Do sexual expectancies and inhibitions predict high-risk sexual behaviours? Evidence from a cross-sectional survey among young psychoactive substance users in informal settlements in Kampala, Uganda

This was an analytical cross-sectional survey conducted between June and July 2019 among young psychoactive substance users in 12 out of the 57 informal settlements in Kampala city, Uganda. We defined an informal settlement as an urban residence characterised by inadequate access to social services and poor structural quality of housing, overcrowding and insecure residential status [44]. Such settlements are characterised by a high prevalence of psychoactive substance use [45‐47] and high-risk sexual behaviours [42]. The study was conducted in four of the five divisions of Kampala (including Kawempe, Makindye, Lubaga, and Nakawa). Kampala has an estimated population of over 1.5 million people, 70% of whom (live in informal settlements [48]. Kampala is Uganda’s main economic hub, generating 65% of the national Gross domestic product (GDP).

The study participants included sexually active young people aged between 18 and 24 years who had a recent history of using at least one or more psychoactive substances, and had resided within the informal settlements of Kampala city for at least 6 months. The most commonly used and accessible psychoactive substances of interest in these settings were alcohol, khat, oral tobacco, marijuana, and heroin. The age group of 18–24 years was considered because it is a critical risk period for the initiation of substance use [1, 49]. We excluded non-residents (e.g., visiting relatives or friends), and those who were not mentally sound or intoxicated during the survey. We trained research assistants to be sensitive to behavioural signs of intoxication such as loss of co-ordination, staggering gait, drowsiness, slurred speech and glazed eyes for alcohol users, and paranoia, anxiety, eye-rolling, pupil dilation/constriction, head movements or jerks for other substances [50]. Research assistants were cautioned to remind the participants of their right to withdraw, particularly when observable signs of intoxication appeared to change [50].

A total of 768 young PSUs were estimated as the appropriate sample size for this survey using the Kish Leslie formula for cross-sectional studies [51]. We assumed a prevalence of high- risk sexual behaviours among young PSUs of 50% because of the limited literature in the study domain. We assumed a 5% margin of error for the estimated prevalence, 95% confidence interval around the estimates, and a design effect of 2.0 to adjust for clustering by informal settlement [52]. A design effect of 2.0 is recommended for behavioural surveys that use respondent driven sampling (RDS) [53]. A final sample of 744 young PSUs was interviewed.

The detailed sampling procedure for the current study has previously been reported [54]. Briefly, we purposively selected 12 informal settlements (four in Nakawa, four in Kawempe, two in Lubaga, and two in Makindye) and these were selected with the aim of geographical representativeness of informal settlements [54]. Respondent-driven sampling was used in the selection of study participants, since it is recommended for hard-to-reach populations such as young psychoactive substance users [53, 55].

During the sampling, four young PSUs were purposefully selected and enrolled as seeds from each of the 12 informal settlements, resulting in a total of 48 seeds. To ensure that the selected seeds were young psychoactive substance users, we sought guidance from peer educators, staff of non-governmental organisations working on HIV and psychoactive substance programs and the local leaders. This increased the chances of selecting seeds with a large social network. We ensured seed diversity by ensuring geographical representativeness, gender, and age group. Selecting diverse seeds was aimed at improving the chances of reaching equilibrium. Equilibrium was established when there was no significant variation in the key study variables per informal settlements i.e., age group, gender, and informal settlement.

Data were collected using an interviewer-administered structured questionnaire developed after a critical review of literature [56‐59]. The questionnaire captured information on socio-demographic characteristics such as sex, age (in years), marital status, highest educational attainment, length of stay in a particular informal settlement, average individual monthly income, history of psychoactive substance use, sexual expectancies and inhibitions. The main outcome variables included multiple sexual partnerships, condom use and having sex under the influence of psychoactive substances. Multiple sexual partnerships were defined as having more than one sexual partner in the last 30 days [42]. Condom use was categorised as consistent and inconsistent. A respondent was considered a consistent user if he/she used condoms each time he/she had sexual intercourse with a non-spousal partner.

The main predictors in this study were sexual expectancies and inhibitions accrued from the use of psychoactive substances. Sexual expectancies were defined as perceptions of positive outcomes accrued from the use of psychoactive substances on an individual’s sexual behaviour [60]. In this study, the sexual expectancies included improved sexual pleasure or euphoria, and performance, and boosting one’s courage or confidence to approach a sexual partner. Inhibitions were defined as behaviours that go against one’s conventional or natural sexual practices [61]. Such behaviours included finding it difficult to use condoms, forgetting to use a condom, a higher likelihood of engaging in sexual intercourse, finding it difficult to decline sexual advance, and finding one’s self-wanting sex while under the influence of psychoactive substances.

The questionnaire (Additional file 1) was uploaded on the Kobo Collect application installed on android phones and tablets and interviews were conducted by well-trained data collectors. In preparation for actual data collection, the structured questionnaire (Additional file 1) was pretested among 20 young PSUs in a similar setting in Wakiso district and the feedback obtained during pretesting was used to refine the questionnaire.

All analyses were done using Stata 14 (StataCorp, Texas, USA). We summarised categorical data using frequencies and proportions. For numerical data, we used the mean and standard deviation to provide summary statistics. We compared the substance use related perceptions on sexual expectancies and inhibitions among males and females using chi-square statistics. To establish the relationship between sexual expectancies and inhibitions, and each of the high-risk sexual behaviours, we employed robust Poisson regression with a robust error variance. We fitted three separate models in order to establish the relationship between the independent variables (sexual expectancies and inhibitions) and each of the outcome variables i.e., engaging in multiple sexual partnerships in the last 30 days, having sex while under the influence of psychoactive substances, and condom use in the last 30 days. We used prevalence ratios (PRs) as more conservative measures of association than odds ratios since the prevalence of the outcomes was greater than 10% [62]. Stepwise backward regression was used, and a cut off p-value of less than 0.2 was set for variables eligible to be included in the multivariable model [63]. We adjusted for known confounders such as age, sex, and level of education. A p-value < 0.05 was considered statistically significant.

A total of 744 young PSUs was studied. The mean age of the respondents was 21.5 years (SD ± 2.1). About 76.6% of the respondents were aged 20–24 years; 78.0% were males; and 85.3% did not live with their parents or guardians. More than three quarters, 78.2% of the respondents were single while the rest were married. Regarding the level of education, more than half, 58.2% had attained post primary education while the remainder had only attained primary education or less. About 74.3% of the respondents had used alcohol in the last 30 days while more than half, 51.9% had used marijuana. (Table 1).

About 40.1% of alcohol consumers used it at least 5 times a week; 63.9% of marijuana users used it at least 5 times a week; 45.6% of khat users used it at least 5 times a week while 42.3% of oral tobacco users used it at least 5 times a week. (Fig. 1).

The median age at sex debut was 16.0 years (IQR = 14, 17), with a median number of sexual partners in the last 12 months at 3 (IQR =1, 9). The median age at sex debut for males was 15.0 years (IQR = 14, 17) while that of females was 16 years (IQR = 14, 17). Nearly a quarter, 23.6% of respondents admitted ever intentionally intoxicating someone with psychoactive substances with the aim of luring them into sex. About 46.2% (43.8% of males and 54.8% of females) of respondents or their partners used a psychoactive substance before the last sexual intercourse. Among those who used psychoactive substances before sex, 55.6% (27.8% of males and 18.3% of females) mentioned that they alone had used the substances; in 39.2% (15% of males and 29.3% of females) both the respondent and the partner used the substances; in 5.2% (1% of males and 7.3% of females) only their sexual partner used the substances. Over half, 55.5% (27.4% of males and 19.5% of females) of the respondents who used the psychoactive substances before sex reported being intoxicated during the most recent sexual intercourse, 36.1% (13.4% of males and 28% of females) indicated that they and their partner were intoxicated during the last sexual intercourse and 5.5% (1.4% of males and 6.7% of females) said only their partner was intoxicated. Only 2.9% (1.5% of males and 0.6% of females) of respondents reported that none of them was intoxicated before the last sexual intercourse. About 36.5% of the 744 respondents had ever had sex with a commercial sex worker or sex client. About 36.8% of the males in the current study had ever had sex with a commercial sex worker while 35.4% of the females had ever had sex with a sex client. Out of the 744 respondents in our study, 32.5% had ever had sex with a commercial sex worker or client while intoxicated. About 41.5% of the females and 30% of the males had ever had sex with a sex client or commercial sex worker while intoxicated.

Over 45% of the respondents believed that psychoactive substance use improves sexual performance, 43.3% believed that sex is more pleasurable when under the influence of psychoactive substances, while more than half (53.1%) believed that using psychoactive substances gives courage or confidence to approach a partner for sex (Table 2).

Results in Tables 3 and 4 show that respondents who endorsed that they were likely to engage in sex when under the influence of psychoactive substances had a 48% higher likelihood of engaging in multiple sexual relationships in the last 30 days compared to those who did not think so (PR 1.48, 95% CI: 1.15–1.90). Respondents who indicated that it is difficult to refuse sex when under the influence of psychoactive substances had a 28% higher likelihood of engaging in multiple sexual relationships in the last 30 days compared to those who did not think so (PR 1.28, 95% CI: 1.06–1.55) (Table 3). Believing that there is a higher likelihood of forgetting to use a condom while under the influence of psychoactive substances was positively associated with inconsistent condom use in the last 30 days (PR 1.26, 95% CI: 1.09–1.45) (Table 4).

The beliefs that psychoactive substance use improves sexual performance (PR 1.14, 95% CI: 1.01–1.30) or increases sexual urge (PR 1.20, 95% CI: 1.04–1.40) were associated with a higher likelihood of having sex while under the influence of psychoactive substances in the last 30 days. The feeling that using psychoactive substances gives courage or confidence to approach a partner increased the likelihood of having sex while under the influence of psychoactive substances in the last 30 days by 21% (PR 1.21, 95% CI: 1.05–1.39). Lastly, respondents who indicated finding themselves wanting to have sex when intoxicated had a 22% higher likelihood of having sex under the influence of psychoactive substances in the last 30 days compared to those who thought otherwise (PR 1.22, 95% CI: 1.03–1.43) (Table 5).

To the best of our knowledge, this is the first study to establish an association between sexual expectancies and inhibitions and high-risk sexual behaviours among young PSUs in informal settlements. However, this was a cross-sectional design and thus, it was not possible to draw causal inferences between the influence of psychoactive substances and high-risk sexual behaviours. Sexual expectancies and inhibitions could be affected depending on the dose of the psychoactive substance used; therefore, these findings cannot be generalised to all PSUs with varied doses. Substance use and sexual behaviours were self-reported, which could be subject to both recall and social desirability biases. Besides, this study did not apply sample weights, making it prone to unequal selection probability [55, 86].