A 4-armed cluster randomised controlled trial conducted among secondary schoolgirls in Siaya, western Kenya, where clusters are the unit of allocation and schoolgirls the unit of measurement. The overall aim of the trial is to inform evidence-based policy to develop intervention programmes which improve adolescent girls' health, school equity and life-chances. The primary objective is to determine the impact of menstrual cups or cash transfer alone, or in combination, compared against controls, on a composite of deleterious outcomes (HIV, HSV-2 infection, and school dropout) over 3 schoolyears follow-up.

Title: Menstrual cups and cash transfer to reduce sexual and reproductive harm and school dropout in adolescent schoolgirls in western Kenya: a cluster randomised controlled trial.

Short Title: Cups or cash for girls (CCg) trial

Background and rationale: Adolescence is a critical time of psychological and biological change, and advocacy has increased to identify interventions that protect against sexual and reproductive health (SRH) harms, which are disproportionately high among adolescent girls in sub-Saharan Africa. In much of eastern and southern Africa including western Kenya, where unprotected transactional sex is common, young females are highly vulnerable to sexually transmitted infections (STIs), including HIV, and pregnancy resulting in school dropout. While the burden of young female SRH harms is high for individuals, communities and health services, sustainable preventive interventions are lacking. Evidence of a positive association between girls' education, health and economic potential has strengthened international resolve to improve educational opportunities for adolescent girls. While SRH education has minimal impact on SRH harms, staying in school has shown to protect girls against early marriage, teen pregnancy, and HIV infection, with schoolgirls reporting less frequent sex, and fewer partners with less age disparity. While MDGs focused on primary school attendance, the post-2015 Sustainability Development Goals continue to encourage investment in secondary, tertiary and vocational education to build human capital, innovation and economic growth, but require the support of cost-effective interventions. Interventions using cash transfer (CT) have demonstrated a protective effect on girls SRH (HIV, HSV-2, sexual behaviours, and school indicators). Menstrual hygiene management (MHM) is a pervasive problem across low middle income countries (LMICs) and a lack of MHM materials and facilities negatively impact girls' school-life. This increases girls' vulnerability to coercive sex, which often creates a pathway to obtain necessities such as soap, sanitary products, and underwear; 10% of 15 year old girls report that they obtain money through sex to purchase sanitary products in western Kenya. To tackle these challenges, our team ran a pilot menstrual study in western Kenya. It provided MHM tools to adolescent girls in the form of reusable menstrual cups and disposable sanitary pads. The results demonstrated a lower prevalence of STI and bacterial vaginosis among girls who were provided with a single menstrual cup (one cup can last up to 10 years), and a lower prevalence of school dropout after 12 months follow-up compared to controls. This pilot requires replication in a larger trial population with longer follow-up. Comparison against CT offers an opportunity to examine the efficacy and cost-effectiveness of these different approaches to improve girls' life-chances in rural western Kenya. The study is designed to inform evidence-based policy to improve girls' health, school equity and their life-chances.

Primary objective: To determine the impact of menstrual cups alone, cash transfer alone, or the two in combination, in secondary schoolgirls on a composite of deleterious outcomes (HIV, HSV-2 infection, or school dropout).

Hypothesis: The investigators postulate the interventions tested will break the cycle of sexual and reproductive ill-health, under-achievement, and poverty which impede girls' successful completion of school.

Overview Study Design: Single site open-label 4-arm, school-cluster randomised controlled superiority trial. Schools are the unit of randomisation (clusters), with girls the unit of measurement. Schools will be randomly allocated into 4 arms using a 1:1:1:1 ratio and permuted block randomisation to minimise bias. Enrollment will be staggered over >=2 school terms if logistically required. Girls will be followed-up through graduation and into employment or up to 10 academic terms to determine if they complete secondary school (Form 4). Sealed, opaque envelopes will be prepared with the study allocation. Counsellors conducting HIV and HSV-2 testing, and laboratory technicians will be blinded to the study arm. Field staff who conduct home visits to confirm dropout will also be masked where feasible.

Sites: The study will be conducted in Siaya County, western Kenya. Depending on the recruitment rate enrolment will be expanded to other neighbouring counties.

Study Population: Secondary schoolgirls who attend eligible schools in the western Kenya study site. Girls will be residents of the area, with a history of established menses (>=3 times), no disability preventing participation, with parent or guardian's consent and girl's assent. Girls attending boarding schools or with visible/declared pregnancy will be excluded at baseline.

Study Interventions: 1. One menstrual cup (Mooncup®) with handwash soap termly; 2. Cash transfer (CT; girls' pocket money) via local community/mobile banking with financial literacy; 3. A combination of cup and CT interventions; 4. 'Usual practice' (control) with handwash soap termly.

Outcome Measures: Primary efficacy outcome: Composite endpoint comprised of incident HIV, HSV-2, and all-cause school dropout, by end of follow-up. Key secondary outcomes include incident HIV, HSV-2, school dropout, BV and reported sexual behaviours including pregnancy, quality of life measures, school indicators (performance, grade repitition, re-enrolment, absence, transfers), and cost-effectiveness. Primary safety outcome measure: toxic shock syndrome (TSS), and severe violence associated with intervention. Key secondary safety outcome include contamination on menstrual cups and other emergent harms associated with the interventions.

Follow-up procedures: HIV and HSV-2 serostatus will be assessed at baseline and around the time of final school term (Form 4), with interim testing or annual testing if funding allows, including for bacterial vaginosis and other STI. School dropout will be assessed every term until the end of Form (class) 4. Other endpoints will be evaluated at baseline, midline, and end of study. Safety monitoring of TSS and physical violence will be conducted throughout by study nurses, supplemented by health facilities and evaluation of HDSS census mortality data.

Sample size: Main trial: 84 schools (clusters) (21 per arm) with an approximate average of 46 girls per school (~3864 participants, 966 per arm). Protocol amendment v7: 96 schools with approximately 41.5 girls per school (3980 participants, 995 per arm).

Data Analysis: Primary trial analyses will be based on the intention to treat principle and a secondary analyses will also be done on the per protocol population. Generalised estimating equation (GEE) log binomial models will be used to analyse the primary endpoint and its components. The GEE model will include the arm as a fixed effect and school as a cluster effect. The RR values for the 5 pre-specified primary comparisons together with their 95% confidence intervals will be derived from the GEE model. The secondary endpoints will be analysed similarly using GEE models. For GEE analysis of a continuous endpoint such as quality of life, normal distribution and identity link functions will be used. For GEE analysis of a binary outcome (such as having an event of STI, HIV, pregnancy, or school dropout), binomial distribution and log link functions will be used; for GEE analysis of recurrence of events (such as number of sexual partners during a specific time), Poisson distribution and log link functions will be used. Covariate adjusted analysis of primary endpoint will be performed within the GEE framework with treatment as the study variable, and other predictors as covariates, and school as cluster effect. For qualitative analysis, FGD recordings will be transcribed verbatim with back translation. Transcripts will analysed using thematic analysis by study group, 2 researchers will separately assign codes for emergent themes, subthemes patterns, and associations using NVIVO with intercoder reliability checked and consensus reached following discussion. As themes emerge, differences and similarities will be compared across trial arms, and between study groups. Qualitative data from in depth interviews will be similarly evaluated. Protocol amendment v10:Secondary analysis on primary and secondary outcomes using 'at the margins' factorial analysis will also be conducted.

Inclusion Criteria:

• Attend secondary day schools in the study area
• Resident of the study area
• Have a history of established menses (>=3 times)
• Have no disability preventing participation
• Have received parent or guardian's consent and themselves assent

Exclusion Criteria:

• Attend boarding schools
• Visibly pregnant or declare pregnancy at baseline (non-declared girls who are pregnant will be excluded from analysis after normal (or otherwise) delivery dates confirm pregnancy started prior to intervention)

Target population for intervention is menstruating schoolgirls.

A link will be provided from LSTM and study-specific website to study resources including the protocol, participant information sheets, SOPs, publications and the database. A data repository is being established at LSTM and will be entrusted with storing data when ready. The investigators policy to data sharing will be published on the study website. A final database containing all research data will be made fully publicly available in an unrestricted format once the findings have been published. The only limits to data sharing will be to safeguard participants' confidentiality. The proposed procedures for data sharing will be set out and explained to the research participants as part of the informed consent process. External users will be bound by data sharing agreements in line with The MRC Data Sharing Policy. External users will be required to accept terms and conditions of use using a declaration tick box.

The investigators cannot mask participants to their treatment (intervention) status due to the nature of interventions provided. However, counsellors conducting the HIV and HSV-2 testing and laboratory technicians will be blinded to the study arm. Investigators and trial statisticians will be masked. Field staff who conduct home visits to confirm dropout will also be masked when feasible. Bias will also be minimised by use of block randomisation stratified by school size and WASH facilities at baseline. An independent person will prepare the sealed envelopes with the study allocation. Study arm allocation will not be recorded in the central database to ensure the trial statistician and data managers remain blinded throughout the study. This information will be recorded separately and only be merged with the main database following approval of the statistical analytics plan (SAP), closure of the databases and submission of a copy to the independent statistician of the DMEC.