Introduction The Sustainable Development Goals, which are grounded in human rights, involve empowering women and girls and ensuring that everyone can access sexual and reproductive health and rights (Goal 5). This is the first systematic review reporting interventions involving rights-based approaches for sexual and reproductive health issues including gender-based violence, maternity, HIV and sexually transmitted infections in low and middle-income countries. Aims To describe the evidence on rights-based approaches to sexual and reproductive health in low and middle-income countries. Methods EMBASE, MEDLINE and Web of Science were searched until 9/1/2020. Inclusion criteria were: • Study design: any interventional study. • Population: females aged over 15 living in low and middle-income countries. • Intervention: a “rights-based approach” (defined by the author) and/or interventions that the author explicitly stated related to “rights”. • Comparator: clusters in which no intervention or fewer components of an intervention were in place, or individuals not exposed to interventions, or exposed to fewer intervention components. • Outcome: Sexual and reproductive health related outcomes. A narrative synthesis of included studies was undertaken, and outcomes mapped to identify evidence gaps. The systematic review protocol was registered on PROSPERO (CRD42019158950). Results Database searching identified 17,212 records, and 13,404 studies remained after de-duplication. Twenty-four studies were included after title and abstract, full-text and reference-list screening by two authors independently. Rights-based interventions were effective for some included outcomes, but evidence was of poor quality. Testing uptake for HIV and/or other sexually transmitted infections, condom use, and awareness of rights improved with intervention, but all relevant studies were at high, critical or serious risk of bias. No study included gender-based violence outcomes. Conclusion Considerable risk of bias in all studies means results must be interpreted with caution. High-quality controlled studies are needed urgently in this area.
The systematic review protocol was registered with PROSPERO on 4/12/2019 (CRD42019158950) [15]. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) were followed, as described in the PRISMA flow diagram (Fig 1) and checklist (S1 File). EMBASE, MEDLINE and Web of Science were searched from inception until 9/1/2020. Search terms included medical subject headings or equivalent and free text terms including sexual health, gender-based violence, maternal health, human rights and more, combined using Boolean operators (S2 File). Search terms were chosen to cover the subject areas of contraception, pregnancy, STIs, awareness of rights, violence and mental health. Reference lists of included studies and relevant identified systematic reviews were screened to identify any further studies not identified in the initial search. There were no language restrictions. Reviews of published studies in languages other than English were conducted by the authors themselves (French, Portuguese) or a University of Warwick colleague, Dr Yen-Fu Chen, (Mandarin). The search strategy was reviewed by a specialist academic librarian. Searches were adapted for each database. Grey literature was not included due to the anticipated lack of detail in the methods of published studies, and anticipated lack of independence of published evaluations (for example, evaluations published by those running an intervention as part of seeking further funding for their organisation or intervention). Following database searches, titles and abstracts were screened by two reviewers independently. Each included full-text article was then reviewed by two reviewers independently. At each state, if there were discrepancies regarding inclusion/exclusion decisions, they were addressed initially via discussion, and if necessary via a third author. Rayyan Systematic Review Web App [17] and Microsoft Excel were used to record decisions. Data were extracted by two reviewers independently onto a data extraction tool developed by the authors which included basic study details (journal, dates, authors, and location, sample size, population, study design, study duration, interventions, statistical analysis, results, ethics and funding). Only relevant results pertaining to populations eligible for inclusion in this systematic review were extracted- for example, where data was collected for males and females, only data presented on females were extracted. A narrative synthesis of included studies was undertaken. Due to study heterogeneity, meta-analysis was not undertaken. Outcomes were mapped to identify evidence gaps. Effectiveness of identified interventions was examined based on the author’s choice of outcomes. Quality and risk of bias was assessed by two reviewers independently using the RoB-2 tool for cluster-randomised studies [18], and the ROBINS-I tool for non-randomised studies [19]. Uncontrolled before-and-after studies based on two cross-sectional surveys were given an overall assessment of serious or critical risk of bias [20]. Disagreements were addressed via discussion and if necessary via a third author. All studies were classed as high, serious or critical risk of bias, meaning that due to issues with the quality of the studies, there was a substantial risk that included studies overestimated or underestimated the true effect of the interventions. The risk of bias in cluster RCTs and non-randomised interventional studies is displayed in S1 and S2 Tables respectively. Uncontrolled cross-sectional before-and-after studies were all classified as serious or critical risk of bias (S3 Table).