SUBSTANTIAL INCREASE IN TRANSACTIONAL SEX AND HIV RISK AMONG ADOLESCENTS DURING THE COVID-19 PANDEMIC: RESULTS FROM A REPEATED CROSS-SECTIONAL STUDY IN BOTSWANA

Abstract

The COVID-19 pandemic and resulting public health response has disrupted the lives of adolescents and their families worldwide. We evaluated the impact of the pandemic on attitudes, beliefs, and sexual risk behavior among adolescents in Botswana. Participants were recruited using household-based sampling across residential districts (blocks) in and around Gaborone, Botswana, and completed surveys on laptop computers at a private, central location. We compared baseline survey data from 380 adolescents who completed the survey pre-pandemic (n = 139) to those who completed the survey intra-pandemic (n = 241). Participants had a mean age of 15.2 years; 58.6% were girls and 41.4% were boys. Intra-pandemic, participants reported greater engagement in transactional sex (38.1% compared to 13.6% pre-pandemic, p < .05), more favorable attitudes toward transactional sex with sugar daddies and sugar mommies (p < .05), greater intentions to remain sexually active in the future (29.2% vs.13.6%, respectively, p < .05), and lower self-efficacy in handling risky sexual situations (p < .01). Public health interventions that lessen these concerning shifts in attitudes and behaviors will be key to protecting the sexual health of adolescents and to supporting their safe transition to adulthood.

INTRODUCTION

The COVID-19 pandemic and resulting public health response has disrupted the lives of adolescents and their families worldwide. Multiple researchers have speculated about the potential impacts of the COVID-19 pandemic on children and adolescents and the need to identify and then address new challenges that arise as a result of the pandemic (Desmond, Sherr, & Cluver, 2021; Lancet, 2020; Silva et al., 2021).

The pandemic threatens the successes in HIV prevention, which is particularly challenging in sub-Saharan Africa, where the greatest number of newly infected people are located (Joint United Nations Programme on HIV/AIDS [UNAIDS], 2021). Current research has focused on how HIV-infected adolescents’ access to services was affected by the pandemic. Results show that adolescents have lost access to critical prevention services (Dourado et al., 2020; Phiri et al., 2022; Stephenson et al., 2021; Vrazo et al., 2020) and that adolescents who are living with HIV report reductions in economic stability, lowered school attendance, and poorer mental health since the onset of the pandemic (Dyer et al., 2021; Enane et al., 2021).

The changes caused by the pandemic may also increase adolescents’ vulnerabilities in acquiring HIV. Little is known, however, about how the COVID-19 pandemic has affected adolescents’ high-risk sexual attitudes, beliefs, and behaviors. The purpose of this paper is to evaluate the impact of the COVID-19 pandemic on sexual risk and behaviors among adolescents in Botswana.

METHODS

PARTICIPANTS AND PROCEDURES

As part of an ongoing behavioral intervention clinical trial to improve HIV prevention among adolescents and their families in Botswana, 456 parent–adolescent dyads were recruited from households in urban and peri-urban residential districts in and around Gaborone, Botswana (for more detail, please see Sun, Seloilwe, Magowe, Dithole, & St. Lawrence, 2020). By the conclusion of the research, half of the adolescents will be 13–15 years old and half will be 16–18 years old, and boys and girls will be equally represented. Parent–adolescent dyads were eligible to participate if both a parent and an adolescent 13–18 provided informed consent and assent respectively, and there was no obvious impairment that would preclude their ability to complete the measures and engage in a group intervention. No participants were excluded based on recruiters’ impressions of any impairment that would preclude participation. When multiple adolescents resided in the same household, one was selected at random to participate in the study. Participants received 60 pula (approximately $5) to compensate their travel expenses between home and the private space at the University of Botswana’s Centre for HIV and AIDS Research, where they completed the ACASI survey on a laptop computer. All study procedures were approved by Institutional Review Boards at the University of Botswana and Portland State University.

Adolescents represented in the current study completed baseline surveys from June 2019 to November 2021. In response to the COVID-19 pandemic, all research efforts were suspended in March 2020 and only allowed to resume after approval of new protocols that protected the health of research staff and participants. No baseline assessments were collected for 7 months. This interval served as a washout period, providing a clear comparison between the pre-pandemic and intra-pandemic participants.

We compared baseline survey data from 380 adolescents who completed the survey pre-pandemic (n = 139) to those who completed it intra-pandemic (n = 241). The sample comprised 58.6% girls and 41.4% boys. Their average age was 15.2 years. The majority (57.1%) were in junior secondary school, 18.5% were in senior secondary school, and 10.0% did not attend school. There were no significant differences in demographic characteristics and perceptions of overall health and mental health between pre-pandemic and intra-pandemic participants (see Table 1).

TABLE 1.

Participant Characteristics

Variable	Pre-pandemic participants (n = 139)	Intra-pandemic participants (n = 241)	Test statistic	p value
	% or M ± SD
Sex			χ2(1) = 0.02	.90
Boy	41.0%	41.7%
Girl	59.0%	58.3%
Age	15.33 ± 1.66	15.15 ± 1.60	t(368) = 0.99	.32
Grade			χ2(3) = 5.26	.15
Primary school	10.2%	14.4%
Junior secondary school	55.5%	58.1%
Senior secondary school	24.1%	15.3%
Not in school	10.2%	12.3%
Parents			χ2(2) = 3.84	.15
Both alive	76.5%	81.7%
One deceased	21.2%	14.0%
Both deceased	2.3%	4.3%
People in household	7.59 ± 4.79	7.78 ± 4.27	t(371) = −0.40	.69
Poor	4.5%	2.1%
Health			N/A1	.30
Excellent	23.1%	27.2%
Very good	18.7%	24.7%
Good	27.6%	25.1%
Not very good	26.1%	20.9%
Days felt anxious or sad in the past 2 months	4.65 ± 7.62	4.16 ± 7.58	t(360) = 0.59	.56

Note. There were no significant differences between the two groups. ¹Fisher’s exact test yields only a p value.

MEASURES

Participants reported their demographic characteristics and perceptions of their overall health (“How would you evaluate your health over the past two months?”) and mental health (“How many days in the past two months did you feel anxious or sad?”). Following the section of the survey that collected demographic information, they completed validated measures that assess constructs that are protective or increase risk for HIV, including HIV/AIDS knowledge (Kelly, Lawrence, Hood, & Brasfield, 1989), attitudes toward condoms (St. Lawrence et al., 1994), barriers toward using condoms (St. Lawrence et al., 1999), self-efficacy beliefs in handling risky sexual situations (Murphy, Stein, Schlenger, & Maibach, 2001), attitudes toward transactional sex (St. Lawrence et al., 2022), and an unpublished measure used in our prior research that assessed their perceptions of sexual behavior community norms. The measures are briefly described below. A more detailed description can be found in Sun et al. (2018). Some measures were used to confirm comparability between the two groups of participants, given that they completed the survey separated by time. Other measures assessed the impact of the pandemic on the attitudes, beliefs, and risk behavior of youth pre-pandemic with those who completed the measures intra-pandemic.

Measures to Assess Comparability of the Subsamples.

AIDS Risk Knowledge Test (Kelly et al., 1989) assesses knowledge of how HIV is transmitted. Possible scores range from 0 to 25. A sample item is “You can get HIV from oral sex.”

Condom Attitude Scale (St. Lawrence et al., 1994) is a measure of attitudes toward condom use. Possible scores range from 0 to 144. A sample item is “Using a condom takes all the pleasure out of sex.”

Condom Barriers Scale (St. Lawrence et al., 1999) assesses perceived barriers to condom use for prevention of HIV and other sexually transmitted infections. Potential scores range from 28 to 144. A sample item is “I can’t afford to get a condom. They cost too much.”

Sexual Behavior Community Norms measures respondents’ perceptions of the prevalence of risky sexual behaviors in the surrounding community. Potential scores range from 0 to 16. A sample item is “Most men have other sexual partners even when they are married.”

Measures Assessing Changes in Attitudes and Behaviors Due to the Pandemic.

Self-Efficacy Beliefs (Murphy et al., 2001) assesses adolescents’ confidence in handling realistic risky situations. Possible scores range from 16 to 160. After reading each scenario, a sample question that participants answer is “How sure are you that you could refuse and walk away?”

Attitudes toward Transactional Sex Scale (St. Lawrence et. al, 2022) measures adolescents’ attitudes toward engaging in transactional sex with a sugar daddy/sugar mommy. Possible scores range from 0 to 96. A sample item is “I think it is okay to have sex with a sugar daddy or sugar mum to get fashionable clothes.”

All participants self-reported lifetime frequencies of sexual behaviors. Participants who were sexually active then completed measures assessing sexual behaviors in the past 2 months. Pre-pandemic and intra-pandemic comparisons reported here are based on self-reports of transactional sex in the past 2 months and their behavioral intentions to remain sexually active in the coming year. Engagement in transactional sex was measured using the item “In the past two months, have you traded money or something else for sex (for example, food, drinks, jewelry, money, clothes, cell phone, grades or transportation)?” Intentions were measured using the item “How likely do you think it is that you will have sexual intercourse in the next year?” Past 2-month data ensured there was no overlap between the pre-pandemic and intra-pandemic periods.

ANALYSIS PLAN

We calculated proportions, means, and standard deviations to describe the sample, followed by t, chi-squared, and Fisher’s exact tests to compare data from participants who completed the survey pre-pandemic (June 2019 to February 2020) with those who completed the survey intra-pandemic (September 2020 to November 2021).

RESULTS

Risk behaviors were significantly higher following the onset of the pandemic (see Table 2). Nearly 4 out of 10 intra-pandemic participants (38.1%) reported engaging in transactional sex within the past 2 months, and this was three times higher than pre-pandemic rates, 13.6%, χ²(1) = 4.14, p < .05. Furthermore, attitudes toward transactional sex were significantly more favorable among intra-pandemic participants (M = 23.58) than among pre-pandemic participants, M = 20.15, t(376) = −2.41, p < .05. Significantly more sexually active adolescents who completed the survey intra-pandemic intended to remain sexually active in the future than participants who completed the survey pre-pandemic (29.2% and 13.6%, respectively, Fisher’s exact test, p < .05). Self-efficacy beliefs were significantly lower among participants who completed survey intra-pandemic (M = 63.97) than among participants who completed survey pre-pandemic, M = 81.50, t(375) = 3.22, p < .01. Finally, a significantly higher proportion of adolescents reported that there was insufficient food for their household intra-pandemic (28.7%) compared to pre-pandemic, 17.0%, χ²(1) = 6.32, p < .05.

TABLE 2.

Comparison of Sexual Risk and Protective Factors Among Participants Who Completed Baseline Survey Before and During the COVID-19 Pandemic

Variable	Pre-pandemic participants (n = 139)	Intra-pandemic participants (n = 241)	Test statistic	p value
	% or M ± SD
Engaged in transaction sex (past 2 months)a	13.6%	38.1%	χ2(1) = 4.14	< .05
Insufficient household food	17.0%	28.7%	χ2(1) = 6.32	< .05
Self-Efficacy Beliefs	81.50 ± 51.12	63.97 ± 50.77	t(375) = 3.22	< .01
Attitudes toward Transactional Sex Scale	20.15 ± 12.74	23.58 ± 13.61	t(376) = −2.41	< .05
Intention to remain sexually activea			N/A1	< .05
Will not	13.6%	22.0%
Probably won’t	31.8%	4.9%
Don’t know	40.9%	43.9%
Probably will	0.0%	14.6%
Sure will	13.6%	14.6%
AIDS Risk Knowledge Test	12.08 ± 4.94	11.02 ± 5.50	t(378) = 1.87	.06
Condom Attitudes Scale	92.04 ± 25.41	89.74 ± 23.50	t(378) = 1.87	.38
Condom Barriers Scale	63.42 ± 18.47	66.80 ±19.10	t(374) = −1.68	.09
Sexual Behavior Community Norms	6.54 ± 3.10	6.41 ± 3.16	t(376) = 0.40	.69

Note. ^aOnly participants who reported they were sexually active (n = 64) responded to these questions. ¹Fisher’s exact test yields only a p value.

On the variables that were unlikely to be affected by the pandemic, there were no significant differences in HIV/AIDS knowledge, attitudes toward condoms, and barriers to condom use between the groups (all p > .05). This finding confirms comparability between the two subsamples on these measures despite the fact that the subsamples completed the surveys separated by 7 months.

DISCUSSION

In this study, we compared the attitudes, behavioral intentions, and sexual risk behaviors reported by Batswana¹ adolescents before and during the COVID-19 pandemic. These findings allowed us to identify changes at the aggregate level. Participants who completed the survey during the pandemic reported greater engagement in transactional sex with sugar daddies and sugar mommies, more favorable attitudes toward transactional sex, greater intentions to remain sexually active in the future, and lower self-efficacy in handling risky sexual situations. Although the two groups completed the surveys separated by at least 7 months, participants were equivalent on measures unlikely to be impacted by the pandemic (i.e., their demographic characteristics, HIV/AIDS knowledge, attitudes toward condom use, barriers toward condom use, and community norms), which suggests that the reported increases in transactional sex and sexual risk during the pandemic are a function of the pandemic.

Nearly 4 out of 10 intra-pandemic participants reported engaging in transactional sex, a behavior that greatly increases their risk for HIV, other sexually transmitted infections, and pregnancy (Choudhry, Ambresin, Nyakato, & Agardh, 2015; Kilburn et al., 2018; Wamoyi, Stobeanau, Bobrova, Abramsky, & Watts, 2016). This finding is particularly worrisome, as Botswana has the third highest HIV rate in the world (UNAIDS, 2021). The rate of transactional sex during the pandemic was both three times higher than the rate from our pre-pandemic sample and higher than other published estimates of transactional sex among adolescents and young adults in Botswana and sub-Saharan Africa (Krisch, Averdijk, Valdebenito, & Eisner, 2019; Wirth et al., 2019). Prevalence of transactional sex during the pandemic could be related to adolescents experiencing a reduction in access to available resources. Our finding that adolescents reported significantly greater food insecurity during the pandemic supports this as a potential explanation. With fewer resources available to them, the adolescents may have turned to transactional sex to help meet basic needs.

The interpretation of these findings should be considered in light of the public health responses to the onset of the pandemic and the changes they produced in the social context of adolescents’ lives. The government of Botswana required and enforced social distancing, school closures, masking, curfews, stay-at-home measures, and travel restrictions within the country and between countries to limit spread of the virus, any of which may have influenced these observed differences.

Data were collected from adolescents who reside in and around Gaborone, where the majority of Batswana reside. It is possible that adolescents in rural areas may be differentially impacted by the pandemic than adolescents in and around Gaborone, the capital city. Finally, with survey data we were only able to observe changes in reported sexual risk behaviors, attitudes, and intentions. We were unable to elaborate on these findings using qualitative data. For example, it would be helpful to understand whether adolescents were trading sex to fulfill basic needs or whether they were motivated to acquire goods that increased their social status (such as mobile phone or more stylish clothing), since these may require different intervention strategies. It would also be interesting to assess how parent–adolescent relationships changed during the pandemic, as those relationships can influence adolescents’ attitudes and behaviors.

CONCLUSIONS

This study identified a critical need to address the observed high rates of transactional sex, increasingly favorable attitudes toward transactional sex, and decreased self-efficacy to extricate oneself from risky sexual situations that appeared after the onset of the pandemic. There is the potential these issues may worsen and accelerate as the pandemic continues. Public health interventions that address these concerning shifts in attitudes and behaviors will be key to protecting the health of adolescents. We likely need to consider multilevel interventions, including economic-based interventions such as cash transfers and educational subsidies that have been effective in reducing transactional sex (Cluver et al., 2013), or interventions that include parents and caregivers who can influence and support their adolescents and their safe transition to adulthood.