Cities for Better Health Childhood Obesity Prevention Initiative

Requirements for interventions:

The interventions in each country will be multi-level multi-component (MLMC) community-based packages. Their exact features will be tailored to the needs and opportunities of each participating country.

Requirements for study design:

The minimum evaluation study design in terms of robustness is a repeated cross-sectional study with a comparison group. This involves measuring outcomes in a cross-section of children living in the intervention areas of a participating city (i.e. the areas in which the CBH COPI interventions were implemented) and a cross-section of children in similar comparison areas in which the interventions were not implemented, before and after the interventions are introduced. The global analysis will then compare trends in outcomes between intervention- and comparison-group areas. Countries are allowed to implement more robust designs (e.g. incorporating randomisation) if this is feasible.

Each country-level study will collect a set of prespecified indicators at baseline and at two subsequent follow-ups (after 10-12 months, and after 20-24 months). The primary outcome indicator is BMI z-score and the required secondary and exploratory outcomes are defined above. A list of sociodemographic indicators (age, sex, ethnicity, household affluence) is also required to be collected (unless this is not feasible in an individual country for legal reasons).

It is anticipated that 2,000 children in the target population will be recruited per data collection point (baseline, first follow-up, second follow-up) in each country. It is recommended (as the default option) that data are collected from 20 'clusters' (schools) from at least 10 'neighbourhoods' (administrative areas or school catchment areas, depending on the country) over these time points. This set up gives a minimum detectable effect size of Cohen's d = 0.15 in country-level analyses. The sample of neighbourhoods/clusters will be chosen according to feasibility and representativeness considerations in each country.

Local evaluations will be adapted to fit with the social, cultural and policy landscape of each implementing site.

To be eligible to be involved in data collection, children must attend the 'clusters' (schools) that are part of the study. In each cluster, country evaluation teams will try to ensure that the sample of children selected for data collection is representative of the total population in that cluster (e.g. the full list of students ages 6-13 who go to that school) - in particular, with respect to age and sex.

All countries will also use the inclusion and exclusion criteria below for children in the study:

Inclusion criteria

• Informed consent obtained by the child's parent or legal guardian before any study-related activities (where study-related activities are defined as any procedure related to recording of data according to the protocol). The specific consent form and procedures for obtaining consent will be outlined by the local project partner.
• Aged 6-13 years (both 6 and 13 allowed) at the time of signing consent. Deviations from or exceptions to this procedure must be agreed between the research team in the relevant country and the GIP, and are subject to the relevant local ethical amendment procedure. The exception must be due to a scientifically sound rationale and must further consider the following guiding principles: "it must be measurable, it must have marginal utility, it must be meaningful (i.e. linked to the ultimate impact and the primary and/or secondary objective)".

Exclusion criteria

• Children will be excluded from the study if their parent(s) do not provide fully informed consent for their data to be collected, if during the measurement day they do not agree to participate, or if the child's parent(s) or legal guardian(s) withdraw their consent for their child participating in the research at any point in line with processes described in local informed consent forms.