Introduction
Gender-based violence (GBV) is a major public health issue. It is projected that one in three women globally will face some form of abuse in childhood, adolescence, or adulthood [
1‐
4]. GBV is now recognized as an important global public health problem because of its acute and chronic impacts on women’s health. GBV includes physical, sexual, and psychological abuse from intimate partners or non-partners [
1‐
4]. The causes of gender-based violence are multi-dimensional, including social, economic, cultural, political, and religious factors [
5].
According to the World Health Organization multi-country study on violence against women, the lifetime and current (past 12 months) prevalence of physical or sexual intimate partner violence ranged from 15 to 71% and 4 to 54%, respectively, and the prevalence of emotional violence ranged from 20 to 75% [
6]. In another study conducted by the World Health Organization, it was estimated that the lifetime prevalence of intimate partner violence among female youths aged 15–19 was 29.4 and 31.6% for ages 20–24. The highest prevalence of intimate partner violence was reported in the African region, particularly in Sub-Saharan Africa (65.64%) [
7]. Evidence from Sub-Saharan Africa (SSA) showed high rates of GBV in educational institutions. Results from the Global Based School Survey (GBSS) revealed that the magnitude of current physical and sexual violence in five African countries ranged from 27–50% and 9–33%, respectively [
8,
9]. In research carried out in South Africa among adolescents aged 10–17 years, it was estimated that the lifetime prevalence (incident) of physical abuse was 56.3% (18.2%), emotional abuse 35.5% (12.1%), and sexual abuse 9% (5.3%) [
10]. Researchers revealed the prevalence of attempted rape (18.7%), actual rape (23.4%), physically violent harassment (8.7%), verbal harassment (24.2%), and forced sexual initiation (11.2%) among female students in Wolaita Sodo University [
11]. In other research, it was shown that the lifetime prevalence of rape was 11% among female secondary students in Arbamich [
12].
In several studies, researchers showed that gender-based violence perpetration and victimization were associated with a combination of different factors. For example, age, rural residence, number of children, having witnessed family violence as a child, educational status, marital conflict, and partner and personal use of alcohol, tobacco products, and illicit drugs were the predictors of gender-based violence [
9,
13‐
15].
GBV has been found to have detrimental effects on women, including injuries, sexual and reproductive health issues, mental health disorders, sexually transmitted infections (STIs), gynecological disorders, adverse pregnancy outcomes, an increased risk of non-communicable disease, and impacts on the health and wellbeing of their children [
16,
17]. Another health effect of GBV is that it increases women’s risk of a number of other health problems, including chronic pain, physical disability, drug and alcohol abuse, and depression [
18]. GBV also has a negative impact on a country’s human, social, and economic development and is an underlying obstacle to eliminating poverty and building peace [
16,
17]. Students who had experienced gender-based violence were more likely to report low school achievement and an increased school dropout rate compared to non-abused youths [
19,
20]. In a study conducted by WHO, it was found that schools and universities were highly vulnerable to GBV [
21]. However, this problem is not well addressed in educational institutions [
22].
As the above research has demonstrated, educational institutions are high risk spaces for GBV. This indicates that urgent intervention is needed to make educational institutions free of violence. To do this, systematically synthesized information is needed to design appropriate interventions and policies that target GBV in educational institutions in SSA.
Furthermore, over the last 20 years, a substantial amount of research on GBV has been conducted and published in SSA. However, the existing studies have not systematically identified and synthesized the prevalence of GBV among female youths in educational institutions [
23]. Therefore, this study aims to produce an overall summary estimate on the prevalence of gender-based violence according to different types and its risk factors among female youths in educational institutions of Sub-Saharan Africa.
Methods
Protocols and registration
This systematic review has been registered in the International Prospective Registry of Systematic Review (PROSPERO registration number
CRD4201073260, November 2, 2017). This systematic review followed the PRISMA guidelines [
24].
Eligibility criteria
The study focused on GBV (sexual, physical, and emotional/psychological violence) among young (10–24 years old) female students in educational institutions. Those studies which clearly reported the prevalence and risk factors for different types of GBV were included. The review included only studies conducted between 2000 and 2017 in SSA because most studies conducted in Africa focused on this area from the late 1990s in order to achieve the Millennium Development Goals, and no systematic review on GBV has been conducted in SSA among female youths in educational institutions. Published and unpublished papers were considered. The review considered studies involving female students in schools or universities or colleges using cross-sectional study design. The outcomes of interest were gender-based violence, sexual violence, physical violence, or psychological/emotional violence. Studies conducted in educational institutions using English were included in the review.
Exclusion criteria
All studies published before 2000 and published in another language were excluded. Articles which did not have a full text or abstract were excluded. Community-based studies were excluded. Studies focusing on both males and females but which did not report separately were excluded. Studies that did not clearly report either prevalence or risk factors for gender-based violence among female youths in educational institutions were excluded.
The sources of the information were identified by searching electronic databases such as MEDLINE, EMBASE, PsychINFO, CINAHL, Google Scholar, and PubMed. The reference lists of identified articles were searched for additional studies. Furthermore, a hand search of key journals was conducted. Unpublished studies were searched in Google Scholar, universities’ online libraries, and government organization’s websites. We also made exhaustive efforts to contact authors to request the articles for which we did not have the full text or abstract or which reported the missing/incomplete data.
Search strategy
The search strategy designed to access published and unpublished materials used the following search key terms and filters: “Gender based violence” OR “sexual violence” OR “physical violence” OR “psychological violence” OR “youth” OR “educational institutions” and “Sub-Saharan Africa” for each database. We consulted a librarian in designing the search strategy and searching the databases. The comprehensive database search was conducted on June 22–29, 2017. Further information regarding the search strategy of the selected databases is attached (see Additional file
1).
Selection of the study
The study selection followed PRISMA flow diagram [
24]. Endnote software was used to organize the papers. The relevance of the topic, objective, and methods of the study were checked. In the first stage, duplicates were removed. In the second stage, the title of the study was screened and those which did not meet the objective were excluded. In the third stage, abstracts of the studies were screened. Lastly, the contents of the remaining articles were assessed against the inclusion criteria.
Data collection process
The Joanna Briggs Institute (JBI) data extraction form [
25] was used by three reviewers (AB, HR, and AM) who independently extracted the data. Disagreement was resolved by discussion and consensus. If this was not possible, the matter was resolved by the fourth reviewer.
Main data items
The extracted data included the authors, year of publication, country, sample size, sampling technique, type of educational institutions, outcomes, tools used to measure outcomes, and risk factors of specific studies.
Quality of study
The quality of the studies and risk of bias were assessed by the checklist guidelines of Loney et al. [
26]. The following criteria were used: (1) specified the target population, (2) used adequate sampling techniques (e.g., random), (3) adequate sample size (> 300 participants), (4) adequate response rate (≥ 80), (5) measurement with valid and tested instruments (Conflict Tact Scales 2 (CTS2) [
27], WHO questionnaires [
28]), (6) reported confidence intervals or standard errors, (7) reported attempt to reduce observer or other form of bias, and (8) study subject described in detail. The reviewers classified the tools into CTS2 and WHO questionnaires for assessing gender-based violence against females, and finally, “own tools” where unknown instruments were used (Additional file
2).
Data synthesis
The individual studies were described concisely using a summary table. The summary table explained the characteristics of included studies and main findings. We used the random-effects model to pool the prevalence. The pooled lifetime prevalence of gender-based violence and type of GBV was computed using STATA version 14. Forest plot graphical representation and Cochran’s
Q test and
I2 were used to detect heterogeneity between the studies. Subgroup analyses were carried out to explore the potential sources of heterogeneity. Publication bias was checked using Eager’s weighted regression test [
29]. Meta-regression analysis was carried out to identify parameters (publication year, quality score, and sample size) associated with GBV. The pooled estimate of lifetime prevalence was reported as an overall for GBV and by type of GBV (however, we decided not to use the result from meta-analysis except for the overall prevalence of GBV, sexual violence, physical violence, and emotional violence, due to high heterogeneity). The results were expressed qualitatively. We did not perform meta-analysis for risk factors due to the factors being varied among the studies. The results were described qualitatively. The findings of the review were presented in summarized tables, text, and a PRISMA flow diagram.
Discussion
In Africa, studies on the prevalence and risk factors for physical and emotional violence are limited but sexual violence has been exhaustively investigated. Studies in Africa are highly heterogeneous due to the definitions used, sample size, tools used, and data collection methods, and more importantly, the culture, norms, and values of the studies varied. Therefore, it is difficult to generalize the results.
Despite these variations, this review found that the overall prevalence of gender-based violence in Sub-Saharan Africa was high, ranging from 42.3 to 67.7%. This finding was relatively similar with the finding from an analysis of WHO data in 2014 among adolescents and young women that ranged from 19.0 to 66.0%) [
54]. This may be due to the sample size and composition of participants. The majority of studies used their own self-administered questions and not standardized cross samples. These questions may not have been validated for use with the populations that were sampled. This may lead to overestimates or underestimates of the prevalence of GBV.
The lifetime prevalence of sexual violence ranged from 4.30 to 76.40%. This wide range may be due to definitions and tools used to measure sexual violence across studies, as well as the differing samples. The majority of studies measured sexual violence using a single item. This may have led to underreporting or overreporting of the prevalence of sexual violence. This finding is higher than a meta-analysis conducted in 2011 that examined the global perspective of child sexual abuse (18.0%) [
55], a worldwide systematic review and meta-analysis conducted on sexual violence (15.0%) [
56], and in research that used a community sample (19.4%) [
57]. The difference may be due to sample size, composition, and definitions and tools used to measure sexual violence. The sampled population varied from high school students to university students; therefore, the prevalence of sexual violence was varied.
The lifetime prevalence of physical violence ranged from 7.4 to 66.1%. However, there was high variation among the studies. These variations may have been due to culture, norms, socioeconomic factors, and definitions and tools used to measure physical violence. These variations may lead to underreporting or overreporting of the prevalence of physical violence. The other possible explanation may be women in developing countries justify the fact that men beating their partner is acceptable; males dominating females is prevalent in Africa [
58]. Additionally, the majority of these systematic reviews used their own self-reporting tools, and the majority of studies used various definitions of physical violence and measured prevalence using single items as well as differing samples.
The lifetime prevalence of emotional violence ranged from 26.1 to 50.8%. However, there was a high degree of variation between studies. The differences between studies might be because of different culture, societal factors, and economic status. Studies used various definitions of emotional violence. Only five studies reported emotional violence. This indicates a lack of studies on this topic.
This systematic review also identified different factors associated with GBV. Living arrangements were associated with GBV [
31,
33]. This is consistent with studies conducted in other parts of the world [
9,
13,
59,
60]. This might be due to rural residents perhaps not having access to health information and infrastructure. This review found that witnessing parental violence was a risk factor for GBV [
31,
32,
36]. This is similar with the studies carried out around the globe [
9,
13,
14,
54,
61].
This review demonstrated that substance use was a contributing factor for GBV [
31,
32,
35,
36]. This is concomitant with the previous studies [
8,
9,
13,
14,
54]. This may be due to alcohol influencing decision making, which may lead to GBV. Marital status was also found to be a risk factor of GBV [
34,
36]. This is in line with previous studies around the globe [
9,
13,
14]. This review also found that educational status was associated with GBV (three studies). This corroborates with previous studies in other parts of the world [
9,
13,
59,
60]. This might be due to educated women having more of an awareness of GBV than illiterate women.
This review showed a strong relationship between witnessing parental violence during childhood, and sexual violence [
31,
32,
41,
50]. This finding corresponds with a study from the USA [
62,
63]. Being a rural resident was found to be a risk factor for sexual violence [
31‐
33,
47]. Alcohol consumption was the most common factor associated with sexual violence [
31,
32,
34,
35,
38,
40,
49‐
51]. This is supported by the evidence from high-income countries [
64,
65]. Being sexually active was associated with sexual violence [
31,
48,
51]. This association was also found in studies in high income countries [
63]. This review also found having peers was a contributing factor for sexual violence [
39,
41,
45,
48]. This corresponds with a study from high-income countries [
63]. One possible explanation may be the persons may not be interested to reject by their peers.
Witnessing parental violence during childhood was found to be associated with physical violence [
31,
35,
37]. This is consistent with a study from a high-income country [
66]. Alcohol drinking was found to be associated with physical violence [
34,
38,
44]. This corroborates with previous studies across the world [
64].
This review also found factors associated with emotional violence. Alcohol consumption was found to be a risk factor for emotional violence [
34,
38]. This is consistent with studies around the globe [
64,
65,
67]. Witnessing parental violence was also found to be a contributing factor for emotional violence [
37]. This was also evidenced by meta-analytic review [
68].
Furthermore, this systematic review showed that witnessing violence inside the home is associated with GBV in the educational setting. Those students who have witnessed violence at home also had increased risks of experiencing sexual, physical violence, and emotional violence in the educational setting. This might be due to continuous disagreement between their parents/caregivers. This has intergenerational consequences. The other possible explanation may be students who have witnessed violence at home have poor parental supervision and less parental support.
Implications of the study
Even though there was a high degree of heterogeneity between studies, GBV is still a significant public health problem. This reveals that GBV is common among youths in educational institutions of SSA. This review has implications for the design of intervention, policy, and programming in SSA. It also suggests that a youth violence prevention policy, intervention strategies, and service provisions are needed in SSA. GBV has a detrimental effect on youths, but especially women and girls, and is an obstacle to achieving the Sustainable Development Goals (SDGs). In order to achieve Sustainable Development Goal 5 (gender equality), eliminating GBV in educational institutions and the community through educating the community/parents and students about the prevalence, causes, and consequences of problem is crucial. This review will also help government policy makers, non-government organizations, and other stakeholders to alleviate the burdens of gender-based violence. All Sub-Saharan African countries should develop a community and school-based intervention program to address gender-based violence against youths at educational institutions and community settings.
Weakness and strengths of study
There was high heterogeneity between studies. This systematic review used institutional-based cross-sectional survey, and so causality cannot be established. The review did not include youths who do not attend school. Additionally, the majority of studies did not use the international standardized questionnaire. The majority of studies were of moderate quality. They may be subject to recall and social desirability bias.
Conclusion and Recommendation
This systematic review found that the prevalence of overall gender-based violence, sexual, physical, and emotional violence was high. However, the results should be interpreted with caution because of high between-study heterogeneity. This review also found that living arrangements, educational status, marital status, witnessing parental violence, substance use, sexual risk factors, and peer pressure was strongly associated with gender-based violence. The results highlight the need for government policy makers, non-governmental organizations, program designers and other stakeholders to develop effective intervention and prevention strategies, and programs to reduce gender-based violence in educational institutions. A comprehensive educational institution-based prevention strategy and effective interventions should be developed to mitigate gender-based violence. We also identified a paucity of studies examining emotional violence in the educational setting in Sub-Saharan Africa. The authors recommend that further studies should be carried out by using a longitudinal study on gender-based violence among in-school and out-school youths in order to establish causality. The future study should focus on ploy-victimization of youths in SSA. Additionally, future research should focus on the culture of the community to identify the risk and protected factors of gender-based violence among youths in SSA. Moreover, due to the high heterogeneity of the included studies, future GBV studies undertaken across SSA should utilize a standardized methodology which could allow for comparisons to be made over time.