Sexual Risk Behaviour Among School-going Adolescents in Sierra Leone and Liberia. A Secondary Analysis of the 2017 Global School Health Survey

Background: There is very little information on sexual risk behaviour among Sierra Leone and Liberia school-going adolescents. The present study assessed the prevalence and determinants of sexual risk behaviours among school-aged adolescents in Sierra Leone and Liberia. Method: We used publicly available nationally representative cross-sectional datasets of the 2017 Sierra Leone and Liberia Global school health survey. The sample consisted of 2798 and 2744 school-going adolescents from Sierra Leone and Liberia, respectively. Results: The majority of adolescents in the two countries were involved in multiple sexual risk behaviour (80.2%), with higher prevalence observed in Sierra Leone (85.2%) than in Liberia (75.3%), Liberian adolescents showed lesser odds of indulging in multiple sexual risk behaviours than their Sierra Leonean counterparts (AOR=0.572; 95%CI: 0.345-0.946). Male compared to females were more likely to engage in multiple sexual risk behaviour (AOR=2.310;95%CI:1.543-3.458), with a similar pattern observed in both countries. Alcohol use was associated with multiple sexual risk behaviour (AOR=3.064; 95%CI: 2.137-4.392). Also, in Sierra Leone and Liberia, adolescents with one and two or more forms of psychological distress were more likely to have ever had sex than those who do show any form of psychological distress. Missing class/school was associated with multiple sexual risk behaviour (AOR=1.655; 95%CI:1.133-2.418). Peer support was only found to be a protective factor against no condom use among Liberian adolescents (AOR=0.608; 95%CI: 0.435-0.850). Less parental support was only associated with ever had sex more likely to have ever had sex as seen among adolescents in Sierra Leone (AOR=2.027; 95%CI: 1.322-3.107) but not Liberia (1.034(0.650-1.644). Conclusion: Our study found a high sexual risk behaviour among school-going adolescents in Sierra Leone and Liberia. Our nding highlights the need to strengthen sexual and reproductive health education in schools and communities that incorporate mental health


Introduction
Youths in Sub-Saharan Africa are prone to risky sexual activities, unwanted pregnancy, and sexual violence [1,2]. Having unprotected sexual intercourse; having multiple sexual partners over one's lifetime; having intercourse with a casual partner; sexual initiation at a young age; sexual intercourse with commercial sex workers; bartering sex for money, goods, or other favours; engaging in sexual activity while under the in uence of alcohol/drugs; and sexualism are all examples of risky sexual behaviours [3].
Adolescence is characterised by greater autonomy, peer in uence, risk-taking behaviours such as initiation of sex and alcohol/drug usage, [1,4]. Compared to adults, adolescents are more likely to have several sexual relationships, participate in unprotected sexual intercourse, and choose high-risk partners [5]. The study of teenage sexual behaviour is crucial because 60% of youths globally are a icted with sexually transmitted infections (STIs), including HIV [6].
Earlier studies in Liberia reported a signi cant prevalence of risky sexual practices among in-school children and young adults. 78% of kids were found to be sexually active; 24.9% of those sexually active said they had sex for money, and 20.9% said they had never used a condom [7]. In the same study, males were also shown to have more sex, have many sexual partners, and start sex earlier than females.
Another study among Liberian youths found that 34% of those who were sexually active did so before the age of 15 (early sexual debut), and 21% of those who were sexually active had several sexual partners, and 26% of sexually active teenagers had never used a condom, 11% had gotten pregnant or helped someone become pregnant one or more times, and 11% had been sexually assaulted [8].A recent study reported that majority of sexually active Sierra Leone youths had condomless sex in their last sexual encounter [9]. A United Nations Population Fund report on the impact of Ebola on adolescent pregnancy in Sierra Leone found that close to half had their rst pregnancy during the Ebola outbreak period and close to a third had used ever use of any kind of family planning [10].
Various sexually related risk behaviours have been observed among adolescents in African countries. In Ghana, 33.5% of adolescents ever had sex, 73.8% had not used a condom at last sex, and 32.5% had multiple sexual partners [11]; in Namibia in 2004, 33.2% ever had sex, and 17.1% had multiple sexual partners [12,13]. Between 2015 and 2017, a community survey of adolescents (15-19 years) in Uganda, Tanzania, Nigeria, Ghana, Eswatini, Ethiopia, and Burkina Faso found that 25.9% had ever had sex.
Among sexually active adolescents, the early sexual debut was 21% for girls and 28% for boys, unprotected last sex was 46% for girls and 40% for boys, and 37% for girls and 8% for boys had made someone pregnant [4].
In a study of 15-year-olds in 30 European countries, Israel, and Canada, 27% had had sexual intercourse, and 14% had not used the contraceptive pill or condoms at their most recent sex [14], and in a study of 15-year-olds in 10 European countries, the prevalence of sexual initiation was 18.8%, and among sexually active, 52.4% had >1 sexual partner [15].
Although the commonness of sexual behaviour varies by country and culture, the relationships between sexual and non-sexual risk behaviours and the function of psychosocial modulators may follow similar patterns. A previous study has identi ed factors associated with sexual risk behaviour among adolescents (ever had sex, early sexual debut, no condom use, and no contraceptive use), and they include male sex, older age, substance use, psychological distress, school truancy, and a lack of parental and peer support [16]. Substance abuse has been connected to a higher chance of youths participating in unsafe sexual practices. When comparing youths who use substances to those who do not, studies show that those who use substances are more likely to engage in early sexual intercourse, have many sexual partners, and use condoms at a lesser rate [17,18]. Few studies have been undertaken among Liberian youths to investigate the link between risky sexual behaviours and substance abuse. One study revealed no link between alcohol consumption and transactional sex [19],while another identi ed a link between alcohol consumption and having several sexual partners-but no other substances were investigated [7]. Other drugs, such as marijuana, cocaine, stimulants such as methamphetamines, have been linked to risky sexual behaviours in adolescence [7,20,21].
Sierra Leone and Liberia are neighbouring countries that have shared unique history. Both countries populations have experienced civil war and, most recently, an Ebola disease outbreak leading to profound economic hardship, psychopathologies such as posttraumatic stress disorder, depression, and psychosis as well family disruption [20,[22][23][24]. These mental health morbidities and family disruption were more profound among adolescents and young people, making them vulnerable to indulge in at risk behaviours such as sexual risk behaviours like early sexual debut, having multiple sexual partners, and not using condoms [10,25,26]. Also, the adolescent birth rate in these countries is reported to be high and above the average sub-Saharan Africa [27]. Given such a unique background of these two countries, it is important to examine how personal, psychosocial, and protective factors in uence sexual risk behaviour, especially among adolescents. Currently, there is limited national data on sexual risk behaviour and related risk factors among only adolescents. Most studies conducted in these countries are either community based or are among adolescents and adults combined [7,19,28,29]. Knowing the prevalence of sexual behaviour and the risk factors associated with it among teenagers in Sierra Leone and Liberia can aid in developing intervention programs aimed at delaying sexual initiation and encouraging "safer sex." As a result, the goal of this study was to assess the prevalence and determinants of sexual risk behaviours among school-aged adolescents in Sierra Leone and Liberia using their 2017 Sierra Leone and Liberia Global school health survey (GSHS).

Sample and procedure
We used publicly available nationally representative cross-sectional datasets of the 2017 Sierra Leone and Liberia Global school health survey [30]. The Sierra Leone and Liberia GSHS employ a two-stage cluster sample design to obtain a nationally representative sample of school-going adolescents. The rst stage involves the selection of schools with probability proportional to enrolment size, whilst the second stage involves randomly selecting classes for which all students have equal chances of being selected. In the Sierra Leone GSHS, the school response rate was 94%, the student response rate was 87%, and the overall response rate was 82% [30]. In the Liberia GSHS, the school response rate was 98%, the student response rate was 73%, and the overall response rate was 71% [30].

Measures
The questionnaire used in this study and the de nition of the variables is shown in Table 1. Sexual risk behaviour was considered as the outcome variable in our study, and it was assessed using the following questions ever having had sexual intercourse, age of sexual debut, number of people who have had sexual intercourse within a lifetime, condom use at last sexual intercourse, and birth control use at last sexual intercourse. Sexual risk behaviour was de ned as ever having had sex, early sexual debut (<14 years), having had two or more sexual partners in a lifetime, non-condom use at last sex and nonbirth control use at last sex. Composite sexual risk behaviour was de ned as having had sex, early sexual debut (<14 years), having had two or more sexual partners in a lifetime and non-condom use at last sex.
As reported in previous studies [31,32], we excluded non-birth control use because of the overlap with non-condom use at last sex. The Independent variables considered in this study are in Table 1. As in a previous study [31], we considered no close friends, loneliness, anxiety, suicidal ideation, and suicide attempt as psychological distress items. Based on similar study by Pengpid and Pelzer [33], we summed these items into three groups -0 = 0, 1 = 1 single and 2-5 = 2 multiple. School attendance, peer and parental or guardian support was considered protective factors and they groups. The four items that measure parental or guardian support were summed and divided into three groups -0-1 as low, 2 as medium and 3-4 as high support.'

Ethical consideration
No formal ethical approval to conduct this study was necessary, given that our study is based on an analysis of a publicly available deidenti ed secondary dataset. Notwithstanding, ethics approval was obtained from the Ministries of Health in Sierra Leone and Liberia prior to conducting the surveys in the two countries.

Data analysis
We analysed our pooled data from the two sets of surveys using SPSS version 27. We employed descriptive statistics to describe our sample. Chi-square statistics were used to compare the independent variables between Sierra Leone and Liberia. Binary regression statistics were used to determine the correlates of individual sexual risk behaviours (non-birth control use at last sex, non-condom use at last sex, multiple sexual partners, early sexual debut, and ever had sex) and a composite measure of multiple sexual risk behaviour. We excluded current tobacco use as an independent variable in our analysis because the data was not available in the Sierra Leone GSHS dataset. We employed complex samples analysis in all statistical procedures to account for the sampling weights and the multi-stage design. Statistical signi cance was set at p<0.05.

Results
Characteristics of the sample and sexual risk behaviour in Sierra Leone and Liberia  Associations with sexual risk behaviour        [31,34] but higher than what was reported in Ghana [11], and in four Caribbean countries [35] and ve East Asia countries [36]. However, in individual countries, the reported prevalence in Mozambique was higher than what is found for Sierra Leone but low regarding our nding for Liberia.
Close to a third had an early sexual debut (<14 years) with a higher prevalence seen among Sierra Leoneans, and this was lower than what other similar studies reported in Caribbean and Asian countries [35,36]. However, the prevalence of early sexual debut for Sierra Leone and Liberia was consistent and higher than what was reported in studies conducted in Ethiopia and Mozambique [31,34], respectively. In addition, close to half (48.7%) did not use a condom in their last sex, with higher prevalence seen among Sierra Leoneans (58.6%) than Liberians (39.8%), and our ndings for both countries are higher than the prevalence reported in studies conducted in Mozambique, Ghana and Ethiopia [11,31,34]. Our study's high prevalence of risky sexual behaviour is consistent with previous community-based studies conducted among adolescents and youths in both countries [7][8][9]29] and similar studies in other African countries [1,4,11,31]. The higher prevalence of risky sexual behaviour in our study may be attributed to increased trauma and economic hardships, changes in parenting styles, breakdown in the social fabric in our society and broken homes experienced by young people, which is due to civil war and lately the Ebola outbreak in these two countries [37][38][39].
Consistent with a previous studies conducted in Ghana [11], Mozambique [31], Fiji, Kiribati, Samoa, and Vanuatu[16], males were more likely to have ever had sex, had an early sexual debut (<14 years), multiple sexual partners and high sexual risk behaviour composite score. Similarly, being 17years and older was associated with being sexually active, having multiple sexual partners, and non-birth control uses other than condoms. Our ndings suggest the need to develop and implement gender and age-speci c interventions that will help prevent adolescents from indulging in risky sexual behaviours. Substance use (alcohol and cannabis use) was associated with sexual risk behaviour among adolescents in Sierra Leone and Liberia. Our nding aligns with previous studies conducted in Liberia [7] and some African countries [11,31]. Substance abuse has been reported to be higher among adolescent in Sierra Leone and Liberia [20,21,40], and such behaviour have been reported to be associated with risky sexual behaviour and teenage pregnancy [7,41].
In contrast to a Mozambican study [31] but consistent with a Ghanaian study [11] and in the Caribbean as well as Asian paci c island countries[16, 35,36], psychological distress was associated with sexual risk behaviours such as ever had sex, early sexual debut and non-condom use. High mental health burden has been reported among adolescents in Sierra Leone and Liberia, and such burden is linked with exposure to trauma during the civil war and Ebola outbreak in these two countries [22,26]. Psychological distress contributes to adolescents being vulnerable, leading to risky sexual behaviour, violence, substance abuse, partly caused by economic hardship, and parental loss [42]. Although peer support was identi ed as a protective factor in only one of the sexual risk behaviour indicators (non-condom use) among Liberian adolescents, previous studies have concluded that negative peer in uence affects adolescents' sexual risk behaviour [23,43]. Low parental support was associated as with ever had sex but was not linked with the other sexual risk behaviour indicators, which is consistent with other studies in which parental support was a protective factor for non-condom use and non-birth control use at last sex but not multiple sexual risk behaviours [11,31]. Our nding may re ect the signi cant trauma due to war, poverty and infectious disease outbreaks in these countries leading to mental health needs that have not been fully addressed. Such an unmet need makes it di cult for adolescent to bene t from protective factors such as peer and parental support.

Policy and Practice Implication
Our ndings underscored the need for strengthening sexual and reproductive health education in schools and communities. Such educational programs should integrate mental health promotion activities targeting adolescent speci c health needs, such as increasing their access to counselling and education. Also, peer support and parental involvement in adolescent daily activities will help reduce adolescent sexual risk behaviour.

Study Limitations
The 2017 Sierra Leone and Liberia GSHS employed a cross-sectional study design, and as such, we cannot infer causal relationships between our dependent and independent variables. Also, our ndings are only applicable to school-going adolescents in these two countries. Future research should focus on both in-school and out-of-school adolescents. There is a tendency for recall bias as responses were based on self-report. GSHS does not provide a clear de nition of 'sexual intercourse, which would have led to the possibility of some respondents misinterpreting some questions.

Conclusion
Our study suggests that most school-going adolescents in Liberia and Sierra Leone have indulged multiple sexual risk behaviours although those in Liberia had fewer odds of being involved in multiple sexual risk behaviours than their Sierra Leone counterparts. Sex, substance use, psychological distress and missing classes were associated with multiple sexual risk behaviours. Peer and parental support were the only protective factors for no condom use among Liberian adolescents and being sexually active among Sierra Leonean adolescents. Our nding highlights the need to strengthen sexual and reproductive health education in schools and communities that incorporate mental health promotion activities.