Study characteristics

Following the study selection process (figure 1), 18 articles were included in this systematic review. Of these, several papers concerned the same intervention, thus, a total of 16 different interventions are reviewed. Articles linked to the same intervention are discussed together, and discussions relate to interventions rather than articles. See table 1 for a summary of the characteristics of the included studies.

Most studies were quantitative (n=11), with quasi-experimental, case-control or cross-sectional design. Five of the included 16 studies used qualitative methods, among which two were mixed methods.

Of the 16 interventions, some studies were assessed to have low risk of bias (31%, n=5), and two (31%) were identified to have moderate risk of bias. Most studies (56%, n=9) were graded as having a high risk of bias—that is, studies on these interventions did not fulfil most of the quality criteria. None of the studies were excluded based on study quality. Instead, this is taken into consideration in the effectiveness assessment (step 2), and subsequent analysis.

The majority of the interventions (69%, n=11) were set in upper middle-income countries—South Africa (n=3), Brazil (n=2), Iran (n=2)and one each in Ecuador, Turkey, Azerbaijan and the Philippines. Lower middle-income countries such as Tunisia, Nigeria and Indonesia each had one study. Only two interventions were based in low-income countries—one in Nepal and the other in Tanzania, indicating a stark lack of focus on the poorest contexts. Similarly, only one intervention was based in a rural setting, while half were in urban areas (50%, n=8) and the remaining in either mixed or semi-urban (44%, n=7).

Table 2 summarises the interventions by type of impairment. Several interventions were not specific to one type of impairment—thus, some studies were coded under more than one category. Thirty-eight per cent of the interventions were related to people with intellectual or cognitive impairments, and it was the same for those with visual impairments. In contrast, only two interventions (13%) each were related to people with mental health conditions or people with physical impairments. It may be that some interventions benefitting people with physical impairments were general but inclusive services (non-targeted, for example, ramps) and may not have been evaluated, while those targeting intellectual impairments may have been additional and targeted (eg, specific information material) and required evaluation before scaling up. Just three interventions (19%) targeted more than one type of impairment, and all three presented findings disaggregated by impartment type.

The majority of the studies (n=11) involved data collected from people with disabilities, while some were based on data from carers or teachers. However, there was variation in disability definitions and assessment. Three used clinical definitions/assessments, one used the Washington Group questions on functional limitations and the remaining studies did not specify, which limits the application of findings. It is notable that among the 16 interventions, only five interventions (31%) involved Disabled People’s Organisations in intervention development, which would have strengthened suitability to context.

Although a number of different search strings were developed for the different components of SRHR, there were no eligible studies of interventions for many key themes, including maternal health, family planning, and abortion. As table 3 shows, 13% of the interventions promote general reproductive health (n=2) and sexual health (n=2) among people with disabilities, and their protection from STIs including HIV/AIDS (n=2). 19% of the interventions (n=3) relate to protection from sexual violence. Nearly half of the included interventions focus on CSE and information provision in and out of schools, though some interventions were coded both as CSE (for delivery method) and content (e.g., violence).

Table 4 summarises the data extracted from the studies, displaying details of intervention, targeted impairment type, SRHR theme, effectiveness before and after considering risk of bias in the evidence and the Levesque dimension of access each intervention targets. Some elements of our interpretation based on these processes are also included. For example, if an intervention was claimed to be effective, but the study making this claim had a high risk of bias, we interpret that there is limited evidence to support that this intervention ‘works’. We then examine the Levesque dimensions to interpret whether the intervention is promising or limited in terms of wider application potential. Combined, each row highlights strengths and weakness of both evidence and intervention, which then frames the subsequent discussion.

Reproductive health

In this systematic review, this theme includes general RH services and programming, RH illnesses, urogenital disorders, and menstrual hygiene management (MHM). Two interventions were identified—one promoting access to MHM and the other on access to obstetric fistulae surgery.

Fiander and Vanneste (2012) report on an intervention that used community-based ‘ambassadors’ to provide transport money to patients who needed surgery for obstetric fistulae and cleft lip in Tanzania.28 In 2010, 239 patients attended 129 obstetric fistula surgeries—four times the scheme target and resulted in a 65% increase in the fistula operations performed. However, this study had a high risk of bias: reviewers note that there was little evidence that potential confounders or intervening factors that may have contributed to the increase in surgeries were considered in interpreting the findings. The intervention may have promise in that it responds directly to patients’ financial barriers (people’s ability to pay, service affordability), and could contribute to improvements in healthcare utilisation. This intervention had potential for wider application, beyond RH surgeries and people with other impairments, though a rigorous evaluation is first needed.

Wilbur and Bright et al (2018) documented the development of an intervention to improve menstrual hygiene management for women with intellectual disabilities in Nepal29 and Wilbur et al (2019) reported on a feasibility study of this intervention conducted with 10 women with intellectual impairment and their 8 carers.30 Pre- and post-survey and process monitoring found that the Bishesta campaign was acceptable, was delivered with fidelity and improved target behaviours. Although some indicators of quality were not reported in Wilbur and Bright et al (2018)29 because it was a methodological paper, the feasibility study (Wilbur et al 2019) was assessed to have low risk of bias, indicating promising evidence that this intervention works. The intervention has promise as it promotes intellectually impaired people’s ability to perceive and seek support services, while also promoting acceptability of services/products related to their menstrual hygiene. Further testing in other settings would confirm wider application.

Sexually transmitted infections

This theme includes HIV/AIDS interventions such as PMTCT (prevention of mother-to-child transmission), testing and treatment. Two interventions were identified for this theme, both of which were assessed to have high risks of bias, indicating limited evidence in this theme.

Oliveira et al (2016) conducted a validation study of an information package with content experts reviewing educational materials (a rhyming approach adapted from breast feeding promotion material) on STIs for visually impaired people in Brazil.31 They reported positive results, with some adaptations needed to ensure accuracy of STI information. Reviewers assessed this study to have a high risk of bias, linked to inadequate sampling technique, unsuitable data collection and analysis methods and interpretation without sufficient detail to convey reliability or rigour. The intervention targeted visually impaired people’s health literacy (ability to perceive need for care)—this is of great importance. However, there was insufficient evidence to indicate that this approach works, and wider application was unclear.

Doherty et al (2016) reported on a situational analysis of a 2-year large-scale HIV prevention intervention for in-school children with impairment in Nigeria.32 They reported that the intervention led to increased SRH knowledge among hearing and visually impaired students, improved health seeking behaviours and access to HIV services, reporting an 80% increase in uptake of HIV counselling and testing services. This study was assessed to have a high risk of bias because of insufficient detail on methodology (data collection methods, analytical approach, sample characteristics) of the end-of-project evaluation that presumably links the intervention activities to increase in uptake. The intervention seems promising given that it targets disabled people’s ability to perceive and seek SRH services, as well as approachability and acceptability of these services through wide stakeholder and community engagement.

Comprehensive sexuality education

For this review, we categorised school-based interventions, peer education, information provision in and out of schools as part of CSE. Of the seven interventions, three were found to have promising evidence, while the remaining had limited evidence supporting their effectiveness. All interventions promoted disabled people’s ability to perceive the need for services, with only one intervention (by Devine et al 2017) also promoting their ability to seek and engage.

Hanass-Hancock et al (2014) reported on the tool (Teachers’ Sexuality Education Questionnaire, TSE-Q) to assess teachers’ needs, knowledge, attitude, practice and self-efficacy in delivering sexuality education to disabled children in South Africa.33 They concluded that this tool has cross-cultural validity, working well to capture educators’ attitudes, practice, self-efficacy and perceived norms, although further work was needed to better capture educators’ knowledge. This study was assessed to have a moderate risk of bias because of limited detail on sample characteristics at individual level. However, this was not expected to significantly change the outcome of the study. This is a promising intervention in that it targets approachability and acceptability of CSE for children with sensory and intellectual impairments and could potentially be applied to other impairment types and in other contexts.

Hanass-Hancock et al (2018a) and (2018b) were two papers derived from the qualitative component of the formative evaluation on the CSE intervention Breaking the Silence implemented in South Africa, in eight schools for learners with special educational needs. Of these, the former study (2018a) explored educator’s perspectives, finding that the intervention can be delivered by educators after a 3-day training, and that it was effective in improving educator’s skills in delivering CSE to intellectually impaired children.34 However, this study was assessed to have moderate risk of bias, as there was little detail on how the sample of educators were recruited (eg, whether it was purposive to capture diversity of opinion). This was not expected to significantly alter the conclusions of the study.

Hanass-Hancock et al (2018b) explored contextual factors that inhibit children with disabilities in getting access to CSE, identifying factors including physical and information access barriers; negative social attitudes, educator and parent discomfort in discussing some topics (eg, genitalia).35 Crucially, they report that a critical mass of staff, including management, needs to be trained for greater effectiveness. This study was assessed to have low risk of bias, indicating promising evidence that this intervention is effective. The intervention targets approachability and acceptability of providing CSE to children with disabilities and shows great promise in terms of wider application to other contexts.

Robles-Bykbaev et al (2019) reported on an intervention for deaf women in Ecuador, creating a web-based platform where SRH information was conveyed through sign language as well as allow users to interact.36 They reported positive feedback and acceptability of the platform among people working with deaf women (mainly deaf educators) as did educators, clinicians and others. Reviewers initially differed in risk of bias assessments: some felt that the purposive sampling approach could have created bias, while others felt that this would not have significantly changed the outcome of the study. The intervention targeted approachability of SRH services for hearing impaired people through engagement with service providers and showed promise in wider application to deliver other SRHR topics and in other settings.

Oliveira et al (2018) evaluated a Braille manual delivering RH information over 3 to 15 days to women with visual impairment in Brazil.37 Through pre-intervention and post-intervention testing, they concluded that the manual improved participants’ knowledge, regardless of sociodemographic characteristics and congenital or acquired blindness. Reviewers assessed this study to have a high risk of bias as a result of their sampling method, sample size and lack of a control group that better determine the effectiveness of the intervention.

Aval et al (2019), too, reported on an educational package delivering RH information (over 2 days) to visually impaired women through Braille in Azerbaijan.38 Following pre-intervention and post-intervention tests, they reported a statistically significant improvement in knowledge regarding menstrual health, RH, STIs and pregnancy care. This study was assessed to have a high risk of bias given that possible confounders (eg, small sample size (n=26), and difficulties recruiting sample) did not seem to be considered in data interpretation. Neither of these interventions addressed any barriers in service provision and both seem limited in scope.

Yildiz and Cavkaytar (2017) conducted a case-control study of a Sexuality Education Program for Mothers of Young Adults with Disability (SEPID), an educational programme for families on how to communicate with their children about sexuality education in Turkey.39 The results showed that SEPID changed attitudes of mothers towards sexuality education for their children and improved perception of social support. However, this study was assessed to have moderate risk of bias as the sample recruitment may have had elements of self-selection. The intervention did not address any barriers from the supply/service provision dimension and was limited in scope.

Devine et al (2017) reported on a qualitative study assessing the effectiveness of a 3-year programme of participatory action research (including peer action groups) to improve access to quality SRH for women with disabilities in Philippines.40 They reported improvements in self-confidence in SRH seeking, SRHR knowledge and social participation among peers. Whether or not these effects are sustained over time was shaped by personal and community level factors. This study had a high risk of bias stemming from various confounders that were not reported in the data interpretation including conflicts of interest (intervention coordinator conducted the evaluation interviews), presumed intervention effect (participants were asked for ‘stories of change’) and insufficient detail on sample recruitment. This intervention held promise because it promoted disabled people’s ability to perceive SRHR needs, to seek services and to engage with services to demand adequacy. It was one of the two interventions in this review with direct links to healthcare utilisation, and a rigorous evaluation is needed to support its effectiveness.

Sexual violence

The theme ‘sexual violence’, including gender-based violence and female genital mutilation, had one intervention with promising evidence and two with limited evidence.

Salahi et al (2018) assessed validity of an intimate partner violence screening tool in women with mental health conditions at a psychiatric hospital in Iran.41 They found that the Farsi version of the Women Abuse Screening Tool (WAST) (and the WAST-Short Form) was easy to implement, suitable for initial screening in busy settings and correlated well with prevalence from the reference standard Conflict Tactics Scale-2. This study was assessed to have low risk of bias, indicating promising evidence to support this intervention’s effectiveness. The intervention, too, has promise because it extends application of a mainstream tool to women with mental health conditions, promoting people’s ability to perceive their SRHR needs as well as enhancing approachability of such services.

Mdikana and Phasha (2012) examined the functionality of school-based support teams (SBSTs) providing information support to children with intellectual disabilities identified to be at risk of or experiencing sexual abuse in South Africa.42 The study concluded that the SBSTs were functioning well despite receiving little support from the district-level teams, mentioning instances where SBSTs had gone beyond their role to provide counselling, accompanying the child to facility, helping with reporting to authorities. However, reviewers noted that this study had a high risk of bias relating to the use of convenience sampling, having a biased sample (only SBSTs were included)and insufficient detail on data collection tool and analysis. There may be promise to this approach involving school-based structures to respond to sexual violence experienced by intellectually impaired children. It promotes people’s ability to perceive the need for care as well as approachability of services. However, a more rigorous evaluation of functionality, barriers and facilitators, incorporating multiple perspectives, is needed.

Neherta et al (2014) reported on an intervention (via slides, videos, discussions) to improve mothers’ knowledge on sexual violence prevention for children with intellectual impairments in Indonesia.43 Using pre-interventions and post-interventions questionnaires, they reported improvements to mothers’ knowledge and attitudes. However, this study had a high risk of bias as several crucial quality criteria were not fulfilled. These included a lack of detail on sample characteristics, response rate, sampling method, whether participants were protected from negative or non-response (all mothers of children receiving care at the facility conducting the study), confounders and intervening factors that may have influenced participant attitudes and knowledge. There was little detail about the intervention, which targeted people’s ability to perceive SRHR needs and did not address any barriers from the supply/service provision dimension, making it limited in scope and application.

Sexual health, sexuality and rights

For this review, we included sexual health, sexuality and sexual dysfunction under one theme, and identified two interventions—one with promising evidence and the other limited.

Ghaderi et al (2017) reported on the effectiveness of a self-encouragement skills training to improve genital self-image in women with physical impairments in Iran.44 After conducting a quasi-experimental study with pre-test and post-test (with 25 women each in intervention and control arms) using the Female Genital Self-Image Scale, they reported improvements to participants’ genital self-image by 61%. However, this study was assessed to have a high risk of bias because of limited detail on control matching, sampling, disability assessments, and in acknowledging intervening factors that may have influenced participant responses. There was little detail about the intervention, which targeted people’s ability to perceive SRHR needs and did not address any barriers from the supply/service provision dimension, making it limited in scope and application.

Nakhli et al (2014) assessed the validity of Arabic version of the Arizona Sexual Experiences Scale (ASEX) that is used to assess sexual dysfunction among patients with schizophrenia in Tunisia.45 Administering the translated tool to patients with schizophrenia (n=100), they found it had ‘highly acceptable’ reliability and validity. The study was assessed to have low risk of bias despite limited information being provided about the sample: it was unlikely that this would alter the outcome of the study. This intervention held promise as the translated tool will extend the use of the ASEX to assess sexual dysfunction among people with schizophrenia in Arab speaking contexts, promoting people’s ability to perceive the need for care, as well as approachability of services.

Dimensions of access

Figure 3 below demonstrates the dimensions of Levesque framework of access25 targeted by the interventions in this review.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Interventions mapped on Levesque access model.

Evidently, there was a disproportionate focus on promoting people’s ability to perceive their SRHR needs: 15 of the 16 interventions in this systematic review focus on this. Though an important entry step in the pathway to achieving optimum SRHR, very few interventions went beyond information provision. Wilbur et al (2018, 2019)29 46 provided information on MHM to women with intellectual impairments as well as promoting their autonomy. Devine et al (2017)40 and Doherty et al (2018),32 despite limited evidence to support its effectiveness, designed interventions that promoted access through multiple domains. Fiander et al (2012),28 though also with limited evidence to confirm its effectiveness, targeted disabled people’s ability to pay for services to achieve better SRH.

Using the Levesque framework,25 we have highlighted that interventions for people with disabilities promoting their SRHR have rarely gone beyond raising awareness.