Study aims and objectives

This trial aimed to evaluate the effectiveness of the multicomponent intervention in improving MH and school attendance among target school girls in rural Gambia. The specific objectives were to:

(1)

Assess the effect of the intervention on menstruation-related school absence during the last period (primary objective) among target school girls.

(2)

Assess the effect of the intervention on reported UTI and RTI symptoms, and biochemical Markers of UTI in urine (nitrites and leucocyte) among target school girls.

(3)

Assess the effect of the intervention on menstruation related wellbeing and social support among target school girls.

(4)

Assess the effect of the intervention on improving menstrual management practices among target school girls.

(5)

Assess the effect of the intervention on improving menstrual needs including reproductive health knowledge and attitudes towards menstrual taboos, among target school girls.

(6)

Assess the cost and cost-effectiveness of the intervention.

(7)

Assess acceptability, coverage, uptake and fidelity of the intervention package.

Study setting

The Gambia is the smallest country in mainland Africa, with the width of the country never exceeding 50 km [29], and is predominantly Muslim (96%) [30]. In 2015, 49% of the population were living below the national poverty lines [31]. Almost 60% of the population is under 25 years [30]. The 2019/2020 Demographic Health Survey found literacy (can read aloud) rates to be higher among males than females at all age groups. The literacy rate for the population between 15 and 24 years among males was 71% and for their female counterparts was 63%. At the 45–49 age group the difference was even larger; males (64%) and female (23%) [32]. The survey also showed the net attendance ratio for primary school is 74% and drastically drops to 46% for secondary school. There is a large variation in secondary school net attendance ratio between urban (51%) and rural (33%) areas [32]. Schools in The Gambia are broadly classified as English-based or Arabic-based schools. The English-based schools are free public schools, whereas Arabic schools are private and mainly focus on Quranic education [33]. English-based schools receive an annual free supply of disposable sanitary pads from the Ministry of Basic and Secondary Education (MoBSE) [25], the schools then decide how to distribute this supply to the girls in the school. However, the Arabic schools do not get such a supply from the government. The education system is classified into Lowe Basic School (LBS) (Grade1-6); Basic Cycle School (BCS) (ECD-Grade 9); Upper Basic School (UBS) (Grade 7–9) Senior Secondary School (SSS) (Grade 10–12) [33]. Many schools may exist as double shift schools, where certain grades attending in the morning and others in the afternoon [34]. The national curriculum contains no teaching specifically on menstruation, though students typically receive some basic information on male and female anatomy in the science and population and family life classes.

The study was conducted in 50 schools across two regions of The Gambia; the Lower River Region (LRR) (Kiang West, Kiang Central, Kiang East, Jarra West, Jarra Central and Jarra East Districts) and North Bank Region (NBR) (Lower Badibou, Central Badibou, Upper Badibou, Jokadou and Lower Niumi Districts) (Fig. 1). These regions are predominantly rural, and agricultural, with most people engaging in production of groundnuts, millet and rice [35].

Fig. 1

Map of The Gambia. The green areas on the map highlight the regions the study was conducted in.

Study design

This study used a cluster randomised control trial design, with two arms, one arm received the MH intervention package (intervention arm) and the other arm did not receive any intervention (control arm). The intervention was delivered at the beginning of the school year, from October-December 2019 and three months later (February 2020) end-line outcome data collection started. Due to COVID-19 trial procedures had to be stopped on 19th March 2020 after completion of data collection in 44 schools. Data for the remaining schools was collected in November–December 2020.

Since the intervention was implemented in schools and communities from the village where the school is placed, the village was chosen as the unit of randomisation, with one school selected per village for the study.

Eligibility criteria for villages and schools

All the villages from LRR and NBR districts with at least one Arabic school, or BCS or UBS or SSS were listed. From the list, 50 villages were randomly selected. One school per village was enrolled. If a village has two or more schools, the school with the largest number of enrolled girls was selected, to ensure sample size was reached. In total, there were 24 schools from the LRR region and 26 from the NBR region. With a total of 14 Arabic-based schools and 36 English-based schools. However, one Arabic-based school from the control arm dropped out due to COVID-19.

Eligibility criteria for participants

For the end line data collection, post menarche girls over 13 years old were enrolled. Only girls that could provide informed consent or assent were enrolled, therefore, participants with mental disabilities were excluded. Physical disability was not an exclusion criterion as MH may be of particular interest in these groups.

The study population reflects the target population. The vast majority of eligible girls in each school were enrolled for outcome assessment. In schools with more than 300 eligible girls, the number of enrolled girls was capped at 300 as it was statistically inefficient to recruit more than that. In schools with fewer than 300 eligible girls, all eligible girls were selected. In schools with more than 300 eligible girls, a representative sample from each grade, excluding exam grades (Grade 9 and 12) were selected, if the eligible girls in the non-exam grades were not enough to reach the 300 cap, then girls from the exam grades were approached till the cap was reached. Girls in exam grades had busy schedules with many extra classes and limited time, therefore we opted to use non-exam grades first. Previous studies [26,27,28] found there was no significant difference in opinions if the exam grades were excluded, as long as there was representation from the other grades. If girls from Grade 9 and 12 were selected, they were approached at a time advised by the teachers and students, in most cases this was after school hours.

Informed consent

Consenting for minors (under 18 years) was done by inviting a parent or guardian to the school, where the study was explained and the information sheet read to them, and it was reiterated to each student that they could withdraw at any time without giving a reason and without having any repercussions. They were then given an opportunity to ask any questions they may have. After which if they accepted, they were asked to sign or put their thumb print on the consent form. An independent witness was present for the consenting discussions. Prior to the survey, participants went through the same process. After agreeing to participate in the study, participants were asked to sign a consent form (Additional file 1: Appendix S1), minors signed the same form their parent or guardian had signed (Additional file 2: Appendix S2).

Intervention development

A stakeholder workshop (with parents, teachers, students, clinicians, regional education officers and NGOs), held in May 2017, in Keneba, The Gambia, discussed initial findings [26, 27] from the formative research on MH in The Gambia, the problems and possible solutions. A four component intervention package was created, that could be included in existing programmes in these communities: (i) Peer Education camps, which would involve girls and boys in schools, discussing puberty and menstrual management issues; (ii) Mother’s outreach sessions to discuss issues linked to puberty, menstruation, menstrual absorbent and hygienic practices with mothers; (iii) Community meetings around puberty and menstruation, to involve the men in the communities; and (iv) minor improvements in school WASH infrastructure.

This intervention package was initially tested in two schools within the West Kiang Region, feedback from this testing was used to refine the material further. After which in 2019, a pilot intervention trial was run to further refine the material, test feasibility and acceptability of components of the intervention, test data collection tools, train implementers and enumerators and inform sample size calculation decisions for the main trial. The pilot was conducted in three intervention schools and three control school, with a total sample size of n = 299. Intervention schools used in the pilot were excluded from the main trial.

Description of the interventionPeer education camps and menstrual hygiene laboratories

Our previous study [26] showed that students gained relatively little knowledge about puberty and menstruation from their teachers and they felt more comfortable to talk with their peers. To address this lack of knowledge, Nova Scotia Gambia Association (NSGA) was subcontracted as a partner, to deliver an innovative two-session MH training per school. NSGA is a Canadian NGO working with adolescents and communities in The Gambia since 1985, it is known for using an interactive and direct engagement approach, dealing with different public health issues (Malaria, HIV, early marriage). The NSGA team went through a four-day training before the pilot intervention and again before the main trial intervention. The training covered aims of the intervention and familiarisation with intervention content.

In preparation for the first session, the school teachers were asked to select 15 boys and 15 girls that were outspoken and respected by other students. The first session was a full day session delivered to the 30 students, it included an introduction to the study, discussion about puberty education affecting both boys and girls, and discussions about menstruation. These 30 students served as menstrual ambassadors, to share information about puberty and MH practices with other students. The session had role-play activities, that showed the students how they can talk to others students about what they have learnt, they were also encouraged to talk to adolescents outside of the school setting. They were given different options of how they could spread the message such as through school assemblies, use of peer educator rooms in school etc. It was explained to the ambassadors that it was their responsibility to share what they learnt in the sessions with those that could not attend. This is the ethos of NSGA and the students are used to this approach. At the second session (the following day), all the school girls aged 13 years and above, were invited to attend a menstrual hygiene laboratory (if the school was large, this ran as multiple sessions, to ensure the group size was manageable). During this session, girls were split up into similar age groups, so as to ensure they were more comfortable discussing and asking questions. They were taught more in detail about menstruation, menstrual management practices, how to prepare for the period, tracking the menstrual cycle, information about absorbent materials and pain management (e.g., exercise, applying heat and or pressure, stretching, hydration, medical-herbal and pharmaceutical). The 15 ambassador girls that attended the first session also attended the menstrual hygiene laboratory session, and supported the NSGA staff in facilitation. One poster was given to each school after the training, with information about menstrual hygiene practices, physiological knowledge and tips for managing pain and how to prepare for your cycle. The same information was also included into a leaflet which was given to each girl that attended the training. The sessions with the students were full day sessions, with many icebreakers and activities throughout to keep the students engaged. Sessions were carried out on weekends, to reduce impact on school learning.

Mother’s outreach

Our previous study showed that mothers were one of the first point of contact for advice about how to manage menstruation for girls [26], however, mothers have limited knowledge about menstruation and also have different perceptions about how menstruation should be managed. At the stakeholder meetings, it was suggested that parents should be involved in the intervention package to be able to support girls with good information and a more open approach. Therefore, we decided to organize mothers’ meetings in each school. The school admin/teachers were asked to identify 20 mothers that were active, confident, vocal and easy for others to approach. These mothers were invited for a two-day meeting. The meeting ran from 9 am to 12/1 pm, we opted to have half day meetings, as we found concentration levels were low in the afternoon, and mothers were rushing to get back to other chores. During the first session, issues about importance of girls finishing school, and information about puberty and reproductive issues were discussed. The second session discussed menstruation in greater detail, role-play approaches of talking with children about menstruation, and menstrual hygiene management (including strategies for using different type of absorbents, absorbent and body hygiene practices, managing pain). Mothers were shown all the types of absorbents and explained their advantages and disadvantages and how to use them. Mothers were encouraged to work with the school administration to help maintain the WASH facilities, examples of assisting with cleaning school facilities or making soap for the school were given. The mothers were also encouraged to share this information with other women and adolescents in the community to increase discussion and awareness about menstruation. During the sessions, the mothers practiced different interactive methods of sharing the information. The sessions were conducted by two female facilitators (one nurse and one traditional influencer) who have experience discussing sensitive topics and have an energetic and engaging approach. The facilitators had three days training to discuss content of training materials and how they could engage the mothers, after which they had a practice session with a group of mothers that were not part of the study, followed by another training session to go over feedback and address any challenges faced.

Community meetings

Following suggestions from the stakeholder workshop to involve men in discussions about puberty, menstruation and the importance of girls schooling, a community meeting in the evening after prayers was considered to be the best opportunity to discuss these topics. One community meeting was conducted in each community, where both men and women attended. The sessions discussed the above mentioned topics and provided an opportunity for community members to ask questions. The sessions were arranged through the village elders, village development community chairman, and a religious elder, to increase acceptance of the program and promote attendance. The meeting lasted an hour or two depending on the level of engagement from the attendees. The sessions were conducted by two facilitators (one female and one male), who have an energetic approach and are used to discussing sensitive topics within communities and are respected by the communities. The facilitators had a day training to go over context of the session, after which they had a practice session in a community that was not part of the study, followed by another training session to go over feedback and address and challenges faced.

Improving school WASH facilities and maintenance

During the stakeholder meeting, representatives from all the schools showed willingness to help with the suggested interventions, and WASH improvements were seen as a major intervention priority. In the intervention schools, the school management team, mother’s outreach group and peer education camps were engaged to implement inexpensive measures including ensuring water storage containers were close to the toilets, and hand-washing stations were present with kettles (water pouring device commonly used in The Gambia), water and soap. After an initial WASH spot check (Additional file 3: Appendix S3), the team discussed with the school management team, simple changes that could be made to the current system where required, e.g., separating toilet facilities for boys and girls where possible, creating simple doors to increase privacy etc. A tailored TOR was created with each school at the beginning of the study with the agreed maintenance tasks. The agreement also included that the study team would provide materials to improve WASH facilities in a step wise manner. First the schools were provided with 100 L water tanks to be placed near the toilets and ensure they were always filled together with two handwashing stations and four kettles. After which at the next visit, they were provided with soap and materials to make tippy taps, which should also be placed near the toilets.

Initially part of the WASH intervention included an aspect of absorbent disposal facility in the schools, as most students dispose their used disposable pads into the pit latrine and the study team recognised that this would cause the pit latrines to filling up quickly. However, after further discussions with the girls and mothers, it was realised there were many differing views on used material disposal, some felt the material should be burnt, while others felt burning would lead to infertility and therefore was not an option. Because an intervention provided by the researchers could not thus be standardised across schools, this aspect was removed. This was replaced by including discussion on disposal options in the outreach sessions, where participants were encouraged to use the method they were most comfortable with, instead of throwing the material into the latrine. During the formative research no girl reported methods of disposal as a barrier to attending school.

All intervention components were simultaneously run in all intervention schools over the intervention period.

Randomisation, allocation and blinding

Villages/schools were randomised within 10 strata, with one stratum with 2 schools, 4 strata with 4 schools, a further 4 strata with 6 schools, and one stratum with 8 schools. Strata were formed by a combination of: (1) region (Lower River vs North Bank), (2) English vs Arabic schools, and (3) level of socio-economic development. These factors were a priori assumed to be associated with menstruation-related school absence and the effect of the intervention on this outcome. The level of socio-economic development was done in two ways: (1) the research team created a level of urbanisation scoresheet and visited all the villages to complete the score, to determine level of development; and (2) Six study staff familiar with the study area were asked to rank the study villages according to the perceived level of development. Ranking from both methods was compared and any ranks that varied were discussed and additional information about the villages was sought or discussed with two other staff in the demographic health survey department. The ranks were collapsed into three equally sized categories of development (high, middle and low).

Randomisation of villages/schools within strata was further restricted by excluding allocations that did not meet pre-specified balance criteria, following methods described by Sismanidis et al. [37]. The variables used for restriction were: access to a tar road (binary), availability of electricity (binary), availability of mobile phone tower (binary), predominant building material of houses, i.e., mud vs other (binary), total number of girls in enrolled school, number of girls of menstruating age in enrolled school, distance to nearest major town (km), number of standpipes in village per number of household compounds, percentage of houses that own a TV dish. For binary variables, randomisations were accepted where arms differed by no more than 5% points. For continuous variables, randomisations were accepted in which the relative difference in means between arms did not exceed 20%, except for the school level variables “total number of girls” and “total number of girls of menstruating age” for which a 10% relative difference was accepted. Randomisations were done 50,000 times, with 20 allocations meeting the criteria. The number of possible allocations was about 1.2 * 107out of 2.9 * 1010 possible allocations—a restriction factor of 99.96%. Of the 20 acceptable randomisations, we chose one at random as final allocation. The validity of the allocation was checked in 281 allocation meetings the balance criteria. Each pair of clusters was assigned together to the same treatment arm with an average of 138 allocations, ranging from 51 to 210 of the 281 allowable allocations (disregarding the stratum with only two villages that could not be allocated to the same treatment).

The trial manager was aware of the allocations for logistic purposes, however the enumerators collecting the end line survey data, and the statistician analysing the data were kept blinded to the allocations. After initial statistical analysis, the allocations would be revealed to the statistician to allow for subgroup analysis.

After randomisation, one village was removed from the study as due to miscommunication, the actual number of eligible girls in the selected school (Arabic) was far lower than reported earlier. This village was replaced by the closest village within the same region with an Arabic school.

A further school dropped out on 17th March after randomisation, just before data collection due to concerns regarding COVID-19 and maintained their refusal status even after data collection was resumed in November.

Sample size

The 2019 pilot data showed a slightly lower prevalence of the primary outcome (Proportion of girls with at least one-day absence during last menstrual period), than previous studies, at 16.8%. We assumed that the intervention would lead to a 33% reduction in the primary outcome, which resulted in a crude sample size (ignoring clustering) of 637 girls per study arm. In the pilot study, the ICC for the primary outcome was 0.026. The number of girls of menstruating age per enrolled school (i.e., cluster size) was 75 girls, resulting in a design effect of 2.9 and an overall sample size of 1862 girls per arm, or 25 schools per arm. An average of 75 menstruating girls per school was used as an estimate to calculate how many schools needed to be enrolled into the study, However, we identified that some schools had less than 75 menstruating girls, to account for this, we allowed to select all the menstruating girls in the school, so that larger schools could compensate for smaller schools while ensuring that the sample size was reached. Although we did introduce a maximum cap of 300 menstruating girls per school, as any more than this would be statistically inefficient and not reduce the number of schools that needed to be visited.

Process evaluation

A detailed mixed-methods process evaluation based on the MRC Process Evaluation Framework [38] was conducted to explore three core functions of the intervention: implementation, mechanisms of impact and context, to provide guidance for sustainable and scalable implementation of the intervention should it prove successful. We collected data on information on fidelity, reach, acceptability, context and mechanism of impact using different tools (Table 1): (i) end-line quantitative survey to assess reach and dose; (ii) interviews with facilitators within a week of completion of intervention delivery to assess feasibility, fidelity (challenges to delivery, barriers to uptake, how the intervention was delivered, assess if implementation varied from what was planned and reasons for variation) and contextual factors (logistical factors impeding or facilitating, engagement level of participants, unexpected participants, adaptations required and factors for these); (iii) in-depth interviews (IDI) with 19 girls, 13 boys and 6 teachers, and focus group discussions (FGDs); 6 with girls and 6 with mothers, conducted after end-line collection to assess acceptability, uptake, and possible mechanisms of impact; (iv) data from three unannounced visits checking on WASH facilities throughout the study to assess changes in WASH over time, and assess if schools were making and maintaining changes discussed, if not, to understand barriers to implementing and maintaining changes; (v) routine data collected as part of the implementation (reports, pictures, registration forms from implementers; school visitation logs); and (vi) unannounced observations of facilitators to ensure schedules were followed and assess quality of intervention delivery and report on context factors (logistic factors, engagement level of participants).

Table 1 Summary of process evaluationTheory of change

The theory of change (Additional file 4: Appendix S4) was based on the Behaviour Centred Design (BCD) framework developed by Aunger and Curtis [39]. It shows the barriers to school attendance that were identified during formative research in The Gambia [26,27,28]. The barriers related to environmental, social and physiological factors, exploring these barriers resulted in ensuring the intervention package addressed all the barrier components identified. Addressing environmental barriers would ensure that girls had access to the hardware that they needed to maintain good MH resulting in an increase chance of attending school while menstruating. Improving social factors would result in more social support, improved communication about menstruation, a place for girls to ask questions about what they were experiencing, reduced teasing creating an overall positive attitude towards menstruation and making girls more comfortable attending school while menstruating. Addressing the physiological barriers such as pain and discomfort and low confidence through the various pathways would increase the ability to walk to school, and concentrate while in school.

Outcome measurement

All outcomes, apart from biomedical markers of UTIs, were measured at end-line through an enumerator-administered survey. After completion of the survey, participants were asked to provide a midstream urine sample to measure biomedical markers of UTIs, through dipstick analysis. All survey questions (Additional file 5: Appendix S5) were translated into the local languages (Mandinka, Wollof and Fula), and back-translated to confirm wording. School girls’ surveys were verbally administered by trained female enumerators. Enumerators received 10 days of training, initially they were trained on the purpose of each question, and ways to ask the questions, after which they had a chance to trial asking the questionnaire to girls from schools that were not part of the study, and further training was conducted to address any challenges that may have been faced during the trial. Data from school girls was collected in a private room on the school grounds. As menstruation is a sensitive topic, female enumerators were selected to facilitate participation in the study. Survey responses were entered on tablets using REDCap software version 8.9.2 [40].

Primary outcomeSchool attendance

The primary outcome for this study was the proportion of girls with at least one-day absence during their last period. We aimed to collect school register data as a gold standard method to measure school attendance and to be able to compare and validate the survey data used to measure the primary outcome. However, we found lots of inaccuracies in register data, and data could not be collected from all the schools. Previous studies have also found registers to be inaccurate, irregularly completed, and difficult to use due to teachers not completing daily, and girls using different names for school enrolment and study enrolment [17, 41, 42]. Thus, we will perform a comparison exercise only with the data from the available schools which have the most complete data to be able to estimate validity of the survey data.

The MENISCUS study in Uganda [15] had success in using self-completed diaries to collect information about school attendance, however formative research in The Gambia found this to be an unsuitable tool for the primary outcome as the diaries are often left empty. This difference could have been due to differences in literacy level, or not being able to have the diaries in the local language, since the local languages are not written, or students in The Gambia had never done such activities. Therefore, for this context, we decided to use self-reported school absence during the last period as the measure to estimate the primary outcome. The question that we used in our questionnaire to measure this outcome was: In the last 30 days how many days of school did you miss because of menstruation? Girls had the option to say they did not menstruate in the last 30 days, in this case, a further question was asked: During your last period, how many days of school did you miss because of menstruation? For the primary outcome the responses from both these questions will be combined to calculate the proportion of girls with at least one-day absence during last period. For the secondary outcome responses from the question “In the last 30 days how many days of school did you miss because of menstruation?” will be used to ascertain the number of days missed in the last 30 days due to period.

Secondary outcomesUrinary tract infections symptoms

This secondary outcome was defined as proportion of girls reporting at least one urinary tract infection symptom over the 7 days preceding the survey. The questions used are listed in Table 2.

Table 2 Questions used in the survey to assess each outcomeBiochemical indicators of UTI

At the end of the survey, participants were requested to provide a mid-stream urine sample in a sterile specimen container. All specimen containers were labelled with participant’s study ID. The urine samples were analysed by the lead researcher (who had been trained to conduct these tests by the clinic staff at MRC Keneba) in the school setting, using urine dipstick analysis with the Combur 9-Test® strips [43]. This is a quick screening test for a UTI in children and adolescents, whose results are interpreted in the context of clinical feature suggestive of a UTI [44].

The guidance for urine collection that was given to the participants Additional file 6: Appendix S6.

Nitrites and leucocyte levels were measured using dipstick tests. A girl was considered to be UTI positive if she has a dipstick test positive for nitrites, or a dipstick test positive for leucocytes and at least one UTI symptom. All the girls that reported at least one UTI symptom or were positive for the nitrites indicator were referred to a clinic.

Reproductive tract infection symptoms

This secondary outcome was defined as proportion of girls reporting at least one RTI symptom over the 7 days preceding the survey. The questions used are listed in Table 2.

Menstruation related wellbeing score while at school

Prevalence of menstruation related wellbeing was measured at end-line among the adolescent girls using a new tool developed in this study. The items for the tool were developed through information gathered in the qualitative discussions during the formative study phase (between 2015 and 2017) and through information from enumerators and clinical staff. The questions were tested during the translation and pilot phase of the trial, and changed according to feedback. The tool was tested by enumerators, clinical staff, community volunteers, and school going adolescent girls. This tool consists of a set of 10 questions with binary response regarding menstruation related wellbeing while at school (happy, unhappy). The answers of these 10 questions will be reported individually as well as collapsed into a score using tetrachoric PCA. The questions used are listed in Table 2. In exploratory factor analysis we will look into the multidimensionality of the variable responses to explore the presence of relevant factor scores.

Social support

A feeling tool developed in this study (that includes questions about comfort, confidence, and support from the mother and community when managing your menstruation) was used to assess the social support impact of the multipack intervention on participating girls. The tool was developed in the same way mentioned above. A set of 7 questions with three-level response regarding social support at home and school will be reported individually as well as collapsed into a score using polychoric PCA (happy, neither happy nor unhappy, or unhappy). Similar to the wellbeing score we will also do exploratory factor analysis for this score. The questions used are listed in Table 2.

Knowledge of menstruation and MH

We evaluated knowledge using end-line questionnaires with closed questions, the outcome was defined as proportion of girls giving correct answers to knowledge based questions. A set of 9 questions with three level responses (yes, no, I don’t know) were used. The questions used are listed in Table 2.

Perceptions or disbelief of taboos towards menstruation and MH

These were assessed through closed questions in the end-line survey. The outcome was defined as proportion of girls giving correct answers and disbelief of common taboos known in this context. A set of 6 questions with three level responses (yes, no, I don’t know) were used. The questions used are listed in Table 2.

MH practices

These were assessed through closed questions in the end-line survey. The questions used are listed in Table 2.

Cost and cost-effectiveness

We will conduct an economic evaluation to estimate the costs and cost-effectiveness of the MEGAMBO intervention. As a within-trial evaluation, the time horizon in the base case will be the length of follow-up, but we will also develop a decision analytic model to extrapolate costs and outcomes over a longer time horizon within which outcomes might be expected to be sustained for the targeted groups (e.g., 5 years). A combination of top-down and ingredients-based costing approaches will be used to generate cost estimates for the whole intervention and for each component. All costs will be estimated from disaggregated social perspective (both provider and societal) and financial and economic costs will be calculated for all inputs. The results will provide the costs of setting up and running the intervention package, describe the distribution of costs across 4 intervention components, the unit cost per student reached and the cost of delivering all activities in intervention schools. Since no intervention is undertaken in control schools, these costs are by definition incremental over control. Primary and selected secondary outcome measures will be used for the cost-effectiveness analysis of the intervention relative to usual practice (represented by the control schools). For example, cost per percentage point change in the primary outcome, and cost per school day gained. If applicable, we will estimate the cost per disability-adjusted life year averted. Since outcomes are across many domains and cannot be aggregated, we will present overall cost-effectiveness results as a cost-consequences analysis.