Tackling reporting issues and variation in behavioural weight management interventions: Design and piloting of the standardized reporting of adult behavioural weight management interventions to aid evaluation (STAR‐LITE) template

In the United Kingdom, the National Institute for Health and Care Excellence make recommendations to guide the local‐level selection and implementation of adult behavioural weight management interventions (BWMIs) which lack specificity. The reporting of BWMIs is generally poorly detailed, resulting in difficulties when comparing effectiveness, quality and appropriateness for participants. This non‐standardized reporting makes meta‐analysis of intervention data impossible, resulting in vague guidance based on weak evidence, reinforcing the urgent need for consistency and detail within BWMI description. STAR‐LITE ‐ a 4‐section, 119‐item standardized adult BWMI reporting template ‐ was developed and tested using a two‐phase process. After initial design, the template was piloted using adult behavioural weight management RCTs and currently implemented UK BWMI mapping information to further refine the template and examine current reporting and variance. Overall, reporting quality of weight management RCTs was poor, and large variance across different components of real‐world BWMIs was observed. Non‐specific guidance and wide variation in adult BWMIs are likely linked to inadequate RCT reporting quality and the inability to perform reliable comparisons of data. Future use of STAR‐LITE would facilitate the consistent, detailed reporting of adult BWMIs, supporting their evaluation and comparison, to ultimately inform effective policy and improve weight management practice.


Summary
In the United Kingdom, the National Institute for Health and Care Excellence make recommendations to guide the local-level selection and implementation of adult behavioural weight management interventions (BWMIs) which lack specificity. The reporting of BWMIs is generally poorly detailed, resulting in difficulties when comparing effectiveness, quality and appropriateness for participants. This non-standardized reporting makes meta-analysis of intervention data impossible, resulting in vague guidance based on weak evidence, reinforcing the urgent need for consistency and detail within BWMI description. STAR-LITE -a 4-section, 119-item standardized adult BWMI reporting template -was developed and tested using a two-phase process. After initial design, the template was piloted using adult behavioural weight management RCTs and currently implemented UK BWMI mapping information to further refine the template and examine current reporting and variance. Overall, reporting quality of weight management RCTs was poor, and large variance across different components of real-world BWMIs was observed. Non-specific guidance and wide variation in adult BWMIs are likely linked to inadequate RCT reporting quality and the inability to perform reliable comparisons of data. Future use of STAR-LITE would facilitate the consistent, detailed reporting of adult BWMIs, supporting their evaluation and comparison, to ultimately inform effective policy and improve weight management practice.

| Intervention guidance and barriers to commissioning
In the United Kingdom, commissioners of these 'Tier 2' multicomponent behavioural interventions have identified a 'lack of clear guidance', indicating that current National Institute for Health and Care Excellence (NICE) best practice guidelines are too broad to effectively assist local-level BWMI selection. 2

NICE recommendations aim
to direct the delivery of high-quality, effective BWMIs, but the supporting evidence -a meta-analysis and systematic review comparing weight management RCTs 3,4 -failed to reliably differentiate between the most effective and ineffective components for weight loss. Authors cited paucity of data and inadequate descriptions of BWMIs as barriers to evaluation and, following this, NICE collated a list of 'knowledge gaps' where evidence lacked, 5 including: • A lack of trials directly comparing BWMIs in the United Kingdom • A lack of evidence on which specific components of a BWMI ensure effectiveness • A lack of evidence on the effect of sexual orientation; disability; religion; place of residence; occupation; education; socioeconomic position; and social capital on the effectiveness of BWMIs and analysis of participants by age and gender • A lack of evidence as to whether any particular type of training for practitioners leads to more effective BWMIs UK weight management mapping efforts have identified considerable variation across nationally implemented BWMIs, with indications that widespread uncertainty regarding best practice amongst those who select interventions for use at local-level is the likely cause. 2,6 The reports highlighted the large inconsistency of outcome reporting by BWMIs, 6 with authors identifying the absence of standardized reporting as problematic for data analysis due to heterogeneity. 2 At present, there are no participant-specific gold standard BWMIs. 7 Given the wide variation between currently implemented interventions, 2,6 the placement of participants into appropriately tailored BWMIs is crucial to maximize individual success. To adequately support informed decision-making regarding the provision of such care, evidence-based guidelines must be drawn from robust analyses of data. To facilitate accurate assessments of intervention effectiveness and identification of the most beneficial components for specific participants, delivery information and outcome reporting must be clear, complete and transparent for the readers. A prominent barrier to drawing reliable comparisons between BWMIs lies within general reporting styles of intervention delivery, in terms of a lack of detail and uniformity -health intervention descriptive reports are often incomplete and widely varying in structure. 7,8 The consistent reporting of BWMIs within both research trial and real-world settings is crucial for successful evaluation. The homogeneous, high-quality reporting of BWMI descriptions would facilitate accurate evaluations of interventions within systematic reviews and meta-analyses -findings of which could inform policy and ultimately improve current clinical practice. Further, consequential resource wastage (ie, time and finances) by the implementation of ineffective interventions following vague recommendations could be mitigated by stronger guidelines.

| Intervention reporting frameworks and templates -development and feedback
Robust frameworks exist within clinical research, created to guide intervention description; tackle low reporting quality within RCTs 8 ; avoid biased reporting of trials 9 ; and address issues of reporting inconsistency (which consequentially hamper comparison efforts), to ultimately facilitate better-informed decisions by policy makers. 10 Numerous tools have attempted to improve the overall poor quality of description within published interventions, present possibly due to little awareness amongst researchers of what constituted adequate reporting. 11 Transparency from authors is encouraged by 'checklists', provided for reporters to follow as guides -however, most tools do not attempt to standardize reporting structure, 8,9,11,12 allowing great variation in content reported. For example, the SPIRIT 2013 Statement (Standard Protocol Items: Recommendations for Intervention Trials) 12 presented a list of minimum items to be addressed within clinical trial protocols, but does not control for variation in depth-of-detail within intervention descriptions. As reporting guidance has developed, more discipline-specific tools have been created -for example, CONSORT-SPI 2018, an extension of CONSORT 2010, expanded on several items to develop checklist relevance for social and psychological RCTs 13 -but a lack of highly specific reporting recommendations for BWMIs persists.
Clinical BWMIs commonly do not publish all outcome or delivery information explicitly and there is an absence of consistency in reporting styles between those that have, limiting accuracy of comparisons. In 2009, the National Obesity Observatory created the 'Standard Evaluation Framework for Weight Management Interventions', a project aiming to facilitate future intervention evaluation. 14 A revised version and online data-collection tool (where intervention leads could submit delivery data to the Public Health England database) was produced in 2018, informed by regionally gathered feedback on the earlier edition from relevant users, that is, BWMI commissioners, providers and researchers. 15 A prominent issue with this tool was the general non-specificity of items included -allowing opportunity for variation in responses. Similar to intervention mapping and NICE guidance knowledge gaps, the Standard Evaluation Framework document cited a need for high-quality evidence regarding BWMI effectiveness. The National Obesity Observatory recommended that to further support Standard Evaluation Framework implementation, standardized reporting templates for BWMIs should be created which would specifically assist the expansion of the current evidence-base of BWMIs and support rigorous evaluations of effectiveness.

| Aims of the current paper
Despite existing tools, reporting quality across weight management interventions remains poor, persistently limiting the effectiveness of comparisons within research and causing authors to call for standardized guidance on reporting. [16][17][18] In order to improve overall BWMI reporting quality with regard to consistency, clarity and completeness, an effective and specific solution must be offered. In 2020, a comprehensive, 24-item 'core outcome and corresponding definition/instrument set' gathered using expert consensus was published to improve BWMI outcome reporting. 19 This list of outcomes (defining which should be measured and how) aimed to resolve uncertainty in decision making by presenting BWMI outcome information equally across all interventions. The current paper describes the development and piloting of a template for the standardized descriptive reporting of adult BWMIs, to complement this core outcome set. Readily available descriptive data for BWMIs is predominantly from lab-based trials or research settings, which may not entirely reflect that of clinical interventions. 20,21 Moreover, this information is found within individual papers and must be deconstructed by readers without a consistently encouraged reporting style or structure. Therefore, the current template will be designed for both clinical BWMIs and behavioural weight management RCTs that are implemented in a real-world setting. Template piloting will provide insight into the current variation and reporting quality seen in both, respectively.

| Phase 1 -initial template design
This phase was designed to produce a preliminary list of items within an initial template draft, which was generated by one researcher and individually checked by the research team. Available research similar in the aim of guiding intervention reporting was examined using online database search engines (PubMed, Google Scholar, ScienceDirect) to identify items for inclusion within the reporting template. Reference lists of relevant papers were hand-searched for related papers to examine.
The initial design phase brought together several published resources -including similar reporting tools, 11,15,22-24 intervention mapping reports, 2,6,15 NICE guidance and related commissioner feedback 5,15 -to identify the key components required for detailed capture of BWMI delivery data (Table 1). Template creation intended to complement a pre-defined core outcome set for BWMI reporting, 19 whilst aiming to address gaps in NICE knowledge 5 and areas of uncertainty via specific item inclusion.

| Phase 2 -piloting
The template was piloted using spreadsheet software for ease-ofdata-entry and analysis (Microsoft Excel 2016). Three types of BWMI reporting data were gathered: • Eleven completed, anonymized Scottish mainland health board  Table 2. BWMI data extraction was undertaken by one researcher. Data was systematically entered into the spreadsheet intervention-by-intervention.
Data gathered were used to refine item inclusion and wording, depending on the item's ability to encourage consistent answer specificity with minimal ambiguity. The same researcher analysed reporting quality in currently available RCTs (examined through reporting frequency and depth-of-description of template-specific items) and variance across real-world BWMIs (relating to delivery-styles and components) by comparing collected data.

| RESULTS
STAR-LITE (STAndardized Reporting of adult behaviouraL weight management InTerventions to aid Evaluation), a BWMI reporting template (Table S1)

| Phase 1 -initial template design
The template included conditional, multiple choice and free-text answers as modes of data-capture.
The 'Referral Pathway' section was designed to capture information regarding how participants entered the intervention, eligibility criteria, referral staff and timescale between referral and active weight loss phase participation. 'Intervention Delivery' included geographical data (ie, total area covered by the intervention, number of bases), delivery setting (ie, primary care, community-based), staff involved and number of sessions (in active weight loss phases and self-defined weight maintenance phases). The third section, 'Intervention Components', dealt with intervention content -specifically, the type of dietary, physical activity and behavioural advice delivered. Questions also aimed to capture whether or not diet and physical activity were monitored, and how. The final section -'Costing'concerned BWMI financial information, specifically the costs for delivering the intervention in a real-world setting (and not including research costs).
Initially, a simple check-list style reporting method was implemented for the description of behaviour change technique (BCT) inclusion using the CALO-RE taxonomy. 24 Upon review, it was decided that a simple 'tick-box' data collection approach elicited minimal detail other than presence or absence of each BCT, and STAR-LITE was refined to require additional delivery information for each technique. As mentioned by the CONSORT statement, rigid reporting guidelines may unintentionally encourage interventions to report fictitious information. 9 As such, users were given a trichotomous 'yes', 'no' or 'unsure' option when reporting technique presence. Identified via Scottish weight management provision mapping, an area of T A B L E 1 Resources used to inform and shape initial template design 1. Template for intervention description and replication (TIDieR) checklist and guide 11 • Items provided a basis for initial template draft to be built upon • For example, 'what', 'who', 'how', 'where' • Layout inspected 2. NICE best practice guidelines for BWMIs 5 • Examined to inform template design and for potential items of inclusion with respect to variation in interventions and areas of uncertainty within reporting

Standard Evaluation Framework 25
• Examined for potential items of inclusion with respect to areas of uncertainty within reporting and variation in interventions • For example, 'essential' and 'desirable' criteria for evaluating a BWMI 4. Standard Evaluation Framework feedback report 15 • Examined to inform template design with respect to variation in interventions, areas of uncertainty within reporting and barriers to uptake • Provided recommendation for standardized data collection tool 5. Two-part NICE-affiliated review of current BWMI evidence 3,4 • Comparisons made within the review used as the basis for NICE BWMI guidance (part 1a and part 1b) informed item inclusion • For example, 'delivery style', 'delivery mode' and intervention content 6. Scottish Tier 2 BWMI mapping survey 6 • Examined for potential items of inclusion, seeking to improve on potential areas of non-specificity relevant to intervention reporting • Layout inspected 7. Public Health England BWMI mapping report 2 • Provided recommendation for standardized data collection tool • Feedback within mapping report informed important items of inclusion • For example, 'costing' 8. Standard Evaluation Framework online data collection tool 22 created by the National Obesity Observatory to allow the collection of intervention summary data by practitioners • Items within the data collection tool were examined for potential inclusion, seeking to improve on potential areas of non-specificity relevant to intervention reporting • For example, 'dietary data collected', 'physical activity data collected' 9. The Coventry, Aberden and London -Refined (CALO-RE) taxonomy 24 • Identified and considered for integration within the template to record behaviour change techniques (BCTs) used within interventions

Taxonomy of BCTs used in interventions 26
• Identified and considered for integration within the template to record BCTs used within interventions 11. The Oxford Food and Activity Behaviours (OxFAB) taxonomy 27 • Identified and considered for integration within the template to record BCTs used within interventions 12. Consensus on Exercise Reporting Template (CERT) 23 • Examined to inform item inclusion for physical activity component description • For example, type of physical activity involved, generalized or personalized physical activity suggested further investigation was 'how, where and by whom' individual BCTs were delivered. 6 Thus, the final template required users to report frequency of and during which intervention week(s) each technique was delivered, how the technique was delivered, and details of staff involved.

| Phase 2 -piloting
Descriptive BWMI data were recorded during template piloting (Table S2). Real-world BWMI reports were examined for areas of variation; RCTs were examined for reporting frequency (quantified within Tables S3 and S4) and general description quality (in terms of depthof-detail) within template items.
Multiple choice and free-text items allowing large response variation were amended to conditional answer format. Almost all multiplechoice items were revised to contain additional answer options according to the most commonly encountered data and variation in intervention description.
Overall, real-world BWMIs and RCTs fit well into STAR-LITE during piloting, aside from 'Costing' (as only one intervention paper 34 reported financial information) and BCT reporting through CALO-RE 24 (as few made use of a recognized taxonomy).

| Referral pathway
Most real-world BWMIs involved self-referral or healthcare professional referral (ie, GP, nurse) and were open to participants ≥18 years, of any gender and ethnicity.
Items related to referral personnel (ie, staff or self-referral) and eligibility criteria were generally well reported by RCTs -of all 39 individually reported intervention arms, 37 reported the referral pathway method (ie, 'self-referral' in response to for example, advertisement flyers; healthcare professional referral). Thirty-eight intervention arms reported specific inclusion criteria, 36 reported exclusion criteria and 29 reported pre-participation assessment methods. Few interventions reported the duration between referral and active weight loss phase initiation (n = 9) or whether incentives for attending the intervention were offered (n = 14).

| Intervention delivery
Real-world BWMIs displayed large variance across delivery and setting, with both group-based and 1-to-1 sessions delivered within primary care  Of the 39 RCT intervention arms, 33 reported BCTs employed, however, only 5 -from one paper 37

| Costs
Costing information could not be adequately collected due to absence of description across all data sources. Three RCT interventions, from one paper, 34 reported estimated costs per participant as estimated by 'the total annual costs of the intervention (per RCT condition), divided by the total number of participants in the group with measured body mass index at 12 months'.

| DISCUSSION
We have used multiple intervention mapping exercises, NICE and Standard Evaluation Framework practice guidelines and previously designed reporting frameworks 5,15,25 to identify and select the critical items required to adequately report BWMIs for the purposes of future analysis, creating STAR-LITE. Through consideration of high-quality, evidence-based tools and pre-existing evidence of a need for a specific BWMI reporting tool, a robust template was produced. 11 4.1 | Phase 1 -initial template design: resources and process STAR-LITE was developed to allow investigation into knowledge gaps identified by NICE through specific item inclusion. 5 For example, evidence surrounding practitioner training is lacking, in relation to which types may lead to more weight loss. NICE recommends that staff are trained prior to intervention implementation and professional staff development sessions are delivered throughout, but fails to make specific qualification recommendations. Therefore, an item included within the template required the description of staff, their qualifications and experience -details commonly ill-defined within weight management RCT reporting, as shown within piloting.
Taxonomies are a recognized method to assist the reporting of (typically complex) behaviour change interventions and their applied BCTs. 24,54,55 Techniques are coded by a corresponding number which can be reported by those who deliver them, facilitating increased clarity and transparency within intervention reporting. 56

| CONCLUSION
STAR-LITE, a specifically designed, developed and tested template, could encourage a higher standard of reporting across adult BWMIs than is currently seen. With effective, evidence-based directions for implementation resulting from robust meta-analysis of data, realworld BWMIs tailored to specific populations would successfully reduce participant obesity prevalence.