Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Low condom use at the last sexual intercourse among university students in sub-Saharan Africa: Evidence from a systematic review and meta-analysis

  • Jonathan Izudi ,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Software, Validation, Writing – original draft, Writing – review & editing

    jonahzd@gmail.com

    Affiliation Faculty of Medicine, Department of Community Health, Mbarara University of Science and Technology, Mbarara, Uganda

  • Gerald Okello,

    Roles Conceptualization, Methodology, Validation, Writing – original draft, Writing – review & editing

    Affiliation Department of Reproductive Health, Ministry of Health, Kampala, Uganda

  • Daniel Semakula,

    Roles Conceptualization, Data curation, Formal analysis, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliation African Centre for Systematic Reviews and Knowledge Translation, Makerere University College of Health Sciences, Kampala, Uganda

  • Francis Bajunirwe

    Roles Conceptualization, Data curation, Formal analysis, Methodology, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliation Faculty of Medicine, Department of Community Health, Mbarara University of Science and Technology, Mbarara, Uganda

Abstract

Background

There is inconsistent data about condom use at the last sexual intercourse (LSI) among university students in sub-Saharan Africa (SSA) and its association with sex, age, and condom negotiation efficacy. The primary objective of this study was to summarize the proportion of condom use at the LSI among university students in SSA. The secondary objective was to determine the association between condom use at the LSI with sex, age, and condom negotiation efficacy among university students in SSA.

Methods

In this systematic review and meta-analysis, two reviewers independently searched electronic databases and grey literature for eligible studies published until July 30, 2020, extracted data, and assessed the risk of bias in the included studies. We used the Dersimonian-Liard random-effects model to pool the proportion of condom use at the LSI and the association between condom use at the LSI with sex, age, and condom negotiation efficacy, reported using risk ratio (RR). We assessed publication bias using funnel plot and Egger’s test, and explored sources of heterogeneity using sub-group and meta-regression analyses.

Results

We meta-analyzed 44 studies with a combined sample size of 27,948 participants.Of 14,778 sexually active participants, 8,744 (pooled proportion, 52.9%; 95% CI, 45.0–60.7; 95% prediction interval, 2.8–98.9; I-squared = 99.0%, p< 0.0001) reported condom use at the LSI and the proportion of condom use at the LSI remained stagnant between 2000 and 2019 (p = 0.512). Condom use at the LSI was not associated with being a female compared to a male (pooled RR, 1.08; 95% CI, 0.68–1.71), being of a younger age (≤24 years old) compared to older age (25 years and more) (pooled RR, 1.16; 95% CI, 0-85-1.57), and having a higher condom negotiation efficacy compared to a lower condom negotiation efficacy (pooled RR, 1.54; 95% CI, 0-81-2.94).

Conclusions

We found a low and heterogenous use of a condom at the LSI among university students in SSA which was not associated with sex, age, or condom negotiation efficacy. Accordingly, context-relevant interventions are needed to improve condom use at the LSI among university students in SSA.

Introduction

In 2019, an estimated 3,400,000 young people (15–24 years old) were living with human immunodeficiency virus (HIV) globally, and two out of every seven new HIV infections were among young people [1]. Every day in 2020, global estimates indicated that 4,000 children and adults were newly infected with HIV and 60% of those infections were in sub-Saharan Africa (SSA) while 90% were among people aged ≥15 years old [2]. Young people comprised at least one-third of the daily new HIV infections among persons aged ≥15 years in 2020 [2].

In SSA, HIV remains the leading cause of morbidity and mortality among young people, and the predominant mode of transmission is condomless sexual intercourse [3]. The notable risk factors for HIV acquisition in young people include early age at sexual debut [3], multiple sexual partnerships [3], misconceptions about condom use which is worsened by negative attitudes [4], and sexual relationships with older men (age-disparate relationships) who are more likely to be HIV infected [4]. Recent data show that age-disparate relationships are associated with condomless sexual intercourse and HIV acquisition among adolescent girls and young women [5]. For that reason, risk reduction measures like abstinence from sexual intercourse, delayed sexual initiation, reduced number of sexual partners, and increased access to and use of condoms is important in preventing young people from acquiring sexually transmitted infections (STIs) as well as HIV and unwanted pregnancies [6].

University students comprise a large proportion of young and sexually active people who are at a higher risk of HIV acquisition [7]. The university environment creates a setting for high-risk sexual behaviors especially condomless sexual intercourse and multiple sexual partnerships [8].

Studies conducted in SSA have reported that a substantial proportion of sexually active university students had not used a condom at their last sexual intercourse (LSI) [7,9,10], placing them at a higher risk of acquisition of STIs including HIV. Studies have also reported that being in a steady sexual relationship [10], having a low perceived risk of HIV acquisition or indulging in unplanned sexual intercourse [9], perceiving that condoms are used solely during fertile days [10], and perceiving that condoms decrease sexual pleasure [9,10] are associated with increased likelihood of condomless sexual intercourse among university students.

Currently, there is a lack of aggregated data on condom use at the LSI among university students in SSA including the associated factors. Also, existing studies on the African continent have shown wide differences in condom use among university students, ranging from below 15% in some countries [1114] to nearly 50% in others [15,16] and more than 75% in other countries, although from fewer studies [17,18]. In addition, the studies have shown inconsistent findings regarding the association between sex, age, and condom negotiation efficacy with condom use at the LSI [1921]. For instance, compared to males, some studies [22,23] have reported that females have a decreased likelihood of condom use at the LSI while other studies [24,25] have reported an increased likelihood of condom use at the LSI. Based on these inconsistencies, we conducted a systematic review and meta-analysis to primarily summarize the prevalence of condom use at the LSI among sexually active university students in SSA. As a secondary objective, we determined the association between younger age (≤24 years), being a female, and having a higher condom negotiation efficacy with condom use at the LSI. This evidence will justify and inform the strategic design of condom promotion campaign programs among university students in SSA, thereby contributing to reducing HIV morbidity and mortality.

Materials and methods

Study design and registration

We adhered to the elements of the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) [26] (S1 Table) and Meta-analysis of Observational Studies in Epidemiology [27]. We registered the study protocol with PROSPERO and the assigned registration number is CRD42020196868 [28].

Inclusion and exclusion criteria for studies

Our eligibility criteria included: 1) Type of participants: studies should have enrolled university students, either undergraduate or postgraduate or both; 2) Exposure of interest: being a female, younger age (≤24 years), and having a high condom negotiation efficacy; 3) Comparison group: being a male, older age (25 years and more), and having a low condom negotiation efficacy; 4) Time: all studies published until July 30, 2020; 5) Study designs: all observational studies, namely cross-sectional, case-control, and cohort; 6) Study setting: both private and public universities in SSA.

We excluded studies with the following characteristics: high risk of bias, not reported in English language, inaccessible full-text articles despite contacting the corresponding author, unclear or incorrect reporting of the primary outcome, primary outcome reported for none sexually active participants, the association between exposure and outcome reported using unadjusted odds ratio (OR), unadjusted risk ratio (RR) and unadjusted hazard ratio (HR), conducted outside SSA, and studies that combined data from a university in SSA with data from another university outside of SSA. We also excluded qualitative studies and conference presentations or abstract papers.

Study outcomes and variables

The primary outcome was condom use at the LSI measured as the number of participants with one or more sexual partner(s) in the past 12 months (sexually active participants) who had used a condom at the LSI, expressed as a percentage. The secondary outcome was the association between being a female, being of a younger age, and having a high condom negotiation efficacy and condom use at the LSI. For the secondary outcome, we considered only the adjusted estimates for OR, RR, and HR. The two primary outcomes, “always used a condom” and “consistently used condom”, were considered synonymous with condom use at the LSI. We defined condom negotiation efficacy as the perceived capability to use condoms consistently, including how difficult a person considers it to realize the desired healthy behavior [4].

Search strategy and process

Two reviewers (JI and DS) developed a search strategy using key concepts in the research question and for each key concept, a Medical Subject Heading (MeSH) term was developed and thereafter combined with Boolean operators: “AND”, “OR”, and “NOT”. The final search strategy used in PubMed/Medline was as follows: (“University student” OR “University students”) AND ("Condom use at last sex" OR "Condom use at last sexual intercourse" OR "Condom use at latest occasion" OR "Condom use at recent sex" OR "consistent condom use" OR "Consistent use of condom" OR "Inconsistent use of condom" OR "Inconsistent condom use" OR "Always use condom"). An example of the full search strategy in PubMed is shown in the supporting information (S1 File). Two reviewers (JI and GO) independently searched MEDLINE/PubMed, EMBASE, Web of Science, GoogleScholar, and Google iteratively for eligible studies between July 20, 2020, and August 30, 2020. The reviewers also hand-searched the reference list of eligible studies to identify other studies which were not identified by the search strategy.

In addition, grey literature like dissertations and reports were searched through LILACS and OpenGrey. We summarized and presented the results of the search strategy in a PRISMA flow chart.

Screening of studies and data abstraction

All the identified citations were exported into EndNote, a bibliographic reference management software, and duplicated citations were excluded. Two reviewers (JI and GO) screened the remaining studies for eligibility using the titles and abstracts and the ineligible citations were excluded. The reviewers retrieved and read the full-text articles of eligible studies and abstracted the following data items using a standardized Microsoft Excel sheet: first author’s last name, publication year, study design, sample size, study setting, country, number of sexually active participants, number of sexually active participants that had used a condom at the LSI, and the adjusted measure of effect (OR, RR, or HR) for each of the exposure variables. In studies where the adjusted measure of effect was reported for the comparison group or no use of a condom at the LSI, we computed the reciprocal of that adjusted measure of effect to obtain that for the exposed group and condom use at the LSI, respectively. For each country, we retrieved the 2019 human development index (HDI), the most recent measure of each country’s development concerning healthcare, education, and life expectancy [29]. The measure categorizes countries into low, medium, and high HDI.

Consensus in data abstraction and risk of bias assessment

We resolved disagreements in data abstraction through consensus with a third (DS) or fourth (FB) reviewer. To assess consistency in the screening of studies and data extraction between the two reviewers (JI and GO), we computed Kappa statistics to ascertain the degree of agreement. This is important because coding behavior changes between and within individuals over time[30].

We assessed the risk of bias in the included studies using a 9-item quality assessment checklist for prevalence studies that had a total of nine scores [31]. This tool is widely used in systematic reviews to assess the risk of bias in cross-sectional studies because it is easy to use, has a high inter-rater agreement, and a nearly perfect agreement between the individual items. The tool measures the risk of bias in each study on a binary scale (yes or no) and the nine items included whether 1) the study’s target population was a close representation of the national population concerning relevant variables like age, sex, and occupation among others; 2) the sampling frame was a true or close representation of the target population; 3) random selection was used to select the sample or a census was undertaken; 4) the likelihood of non-response bias was minimal; 5) data were collected directly from the subjects as opposed to a proxy; 6) an acceptable case definition was used in the study; 7) the study instrument used to measure the parameter of interest (condom use at the LSI) was shown to have reliability and validity; 8) the same method of data collection was used for all the participants, and 9) the numerator and denominator for condom use at the LSI were appropriate. The risk of bias in each study was summarized as follows: high when the total score was 7–9, moderate when the score was 4–6, and low when the score was 0–3.

Data analysis

Analysis was performed in R programing language and statistical software version 3.5.2 [32]. In the primary analysis, we summarized and displayed the characteristics of the included studies in an evidence table. The primary outcome was condom use at the LSI computed as a proportion. The numerator was the number of participants that had one or more sexual partner(s) in the past 12 months (sexually active participants) and had used a condom at the LSI. The denominator was the number of sexually active participants in the primary study. In pooling the proportion of condom use at the LSI, we used the DerSimonian-Laird random-effects model, allowing for Freeman-Tukey double arcsine transformation to stabilize the variances [33].

We reported the pooled proportion of condom use at the LSI with the subsequent 95% CI and the 95% prediction interval (PI) in a forest plot. We included a 95% PI to demonstrate the variation in condom use at the LSI in different settings including the direction of evidence in future studies [34]. In the secondary outcome analysis, we applied the DerSimonian-Laird random-effects model for binary outcomes to determine the association between sex (female versus male), age (≤24 years old versus 25 years and older), and condom negotiation efficacy (high versus low) with condom use at the LSI, expressed using risk ratios (RR) with 95% CI. Similarly, we displayed the results graphically using a forest plot.

Assessment of statistical heterogeneity

We assessed heterogeneity among the included studies using Cochran’s (Q) test and quantified using the I-squared statistics [35]. To investigate the sources of heterogeneity, we performed sub-group and random effects meta-regression analyses [36]. Here, we used the primary study characteristics, namely the study design, year of publication, country, HDI, risk of bias, and sample size.

Publication bias assessment

We used a funnel plot and Egger’s test to assess the studies for possible publication bias. An asymmetrical funnel plot was interpreted as suggestive of publication bias [37,38]. We confirmed the asymmetry using a probability value of less than 0.1 (p<0.1) for Egger’s test [37]. We used a contour-enhanced funnel plot to aid the interpretation of funnel plot asymmetry and to differentiate between genuine publication bias and small study effect [39]. For confirmed publication bias, we applied the trim and fill analysis to estimate the number and outcome of missing studies [40].

Sensitivity analysis

To examine the influence of the included studies on the methodological robustness, meta-analytic results, and conclusions including the study quality and sample size, we performed sensitivity analysis [41].

Results

Summary of study identification and screening

We identified 3,014 citations: 344 through electronic databases and 2,670 through other sources. Of the identified citations, 17 duplicates were excluded leaving 2,997citations. Another 2,924 irrelevant citations were excluded after screening the titles and abstracts. Therefore, we assessed 73 full-text articles for eligibility and excluded 39 of them with reasons. From the reference list of the remaining 34 eligible full-text articles, we identified 10 additional studies. Overall (Fig 1), we meta-analyzed 44 studies [7,1118,2225,4272].

thumbnail
Fig 1. A PRISMA flow chart showing the identification, screening, and selection of primary studies.

https://doi.org/10.1371/journal.pone.0272692.g001

Characteristics of included studies

Table 1 presents the characteristics of the included studies. Most of the studies were from Ethiopia (n = 16) and Nigeria (n = 7). The least number of studies (n = 1) were from the Democratic Republic of Congo, Kenya, and Madagascar. Geographically, there were 23 studies from East Africa, 12 from West Africa, 8 from Southern Africa, and 1 from Central Africa. The number of studies from countries with low, medium, and HDI were 30, 10, and 4, respectively. Of the 44 studies, 40 were conducted among students at public universities, 35 involved undergraduate students, 6 included both undergraduate and postgraduate students, and 3 did not report the level of the students. Condom use at the LSI was measured in three ways: 1) “Condom use at last sex” in 8 studies; 2) “Consistent condom use” in 14 studies; and, 3) “Always used a condom” in 12 studies.

thumbnail
Table 1. Evidence table summarizing the characteristics of included studies.

https://doi.org/10.1371/journal.pone.0272692.t001

Risk of bias and percentage agreement between the reviewers

Of the 44 studies, 38 had a low risk of bias while 6 had a moderate risk of bias (S2 Table). The percentage agreement between the two reviewers was 83.3%, which is significantly higher than the expected agreement of 62.0% (Kappa statistics = 0.46, p<0.0001).

Condom use at the LSI: Prevalence and time trend

The combined sample size for all the 44 studies [7,1118,2225,4272] was 27,948 participants. Of 14,778 sexually active participants, 8,744 had used a condom at the LSI (Fig 2). Accordingly, the pooled proportion of condom use at the LSI was 52.9% (95% CI, 45.0 to 60.7%; 95% PI, 2.8 to 98.9%; I-squared = 99.0%, p< 0.0001). A time-trend analysis showed a gradual increase in condom use at the LSI between 2000 and 2019 although the increase was not statistically significant (p = 0.4572).

thumbnail
Fig 2. Forest plot showing the pooled proportion of condom use at the last sexual intercourse among university students in SSA.

https://doi.org/10.1371/journal.pone.0272692.g002

Sub-group analysis and sources of heterogeneity.

Table 2 presents the results of the sub-group analysis. A higher proportion of condom use at the LSI was found among university students in West Africa (58.6%) than any other region in SSA: Southern Africa (50.5%), Central Africa (43.9%), and East Africa (40.5%). The proportion of condom use at the LSI using a 5-year range was as follows: 27.0% between 2000 and 2004, 40.7% between 2005 and 2009, 62.7% between 2010 and 2014, and 48.4% between 2015 and 2019. We deemed the interval sufficient to demonstrate a credible change in condom use at the LSI.

thumbnail
Table 2. Sub-group analysis and sources of heterogeneity in included studies.

https://doi.org/10.1371/journal.pone.0272692.t002

The proportion of condom use at the LSI was 80.5% for studies that combined data on students in private universities with data on students in public universities, 36.1% for studies conducted among students in private universities, and 52.1% for studies conducted among students in public universities. The proportion of condom use at the LSI was almost comparable for studies conducted among students in countries with low and medium HDI: 52.6% versus 55.6%, respectively. However, for studies conducted among students in countries with high HDI, the proportion of condom use at the LSI was 46.8%.

Concerning heterogeneity (Table 2), the year of publication (p = 0.034), risk of bias (p = 0.01), and measure of outcome (p = 0.013) showed statistically significant heterogeneity while the study sample size (p = 0.056) and number of sexually active participants (p = 0.051) showed borderline statistical significance.

Meta-regression analysis results.

Table 3 shows the univariate and multivariate meta-regression analysis results. In the univariate analysis, studies published between 2010 and 2014 (Beta-coefficient (β) = 0.36, 95% CI, 0.06–0.67), studies with a lower risk of bias (β = 0.27, 95% CI, 0.06–0.49), and studies with primary outcome reported as “condom use at the last sex” (β = 0.22, 95% CI, 0.04–0.41) or “consistently used condom” (β = 0.23, 95% CI, 0.04–0.43) demonstrated statistically significant heterogeneity. In the multivariate meta-regression analysis, none of the factors was statistically significant.

thumbnail
Table 3. Univariate and multivariate meta-regression analysis.

https://doi.org/10.1371/journal.pone.0272692.t003

Publication bias and small study effect findings.

We found an asymmetrical distribution of studies in a funnel plot suggesting possible publication bias. However, in a contour-enhance funnel plot, smaller studies were distributed in the regions of both statistical and non-statistical significance, suggesting that the asymmetry might have resulted from several factors but not solely publication bias. Egger’s test was statistically significant, with the 95% CI of the intercept excluding zero (Bias (intercept), -6.09; 95% CI, -9.90, -2.28; p = 0.002). This meant that the asymmetry resulted from a small study effect. Accordingly, trim and fill analysis was not performed.

Sensitivity analysis findings.

The exclusion of one study at a time generated a new pooled proportion of condom use at the LSI that fell within the 95% CI of the original pooled result. This meant that the meta-analytic results and conclusions are robust to methodological quality, meta-analytic approach, study quality, and the sample size.

Association between sex, age, and condom negotiation efficacy with condom use at the last sexual intercourse (LSI)

Fig 3 summarizes the association between sex (5 studies), age (5 studies), and condom negotiation efficacy (3 studies) with condom use at the LSI. Fig 3A shows that condom use at the LSI was more likely among female than male students (pooled RR = 1.08; 95% CI, 0.68–1.71; I-squared = 73%, p<0.01). Fig 3B shows that condom use at the LSI was more likely among younger students (≤24 years old) compared to students aged 25 years and more (pooled RR, 1.16; 95% CI, 0-85-1.57; I-squared = 42%, p = 0.14).

thumbnail
Fig 3. Panel of three forest plots showing the pooled association between sex, age, and condom negotiation efficacy with condom use at the LSI among university students in SSA.

Fig 3A shows condom use at the LSI among female versus male students. Fig 3B shows condom use at the LSI among younger (≤24 years old) versus older students (25 years and more). Fig 3C shows the condom use at LSI among students with a high versus low condom negotiation.

https://doi.org/10.1371/journal.pone.0272692.g003

Fig 3C indicates that students with a higher condom negotiation efficacy were more likely to use a condom at the LSI compared to students with a lower condom negotiation efficacy (pooled RR, 1.54; 95% CI, 0-81-2.94; I-squared = 80.0%, p<0.01).

Discussion

We studied condom use at the LSI among university students in SSA. Condom at the LSI measures progress towards preventing exposure to HIV through condomless sexual intercourse among people at a higher risk of STIs, such as university students. Improving access to condoms through condom availability programs (defined as interventions at school, city, state, or federal levels that provide improved access to condoms) is important in achieving better sexual health outcomes among university students. Previous studies have reported that condom availability programs that target students lower the age at sexual initiation, increase condom use at the LSI [73], and reduce the risk of STI acquisition [74] including the prevalence of STI symptoms [75,76], and importantly, by no means increase sexual activity and multiple sexual partnerships. We found that merely 52.9% of university students in SSA had used a condom at the LSI. The study further found a tendency towards higher condom use at the LSI among female than male students, younger than older students, and students with a higher condom negotiation efficacy compared to those with a lower condom negotiation efficacy.

Our data also show that condom use at the LSI among university students in SSA increased between 2000 and 2019. The pooled proportion of condom use at the LSI in this study was remarkably lower than the 80–90% condom use at the LSI observed across most countries globally [77]. Although several reasons might plausibly explain the low use of a condom at the LSI, we present a few of them. First, across most countries in SSA, young people have poor access to condoms and face stigma including age restrictions, and religious and gender barriers to accessing condoms. Second, some countries have laws that prohibit young people from carrying condoms and people from promoting and distributing condoms at venues where young people usually socialize. Third, besides the fall in international funding for condom procurement, most countries in SSA have limited or no domestic funding for condom procurement leading to a scarcity of condoms. For instance, in 2015, six billion condoms were needed in SSA but only 2.7 billion were distributed, representing at least a 50% shortfall [77]. The low condom use at the LSI has important consequences among sexually active university students as it results in STI acquisition including HIV [78,79] and unwanted pregnancies.

Our study found that sex, age, and condom negotiation efficacy were not associated with condom use at the LSI, implying that HIV prevention strategies such as risk reduction counseling and condom campaigns should target all sexually active university students to reduce the risk of acquisition of STIs including HIV [80]. Our findings also underscore the importance of improving access to HIV pre [81] and post-exposure prophylaxis to prevent HIV acquisition among university students in SSA. The findings further emphasize the importance of access to emergency contraception among university students in SSA to prevent unwanted pregnancies. Additionally, there is a need to tackle the barriers to and inequities in condom access and use among university students in SSA.

We found increased use of a condom at the LSI between 2000 and 2019, indicating that condom promotion campaigns are achieving the desired effect of reaching individuals who engage in high-risk sexual relationships [82]. Therefore, universities in SSA need to sustain the increase in condom use at the LSI.

Study strengths and limitations

Our study has some important strengths. To the best of our understanding, this is the first study to highlight condom use at the LSI among university students in SSA. The data analyzed were from almost all the regions in SSA. We meta-analyzed data for sexually active students in pooling the proportion of condom use at the LSI, making our results less biased. Our findings are robust to the methodological quality and analytic approach since no study had a significant influence on the overall meta-analytic results and conclusions. The risk of bias was low in the majority of the studies. Nonetheless, there are limitations to consider. All the 44 studies meta-analyzed employed a cross-sectional design which is characterized by limitations of selection bias and confounding. However, this limitation was mitigated by abstracting data for only adjusted measures of effect. Also, there is a likelihood that the outcome measures “Consistent condom use” and “Always used condom” might have been undermined by recall and reporting biases. For instance, male students might perceive condom use at the LSI as suggestive of low masculinity. Relatedly, in settings where condom promotion campaigns exist, there is a possibility that male students are more likely to report condom use at the LSI even when this was not the case. Nonetheless, both measures of outcome signify condom use at the LSI. The exclusion of studies published in non-English languages such as French, Portuguese, and Arabic among others might have led to an inaccurate measure of condom use at the LSI.

The studies meta-analyzed were statistically heterogeneous despite the use of an appropriate analytic approach (random-effects model) to pool the data and additional measures (sub-group and meta-regression analyses) taken to investigate and explain the sources of statistical heterogeneity. We emphasize a need to cautiously interpret the results as the analytic approaches do not eliminate statistical heterogeneity. Also, several factors that contribute to heterogeneity such as time since LSI, preventive measures other than a condom, type of sexual activity, urban versus rural residence, power relationship with a sexual partner, covariates used for statistical adjustments, and methods used for collecting information on condom use among others, were not measured.

Conclusions and recommendations

Our study found a 52.9% prevalence of condom use at the LSI among sexually active university students in SSA, which is considerably lower than expected. The low proportion of condom use at the LSI places these predominantly young adults at a higher risk of acquisition of STIs including HIV. With regards to sex, age, and condom negotiation efficacy, we found similar use of a condom at the LSI. Therefore, there is a need for context-relevant campaigns to advocate for use of condoms among university students in SSA.

Supporting information

S2 Table. Detailed summary of the risk of bias findings.

https://doi.org/10.1371/journal.pone.0272692.s002

(DOCX)

Acknowledgments

We thank the primary authors of the studies meta-analyzed for providing additional data and clarifications when needed.

References

  1. 1. Joint United Nations Programme on HIV/AIDS (UNAIDS). Young people and HIV. Geneva, Switzerland: 2021.
  2. 2. Joint United Nations Programme on HIV/AIDS (UNAIDS). UNAIDS Data 2021. Geneva, Switzerland: 2021.
  3. 3. AVERT. Young people, HIV and AIDS 2020 [updated 23 Apr 2020; cited 2020 1 Oct]. Available from: https://www.avert.org/professionals/hiv-social-issues/key-affected-populations/young-people.
  4. 4. Elshiekh HF, Hoving C, de Vries H. Exploring Determinants of Condom Use among University Students in Sudan. Archives of Sexual Behavior. 2020:1–13.
  5. 5. Bajunirwe F, Semakula D, Izudi J. Risk of HIV infection among adolescent girls and young women in age-disparate relationships in sub-Saharan Africa: a systematic review and meta-analysis. AIDS (London, England). 2020.
  6. 6. WHO. HIV and Youth: WHO; 2020 [cited 2020 June 11]. Available from: www.who.int/maternal_child_adolescent_/topics/adolescence/hiv/en/#:~test=HIV%20and%20youth,million%20youth%20living%20with%20%HIV.
  7. 7. Lake Victoria Basin Commission. HIV Sero-behavioural Study in Six Universities in Uganda. 2010.
  8. 8. Adefuye AS, Abiona TC, Balogun JA, Lukobo-Durrell M. HIV sexual risk behaviors and perception of risk among college students: implications for planning interventions. BMC Public Health. 2009;9(1):281. pmid:19653901
  9. 9. Adebiyi A, Asuzu M. Condom use amongst out of school youths in a local government area in Nigeria. African Health Sciences. 2009;9(2).
  10. 10. Rahamefy OH, Rivard M, Ravaoarinoro M, Ranaivoharisoa L, Rasamindrakotroka A, Morrisset R. Sexual behaviour and condom use among university students in Madagascar. Journal of Social Aspects of HIV/AIDS. 2008;5(1):28–35. pmid:18496617
  11. 11. Rahamefy OH, Rivard M, Ravaoarinoro M, Ranaivoharisoa L, Rasamindrakotroka AJ, Morisset R. Sexual behaviour and condom use among university students in Madagascar. Sahara Journal. 2008;5(1):28–35. pmid:18496617
  12. 12. Fitaw Y, Worku A. High-risk sexual behavior and pattern of condom utilization of the Gondar Collage of Medical Sciences (GCMS) Students, North-west Ethiopia. Ethiopian Journal of Health Development. 2002;16(3):335–8.
  13. 13. Muiga M. Relationship Between Knowledge Of Proper Condom Use And Behaviour Among Students Of Kirinyaga University: University of Nairobi; 2017.
  14. 14. Yarinbab T, Tawi N, Darkiab I, Debele F, Ambo W. Determinants of Risky Sexual Behaviors among Students of Mizan Aman College of Health Science, Southwest Ethiopia: Cross-Sec-tional Study. Int J Womens Health Wellness. 2018;4(074):2474–1353.
  15. 15. Tura G, Alemseged F, Dejene S. Risky Sexual Behavior and Predisposing Factors among Students of Jimma University, Ethiopia. Ethiop J Health Sci. 2012;22(3):170–80. pmid:23209351
  16. 16. Teferra TB, Erena AN, Kebede A. Prevalence of premarital sexual practice and associated factors among undergraduate health science students of Madawalabu University, Bale Goba, South East Ethiopia: institution based cross sectional study. The Pan African Medical Journal. 2015;20:209. pmid:26113940
  17. 17. Muhindo R, Sanyu B, Nankumbi JO. Factors Associated with Condom Use among University Female Students in Uganda. J Public Health Dis Prev. 2018;1(1).
  18. 18. Akpan U, Ekott M, Udo A. Male condom: Knowledge and practice among undergraduates of a tertiary institution in Nigeria. International Journal of Obstetrics and Gynaecology. 2014;2(1):032–9.
  19. 19. Mehra D, Östergren PO, Ekman B, Agardh A. Inconsistent condom use among Ugandan university students from a gender perspective: a cross-sectional study. Glob Health Action. 2014;7:22942. pmid:24725363
  20. 20. Choudhry V, Agardh A, Stafström M, Östergren PO. Patterns of alcohol consumption and risky sexual behavior: a cross-sectional study among Ugandan university students. BMC Public Health. 2014;14:128. pmid:24502331
  21. 21. Othieno CJ, Okoth R, Peltzer K, Pengpid S, Malla LO. Risky HIV sexual behaviour and depression among University of Nairobi students. Annals of General Psychiatry. 2015;14(1):16. pmid:25873984
  22. 22. Agardh A, Emmelin M, Muriisa R, Ostergren PO. Social capital and sexual behavior among Ugandan university students. Glob Health Action. 2010;3. pmid:21042434
  23. 23. Heeren GA, Jemmott JB 3rd, Mandeya A, Tyler JC. Sub-Saharan African university students’ beliefs about condoms, condom-use intention, and subsequent condom use: a prospective study. AIDS Behav. 2009;13(2):268–76. pmid:18600442
  24. 24. Terefe G, Antehun A. Consistent condom use and preference of HIV risk preventive interventions among undergraduate university students in Southern Ethiopia: a cross sectional study. Journal of AIDS and Clinical Research. 2015;6(4).
  25. 25. Haffejee F, Koorbanally D, Corona R. Condom use among South African university students in the province of KwaZulu-Natal. Sexuality & Culture. 2018;22(4):1279–89.
  26. 26. Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Systematic reviews. 2015;4(1):1. pmid:25554246
  27. 27. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000;283(15):2008–12. pmid:10789670
  28. 28. Izudi J, Okello G, Semakula D, Bajunirwe D. Condom use at last sex among university students in Africa: a systematic review and meta-analysis 2020. Available from: https://www.crd.york.ac.uk/PROSPERO/#recordDetails.
  29. 29. Conceição P. Human development report. 2019: beyond income, beyond averages, beyond today: inequalities in human development in the 21st century. 2019.
  30. 30. Belur J, Tompson L, Thornton A, Simon M. Interrater reliability in systematic review methodology: exploring variation in coder decision-making. Sociological Methods and Research. 2021;50(2):837–65.
  31. 31. Hoy D, Brooks P, Woolf A, Blyth F, March L, Bain C, et al. Assessing risk of bias in prevalence studies: modification of an existing tool and evidence of interrater agreement. J Clin Epidemiol. 2012;65(9):934–9. pmid:22742910
  32. 32. R Core Team. A language and environment for statistical computing. R Foundation for Statistical Computing. Viena, Austria2018.
  33. 33. Nyaga VN, Arbyn M, Aerts M. Metaprop: a Stata command to perform meta-analysis of binomial data. Archives of Public Health. 2014;72(1):39. pmid:25810908
  34. 34. IntHout J, Ioannidis JPA, Rovers MM, Goeman JJ. Plea for routinely presenting prediction intervals in meta-analysis. BMJ Open. 2016;6(7). pmid:27406637
  35. 35. Higgins JP, Green S. Cochrane handbook for systematic reviews of interventions: John Wiley & Sons; 2011.
  36. 36. Thompson SG, Higgins J. How should meta‐regression analyses be undertaken and interpreted? Statistics in Medicine. 2002;21(11):1559–73. pmid:12111920
  37. 37. Egger M, Smith GD, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. Bmj. 1997;315(7109):629–34. pmid:9310563
  38. 38. Song F, Gilbody S. Bias in meta-analysis detected by a simple, graphical test. Increase in studies of publication bias coincided with increasing use of meta-analysis. British Medical Journal. 1998;316(7129):471. pmid:9492690
  39. 39. Palmer TM, Sutton AJ, Peters JL, Moreno SG. Contour-enhanced funnel plots for meta-analysis. The STATA journal. 2008;8(2):242–54.
  40. 40. Duval S, Tweedie R. A non-parametric ‘trim and fill’ method of assessing publication bias in meta-analysis. Journal of the American Statistical Association. 2000;95.
  41. 41. Mavridis D, Chaimani A, Efthimiou O, Leucht S, Salanti G. Addressing missing outcome data in meta-analysis. Evidence Based Mental Health. 2014;17(3):85–9. pmid:25009175
  42. 42. Peltzer K. Factors affecting condom use among South African university students. East Afr Med J. 2000;77(1):46–52. pmid:10944840
  43. 43. Olley B, Rotimi O. Gender differences in condom use behaviour among students in a Nigerian University. African Journal of Reproductive Health. 2003:83–91. pmid:12816316
  44. 44. Okafor II, Obi SN. Sexual risk behaviour among undergraduate students in Enugu, Nigeria. Journal of Obstetrics and Gynaecology: the journal of the Institute of Obstetrics and Gynaecology. 2005;25(6):592–5. pmid:16234148
  45. 45. Olley B. Child sexual abuse, harmful alcohol use and age as determinants of sexual risk behaviours among freshmen in a Nigerian University. African Journal of Reproductive Health. 2008;12(2):75–88. pmid:20695043
  46. 46. Tagoe M, Aggor RA. Knowledge, behaviour, perceptions and attitudes of University of Ghana students towards HIV/AIDS: what does behavioural surveillance survey tell us? Journal of health and human services administration. 2009;32(1):51–84. pmid:19558033
  47. 47. Mubita-Ngoma CA. Gender differences in HIV/AIDS Knowledge and sexual behavior among University students in Lusaka, Zambia. Africa Journal of Nursing and Midwifery. 2010;12(2):27–35.
  48. 48. Agardh A, Odberg-Pettersson K, Ostergren PO. Experience of sexual coercion and risky sexual behavior among Ugandan university students. BMC Public Health. 2011;11:527. pmid:21726433
  49. 49. Berhan Y, Hailu D, Alano A. Predictors of sexual-risk behaviour and HIV-preventive practices among university students in Ethiopia. Afr J AIDS Res. 2011;10(3):225–34. pmid:25859791
  50. 50. Fiaveh DY. Male perspective (s) on condom use: Context of STI/HIV prevention in the University of Ghana community. Journal of Public Health and Epidemiology. 2011;3(1):17–27.
  51. 51. Agardh A, Cantor-Graae E, Ostergren PO. Youth, sexual risk-taking behavior, and mental health: a study of university students in Uganda. Int J Behav Med. 2012;19(2):208–16. pmid:21590465
  52. 52. Dingeta T, Oljira L, Assefa N. Patterns of sexual risk behavior among undergraduate university students in Ethiopia: a cross-sectional study. The Pan African Medical Journal. 2012;12:33. pmid:22891091
  53. 53. Iliyasu Z, Abubakar IS, Sani IH, Jibo AM, Karaye IM, Salihu HM, et al. Male circumcision and HIV risk behavior among University students in Northern Nigeria. American Journal of Men’s Health. 2013;7(2):94–101. pmid:22892146
  54. 54. Masoda M, Govender I. Knowledge and attitudes about and practices of condom use for reducing HIV infection among Goma University students in the Democratic Republic of Congo. Southern African Journal of Epidemiology and Infection. 2013;28(1):61–8.
  55. 55. Mengistu TS, Melku AT, Bedada ND, Eticha BT. Risks for STIs/HIV infection among Madawalabu university students, Southeast Ethiopia: a cross sectional study. Reproductive Health. 2013;10:38. pmid:24069905
  56. 56. Nkomazana N, Maharaj P. The prevalence of condom use among university students in zimbabwe: implications for planning and policy. J Biosoc Sci. 2013;45(5):643–59. pmid:23194440
  57. 57. Wells CJ, Alano A. Prophylactic procurement of university students in Southern Ethiopia: stigma and the value of condom machines on campus. PLoS One. 2013;8(4):e60725. pmid:23565272
  58. 58. Asante KO, Boafo IM, Nyamekye M. Identifying gender differences in Ghanaian university students’ sexual practices, attitudes and knowledge regarding HIV. Journal of Psychology. 2014;5(1):9–18.
  59. 59. Ngoma M, Menon J, Malungo J, Siziya S, Nkumbula T, Musepa M, et al. Appropriate HIV and AIDS Interventions drawn from Baseline Knowledge Attitude and Behaviour Surveys of University Students. Medical Journal of Zambia. 2014;41(3):109–14.
  60. 60. Negeri EL. Determinants of risky sexual behavior, relation between HIV risk perception and condom utilization among Wollega University students in Nekemte town, Western Ethiopia. Science, Technology and Arts Research Journal. 2014;3(3):75–86.
  61. 61. Sendo E, Bedada W. Risky sexual behavior for STIs/HIV infection among Alkan University College students in Ethiopia. Science Postprint. 2014;1(1):e00031.
  62. 62. Shiferaw Y, Alemu A, Assefa A, Tesfaye B, Gibermedhin E, Amare M. Perception of risk of HIV and sexual risk behaviors among University students: implication for planning interventions. BMC Res Notes. 2014;7:162. pmid:24642193
  63. 63. Tobin-West CI, Maduka O, Onyekwere VN, Tella AO. Awareness, acceptability, and use of female condoms among university students in Nigeria: implications for STI/HIV prevention. Int J Adolesc Med Health. 2014;26(2):259–65. pmid:24114893
  64. 64. Manyumwa YT. Prevalence and predictors of high risk sexual behaviours among undergraduate university students at Midlands State University. Midlands State University, Zimbabwe: Zimbabwe; 2015.
  65. 65. Mavhandu-Mudzusi AH, tesfay Asgedom T. The prevalence of risky sexual behaviours amongst undergraduate students in Jigjiga University, Ethiopia. Health Sa Gesondheid. 2016;21(1):179–86.
  66. 66. Asante KO, Osafo J, Doku PN. The role of condom use self-efficacy on intended and actual condom use among university students in Ghana. Journal of Community Health. 2016;41(1):97–104. pmid:26233713
  67. 67. Regassa T, Chala D, Adeba E. Premarital sex in the last twelve months and its predictors among students of Wollega University, Ethiopia. Ethiopian journal of health sciences. 2016;26(4):351–8. pmid:27587933
  68. 68. Mamo K, Admasu E, BERTA MM. Prevalence and Associated Factors of Risky Sexual Behavior among Debremarkos University Regular Undergraduate Students, Debremarkos Town North West Ethiopia, 2016. Journal of Health, Medicine and Nursing. 2016.
  69. 69. Hoffman S, Levasseur M, Mantell JE, Beksinska M, Mabude Z, Ngoloyi C, et al. Sexual and reproductive health risk behaviours among South African university students: results from a representative campus-wide survey. Afr J AIDS Res. 2017;16(1):1–10. pmid:28367750
  70. 70. Akibu M, Gebresellasie F, Zekarias F, Tsegaye W. Premarital sexual practice and its predictors among university students: institution based cross sectional study. Pan African Medical Journal. 2017;28(1). pmid:29881479
  71. 71. Sakeah JK. Determinants of condom use intentions among university students in Ghana: Lethbridge, Alta: University of Lethbridge, Faculty of Health Sciences; 2017.
  72. 72. Ajayi AI, Ismail KO, Akpan W. Factors associated with consistent condom use: a cross-sectional survey of two Nigerian universities. BMC Public Health. 2019;19(1):1207. pmid:31477068
  73. 73. Algur E, Wang E, Friedman H, Deperthes B. A systematic global review of condom availability programs in high schools. J Adolesc Health. 2019;64(3):292–304. pmid:30819331
  74. 74. Wretzel SR, Visintainer PF, Pinkston Koenigs LM. Condom availability program in an inner city public school: Effect on the rates of gonorrhea and Chlamydia infection. J Adolesc Health 2011;49(3):324–6. pmid:21856527
  75. 75. Doyle AM, Ross DA, Maganja K, Baisley K, Masesa C, Andreasen A, et al. Long-term biological and behavioural impact of an adolescent sexual health intervention in Tanzania: Follow-up survey of the community-based MEMA kwa Vijana trial. PLoS Med. 2010;7(6):e1000287. pmid:20543994
  76. 76. Ross D, Changalucha J, Obasi AI, Todd J, Plummer LM, Cleophas-Mazige B, et al. Biological and behavioural impact of an adolescent sexual health intervention in Tanzania: A community-randomized trial. AIDS. 2007;21(14):1943–55. pmid:17721102
  77. 77. UNAIDS. Update: A condom crisis at the center of the HIV prevention crisis. 2018 [cited 2020 Oct 3]. Available from: https://www.unaids.org/en/resources/presscentre/featurestories/2018/july/20180723_condoms-AIDS2018.
  78. 78. Rutherford GW, Anglemyer A, Bagenda D, Muyonga M, Lindan CP, Barker JL, et al. University students and the risk of HIV and other sexually transmitted infections in Uganda: the Crane survey. International Journal of Adolescent Medicine and Health. 2014;26(2):209–15. pmid:24762640
  79. 79. Kassie BA, Yenus H, Berhe R, Kassahun EA. Prevalence of sexually transmitted infections and associated factors among the University of Gondar students, Northwest Ethiopia: a cross-sectional study. Reprod Health. 2019;16(1):163. pmid:31703688
  80. 80. Heeren GA, Jemmott JB III, Mandeya A, Tyler JC. Theory–based predictors of condom use among university students in the United States and South Africa. AIDS Education & Prevention. 2007;19(1):1–12. pmid:17411385
  81. 81. Saul J, Bachman G, Allen S, Toiv NF, Cooney C, Beamon T. The DREAMS core package of interventions: A comprehensive approach to preventing HIV among adolescent girls and young women. PloS One. 2018;13(12):e0208167. pmid:30532210
  82. 82. MEASURE Evaluation. Condom use at last high-risk sex. University of North Carolina at Chapel Hill: Carolina Population Center; 2020 [updated 2020; cited 2020 Aug 17]. Available from: https://www.measureevaluation.org/prh/rh_indicators/womens-health/stis-hiv-aids/condom-use-at-last-high-risk-sex.