I have a research assignment titled “Community-Based Multi-Component Interventions for Obesity Prevention in Young Adults (<30 years)” The following edits are needed in the methodology section 1. Methodology edits (the length of this can be increased as the background should be reduced) - Narrowing objections- Objective 1 as primary objective and Objective 2-4 as secondary or exploratory objective - Ensure every research question maps directly to an outcome measure and analysis plan. For each research question, it should be clear what data will be collected to answer it and how it will be analysed statistically. For example, if you ask whether the intervention reduces obesity risk, you need a corresponding outcome measure (e.g., change in BMI at 12 months) and a planned statistical analysis (e.g., mixed-effects regression comparing intervention and control groups). Research questions that do not have a clearly defined outcome and analysis method should be revised or removed. - Justification for cluster randomization and clarity on which cluster to choose - Addition of subsection describing allocation concealment and randomization procedure - Specifying whether outcome assessors will remain blinded through follow up - Justification on why Lusaka was selected - Provision of information regarding population demographics and obesity prevalence in the selected setting and clarifying the number of communities expected to participate - Target population: There is inconsistency between the stated target population (BMI ≥25 kg/m²) and inclusion criteria (BMI ≥18.5 kg/m²), Clarify whether the study is prevention, treatment, or both And Provide justification for selecting the 18–29 year age group. - Inclusion criteria: explaining why normal weight individuals are included if obesity prevention is the main aim, consider excluding individuals with conditions that significantly affect weight trajectories, clarify management of participants who become pregnant during follow-up - Sample size: Full sample size calculation, Specify ICC, average cluster size, attrition assumption and final number of clusters per arm. State primary end point clearly and final planned sample size - Intervention components: o After each intervention level, specify frequency, duration and mode of delivery. o Describe training of community health workers and peer educators. o Add a TIDieR-style description to improve reproducibility. o Explain whether the mobile application already exists or will be developed. - Comparator: discuss possible contamination between intervention and control communities and how this will be minimized. Describe what participants will receive - Outcomes: Clearly define measurement methoods for anthropometric behavioral and biochemical outcomes. Specify validated tools for quality of life and mental health assessment. Select one primary outcome rather than BMI and waist circumference together with justifying why this outcome is clinically meaningful. - Data collection: Describe assessor training and standardisation procedures. Provide details regarding data management, storage and quality assurance. Explain how blinding will be maintained. - Data analysis (needs expansion) o Explain how clustering and repeated measures will be incorporated. o Specify adjustment variables (age, sex, baseline BMI, socioeconomic status). o Describe handling of missing data. o Clarify whether subgroup analyses are exploratory and whether they are powered. o specify the primary statistical model. 2. Addition mission section - Adverse Events and Safety Monitoring (A dedicated section is currently missing). o Include definitions of adverse events, serious adverse events and reporting procedures. o Explain monitoring of exercise-related injuries and psychological distress. o Consider establishment of an independent monitoring committee if appropriate. 3. Ethical consideration: Should be tailored to Zambian context and briefly clarify management of abnormal screening findings and participant reimbursement arrangements if applicable- Be mindful of word count 4. Dissemination plan: needs to be expanded - Expand to include peer-reviewed publications, conference presentations, stakeholder reports, policy briefs, community dissemination events and participant summaries. - Explain how findings will influence future public health policy. Here is the methodology I have so far: OBJECTIVES AND RESEARCH QUESTIONS RESEARCH OBJECTIVES 1. Evaluate intervention effectiveness: To determine the effectiveness of community- based multi-component interventions (dietary, physical activity, behavioral, and environmental) in preventing obesity compared to single-component interventions. 2. Identify behavioral mediators: To explore behavioral mechanisms (self-monitoring, feedback, personalized goal-setting) that enhance adherence and sustainability of obesity prevention strategies. 3. Examine environmental supports: To investigate how environmental and policy- level supports (food access, built environment, workplace/school programs) influence obesity prevention outcomes in young adults. 4. Assess long-term outcomes: To evaluate the long-term impact of community-based interventions on BMI trajectories, incidence of obesity, and cardiometabolic risk factors in young adults. RESEARCH QUESTIONS 1. How do multi-component community-based interventions compare with single- component interventions in reducing BMI and preventing obesity in young adults? 2. Which behavioral strategies (self-monitoring, feedback, personalized goal-setting) are most effective in sustaining lifestyle changes among young adults? 3. How do environmental modifications (healthy food access, active transport infrastructure, workplace/school initiatives) contribute to obesity prevention outcomes? 4. What is the long-term impact of community-based multi-component interventions on obesity incidence and cardiometabolic health in young adults? METHODOLOGY STUDY DESIGN A Cluster randomized controlled trial (RCT) will be conducted within communities randomized to intervention or control arms. This study design is slected so as to avoid contamination between intervention and control participants within the same community setting. It will allow for evaluation of community level effects (Willmott et al., 2022). Community clusters (e.g., university campuses, workplaces, recreational facilities, and residential facilities) will be the unit of randomisation with young adults (aged 18-29 years). The intervention will last 12 months, with baseline, 6 month and 12 month follow up assessments. This design aligns with recommendations for multi-level, multicomponent approaches that address individual, social and environmental determinants of obesity. STUDY SETTING The study will be conducted in selected urban and peri-urban communities in Lusaka Province, Zambia. Communities will be matched on socio-economic status and randomised to intervention or control (usual care/health information only) arms using computer-generated randomisation by an independent statistician. TARGET POPULATION Individuals aged 18-29 years residing within the defined community cluster areas who are at risk of or currently experiencing overweight or obesity (defined as BMI ≥ 25 kg/m²) are the target population. The age bracket represents a critical window as obesity trajectories established in young adulthood strongly predict cardiometabolic risk across the life course (Wong et al., 2020) with this cohort being underserved by existing adult weight management services primarily designed for middle-aged populations (NiCE, 2024). INCLUSION CRITERIA 1. Age 18-29 years (inclusive) at the time of enrolment 2. BMI ≥ 18.5 kg/m² residing in the selected communities for at least 6 months (to include normal weight individuals for prevention) 3. Residing in or regularly attending a selected cluster site for at least 6 months 4. Able to provide written informed consent 5. Able to communicate in the study language (with interpreter access for linguistic minorities) EXCLUSION CRTIERIA 1. Pregnant or breastfeeding women (current) 2. Diagnosed and receiving active treatment for a severe eating disorder 3. Diagnosed condition precluding physical activity participation without medical supervision e.g., uncontrolled cardiac disease 4. Individuals enrolled in weight management programmes or clinical trials 5. Plans to relocate outside study area within 12 months SAMPLE SIZE Sample size calculations are based on detecting a clinically meaningful difference in BMI of 0.5-1.0 kg/m² (standard deviation 2.5 kg/m²) between intervention and control arms at 12-month follow-up (WHO, 2000). Assuming a detectable difference of 0.5–1 kg/m² in BMI change, intra-cluster correlation of 0.02–0.05 (Donner and Klar, 2000; Hemming et al., 2011), 80% power, and 5% significance level, approximately 800–1,200 participants across 20–30 clusters will be required (accounting for 20% attrition). Calculations will be finalised using simulation methods or standard cluster trial software (Murray, 1998; Campbell et al., 2004). INTERVENTIONN COMPONENTS The multi-component programme will include follow an ecological framework, targeting individual, interpersonal, organisational, and community levels  Individual level: Personalised behavioural coaching via mobile app and peer- led sessions on nutrition, physical activity, and goal-setting (e.g., using self- monitoring tools).  Interpersonal level: Family and peer support groups, including cooking demonstrations and group physical activities (Group workshops on portion control and reducing ultra-processed foods).  Organisational/Environmental level: Partnerships with local gyms, markets, and workplaces to improve access to healthy foods (e.g., subsidies or point-of- choice labelling) and safe spaces for activity.  Community level: Mass media campaigns, community events, and policy advocacy for healthier food environments, delivered through community coalitions. Trained community health workers and peer educators will facilitate delivery. Fidelity will be monitored through attendance logs, session checklists, and process evaluations. The control arm will receive standard public health information (e.g., pamphlets on healthy living) but no structured multi-component support. COMPARATOR Control clusters will receive usual care: standard health education leaflets without structured intervention. OUTCOMES Primary outcome: Change in BMI (kg/m²) and waist circumference at 12 months, measured using standardised protocols. Secondary outcomes: • Physical activity (measured by IPAQ or accelerometers). • Dietary intake (food frequency questionnaire or 24-hour recalls). • Behavioural mediators (e.g., self-efficacy, social norms). • Quality of life and mental health indicators. • Metabolic markers (HbA1c, lipid profile) Process measures will assess reach, dose delivered, and acceptability. DATA COLLECTION Trained research assistants blinded to allocation will collect data at all time points.  Baseline: Anthropometric, biochemical, and behavioural data.  Follow-up: At 6 and 12 months.  Sources: Clinical measurements, questionnaires, app usage logs. Digital tools will minimise data entry errors. DATA ANALYSIS Data will be analysed using intention-to-treat principles. Mixed-effects models or generalised estimating equations (GEE) will account for clustering and repeated measures.  Subgroup analyses: Gender, socioeconomic status, baseline BMI.  Sensitivity analyses: Address missing data and attrition. Qualitative process data will be thematically analysed. Statistical significance will be set at p<0.05, with adjustments for multiple comparisons where appropriate. Adverse Events and Safety Monitoring (A dedicated section is currently missing). • Include definitions of adverse events, serious adverse events and reporting procedures. • Explain monitoring of exercise-related injuries and psychological distress. • Consider establishment of an independent monitoring committee if appropriate. ETHICAL CONSIDERATION • Study risk and sensitivity: This project uses a community-based approach to reduce obesity risk among young adults. While the intervention itself is not expected to cause physical harm, it does involve collecting personal details such as body measurements and lifestyle habits. Because this information can feel private, it must be handled with care. • Ethical approval process: Approval will be sought from a university ethics committee before the study begins. If the project is linked to healthcare settings, guidance from the Health Research Authority will also be followed. This step ensures the research is carried out responsibly and meets recognised standards. • Participant understanding and consent: Before taking part, individuals will be given clear and simple information about what the study involves. This will help them decide whether they want to participate. Written consent will be obtained, confirming that they understand the study and agree to take part. • Freedom to participate or withdraw: Taking part in the study will be entirely optional. Participants will not be under any pressure to join, and they can leave the study at any point without needing to explain their decision. • Managing sensitivity around weight-related topics: Discussions about weight and lifestyle can sometimes make people feel uncomfortable. To reduce this, the study will avoid judgmental wording and instead focus on positive health behaviours. Care will be taken to ensure participants feel respected throughout. • Protecting personal information: Any information collected will be stored securely and handled in a way that prevents the identification of individuals. Data will be coded or anonymised where possible, and only the research team will have access to it. All procedures will follow UK data protection standards. • Limiting unnecessary data collection: Only information that is directly relevant to the research question will be collected. This helps reduce any unnecessary intrusion into participants’ personal lives. • Ensuring participant wellbeing: The study will not involve any harmful procedures. If a participant feels uncomfortable at any stage, they can stop immediately. Where appropriate, general health advice or signposting to support services can be provided. • Inclusive and fair recruitment: Efforts will be made to recruit participants from a range of backgrounds, rather than focusing on a single group such as university students. This makes the findings more

Searching the Web

obesity prevalence Lusaka Zambia young adults urban statistics

Searching the Web

TIDieR checklist intervention description template components

Searching PubMed

obesity prevention young adults community intervention Zambia Africa

Searching PubMed

cluster randomized trial intraclass correlation obesity community intervention

Searching PubMed

cluster randomized trial ICC obesity lifestyle intervention

Searching the Web

Zambia Lusaka urban population demographics overweight obesity 2020 2024 young adults DHS

Reading File

Running Command

Writing File

Running Command

The revised methodology document is ready. Download it here:

Obesity Prevention Revised Methodology

Word Document · DOCX

Here is a full summary of every change made and why:

Summary of All Changes Made

Objectives and Research Questions

Objective 1 made the sole primary objective, explicitly labelled as primary. Objectives 2–4 are labelled secondary/exploratory with rationale.
Every research question now maps to a named outcome measure and a named statistical analysis. RQ1 → BMI change → LMER; RQ2 → IPAQ/dietary recall/GSE-10 → causal mediation analysis; RQ3 → environment audit scores → cross-level interaction; RQ4 → incidence obesity + metabolic markers → log-binomial/LMER. Each mapping is stated explicitly.

Study Design

Justification for cluster RCT added (three reasons: community-level intervention, contamination risk, precedent in comparable trials).
Cluster unit specified as geographic community zones with justification for why geographic clusters were chosen over institutional ones.
New subsection: Allocation Concealment and Randomisation Procedure - describes matched-pair design, independent statistician, computer-generated sequence, withholding allocation until after baseline data collection.
Blinding subsection - specifies that outcome assessors remain blinded at baseline, 6-month, AND 12-month follow-ups; blinding verification procedure added; analyst blinding described.

Study Setting

Lusaka selection justified across four dimensions: epidemiological burden (DHS 2024 data cited), double burden context, infrastructure/CHW network, and existing research environment.
Population demographics added: Lusaka Province population ~3.7 million, 80% urban, 40%+ aged 15–34, linguistic diversity, socioeconomic heterogeneity.
Obesity prevalence data grounded in real evidence: 14.2% obesity prevalence (Rudatsikira et al., 2012); urban women overweight/obesity 35–40% (ZDHS 2024); rising trend among women aged 20–29.
Number of communities clarified: 24–30 clusters across both arms (12–15 per arm), each cluster approximately 500–1,000 young adults.

Target Population and Inclusion Criteria

BMI inconsistency resolved: Study is explicitly framed as a combined prevention + early intervention trial. Both normal-weight (18.5–24.9) and overweight/obese individuals are included, with clear justification for each group.
Justification for 18–29 age group: Three reasons stated - high-risk transition period, underserved group, and life-course leverage.
Inclusion criteria tightened and clarified (removed ambiguous second bullet combining two criteria).
New exclusion criterion added: BMI < 18.5 kg/m² (underweight).
Management of participants who become pregnant during follow-up explicitly addressed: not withdrawn; dietary/PA guidance adapted; measurements flagged; included in ITT with sensitivity analysis.

Sample Size

Full sample size calculation provided step-by-step: effect size (0.5 kg/m²), SD (2.5), power 80%, α 0.05.
ICC specified: 0.02 with cited source (Gray et al., 2016; Donner & Klar, 2000) and sensitivity calculation at ICC = 0.05.
Average cluster size specified: 40 participants.
Design effect calculated explicitly: DEFF = 1.78.
Attrition assumption: 20%, with adjusted total ~877.
Number of clusters per arm: 12–15 (24–30 total).
Primary endpoint clearly stated: Change in BMI at 12 months.
Final planned sample: 960–1,200; confirmed via simulation (clusterPower/Stata).

Intervention Components

TIDieR-style description applied across all 12 items.
Frequency, duration, and mode of delivery specified for each of the four ecological levels (individual, interpersonal, organisational, community).
CHW training described: 5-day standardised programme with competency assessment and refresher at month 6.
Mobile app clarification: Does NOT currently exist; will be purpose-built using React Native; pilot tested 3 months before trial; paper backup provided.
Fidelity monitoring procedures detailed.

Comparator

Contamination risk addressed: Geographic separation (≥ 2 km buffer), no shared spaces policy, participant instruction, contamination monitoring survey, sensitivity analysis.
What control participants receive is clearly described: single 15-minute session with standard MOH health information leaflets only.

Outcomes

Single primary outcome selected: BMI change at 12 months (waist circumference demoted to secondary). Justification provided.
Measurement methods defined for every outcome: calibrated stadiometer/scale protocols for anthropometrics; WHO protocol for waist; IPAQ short form for PA; 24-hour recall for diet; GSE-10 for self-efficacy; SF-12 for QoL; PHQ-9 and GAD-7 for mental health; fasting blood panel for metabolic markers.
Validated tools named for quality of life and mental health.

Data Collection

Assessor training and standardisation described: 2-day standardisation training, TEM calculation, equipment calibration checks.
Blinding maintenance through follow-up explained.
Data management: REDCap, role-based access, encryption, weekly quality audits, anonymisation.

Data Analysis (Expanded)

Primary statistical model specified in full (LMER equation written out).
Clustering handled via random cluster intercept; repeated measures via nested participant-level slope.
Adjustment variables listed with rationale: age, sex, baseline BMI, SES.
Missing data: MICE (50 datasets, MAR assumption), with complete-case and pattern-mixture sensitivity analyses.
Subgroup analyses: Explicitly labelled exploratory and not independently powered; Bonferroni correction applied.
GEE listed as sensitivity analysis.

New Section: Adverse Events and Safety Monitoring

Definitions of AE, SAE, and intervention-related AE provided.
Reporting procedures: 24-hour SAE reporting to PI; 7/15-day reporting to UNZABREC/ZNPHI; SUSAR pathway.
Exercise injury monitoring: PAR-Q screening; low-to-moderate intensity protocols; CHW injury recognition training.
Psychological distress monitoring: PHQ-9/GAD-7 screening; referral pathway; community mental health resource list.
DSMC established: Independent statistician, clinician, lay member; 6-month review; pre-specified stopping rules.

Ethical Considerations

Tailored to Zambian context: UNZABREC, National Health Research Act, ZNPHI, PACTR registration, Zambia Data Protection Act 2021.
Management of abnormal findings: Referral with standardised letter; referral log maintained.
Participant reimbursement: Transport cost reimbursement at each visit, at UNZABREC-approved level, non-coercive.
Community engagement and non-stigmatising communication addressed.
Removed UK-specific references (HRA, UK data protection).

Dissemination Plan (Expanded)

Peer-reviewed publications: Open-access, AllTrials commitment, separate papers for primary, process, and secondary outcomes.
Conferences: International (ICO, ECO, Africa Health) and national (ZNPHI Symposium).
Policy briefs for MOH, Lusaka Provincial Health Office, local government.
Stakeholder reports with plain language; workshops with policymakers and community leaders.
Community dissemination events in each cluster; participant summaries in English and Nyanja.
Influence on future policy: Integration into Zambia NCD Policy, WHO LMIC guidance, scale-up design, proactive ZNPHI engagement.
Data sharing statement included.

GIVE ME THE FULL TEXT WITH ALL THE CHANGES

Here is the complete revised methodology in full:

Community-Based Multi-Component Interventions for Obesity Prevention in Young Adults (<30 years)

REVISED METHODOLOGY SECTION

OBJECTIVES AND RESEARCH QUESTIONS

Research Objectives

This study adopts a hierarchical objective structure consistent with best practice in clinical trial design (Schulz & Grimes, 2005). One primary objective drives the primary analysis and sample size calculation; secondary and exploratory objectives address mechanistic, contextual, and longer-term questions.

Primary Objective

To determine the effectiveness of a community-based multi-component intervention (combining dietary, physical activity, behavioural, and environmental components) in preventing progression to overweight or obesity compared with a health-information-only control arm, as measured by change in BMI (kg/m²) from baseline to 12 months in young adults aged 18–29 years in Lusaka, Zambia.

Secondary Objectives

Objective 2 - Behavioural mediators: To explore whether self-monitoring, personalised goal-setting, and peer feedback mediate intervention-related changes in BMI and lifestyle behaviours.
Objective 3 - Environmental supports: To investigate how environmental and policy-level supports (food access, built-environment features, workplace/campus initiatives) modify obesity prevention outcomes at the community level.
Objective 4 - Long-term outcomes: To evaluate the sustained impact of the intervention on BMI trajectories, incidence of obesity (BMI ≥ 30 kg/m²), and cardiometabolic risk markers at 12 months and, where retention permits, beyond.

Research Questions, Outcome Measures, and Planned Analyses

Each research question is mapped directly to a pre-specified outcome measure and a statistical analysis plan, in accordance with SPIRIT 2013 and CONSORT 2010 extension for cluster trials (Chan et al., 2013; Campbell et al., 2012).

RQ1 (Primary): Does the multi-component community-based intervention reduce BMI at 12 months compared with health-information-only control?

Outcome measure: Change in BMI (kg/m²) from baseline to 12 months, measured by calibrated stadiometer and digital scale using WHO standardised protocols.
Analysis: Linear mixed-effects regression (primary statistical model) with BMI change as the dependent variable, treatment arm as the fixed effect, and cluster as a random intercept to account for within-cluster correlation. Baseline BMI, age, sex, and socioeconomic status will be included as covariates. Intention-to-treat (ITT) principle applied throughout.

RQ2 (Secondary): Which behavioural strategies are most effective in sustaining lifestyle changes, and do self-monitoring and goal-setting mediate BMI reduction?

Outcome measures: (a) Self-reported physical activity (International Physical Activity Questionnaire - IPAQ short form, validated in sub-Saharan Africa); (b) dietary quality score derived from 24-hour dietary recall; (c) self-efficacy assessed by the General Self-Efficacy Scale (Schwarzer & Jerusalem, 1995).
Analysis: Causal mediation analysis (Baron & Kenny approach with bootstrapped confidence intervals) applied to the ITT sample. Mediation analyses are treated as exploratory and are not independently powered for mediation pathways.

RQ3 (Secondary): Do environmental and policy-level supports modify the effect of the intervention on BMI outcomes?

Outcome measures: Intervention effect size (BMI change) stratified by community-level measures: (a) distance to nearest food market offering fresh produce (GPS-measured); (b) availability of safe walking or exercise space (community audit score); (c) presence of workplace or campus wellness policy.
Analysis: Exploratory moderation analysis using cross-level interaction terms (individual BMI change × community environment score) within the mixed-effects model. These analyses are hypothesis-generating and are not independently powered.

RQ4 (Secondary/Exploratory): What is the long-term impact of the intervention on obesity incidence and cardiometabolic markers?

Outcome measures: (a) Incidence of obesity (BMI ≥ 30 kg/m²) at 12 months; (b) waist circumference (cm); (c) fasting plasma glucose (mmol/L); (d) HbA1c (%); (e) fasting lipid profile (total cholesterol, LDL, HDL, triglycerides in mmol/L).
Analysis: Incidence of obesity will be compared between arms using log-binomial regression with cluster-robust standard errors to estimate risk ratios. Continuous cardiometabolic outcomes will use the same mixed-effects regression framework as the primary analysis. These are secondary endpoints; results will be interpreted descriptively.

METHODOLOGY

Study Design

This study will employ a two-arm parallel cluster randomised controlled trial (cluster RCT) design, following the CONSORT 2010 extension for cluster randomised trials (Campbell et al., 2012) and SPIRIT 2013 guidance for trial protocols (Chan et al., 2013).

Justification for Cluster Randomisation

Cluster randomisation was selected for three reasons. First, the intervention operates at the community level through environmental modifications, mass media campaigns, and community health worker (CHW) delivery; these elements cannot be meaningfully provided to individuals within the same community without contaminating the control arm. Second, randomising individuals within the same setting to different arms would create co-intervention bias, as participants share social networks, food environments, and physical infrastructure (Donner & Klar, 2000). Third, cluster designs are well-established in community-based obesity prevention research (Adab et al., 2015; Li et al., 2017), providing a strong methodological precedent for this approach.

Choice of Cluster Unit

Geographically defined community zones within Lusaka Province (e.g., distinct residential compounds, peri-urban wards, or urban neighbourhood units) will serve as the unit of randomisation. These natural geographic clusters were chosen over institutional clusters (e.g., individual workplaces or campuses) because: (a) they capture the full ecological context of the intervention including food retail, built environment, and social norms; (b) they reduce the risk of cluster-level selection bias that arises when clusters are defined by voluntary institutional enrolment; and (c) geographic clusters are a standard unit in community-level NCD prevention trials in sub-Saharan Africa (Willmott et al., 2022). Approximately 24–30 community clusters are planned across intervention and control arms (see Sample Size section).

Allocation Concealment and Randomisation Procedure

Cluster-level randomisation will be conducted by an independent statistician with no involvement in participant recruitment or outcome assessment. Eligible community clusters will be identified and matched on the following variables prior to randomisation: (a) urban/peri-urban classification; (b) approximate socioeconomic profile (tertile of household poverty index derived from Zambia DHS 2024 data); and (c) estimated adult population size.

Within each matched pair, one cluster will be randomly allocated to the intervention arm and the other to the control arm using a computer-generated random number sequence (e.g., R function sample()). The allocation sequence will be held by the independent statistician and released to the field team only after baseline data collection for all clusters in a matched pair is complete, ensuring that baseline assessments are conducted without knowledge of assignment. Cluster allocations will be documented in a secure, password-protected file with a full audit trail.

Blinding

Due to the nature of community-level behavioural interventions, participants and intervention facilitators cannot be blinded to arm assignment. However, the following blinding procedures will apply:

Outcome assessors: Research assistants responsible for anthropometric, biochemical, and questionnaire data collection will be blinded to cluster allocation at all time points - baseline, 6 months, and 12 months. They will not be involved in intervention delivery and will be explicitly instructed not to enquire about or record arm assignment. Blinding will be verified at each follow-up visit by asking assessors to record their guess of cluster assignment; concordance with actual assignment will be reported in the trial publication.
Data analysts: The primary outcome analysis will be conducted on a blinded dataset (arms coded as A and B) until the statistical analysis plan is finalised and signed off, at which point the allocation key will be released (Schulz & Grimes, 2005).

Study Setting

Selection of Lusaka and Justification

The study will be conducted in Lusaka Province, the capital and most populous province of Zambia (population approximately 3.7 million as of 2022 Census projections). Lusaka was selected for the following reasons:

Epidemiological burden: Lusaka has the highest urban concentration of overweight and obesity in Zambia. Data from the 2024 Zambia Demographic and Health Survey (ZDHS 2024) indicate that among urban women aged 20–49, the combined prevalence of overweight and obesity substantially exceeds rural equivalents (urban approximately 35–40%; rural approximately 15%). Community-based surveys in Lusaka have previously recorded an overall obesity prevalence of 14.2% (5.1% in men, 18.6% in women), with overweight rates adding a further 20–25% of the population (Rudatsikira et al., 2012). Among women aged 20–29 specifically, the ZDHS 2024 shows that overweight/obesity prevalence is rising compared with the 2013–14 survey, signalling an accelerating transition in young adults.
Double burden context: Zambia faces a nutritional double burden, with undernutrition persisting alongside rising rates of overweight and obesity, particularly in urban settings. Lusaka is the epicentre of this transition and therefore offers high scientific and policy relevance (Global Nutrition Report, 2024). National data indicate that 14.7% of adult women and 4.4% of adult men are living with obesity nationally, with urban prevalence markedly higher.
Infrastructure and feasibility: Lusaka has an established network of urban health centres, Community Health Assistants (CHAs, who are formally trained and government-registered), and NGO health platforms that can support CHW training and programme delivery. The city hosts the University of Zambia School of Public Health and the Zambia National Public Health Institute (ZNPHI), providing academic and logistical infrastructure for the trial.
Research environment: Lusaka has hosted previous community-based NCD intervention research, providing institutional familiarity with research ethics procedures at the University of Zambia Biomedical Research Ethics Committee (UNZABREC) and the ZNPHI research governance pathway.

Population Demographics and Number of Communities

Lusaka Province is predominantly urban (approximately 80% urban residence), with a high proportion of young adults: the 15–34 age group constitutes over 40% of the provincial population (Zambia Statistics Agency, 2022). The province is linguistically diverse, with Nyanja and English as the primary languages of urban communication. Socioeconomic heterogeneity is marked, ranging from high-income formal residential areas to densely populated low-income compounds (e.g., Mtendere, Kanyama, Chaisa).

Approximately 24–30 geographically defined community clusters are planned across the two arms (12–15 per arm). Each cluster will be a defined neighbourhood or ward unit with an estimated catchment of 500–1,000 young adults aged 18–29 years, from which a target sample of 35–50 eligible participants per cluster will be recruited. The final number of clusters will be confirmed following a pre-trial community mapping exercise and feasibility assessment.

Target Population and Inclusion Criteria

Clarification of Study Aim and BMI Eligibility

This study is designed as a combined prevention and early intervention trial addressing both primary prevention (preventing normal-weight individuals from transitioning to overweight) and secondary prevention (halting or reversing the progression from overweight to obesity). This dual framing directly resolves the apparent inconsistency in the previous protocol between the stated target population (BMI ≥ 25 kg/m²) and the inclusion criterion (BMI ≥ 18.5 kg/m²).

The rationale for including normal-weight individuals (BMI 18.5–24.9 kg/m²) is threefold: (a) the intervention is community-delivered and therefore operates at population level across a defined geographic cluster, making selective screening by BMI category logistically impractical; (b) young adulthood (18–29 years) is a period of rapid weight gain, and intervening before BMI trajectories become established offers the greatest long-term prevention benefit (Wong et al., 2020); and (c) a population-level community approach to NCD prevention is consistent with WHO guidance recommending that upstream interventions target whole communities rather than individuals already meeting clinical thresholds (WHO, 2000). The study is therefore best characterised as a combined primary and secondary prevention trial.

Justification for the 18–29 Year Age Group

The 18–29 year age group was selected for three reasons:

High-risk transition period: Cross-sectional and longitudinal data consistently show that mean BMI increases most steeply between ages 18 and 35, driven by changes in diet, physical activity, occupational patterns, and social environment during the transition from adolescence to adulthood (NICE, 2024; WHO, 2000).
Underserved group: Existing weight management services in Zambia and most low- and middle-income countries are primarily oriented towards middle-aged and older adults with established cardiometabolic disease. Young adults are rarely targeted by dedicated obesity prevention programming, representing a gap in public health provision.
Life-course leverage: BMI established by age 25–30 is a strong predictor of cardiometabolic disease risk, premature mortality, and quality of life across the life course (Wong et al., 2020). Intervening before these trajectories consolidate offers maximum long-term population health benefit.

Inclusion Criteria

Age 18–29 years (inclusive) at enrolment.
BMI ≥ 18.5 kg/m² at baseline screening (individuals with BMI < 18.5 kg/m² will be excluded as the intervention is not appropriate for underweight individuals).
Residing in, or regularly present in (defined as ≥ 4 days per week), a selected cluster community for at least 6 months prior to enrolment.
Able to provide written informed consent in English or Nyanja.
Willing to participate for the full 12-month study duration.

Exclusion Criteria

Currently pregnant or breastfeeding at baseline. (See management note below for pregnancies occurring during follow-up.)
Diagnosed severe eating disorder (anorexia nervosa, bulimia nervosa, or binge eating disorder) currently under active treatment.
Medical condition that precludes participation in moderate physical activity without direct medical supervision (e.g., uncontrolled cardiac disease, severe respiratory disease, or musculoskeletal condition assessed by a clinician as precluding exercise).
BMI < 18.5 kg/m² (underweight) at baseline screening.
Currently enrolled in a separate structured weight management programme or clinical trial.
Planning to relocate outside the study area within 12 months.
Any condition that, in the opinion of the principal investigator, would compromise participant safety or data integrity.

Management of Pregnancy During Follow-Up

Participants who become pregnant during the 12-month follow-up period will not be withdrawn from the study. Management will proceed as follows: (a) dietary and physical activity coaching will be adapted to pregnancy-safe guidelines by a qualified dietitian or midwife; (b) anthropometric measurements will continue to be collected but pregnancy-period measurements will be flagged in the dataset; (c) flagged measurements will be excluded from the primary per-protocol analysis but retained in the ITT analysis with last-observation-carried-forward imputation; (d) a pre-specified sensitivity analysis will be conducted excluding all participants who became pregnant. This approach ensures participant safety, maintains retention, and allows transparent reporting of any differential pregnancy rates between arms.

Sample Size Calculation

Primary Endpoint and Effect Size

The primary endpoint is change in BMI (kg/m²) from baseline to 12 months. Based on evidence from community-based obesity prevention trials in young adults and comparable sub-Saharan African settings, a minimum clinically important difference of 0.5 kg/m² is considered meaningful (WHO, 2000), with an assumed standard deviation of 2.5 kg/m² for BMI change.

Calculation Parameters

Effect size: 0.5 kg/m² difference in mean BMI change between arms (SD = 2.5 kg/m²).
Power: 80%.
Two-sided significance level: α = 0.05.
Intraclass correlation coefficient (ICC): 0.02 (based on published values from community-based obesity and lifestyle intervention trials; Gray et al., 2016; Donner & Klar, 2000). A sensitivity calculation using ICC = 0.05 will also be presented to examine the effect of a higher ICC on required sample size.
Average cluster size (m): 40 participants per cluster (based on community mapping estimates).
Design effect (DEFF): DEFF = 1 + (m - 1) × ICC = 1 + (39 × 0.02) = 1.78.
Individual-level sample size (ignoring clustering): n = 2 × [(z_α/2 + z_β)² × σ² / δ²] = 2 × [(1.96 + 0.842)² × 6.25 / 0.25] ≈ 394 participants total (197 per arm).
Cluster-adjusted sample size: 394 × 1.78 ≈ 701 participants.
Attrition adjustment: Assuming 20% dropout over 12 months: 701 / 0.80 ≈ 877 participants.
Number of clusters required: 877 / 40 ≈ 22 clusters (11 per arm). To provide an additional power margin and account for possible cluster dropout, 24–30 clusters (12–15 per arm) are planned.

Final Planned Sample

Target enrolment: 960–1,200 participants across 24–30 clusters (minimum 800 for 80% power; 1,200 provides a buffer against higher-than-anticipated attrition or ICC). Final parameters will be confirmed using simulation methods in R (package clusterPower) or Stata (xtmixed-based simulation) prior to trial registration. All parameters will be pre-specified in the trial registration record and final protocol.

Intervention Components

The multi-component intervention follows a Social-Ecological Model framework (Stokols, 1996), targeting determinants of obesity at individual, interpersonal, organisational, and community levels. The intervention is described in accordance with the Template for Intervention Description and Replication (TIDieR) checklist (Hoffmann et al., 2014).

TIDieR Item 1: Brief Name

Community-Based Multi-Component Obesity Prevention Intervention (CMOPI-ZM).

TIDieR Item 2: Why - Theoretical Rationale

The intervention is grounded in Social Cognitive Theory (Bandura, 1986) and the Behaviour Change Wheel (Michie et al., 2011), targeting the mechanisms of self-efficacy, outcome expectancies, social reinforcement, and environmental opportunity. Obesity in young adults in urban Lusaka is driven by a combination of individual behaviour, social norms, and obesogenic environmental features (easy availability of ultra-processed foods; limited safe spaces for physical activity). The multi-level approach addresses all three domains simultaneously.

TIDieR Item 3: Materials

Printed and illustrated dietary guidance materials available in English and Nyanja.
A mobile health (mHealth) application to support self-monitoring of diet, physical activity, and goal progress. This application does not currently exist and will be developed prior to the trial as a purpose-built tool using open-source frameworks (e.g., React Native). A 3-month pilot feasibility phase will test usability and acceptability before full trial launch. The app will be Android-compatible and will include an offline-capable version for low-connectivity settings. A paper-based equivalent (goal-tracking diary) will be provided to participants without smartphone access.
Group session facilitator manuals and fidelity checklists.
Community environment audit tools and poster materials for community-level campaigns.

TIDieR Item 4: Procedures - Individual Level

Content: Personalised dietary coaching (portion control, reducing ultra-processed food consumption, increasing fruit and vegetable intake); physical activity goal-setting; self-monitoring of diet and activity via mobile app or paper diary; motivational interviewing techniques.
Frequency and duration: 12 individual coaching contacts over 12 months (monthly), each lasting approximately 30 minutes.
Mode of delivery: In-person at community meeting points or health posts, supplemented by SMS reminders and in-app notifications.

TIDieR Item 5: Procedures - Interpersonal Level

Content: Monthly peer support group sessions (groups of 8–12 participants), including cooking demonstrations, group physical activity sessions (e.g., walking groups, group aerobics), and peer-led discussions on behaviour change and social norms.
Frequency and duration: Monthly sessions of 60–90 minutes over 12 months (12 sessions total); supplementary informal peer contact facilitated through a closed community messaging group.
Mode of delivery: In-person at community venues (churches, community halls, open spaces); facilitated by trained peer educators drawn from within the community.

TIDieR Item 6: Procedures - Organisational and Environmental Level

Content: Partnerships with local food markets to improve healthy food access (subsidised fruit and vegetable stalls or point-of-choice labelling); partnerships with local gyms, recreational facilities, and employers/schools to provide low-cost or free access to physical activity spaces; community environment audits to identify and advocate for improvements (e.g., safe walking routes).
Frequency and duration: Environmental partnerships established at trial start and maintained throughout 12 months; formal review of environmental changes at 6 and 12 months.
Mode of delivery: Multi-sector coordination meetings with community leaders, market operators, and workplace/campus managers; written memoranda of understanding where applicable.

TIDieR Item 7: Procedures - Community Level

Content: Mass media campaigns (community radio, posters, social media); monthly community health events with free anthropometric screening and nutrition education; policy advocacy through community coalitions targeting local government food and environment policies.
Frequency and duration: Monthly community health events; radio campaign aired weekly for the first 6 months and fortnightly thereafter; policy advocacy activities ongoing throughout the trial.
Mode of delivery: Community radio broadcasts, printed and digital posters, and community coalition meetings led by CHWs and community leaders.

TIDieR Items 8 and 9: Who Provides - Training of Community Health Workers and Peer Educators

Trained Community Health Assistants (CHAs, formally trained and government-registered in Zambia) will lead individual coaching and community-level activities. Peer educators (lay community members) will facilitate group sessions and peer support activities. All CHWs and peer educators will complete a standardised 5-day pre-trial training programme covering:

Behaviour change communication techniques, including motivational interviewing.
Nutrition education and dietary coaching.
Physical activity promotion and group facilitation.
Use of the mobile application and paper-based self-monitoring tools.
Adverse event recognition and participant referral procedures.
Ethical principles, including confidentiality and non-stigmatising communication about weight.

Competency will be assessed by structured role-play assessment before deployment. A 1-day refresher training will be provided at month 6. CHW performance will be monitored monthly through supervision visits by a trial coordinator.

TIDieR Items 10–12: Fidelity, Tailoring, and Modifications

Intervention fidelity will be monitored through: (a) session attendance logs; (b) session delivery checklists completed by CHWs after each contact; (c) bi-monthly supervision visits with direct observation of a randomly selected 10% of sessions; (d) mobile app usage logs (frequency of logins and goal entries). Minor adaptations to content may be made for individual health profiles (e.g., co-existing hypertension) but the core four-level component structure will remain consistent across all clusters. All protocol deviations will be documented and reported.

Comparator (Control Arm)

Control clusters will receive usual care: a single provision of standard public health information leaflets on healthy eating and physical activity (equivalent to materials currently available at Zambia Ministry of Health community health posts), with no structured multi-component support. No active coaching, peer support groups, environmental modifications, or application access will be provided. The information will be delivered by a CHW in a single session of approximately 15 minutes at enrolment, with no follow-up contacts.

Minimising Contamination Between Arms

The risk of contamination between intervention and control communities is directly addressed through geographic cluster randomisation. Additional measures include:

Communities will be selected to be spatially separated by a minimum inter-cluster buffer distance of 2 kilometres (confirmed during pre-trial community mapping).
Clusters will be selected to minimise shared social or commercial spaces (e.g., the same market or workplace should not span both an intervention and a control cluster).
Participants will be advised at enrolment not to share intervention materials with individuals outside their cluster; this will be reinforced at each contact.
Contamination will be monitored through a process evaluation survey at 6 and 12 months asking control arm participants whether they received or sought out additional dietary or physical activity support beyond usual health services.
A pre-specified sensitivity analysis excluding any clusters identified as contaminated will be conducted and reported alongside the primary analysis.

Outcomes

Primary Outcome

Change in BMI (kg/m²) from baseline to 12 months. BMI is selected as the single primary outcome because it is a continuous, internationally standardised measure of adiposity with established clinical cut-points, it is sensitive to change over a 12-month period, and it directly underpins the sample size calculation. Waist circumference is classified as a secondary outcome rather than co-primary, given the practical advantage of a single pre-specified primary endpoint for trial registration and regulatory clarity.

Justification for selecting BMI as the primary outcome: Although waist circumference may be a more sensitive marker of cardiometabolic risk at a given BMI (particularly in African populations where adiposity distribution differs from European reference populations), BMI was selected as the primary endpoint because: (a) it is the most widely used endpoint in comparable multi-component obesity prevention trials, facilitating future meta-analytic comparison; (b) standardised BMI measurement has lower inter-observer variability than waist circumference in field settings; and (c) it directly links to the study's prevention framing through established BMI classification thresholds (WHO, 2000).

Secondary Outcomes

Waist circumference (cm): Measured at the midpoint between the lower costal margin and iliac crest using a non-elastic tape measure, participant standing in light clothing (WHO protocol).
Physical activity level: IPAQ short form (7-day self-report recall), validated and widely used in sub-Saharan Africa. Accelerometry (wrist-worn device) will be used in a 20% subsample for objective validation.
Dietary intake: Interviewer-administered 24-hour dietary recall (multiple-pass method) at each time point, analysed using the Zambia-specific food composition database or FAO/INFOODS. A validated food frequency questionnaire (FFQ adapted for Zambian dietary patterns) will supplement this at 12 months.
Behavioural mediators: Self-efficacy assessed using the General Self-Efficacy Scale (GSE-10; Schwarzer & Jerusalem, 1995); social norm perceptions assessed by a 5-item adapted scale.
Health-related quality of life: Short Form-12 (SF-12; Ware et al., 1996), validated for use in sub-Saharan African contexts, administered at baseline and 12 months.
Mental health: Patient Health Questionnaire-9 (PHQ-9) for depression screening and the Generalised Anxiety Disorder 7-item scale (GAD-7) for anxiety, both of which have demonstrated validity in Zambian and broader African populations. Administered at baseline, 6 months, and 12 months.
Cardiometabolic markers: Fasting venous blood sample for HbA1c (%), fasting plasma glucose (mmol/L), total cholesterol (mmol/L), LDL-cholesterol (mmol/L), HDL-cholesterol (mmol/L), and triglycerides (mmol/L). Collected at baseline and 12 months only.
Incidence of obesity: New-onset BMI ≥ 30 kg/m² among participants with baseline BMI < 30 kg/m².

Process Measures

Reach: Proportion of eligible community members enrolled.
Dose delivered: Number and proportion of planned sessions completed per participant.
Acceptability: Post-intervention satisfaction survey (5-point Likert scale).
App engagement: Frequency of logins and goal entries per participant per month.

Data Collection

Timeline and Sources

Data will be collected at three time points: baseline (T0), 6 months (T1), and 12 months (T2). Biochemical measures will be collected at T0 and T2 only to minimise participant burden.

Anthropometric: Height (calibrated stadiometer, Seca 213), weight (calibrated digital floor scale, Seca 876), and waist circumference (non-elastic tape measure). All measurements conducted in duplicate; the mean of two readings is used; a third measurement is taken if duplicates differ by more than 0.5 kg (weight) or 1 cm (waist circumference).
Biochemical: Fasting venous blood drawn by a trained phlebotomist. Samples processed within 2 hours at a certified laboratory (ZNPHI or University Teaching Hospital laboratory, Lusaka).
Questionnaire data: Administered by trained research assistants using electronic data capture (Open Data Kit, ODK) on tablet devices. Paper backup available in case of device failure.
App usage data: Automatically logged server-side with daily extraction to the study database.

Assessor Training and Standardisation

All research assistants conducting outcome assessments will attend a 2-day standardisation training prior to data collection, covering: (a) anthropometric measurement protocols following WHO methods, including duplicate measurement procedures and equipment calibration checks; (b) phlebotomy and sample handling procedures; (c) questionnaire administration using scripted procedures to minimise interviewer bias; (d) data entry using ODK including use of built-in validation rules. Intra-observer and inter-observer technical error of measurement (TEM) will be calculated during training; assessors with TEM exceeding the acceptable threshold will receive additional training before being authorised to collect data. Equipment calibration will be checked at the start and end of each data collection day.

Blinding of Assessors

As described under Study Design, outcome assessors will remain blinded to cluster allocation at baseline, 6-month, and 12-month follow-up visits. Assessors will not be involved in intervention delivery and will be deployed to assigned clusters only. If an assessor inadvertently learns the assignment of a cluster, this will be documented as a protocol deviation and a different assessor will be deployed for subsequent visits to that cluster where feasible.

Data Management, Storage, and Quality Assurance

Electronic data will be managed using REDCap (Research Electronic Data Capture) hosted on a secure server at the University of Zambia or a Zambian data protection-compliant cloud server. Access will be restricted to named trial team members by role-based permissions. All data will be encrypted in transit and at rest. Paper records (where used) will be stored in locked cabinets at the trial coordinating centre. Data quality checks built into ODK (skip logic, range checks, mandatory fields) will flag potential errors at the point of entry. A dedicated data manager will perform weekly data quality audits including checks for duplicate entries, out-of-range values, and missing data patterns. An anonymised dataset will be used for all analyses; participant identifiers will be held separately by the trial manager.

Data Analysis

Primary Statistical Model

The primary analysis will use a linear mixed-effects regression model (LMER) implemented in R (lme4 package) or Stata (mixed command). The model specification is:

BMI_change_ij = β0 + β1(Treatment_j) + β2(Baseline_BMI_ij) + β3(Age_ij) + β4(Sex_ij) + β5(SES_ij) + u_j + ε_ij

Where i indexes individuals and j indexes clusters; u_j is the cluster-level random intercept capturing within-cluster correlation; ε_ij is the individual-level residual. The coefficient β1 represents the estimated treatment effect - the mean difference in BMI change between intervention and control arms - adjusted for the pre-specified covariates. A random slope for time will be added if model fit statistics (AIC/BIC) support this.

Intention-to-Treat Analysis

The primary analysis will follow ITT principles: all participants enrolled and randomised (at the cluster level) will be included regardless of adherence to the intervention or completeness of follow-up data. A per-protocol sensitivity analysis will be conducted restricting to participants who attended at least 75% of scheduled intervention contacts.

Incorporating Clustering and Repeated Measures

The mixed-effects model accounts for within-cluster correlation through the random cluster intercept. Repeated measures at 6 and 12 months will be incorporated using a longitudinal extension of the LMER model, with time as a fixed effect and a random intercept-and-slope structure at the participant level nested within clusters. Generalised Estimating Equations (GEE) with an exchangeable correlation structure will be used as a sensitivity analysis to verify robustness of the primary result.

Adjustment Variables

The following variables will be included as fixed covariates in the primary model, selected a priori based on known determinants of BMI change and potential confounders: age (continuous), sex (binary), baseline BMI (continuous), and socioeconomic status (standardised household asset index score derived from the baseline questionnaire, consistent with the Zambia DHS wealth quintile approach).

Handling of Missing Data

Missing data will be investigated for patterns (missing completely at random [MCAR], missing at random [MAR], or missing not at random [MNAR]) using Little's MCAR test and visual inspection of missingness patterns. The primary analysis will use multiple imputation by chained equations (MICE) assuming MAR, with 50 imputed datasets pooled using Rubin's rules. Imputation models will include all analysis variables plus auxiliary variables predictive of missingness (e.g., attendance records, baseline health status). Complete-case analysis will be presented as a sensitivity analysis. Where MNAR is suspected (e.g., differential dropout related to BMI trajectory), a further sensitivity analysis using pattern-mixture models will be conducted.

Subgroup Analyses

Pre-specified exploratory subgroup analyses will examine treatment effect modification by: (a) sex (male vs. female); (b) baseline BMI category (normal weight 18.5–24.9 vs. overweight 25.0–29.9 vs. obese ≥ 30 kg/m²); and (c) socioeconomic status (lower two quintiles vs. upper three quintiles). These analyses are explicitly exploratory, hypothesis-generating, and not independently powered. Effect estimates will be presented with confidence intervals and interaction p-values and interpreted descriptively. A Bonferroni correction will be applied across the three pre-specified subgroup interactions to reduce the risk of false positives.

Secondary Outcome Analyses

Binary outcomes (incidence of obesity) will be analysed using log-binomial regression with cluster-robust standard errors to estimate risk ratios. Count or ordinal behavioural outcomes will use generalised linear mixed models (GLMM) with appropriate link functions. Process measures will be reported descriptively. Statistical significance will be set at p < 0.05 (two-tailed) for all primary analyses; secondary and subgroup analyses will be interpreted in the context of both effect size and clinical meaning. All analyses will be conducted in R (version ≥ 4.4) or Stata (version ≥ 18). The statistical analysis plan (SAP) will be finalised, registered, and locked prior to unblinding of outcome data.

Adverse Events and Safety Monitoring

Definitions

Adverse Event (AE): Any untoward medical occurrence in a participant during the study period, regardless of causal relationship to the intervention. Examples include exercise-related musculoskeletal injuries, psychological distress, hypoglycaemic episodes, or cardiovascular events.
Serious Adverse Event (SAE): An AE that results in death, is life-threatening, requires hospitalisation or prolongation of existing hospitalisation, results in persistent or significant disability, or is deemed medically significant by the principal investigator. Examples include acute myocardial infarction, severe exercise-induced injury requiring hospital admission, or a severe depressive episode requiring psychiatric care.
Adverse Event Related to Intervention: An AE judged by the principal investigator as possibly, probably, or definitely related to study participation or the intervention activities.

Reporting Procedures

All AEs will be recorded by CHWs or research assistants on a standardised AE reporting form at each participant contact and whenever a participant presents with a concern. SAEs must be reported to the principal investigator within 24 hours of the trial team becoming aware of the event. SAEs will subsequently be reported to UNZABREC and ZNPHI within 7 calendar days (or 15 days for non-fatal, non-life-threatening events) in accordance with Zambian research ethics regulations. A serious unexpected suspected adverse reaction (SUSAR) will be reported to all relevant regulatory authorities within 7 days.

Monitoring of Specific Risk Areas

Exercise-related injuries: All participants will be screened at baseline for conditions that could increase exercise injury risk using the Physical Activity Readiness Questionnaire (PAR-Q). Physical activity goals will be set at low-to-moderate intensity (e.g., 150 minutes per week of brisk walking, in line with WHO guidelines), escalating only where clinically safe to do so. CHWs will be trained to identify musculoskeletal injury symptoms and refer participants to the nearest health facility promptly.
Psychological distress: Mental health screening will be conducted at baseline and 12 months using the PHQ-9 and GAD-7. Participants who screen positive for moderate-to-severe depression (PHQ-9 ≥ 10) or moderate-to-severe anxiety (GAD-7 ≥ 10) at any time point will be referred to a mental health nurse or counsellor via the health facility referral pathway. A list of local mental health resources (community counsellors, ZNPHI mental health helpline) will be provided to all participants at enrolment.
Weight-related distress and disordered eating: All CHWs will be trained in non-stigmatising, weight-inclusive communication. Participants showing signs of disordered eating cognition (identified through regular monitoring during coaching sessions) will be referred to a dietitian. Any new diagnosis of an eating disorder during the study will be treated as an SAE.

Data Safety and Monitoring Committee

An independent Data Safety and Monitoring Committee (DSMC) will be established, comprising: (a) an independent biostatistician; (b) a clinician with expertise in NCD prevention; and (c) a community representative or lay member. The DSMC will review unblinded accumulating outcome and safety data at 6 months (after approximately 50% of primary outcome data have been collected) and will report to the trial steering committee. The trial will be stopped early for harm if there is evidence that the intervention arm is associated with a significantly higher rate of SAEs than the control arm (p < 0.01, two-sided, unadjusted), or if an imbalance in severe adverse events raises unacceptable safety concerns. The DSMC charter will be finalised and agreed prior to trial commencement.

Ethical Considerations

Ethical Approval

Ethical approval will be sought from the University of Zambia Biomedical Research Ethics Committee (UNZABREC) prior to any recruitment or data collection. In accordance with the Zambia National Health Research Act and ZNPHI research governance procedures, the protocol will also be registered with the National Health Research Authority. The trial will be registered prospectively on the Pan African Clinical Trials Registry (PACTR) or ClinicalTrials.gov, and the registration number will be cited in all publications and communications.

Informed Consent

Written informed consent will be obtained from all participants prior to enrolment. Consent forms will be available in English and Nyanja and will be reviewed using a "teach-back" method to verify comprehension. Consent will cover the study purpose and procedures, the voluntary nature of participation, the right to withdraw at any time without penalty or effect on access to health services, data confidentiality arrangements, and contact information for the principal investigator and UNZABREC. Where potential participants have low literacy, the consent process will be conducted verbally in the presence of a literate witness, with a thumbprint signature.

Community Engagement and Cultural Sensitivity

Community entry will be preceded by engagement with community leaders (Neighbourhood Health Committees, local councillors, and chiefs where applicable) to obtain community-level assent before individual participant recruitment begins. Weight-related discussions will use non-stigmatising, culturally appropriate language. Given the documented association between food insecurity and overweight in Zambian urban settings, intervention messaging will be framed around positive health behaviours and wellbeing rather than body weight per se.

Management of Abnormal Screening Findings

Participants identified at baseline screening as having significantly abnormal findings (e.g., fasting glucose ≥ 7.0 mmol/L, blood pressure ≥ 160/100 mmHg, or BMI ≥ 40 kg/m² with comorbidities) will be referred promptly to the nearest primary health care facility with a standardised referral letter. These participants will not be excluded from the study unless their condition warrants exclusion on safety grounds. A referral log will be maintained, including documentation of whether the participant subsequently attended the referral facility.

Participant Reimbursement

Participants will receive reimbursement for time and transport costs associated with each study visit (baseline, 6-month, and 12-month assessments). Reimbursement will be set at a level approved by UNZABREC and consistent with Zambian guidelines for research participant reimbursement - sufficient to cover reasonable transport costs without being coercive. No conditional or performance-related payments will be made.

Data Protection and Confidentiality

All data will be treated in strict confidence. Participant data will be pseudonymised using a unique study ID; identifiable information will be held separately by the trial manager and not shared with the analysis team. Data storage and processing will comply with the Zambia Data Protection Act 2021. Data will be retained for a minimum of 10 years after study completion per UNZABREC and institutional requirements, then securely destroyed.

Dissemination Plan

Academic Dissemination

Results will be submitted for publication in peer-reviewed, open-access journals with broad global and sub-Saharan African public health readership (e.g., PLOS Medicine, The Lancet Global Health, BMC Public Health, or Obesity Reviews). Both positive and null results will be submitted in line with the AllTrials commitment to reduce publication bias. Primary results (12-month BMI outcomes), process evaluation findings, and secondary outcome data will each be reported in separate publications. Authorship will follow ICMJE guidelines and will include community representatives and CHW supervisors where their contribution meets authorship criteria.

Findings will be presented at relevant international conferences (e.g., International Congress on Obesity, European Congress on Obesity, Africa Health Conference) and national Zambian research forums (e.g., ZNPHI Annual Research Symposium, Zambia National Health Research Conference).

Policy and Stakeholder Dissemination

Policy briefs summarising key findings and actionable recommendations will be prepared for: (a) the Zambia Ministry of Health Non-Communicable Diseases Unit; (b) the Lusaka Province Health Office; and (c) local government authorities in the study area. Stakeholder dissemination workshops will be held in Lusaka at the close of the trial, bringing together MOH representatives, community leaders, NGO partners, and the trial team to discuss findings, implications, and scale-up considerations. A plain-language stakeholder report will be submitted to the Ministry of Health and made publicly available on the study website within 3 months of trial completion.

Community and Participant Dissemination

Community dissemination events will be held within each participating cluster at the end of the trial to present aggregated findings to community members in accessible formats (oral presentations, visual summaries, community radio broadcasts). Plain-language participant summaries of study findings, available in English and Nyanja, will be distributed to all enrolled participants. These activities fulfil the ethical obligation to return findings to communities that contributed their time and data, and reinforce the study's community partnership model.

Influence on Future Public Health Policy

The trial is designed to generate evidence directly applicable to Zambian and wider sub-Saharan African obesity prevention policy. If the intervention demonstrates effectiveness, findings will be used to: (a) advocate for inclusion of community-based young adult obesity prevention programmes in the Zambia National NCD Policy and Strategic Plan; (b) inform the design of a national scale-up programme with particular attention to integrating delivery within the existing CHW workforce; and (c) contribute to WHO guidance on community-based NCD prevention in low- and middle-income country settings. The trial team will engage proactively with the ZNPHI Health Policy Evidence Unit throughout the trial to facilitate timely translation of findings into policy guidance. A five-year follow-up observational study will be proposed if resources permit, to assess the durability of intervention effects and long-term cost-effectiveness.

Data Sharing

Anonymised individual participant data will be made available to qualified researchers on reasonable request following publication of the primary results, in accordance with UNZABREC data sharing provisions and the Zambia Data Protection Act 2021. A data sharing statement and formal request procedure will be included in all publications.

All sections above incorporate the requested revisions. The downloadable Word document is also available at the link provided in the previous message.

This is a shared conversation. Sign in to Orris to start your own chat.