare K, Bakhai C, Ells LJ. A cross-sectional analysis of emotional and binge eating in UK adults enrolled on the NHS low-calorie diet pilot for type 2 diabetes. Clin Obesity 2025;e70025. https://doi.org/10.1111/cob.70025

RA13_WP3: Radley D, Drew KJ, Homer C, Brown T, Kinsella K, Clare K, et al. Participant experiences during the NHS Low Calorie Diet Programme Pilot: findings from an online survey. Br J Diabetes 2024;24:88–94. https://doi.org/10.15277/bjd.2024.431

RA14_WP4: Martin A, Zabula T, Tebaldi D, Yang M, Brown T, Ells LJ. Economic evaluation of NHS England’s Type 2 Diabetes Path to Remission pilot scheme: evidence from real-world data and patient simulation modelling. Submitted Pharmacoecon Open 2024 (under review).

RA15_WP5: Burton W, Padgett L, Nixon N, Ells L, Drew KJ, Brown T, et al. Transferability of the NHS Low Calorie Diet Programme: a qualitative exploration of factors influencing the programme’s transfer ahead of wide-scale adoption. Diabet Med 2024;41:e15354. https://doi.org/10.1111/dme.15354

RA16_WP3: Drew KJ, Homer C, Radley D, Bakhai C, Ells L. A qualitative study of the experiences of individuals who did not complete the NHS Low Calorie Diet Programme Pilot. Br J Diabetes 2024;24:81–7. https://doi.org/10.15277/bjd.2024.434

RA17_WP2: Drew KJ, Homer C, Radley D, Jones S, Freeman C, Bakhai C, Ells L. Normalisation and equity of referral to the NHS Low Calorie Diet Programme Pilot: a qualitative evaluation of the experiences of health care staff. BMC Public Health 2024;24:152. https://doi.org/10.1186/s12889-023-17526-2

RA18_WP1: Dhir P, Evans TS, Drew KJ, Maynard M, Nobles J, Homer C, Ells L. Views, perceptions, and experiences of type 2 diabetes or weight management programmes among minoritised ethnic groups living in high-income countries: a systematic review of qualitative evidence. Obes Rev 2024;25:e13708. https://doi.org/10.1111/obr.13708

RA19_WP3: Homer C, Kinsella K, Brown T, Marwood J, Drew K, Radley D, et al. ‘Life changing or a failure’? Qualitative experiences of service users from the Weight Maintenance Phase of the NHS Low Calorie Diet Programme Pilot for Type 2 Diabetes. Br J Diabetes 2024;24:74–80. https://doi.org/10.15277/bjd.2024.432

RA20_WP1: Homer C, Kinsella K, Ells L, Marwood J, Brown T, Radley D, et al. The Re:Mission study. Evaluating the NHS Low Calorie Diet pilot – an overview of service user data collection methods. Br J Diabetes 2024;24:56–9. https://doi.org/10.15277/bjd.2024.433

RA21_WP2: Dhir P, Maynard M, Drew KJ, Homer C, Bakhai C, Ells L. A qualitative evaluation in community settings in England exploring the experiences of coaches delivering the NHS Low Calorie Diet programme pilot to ethnically diverse participants. BMJ Open 2024;14:e085200. https://doi.org/10.1136/bmjopen-2024-085200

RA22_WP3: Homer C, Kinsella K, Brown T, Marwood J, Drew K, Radley D, et al. Trying to make healthy choices’: The challenges of the food reintroduction phase of the NHS Low Calorie Diet Programme Pilot for Type 2 Diabetes. Br J Diabetes 2024;24:67–73. https://doi.org/10.15277/bjd.2024.436

RA23_ WP3: Flint SW, Goldberg E, Kaykanloo M, Sherman S, Radley D, Kingsbury SR, Ells L. Is personality associated with the lived experience of the NHS England low calorie diet programme: A pilot study. Clin Obes 2025 Feb 25:e70003. https://doi.org/10.1111/cob.70003

Service delivery and fidelity (RQ1–4)

Our service delivery and fidelity assessment were undertaken on the four providers responsible for the first wave of service delivery. To understand the underpinning behaviour change theory used by each provider, we used theory coding analysis of providers programme plans and training materials. Unfortunately, it was not possible to observe and code live staff training as this had been largely completed prior to the study commencement. Our analyses revealed that the application and type of behaviour change theory varied significantly across all providers (RA2). The analyses we were able to undertake found that SP1 included the Transtheoretical Model (Stages of Change) within their training content, while SP2 described multiple theories underpinning their programme design [including Capability, Opportunity and Motivation - Behaviour (COM-B) model, Health Beliefs Model, Transtheoretical Model and the Cognitive Behavioural Model Service]. SP3 also included the Cognitive Behavioural model, in addition to the Antecedents–Behaviour–Consequence model, while SP4 included COM-B and Social-Cognitive Theory in their programme logic model. When BCTs in providers’ programme designs were analysed for the extent to which they linked to theory, constructs or predictors revealed that SP1 linked no BCTs, SP2 linked 63% of BCTs, SP3 linked 70% of BCTs and SP 4 linked all 100% BCTs.

Our assessment of the extent to which the NHS LCD is delivered with fidelity to the specification as set out by NHSE was reported across four publications:

RA3 described and compared the different provider programme designs against the NHSE service specification using the Template for Intervention Description and Replication Framework and Behaviour Change Technique (BCT) taxonomy, to assess drift in fidelity of service parameters and BCTs. This study found that all four providers demonstrated fidelity to most but not all of the stipulated service parameters. Providers included between 74% and 87% of the 23 BCTs identified in the NHS specification: 12 of these BCTs were included by all 4 providers; 2 BCTs were consistently absent. An additional 7–24 BCTs (not in the NHS specification) were also included across providers.

Fidelity of BCT delivery was reported in two papers, both of which used the BCT checklist developed for RA3. RA9 examined the BCT delivery across the digital materials (online/app content and supporting app chat – SP2 and phone call transcripts – SP4) delivered by the two digital providers. Fidelity of BCT delivery was found to be high with 78% and 83% of the BCTs identified in the NHS service specification delivered by the digital providers, and the fidelity of BCT delivery to those specified in providers’ programme plans was 60–65% for SP2 and 82% for SP4. RA8 reported on observed remote delivery of 1 : 1 and group-based programmes across the different providers. A total of 122 sessions were observed across eight samples and two providers completed (despite extensive efforts, the remaining SP delivering 1 : 1 sessions did not engage in the evaluation process). Our findings demonstrated that the complete delivery of the programme was undertaken in five of the eight samples. Fidelity ranged from 33% to 70% across samples and was higher for group-based delivery models (64%) compared with 1 : 1 models (46%). Although this finding should be interpreted within the context of each provider’s delivery style, since the provider with predominantly group-based programmes delivered content through structured activities that appeared to support the delivery of BCTs. A qualitative analysis found facilitators to effective BCT delivery to include alignment with the programme’s target behaviours and outcomes, structured session content, enough available time, effective time management, and not deviating from the session plan.

Learning from RA8 was further supported by the qualitative observations from across the 122 sessions, and how they compared to key delivery criteria within the NHS service specification findings (RA7). These findings demonstrated that adherence to the service specification varied across delivery models and providers, but overall was stronger for more measurable outcomes such as weight and blood glucose, while less tangible elements of the specification, such as empowering service users, reducing health inequalities and person-centred delivery, were less consistently observed. The more interactive and diverse delivery observed in the providers delivering more BCTs seemed to enhance participant engagement. Other observed variations between providers included: (1) level of cultural adaptation of the programme; (2) discrepancies in advocacy of non-starchy vegetables (intended as an option to supplement TDR for service users experiencing difficulty with compliance) and physical activity promotion (which should not be actively advocated during the TDR phase of the programme). Overall, 1 : 1 sessions were more successful in their person-centred delivery, and the skills of, and continuity in, the coaches delivering the sessions had a strong impact on adherence to the service specification. Critical to effective person-centred delivery were friendly and accessible communication, an ability to provide positive feedback, and dedicated efforts to establish connections and build relationships, with some groups using platforms such as WhatsApp to promote peer support. The observations also highlighted some gaps in the support needs of service users in relation to emotional eating and psychological support during the programme, although not a requirement within the specification, it identified an area for future service specification and delivery development.

The experiences and perceived barriers and facilitators to effective delivery, integration and normalisation into routine care were reported for providers (RA4 and RA21), NHS referral staff (RA17) and NHS locality lead staff (RA6).

RA4 found that the programme did fulfil requirements for normalisation from the providers' perspective. Identified facilitators to effective delivery included good internal teamwork, trusting coach and service user relationships, and a wider choice of TDR products. Barriers were identified in engaging general practices (GPs) and receiving sufficient and appropriate referrals, in addition to supporting service users through challenges to remain compliant. Reasons for non-compliance that were discussed by providers were associated with personal characteristics, environment, skills and preferences (such as psychological support needs, multiple life events, busy lifestyles, work commitments that revolve around food, people with larger families, or who do not work but have a lot of ‘thinking time’). The providers reported the main barriers to effective delivery disproportionately impacting diverse ethnic groups, with multiple and intersecting barriers including ethnicity, culture, language and translation requirements, digital competency, dietary requirements, taste preferences, needs of the family and time to attend sessions.

Given the significant barriers identified around the support needs of ethnically diverse communities, a further study was undertaken to interview coaches either from, or with responsibility for delivery to, diverse ethnic communities (RA21). This study identified variations in communication and training between provider sites. Areas for learning and improvement included adapting systems and processes and closing the gap where the needs of service users are not fully met. The themes identified highlighted the varying cultural competency of coaches and the potential impact of knowledge gaps on programme delivery, and the need for comprehensive cultural competency training, to address language barriers, utilise culturally tailored resources, understand diverse cultures and implement effective cultural tailoring strategies with an understanding of cultural nuances.

Interviews with healthcare professionals involved in referring service users to the first wave of the programme (RA17) highlighted the impact of COVID-19 and competing demands on primary care; differences in the expertise and knowledge of referrers; variation in patient identification and the referral process; and barriers to referrals. These barriers included the impact of referrals on workload, confidence in deprescribing, and the intensity and level of commitment required. When equity of referrals was discussed, nearly half of the referral staff felt the referral process was inclusive, although some did report on the potential inequalities of remote delivery and digital exclusion. Language was identified as a potential barrier to referrals but also a facilitator in areas where the programme referral and delivery could occur in another language. Our findings demonstrated that at the time of interview the referral process had not yet been normalised into routine care, with variation and potential inequities in the referral processes across settings identified.

Themes from the referral staff also aligned with those arising from the locality lead interviews (RA6) which highlighted the impact of COVID-19 and primary care capacity and engagement; different methods of communication; and variation in approaches to training, incentivisation and referrals. Barriers to referrals were discussed and included ineligible referrals and time taken to refer, staff turnover in the local health system (LHS) and referrer confidence and expertise. A key facilitator to effective mobilisation of the programme was effective collaboration across all stakeholders.

Service user experience and inequalities (RQ5–8)

The extent to which the content of the NHS LCD programme is understood and applied by service users, and how sociodemographic characteristics and delivery model impact access, uptake, compliance and success on the programme, builds on learning from RQ1 to RQ4 and the quantitative analysis of the LCD MDS undertaken by NHSE (see Box 1).17 We conducted our own service user participation survey (RA12, RA13, RA23) and service user interviews (RA10, RA11, RA16, RA19, RA20, RA22), to provide understanding and context to the NHS process and clinical data provided by the MDS. We also invited our longitudinal interviewees to complete the MyFood24 dietary assessment; however, only 16 out of 30 participants opted to complete this at 12 weeks, which dropped to 11 and 8 in weeks 18 and 52, respectively. This, therefore, limited the utility of the dietary data collected (which is summarised in Appendix 3), but for those who completed the assessment, the only significant findings were an increase in fat over the 52 weeks, while fibre intake remained consistently low, and fruit and vegetable intake remained below recommended intake throughout. At 52 weeks, the average energy intake remained below standard recommended values.

A summary of key learning from across the service user interviews and surveys, alongside NHS quantitative findings, is provided by programme stage:

Referral

Early insights from the first 7540 people referred to the NHS LCD programme (September 2020 to December 2022) reported that 43% of those referred were men, 55% were in the oldest age bracket (50–65 years), and were of Asian (19%), black (8%), mixed (3%), other (1%) and white (64%) ethnicities. A higher proportion of referrals were from service users living in the most deprived areas (26%) compared to the least deprived areas (14%), and around half (47%) had a T2D diagnosis of < 1 year at referral.17 Average weight at referral was 109.2 kg, average BMI was 38.0 kg/m² and average HbA1c was 58.5 mmol/mol.17 For the 66% of people taking glucose-lowering medication at referral, the most common medications were: metformin followed by sodium-glucose co-transporter protein 2 inhibitors, dipeptidyl peptidase 4 inhibitors, glucagon-like peptide-1 analogues and sulfonylureas.17 Our service user survey (RA13) and initial interview data (RA10) showed that service users were motivated to be referred in order to improve their T2D, weight and health. Although most were satisfied with the referral process, the survey (RA13) and interview data (RA10) highlighted some inconsistencies in referral knowledge, with suggestions that more programme information at referral would have been helpful.

Initial assessment and starting total diet replacement

Of those referred with sufficient time to start the programme, 5115 attended an IA and 4340 started TDR,17 with those who were more affluent, younger, of mixed ethnicity, or with a T2D diagnosis 4–6 years prior to referral, more likely to start TDR, while males, older adults, those with a lower starting BMI (25–29) or of Asian, black or other ethnicity were least likely to start TDR. The majority of our survey respondents reported that the information provided at IA was clear (RA13), although some survey respondents and interviewees (RA10) identified some inconsistencies between providers in the IA information provided, and the time point at which service users received their starter packs/course materials. Concerns were also raised by some of our interviewees around the cultural competence of the programme (RA11), which may explain the reduced uptake across some ethnically diverse communities. The perceived service user costs may also have deterred participants living in areas of socioeconomic deprivation. The higher uptake in those who have lived with T2D for longer may also reflect the high motivation to improve health when starting the programme (RA13). This concurs with additional artificial intelligence analysis of the survey and linked MDS data which concluded that personality rather than demographic characteristics determine motivation to continue the programme (RA23).

Food reintroduction and weight maintenance

Of the 1740 participants who started TDR and had sufficient time to finish the programme, 84% were active at the end of TDR, 79% were still active at the end of FR and 55% completed WM (i.e. the whole 12-month programme). Those who were least likely to complete were living in the most deprived quintile, had a BMI of 40+, or were younger (18–39 years).17 While there were some differences between providers, there were no significant differences by delivery method, sex, years since diagnosis or ethnicity identified.17 However, in terms of clinical outcomes service users of Asian, mixed and black ethnicities lost a lower percentage of their baseline weight compared to those of white ethnicity, with no other significant differences by other sociodemographic characteristics. Where data were available to measure remission, 32% achieved it, which was more likely in those who had a more recent T2D diagnosis.17

It is clear that while the LCD programme is successful in achieving the desired improvements in weight, glycaemic control and remission in many people living with T2D, the full 12-month programme was only completed in just over half of all those who started. There are also some population groups: younger adults, those living with severe obesity, or living in areas of socioeconomic deprivation who may need additional support to complete the programme. The clinical outcome data also highlight a need for further adjustments to help improve the impact of the programme for service users of Asian, mixed and black ethnicities. Aggregated findings from our service user survey (RA12, RA13, RA23) and interviews (RA10, RA11, RA16, RA19, RA20, RA22), provide critical insights to help understand the observed attrition and sociodemographic inequities, and inform future service development. These key learning points are summarised below:

• Range and delivery of TDR products was an issue for service users where products were delivered in bulk which created storage issues and a lack of flexibility in changing products, most TDR products were also very tailored to a Western palate.
• Allowance of supplementary foods such as non-starchy vegetables, or a day ‘off’ (i.e. a short return to normal foods) during TDR, supported adherence to the programme for some, although the use of this approach varied across providers.
• Social impact of TDR was reported by some service users, who found going out, socialising, and family meals and celebrations involving food challenging during TDR. This was particularly important if there was a cultural, employment or family role linked specifically to food or food preparation, or where socialising was an important part of everyday life.
• Variation in cultural competence of programme content and delivery was apparent. Examples of good practice were provided that included session delivery in other languages, matching the background of coaches to service users, and the use of tailored resources such as supporting guidance for TDR during religious festivals, which were reported to improve engagement and success across ethnically diverse participants. Additional culturally tailored recipes, adaptations of traditional cooking methods and culturally adapted meal plans were identified as a gap.
• Person-centred delivery and good service user–coach relationships helped tailor the programme to individual needs and were seen by service users as key to success on the programme, yet the degree of person-centredness varied, and was more difficult to achieve in group and digital delivery. This is important given assumed health literacy did not always align with that of the service user. Service users from ethnically diverse backgrounds sometimes reported challenges in understanding, due to sociocultural barriers that can exist with a coach from a different background to their own.
• Increased support and availability of coaches was a need expressed by some service users, particularly during FR and WM when session numbers decreased and service users experienced difficulties in contacting their coach or provider out of set session times.
• Additional psychological support needs were reported by some service users experiencing mental health and or life event challenges, which were frequently reported as reasons for non-completion, missing sessions or difficulties engaging with the programme. Service users felt more flexibility in programme attendance may overcome the impact of life events.
• Family and peer support was frequently reported as critical to the success of the programme, with some service users finding support from WhatsApp groups. Service users from different cultural and ethnic backgrounds found peer support and exchange of cultural adaptations helpful.
• Emotional and disordered eating support needs were reported by several interviewees and were reflected in the large proportion of survey respondents, demonstrating high scores for emotional and disordered eating [just less than a quarter of the sample screened positive on the Binge Eating Disorder Screener (BEDS-7) for likely binge eating disorder (BED), and the mean emotional eating subscale score from the Three Factor Eating Questionnaire-R21 was 2.58 (out of a range of 1–4)].
• Concern about FR was frequently raised by interviewees, and survey respondents many of whom reported a continued reliance on TDR products (outside of the service provision – nearly 20% of survey respondents at 52 weeks reported purchasing their own TDR products), with calls for more structured resources to support FR and the consideration for a longer FR phase with more frequent sessions than bi-weekly.
• Wider benefits of the programme, where reported, with outcomes beyond weight and glycaemia, important to service users, these included improved appearance, daily physical functioning, improved social relationships and the extended benefit on the health of friends and family.

Economic evaluation (RQ9–12)

When service user costs were assessed through the evaluation survey, we found that for those who reported spending money on resources they attributed to enrolling on the programme, the average spend was £125.99 over the whole programme (RA14). There were no statistically significant differences in terms of additional expenditure that service users incurred during the programme by gender, ethnic group, or socioeconomic status. The main component of this additional expenditure was purchases of additional TDR products (during the TDR phase, e.g. to replace flavours they didn’t like and couldn’t exchange, or to continue the use of TDR ‘off programme’ during WM). Other items of expenditure listed by patients included extra glucose monitoring strips, diet and cookbooks.

Data were gathered on costs of delivering the programme from NHSE records of SP payments and cross-checked against other sources: NHS weight management guidelines, our own survey of SPs using an adapted STAR-LITE questionnaire18 and the DROPLET and DiRECT trials. Despite small differences in how the costs were estimated, the majority fell within the range of costs reported in the DiRECT trial paper as represented by the 95% confidence interval (CI) between £1147 and £1294 per service user. None of this evidence indicated substantial or consistent differences in costs by delivery model; however, since most services had been disrupted by the COVID-19 pandemic and were being delivered remotely, there remains a need for future studies to assess these cost differences to ensure that value for money is achieved.

Real-world clinical data were collected by commercial SPs and GPs and then linked to National Diabetes Audit data and cost data. A commercially available, pre-built patient-level simulation model [United Kingdom Prospective Diabetes Study (UKPDS) Outcomes Model] utilised data collected pre-referral and at 12 months to predict health economic outcomes for the service users at 24 months and annually until death. Additionally, a counterfactual scenario portraying potential outcomes for the same service users without the intervention was generated utilising only pre-referral data. It was necessary to test various different assumptions regarding weight regain and HbA1c trajectories among service users at 24 months onwards, relative to the counterfactual scenario. These scenarios indicated that the programme could be cost-effective against a £20,000/QALY threshold if weight and HbA1c trajectories remain below those in the counterfactual scenario for at least 6–7 years (£19,759.80/QALY assuming 7 years) and against a £30,000/QALY threshold if those trajectories remain apart for 4–5 years (£27,625.99/QALY assuming 5 years). More rapid weight and HbA1c trajectories would mean the intervention is less likely to be cost-effective. This finding does not vary by ethnicity, area-level deprivation nor in sensitivity analyses, which included: (1) running the model across a range of different time horizons, including 10 years, 15 years and 20 years; and (2) altering the intervention costs to reflect possible differences in cost by delivery model.

This overall finding contrasts with the results of the DiRECT study which found the comparable Counterweight-Plus intervention to be cost saving, in the sense that it improved health outcomes and decreased (rather than increased) costs to the NHS.19 In our study, this would be the case if a societal perspective were adopted, due primarily to the money service users are likely to save on their usual grocery shopping. However, it seems unreasonable to include these in healthcare commissioning decisions.

Compared to existing clinical trials of similar interventions, our sample was larger (n = 838 vs. e.g. 149 in the DIRECT study intervention arm) and more sociodemographically diverse. While our sample had a comparable mean HbA1c to DiRECT, the mean baseline weight was higher (+ 10 kg). After 12 months, the reduction in HbA1c in our study was smaller (2.3 vs. 9.8 mmol/mol in DiRECT), but mean weight loss was similar (9.9 kg vs. 10.6 kg in DiRECT). Differences in HbA1c and weight reductions were noted across ethnic groups and, to a lesser extent, by area-level deprivation. In the DiRECT trial, 46% of participants in the intervention arm achieved T2D remission after 1 year compared to 4% in the usual care arm, with an intervention cost of £2939 per incremental case of remission.19 In comparison, 27% of participants achieved remission in the real-world data used in our study, which was also reported in a cohort analysis by NHSE.17 Over a longer period, DiRECT showed that 36% of LCD participants achieved remission at 2 years,7 and after 5 years, 10% (n = 12) achieved remission20 (compared to 5% in the control group), with an intervention cost exceeding £20,000 per incremental case of remission.

Transferability assessment (RQ13–15)

Our transferability assessment (RA15) determined that the core elements required to achieve impact included confidence in the programme, multidisciplinary working and good communication, across all stakeholders and a choice of delivery model to promote acceptability and accessibility. Local adaptations to referral strategies were also deemed necessary such as utilising local population characteristics data on deprivation and ethnicity to inform efforts to drive equitable uptake, and referrals that reflect the local target population. Adaptations to programme delivery such as ensuring a person-centred approach and incorporating cultural tailoring were again highlighted to ensure the needs of individual patients were met. Policy implications for widespread adoption should include referral strategies to reach under-represented groups, a choice of delivery model to optimise uptake, and the provision of timely data from SPs on access and programme benefits.