Both internal and external stakeholder will be recruited. Internal stakeholders will be adolescents, teachers, and parents.

For adolescents, the inclusion criteria will be aged between 10 and 19 years, studying at any of the selected secondary schools, and written informed consent for participating in the study, and if under 18 years, written informed consent from their parents or appointed guardians as well.

If parents are illiterate the researcher will verbally read the participant information sheet and consent form to them and take verbal consent for each item on the consent form. Their statement/s and verbal consent will be witnessed by the local supervisor or a research assistant.

For teachers, the inclusion criteria will be currently employed at one of the selected secondary schools and having completed the two years of probation period.

For parents/appointed guardians, the inclusion criteria will be living with an adolescent studying at one of the selected secondary schools.

External stakeholders will be included based on their expertise and experience.

We will recruit policy actors who are involved in designing or implementing policies on healthy eating among adolescents in Bangladesh and experts on healthy eating among school-age adolescents working in non-government organisations (NGOs), international non-government organisations (INGOs), schools, and health providers.

All the stakeholders will need to provide written informed consent to participate in the study.

To recruit internal stakeholders, we will approach the head teachers of the selected schools.

After seeking their permission flyers about the study will be posted on school notice boards and presented at school events and communication channels.

Stakeholders willing to join the workshops will be asked to contact the author (NS) directly. Key external stakeholders known to the study team will be approached via phone calls and emails by two authors (NS and FA).

Snowball sampling techniques will be used thereafter to identify other external stakeholders.

The co-design process (Figure 1) will be informed by the four stages of the Trochim’s concept mapping approach (18).

All workshops will be conducted in-person.

In stage 1 and stage 2, the workshops will be conducted separately with the adolescent boys, adolescent girls and the adult stakeholders, while in stage 4, all stakeholder groups will be brought together in the workshop.

We will conduct 1h 30m 0s workshops with the stakeholders that will involve brainstorming around the theme of promoting healthy eating among adolescents in urban Bangladesh.

To guide the discussions, we will use a single focus prompt “The intervention in school that will promote healthy eating among adolescents in urban Bangladesh is…” and the supporting prompt questions, such as “What will be the intervention component(s)? and “How do you suggest delivering the intervention?”.

To stimulate the participants' ideas, we will share findings from our exploratory study, which highlights the drivers of unhealthy eating among the adolescents. Hence, the aim of the workshop will be to generate statements regarding school-based healthy eating interventions for adolescents in urban Bangladesh.

This workshop format will be replicated three times.

We will conduct two single-sex workshops for adolescent boys and girls, while combining adult stakeholders such as parents, teachers, and external contributors in another workshop.

Each workshop will involve 10-15 stakeholders (18), and given that we are planning three workshops, we anticipate approximately 45 participants in total (15 adolescent girls, 15 adolescent boys, and 15 adult stakeholders).

Moreover, we will conduct individual discussions with external stakeholders who will consent to participating in the study but may not be able to join the workshops in person.

Their statements will be integrated with the statements from workshop with the adult stakeholders. As a result, we will have three distinct sets of statements derived from the three workshop sessions.

We will clean the statements (18), by editing them for clarity and by eliminating duplicates.

The research team will validate the final statements which will then be used for the subsequent sorting and rating exercises during the second workshop.

The aim of the second workshop will be to sort and rate the statements. The same workshop will be repeated three times similar to the first workshop.

Each of the three stakeholder groups will be provided with their corresponding sets of statements.

Each workshop will be for 1h 30m 0s and will consist of two segments: sorting and rating the statements.

Sorting :

Stakeholders will be given sticky notes, pens, and a large chart paper to sort statements. They will consider statement pertaining to intervention components and delivery modes separately.

Stakeholders will be asked to group statements that are similar and give this group or category a label.

They will be instructed that each statement can only belong to one category and grouping all statements in one category is not allowed nor having a category with just one statement (18).

Before proceeding to the rating activity, stakeholders can discuss and add additional statements or modify existing ones.

Rating : Stakeholders will be provided with statements in stick notes again and this time, they will be asked to discuss and rate each statement on a Likert scale for feasibility (1 = Not at all feasible to 5 = Extremely feasible) and importance (1=Relatively unimportant to 5=Extremely important) (18).

The sorted and rated data will be analysed using multidimensional scaling (MDS) (19) and hierarchical clustering analysis (HCA) (20) using R-CMap package, implemented in R (21) to create maps that will represent clusters of the statements.

Next, we will employ bridge analysis to compare average cluster ratings between two variables, such as importance versus feasibility (18).

We will employ a bi-variate plot, such as a go-zone map (18), which will identify the priority items for action, such as the healthy eating intervention components and modes of delivery, from all the clusters.

The priority items, such as the interventions components and the delivery modes will be determined to be those that surpass both the mean ratings of importance and feasibility (18).

A workshop will be conducted subsequent to the analysis of the concept mapping data, wherein representatives from each stakeholder group will be invited together to deliberate upon the concept maps and provide their feedback. It is plausible that there could be disagreements regarding the interventions that have received the highest importance and feasibility scores in the analysis.

Thus, by reaching a consensus, the intervention(s) deemed to be of the utmost importance and feasibility will be ultimately determined. Additionally, the workshop will address how the findings can contribute to the development of healthy eating policies and interventions for adolescents in urban Bangladesh.