Garazi Zulaika, Elizabeth Nyothach, Anna Maria van Eijk, Duolao Wang, Valarie Opollo, David Obor, Linda Mason, Tao Chen, Emily Kerubo, Boaz Oyaro, Alex Mwaki, Alie Eleveld, Isaac Ngere, Eunice Fwaya, Feiko O. ter Kuile, Daniel Kwaro, and Penelope A. Phillips-Howard
Summary: Background: High rates of sexual and reproductive health (SRH) harms and interrupted schooling are global challenges for adolescent girls, requiring effective interventions. We assessed the impact of menstrual cups (MCs) or cash transfers conditioned on school attendance (CCTs), or both, on SRH and schooling outcomes in western Kenya. Methods: In this cluster-randomised Cups or Cash for Girls (CCG) trial, adolescent girls in Forms two and three at 96 secondary schools in Siaya County (western Kenya) were randomised to receive either CCT, MC, combined CCT and MC, or control (1:1:1:1) for an average of 30 months. The CCT intervention comprised 1500KES (US$15 in 2016) via a cash card each school trimester. All four treatment groups received puberty and hygiene training. Assenting girls with parent or guardian consent who were post-menarche, not pregnant, area residents, not boarding, and had no disabilities precluding participation were eligible. Socio-behavioural risk factors and incidence of HIV and herpes simplex virus type 2 (HSV-2) were measured annually. School retainment and adverse events were monitored throughout. The primary outcome comprised a composite of incident HIV, HSV-2 and/or all-cause school dropout by school exit examination. The primary analysis was by intention-to-treat (ITT) using generalised linear mixed models, controlling for a priori selected baseline covariates. The trial is registered with ClinicalTrials.gov, NCT03051789. Findings: Between February 28, 2017 and June 30, 2021, 4137 girls (median age 17.1 [interquartile range (IQR): 16.3–18.0]) were enrolled and followed annually until completion of secondary school (median 2.5 years [IQR: 2.4–2.7]); 4106 (99.3%) contributed to the ITT analysis. No differences in the primary composite outcome between intervention and control groups were seen (MC: 18.2%, CCT: 22.1%, combined: 22.1%, control: 19.6%; adjusted risk ratio [aRR]: 0.97, 95% confidence interval 0.76–1.24; 1.14, 0.90–1.45; and 1.13, 0.90–1.43, respectively). Incident HSV-2 occurred in 8.6%, 13.3%, 14.8%, and 12% of the MC, CCT, combined and control groups, respectively (MC: RR: 0.67, 0.47–0.95, p = 0.027; aRR: 0.71, 0.50–1.01, p = 0.057; CCT: aRR: 1.02, 0.73–1.41, p = 0.92; combined aRR: 1.16, 0.85–2.58, p = 0.36). Incident HIV was low (MC: 1.2%, CCT: 1.5%, combined: 1.0%, and control: 1.4%; aRR: 0.88, 0.38–2.05, p = 0.77, aRR: 1.16, 0.51–2.62, p = 0.72, aRR: 0.80, 0.33–1.94, p = 0.62, respectively). No intervention decreased school dropout (MC: 11.2%, CCT: 12.4%, combined: 10.9%, control: 10.5%; aRR: 1.16, 0.86–1.57; 1.23, 0.91–1.65; and 1.06, 0.78–1.44, respectively). No related serious adverse events were seen. Interpretation: MCs, CCTs, or both, did not protect schoolgirls against a composite of deleterious harms. MCs appear protective against HSV-2. Studies of longer follow-up duration with objective measures of health impact are needed in this population. Funding: Department of Health and Social Care, the Foreign, Commonwealth & Development Office, the Medical Research Council and Wellcome.