Substance Use and Testing Sexual Partners using HIV Self-Tests

Abstract

HIV self-tests (HIVST) provide the possibility of testing partners prior to sexual activity. This study examines the role of substance use among men who have sex with men (n=123) and transgender women (n=13) who were provided HIVST to use with potential sex partners. Several reported avoiding/delaying alcohol (44%) or drug use (27%) because they intended to use HIVST. Those who used HIVST with alcohol (37%) or drugs (24%) did not differ from those who did not on number of partners asked, proportion who agreed, or intentions to use HIVST. A minority reported problems caused by substances. Ten did not ask someone to test because they were too drunk/high. Fourteen said it was fairly or very hard to use HIVST when under the influence. Eleven reported substances caused problems when discussing or administering HIVST, but only 2 of those felt the problems were major.

Introduction

The OraQuick™ In-Home HIV Test was approved by the FDA (US Food and Drug Administration) in 2012 (1). This product provides a means for individuals to test themselves for HIV in any setting they wish and receive results within 20 minutes (2). The test, which requires only an oral swab, is easy to use and interpret(3). The convenience and privacy of the test allows for more frequent testing, especially for those who live in rural areas, and for those who wish to avoid medical/professional settings, such as clinics or hospitals out of concerns about anonymity or stigma (4, 5). Early detection of HIV infection is critical for getting newly HIV positive individuals into treatment and modifying behavior to prevent subsequent transmission (6, 7).

HIV self-tests (HIVST) have been shown to be highly acceptable and easy to perform accurately among a wide range of populations (8–10). The availability of HIVST has also been associated with increased HIV testing frequency and a reduction in the number of additional exposures (11–13). Furthermore, having HIVST available also raises the possibility of testing potential sex partners prior to engaging in intercourse. A prior study examined hypothetical use of HIVST among a sample of men who have sex with men (MSM) at risk for HIV and found that, among 60 MSM, 80% said they would use HIVST to test potential partners if available (14). In a subsequent study, Carballo-Diéguez et al (2012) recruited a small sample (n=27) of MSM who reported having multiple partners and condomless anal sex, and provided them with HIVST to use themselves and with potential sex partners. The results showed high HIVST acceptability among participants, who were also able to use the tests with potential partners, leading to the detection of previously unknown infections and preventing exposure (15). Although having HIVST available did not eliminate all risk behavior, many participants did modify their behavior as a result, mainly abstaining from having sex or choosing to use condoms with partners who refused to test or tested HIV positive (16). In addition, participants reported increased awareness of risk, more frequent discussion of HIV/STI protection, and changes in partner selection (17).

Though MSM appear to be able to use HIVST correctly when sober, it is unknown whether this is also the case when individuals are under the influence of substances. Alcohol and drug use have been associated with negative outcomes in many aspects of HIV. Specifically, substance use can result in riskier sexual behavior, delays in HIV testing, delays in entering treatment, and poorer adherence to HIV medication (18). It is possible that substance use may also be a barrier to HIVST. In qualitative interviews with MSM regarding using HIVST on themselves and sex partners, substance use was mentioned as a possible impediment (15). Some felt they might forego using the test altogether if they were too drunk or high, while others questioned if they would be able to use the test properly.

To understand how substance use may affect HIVST, MSM and transgender women (TGW) participants in the current study were provided with HIV test kits to use with potential sex partners. We explored whether there were any indications that substance use might impact HIVST use or intentions. Participants were also asked specifically whether their own or their partners’ substance use caused any problems in the use of HIVST. Findings from this work could help to inform future interventions involving HIVST, particularly among substance using populations.

Methods

ISUM (“I’ll show you mine”) was a 5-year, randomized trial exploring the use of HIVST as a risk-reduction tool for individuals at high-risk of HIV infection. Eligible participants had to be HIV-negative, 18 years of age or older, identify as a man or transgender woman who has sex with men, report three or more occasions of condomless anal intercourse with sero-discordant or unknown status partners in the prior three-months, report two or more sexual partners in the previous three months, and not be on oral PrEP at the time of recruitment. The study took place in New York City (NYC), NY, and San Juan (SJ), Puerto Rico. Participants were recruited between 2014 and 2017 (detailed elsewhere – Iribarren et al. 2018) (19) through a wide variety of sources and methods, primarily by Master’s-level staff members: online recruitment via social media sites, dating sites, and dating apps; in person at gay pride events, LGBT non-profit organizations, bars and clubs; and through study participants who were offered a $10 incentive for referring friends who enrolled in the study. There was a concentrated effort to recruit transgender women, including contacting TGW who participated in other studies at our institution, targeting transgender social media sites, and encouraging TGW participants to tell others about the project.

Participants came to our project offices to complete study procedures. Informed consent was obtained by primarily Master’s-level staff with training and experience in similar research projects. Participants could choose English or Spanish when responding to surveys and interviews. They completed a baseline behavioral questionnaire via computer administered self-interview (CASI), and completed a rapid HIV self-test (OraQuick® In-home HIV Test) correctly without direction from staff (within 2 attempts), followed by a confirmatory capillary (fingerprick) whole blood test (Alere Determine™ HIV-½ Ag/Ab Combo Test) administered by staff. Eligible individuals were randomized to either intervention or control group. The intervention group participants received ten rapid oral HIV self-test kits to take home and viewed a video that included key points to consider when using the tests to screen sexual partners or clients (https://www.youtube.com/watch?v=uq6Qb4BJLdM). This initial visit took 1–2 hours. After three months, participants returned for a follow-up visit (Visit 3), in which they were re-tested for HIV and completed a follow-up CASI. The current analyses include intervention participants only because these individuals had the opportunity to use the test kits in the context of substance use-associated sexual encounters. All procedures were reviewed and approved by the Institutional Review Boards at New York State Psychiatric Institute and the University of Puerto Rico Medical Sciences Campus.

Assessment instruments

At baseline, the CASI survey included demographics such as age, education, income, ethnicity, race, gender identity, sexual orientation, employment status, and whether currently a student. Frequency of alcohol and drug use (13 different substances) in the previous three months was assessed using a 7-point scale (1=I never use this substance; 7=I use this substance more than once a day). These were dichotomized to indicate any use of the given substance vs. none in the past 3 months.

At follow-up, participants were asked various questions about HIVST use and substances: whether they avoided or delayed substance use because they were going to use the tests, how often they or their partners used substances prior to test use, whether they ever avoided asking a partner to use HIVST because they were too drunk/high, and how easy it was to use HIVST when they were under the influence of substances. Such questions were usually asked separately for alcohol and drugs and utilized 4- or 5-point Likert response scales (e.g., 1=“very easy” to 4=“very hard”).

Those who reported substance use prior to HIVST use were asked whether they felt alcohol or drug use caused any problems (minor or major) when they tried to raise the topic of HIVST with a potential sex partner or in the actual administration of HIVST use. Those who reported problems were then asked to describe what happened.

Participants were also asked general questions regarding HIVST: count variables included how many potential sex partners they asked to use the test and how many agreed to be tested. They were also asked how likely they were to use HIVST to test themselves or sex partners in the future (1= “There is no way…” to 4 “I would definitely use…”).

Analyses

Data presented are primarily descriptive. T-tests were used to compare those who used HIVST when 1) alcohol or 2) drugs were involved (self or partner) to those who did not on several continuous variables regarding HIVST use and intentions. Too few participants reported problems caused by substance use to warrant statistical comparisons with those who did not.

Results

Of the 272 participants who enrolled in the study, 136 were randomized to the intervention condition (83 in New York City and 53 in San Juan). Table 1 shows their baseline demographics. Participants were 18–69 years old (mean = 33.85) and most had some college education. Mean annual income was nearly $25,000. Ten percent of the participants identified as women or transgender.

Table 1.

Sample description (N=136)

	Mean (SD)
Age	33.85 (11.12)
Education1	4.36 (1.26)
Income	$24,668 (29,876)
Ethnicity	N (%)
Hispanic/Latino	76 (56%)
Race
White	39 (29%)
Black/African-American	64 (47%)
Asian	3 (2%)
Native American	1 (1%)
Other/More than one	28 (21%)
Gender
Man	123 (90%)
Woman/Transgender	13 (10%)
Sexual orientation
Gay/Homosexual	102 (75%)
Bisexual	26 (19%)
Straight/Heterosexual	4 (3%)
Other	4 (3%)
Occupation
Employed	94 (70%)
Student	22 (16%)

¹Education measured on 7-point scale; 4=partial college

Table 2 shows substance use in 3 months prior to baseline. Alcohol and marijuana were the most commonly used substances, with 86% reporting alcohol use and 54% reporting marijuana use the past 3 months. In addition, use of poppers (42%), cocaine (23%), and erectile disfunction (ED) drugs (18%) were fairly common. Regarding frequency of use, 42% of those who used alcohol did so more than once a week and 41% of those who used marijuana did so more than once a week. The only other substance that more than 5 participants used more than once per week was poppers (25% of those who used).

Table 2.

Substance use in prior 3 months (N=136)

Alcohol	116 (86%)
Marijuana	74 (54%)
Poppers	57 (42%)
Cocaine	31 (23%)
Viagra/Cialis/Levitra	24 (18%)
Other pharmaceutical	12 (9%)
MDMA	10 (7%)
Speed	10 (7%)
Ketamine	5 (4%)
Crack	5 (4%)
GHB	4 (3%)
Hallucinogens	3 (2%)
Heroin	3 (2%)

One hundred and thirty intervention participants (96%) returned for the 3-month follow-up visit. Regarding HIVST use, 111 (85%) asked at least 1 potential sex partner to be tested and 104 (80%) had at least one partner agree (71% had more than one partner agree to be tested). On average, 81% of those who were asked agreed to be tested. Table 3 shows participants’ responses to various questions about HIVST use and substances. Note that Ns are typically less than 130 because of skip patterns (e.g., some questions only applied to those who actually used the tests) or responses such as “don’t know” or “refuse to answer”, which were treated as missing data.

Table 3.

Substance and Home Test use.

	N with data	N (%)
Avoided or delayed drinking alcohol to use test	119
Never		66 (56%)
Once or twice		30 (25%)
Three or four times		13 (11%)
Five or more times		10 (8%)
Avoided or delayed using drugs to use test	116
Never		85 (73%)
Once or twice		15 (13%)
Three or four times		8 (7%)
Five or more times		8 (7%)
Substance use in 30 minutes prior to using test
Participant drank alcohol	118
None of the time		83 (70%)
Some of the time		24 (20%)
About half of the time		5 (4%)
Most of the time		1 (1%)
Every time		5 (4%)
Partner drank alcohol	108
None of the time		68 (63%)
Some of the time		30 (28%)
About half of the time		5 (5%)
Most of the time		2 (2%)
Every time		3 (3%)
Participant used recreational drugs	117
None of the time		98 (84%)
Some of the time		17 (15%)
About half of the time		1 (1%)
Most of the time		1 (1%)
Every time
Partner used recreational drugs	102
None of the time		77 (76%)
Some of the time		21 (21%)
About half of the time		1 (1%)
Most of the time		2 (2%)
Every time		1 (1%)
Did not ask partner to use test because too drunk/high	127	10 (8%)
How easy was it to use the test…
…when you were under influence of substances	127
Very easy		20 (16%)
Fairly easy		26 (21%)
Fairly hard		11 (9%)
Very hard		3 (2%)
Not applicable, I never did it		67 (53%)
…when your partner was under the influence of substances	127
Very easy		18 (14%)
Fairly easy		29 (23%)
Fairly hard		17 (13%)
Very hard		2 (2%)
Not applicable, I never did it		61 (48%)
How likely to use test to test self?	130
No way		2 (2%)
Unlikely		6 (5%)
Likely		21 (16%)
Definitely		101 (78%)
How likely to use test to test a partner?	130
No way		2 (2%)
Unlikely		8 (6%)
Likely		24 (19%)

Nearly half (44%) of participants reported avoiding or delaying drinking alcohol because of their intentions to use HIVST, and over a quarter (27%) reported avoiding drugs. The majority of participants did not mix substance use and test use. Thirty percent of participants reported that they drank alcohol and 16% used drugs prior to test use. Rates of partner substance use were somewhat higher. Thirty-seven percent used the test with at least one partner who had been drinking and nearly a quarter used the test with at least one partner who had been using drugs.

Ten participants (8%) said they didn’t ask a partner to use the test because they were too drunk or high. Eleven participants (9%) said it was “fairly hard” to use the test when they were under the influence of substances and three (2%) said it was “very hard.” Similar low rates were found for difficulties raised by a partner’s substance use. However it should be noted that when asked how hard it was to use the test when substances were involved, approximately half of participants replied “not applicable” because they did not use the test under those circumstances.

Based on their reports of substance use prior to testing, two dichotomous variables were calculated (results not shown). First, 44 participants who used the test when either themselves or a partner had been drinking were compared to 74 who did not. These groups were compared via t-tests on number of potential sex partners they asked to use the test (7.4 vs. 6.0; t=−1.1; df=116; p=.263), the proportion of asked partners who agreed to take the test (87% vs. 78%; t=−1.5; df=103; p=.145), the likelihood that they would use home tests in the future to test themselves (3.7 vs. 3.7; t=−0.1; df=116; p=.926), and the likelihood of using the test on sexual partners in the future (3.7 vs. 3.7; t=−0.6; df=116; p=.593). Likewise, 28 participants who used the test when either themselves or a partner had been using drugs were compared to 89 who did not. These groups were also compared on number of potential sex partners they asked to use the test (6.5 vs. 6.6; t=0.1; df=115; p=.952), the proportion of asked partners who agreed to take the test (79% vs. 82%; t=0.5; df=105; p=.615), the likelihood that they would use home tests in the future to test themselves (3.5 vs. 3.8; t=1.7; df=40; p=.092), and the likelihood of using the test on sexual partners in the future (3.5 vs. 3.8; t=1.7; df=41; p=.091). No significant differences were found.

The 44 participants who reported that either themselves or their partner had drunk alcohol prior to test use were asked whether alcohol caused any problems, either in raising the issue of testing with a partner, or actually using the test (Table 4). Five said alcohol caused some minor problems in talking about the test and one said “major problems.” Specifically, that participant said “Sí, la persona se puso agresivo” (“Yes, the person got aggressive”). None felt that alcohol caused major problems when actually using the test. The 28 participants who reported self or partner drug use prior to testing were also asked if drug use caused any problems. Two said drug use caused minor problems in raising the issue of testing and one said “major problems.” This participant reported, “Por desconocimiento, esperé un largo rato para realizarle la prueba casera a la pareja potencial” (“Because I was unaware, I ended up waiting a long time to do the home test with my potential partner”). No one felt that drug use caused any major problems when actually using the test. Their reasons for problems rarely included incapacitation due to substance use (e.g., “The entire process was much slower and more laborious in order to avoid making errors”) but typically referenced social/interpersonal issues (aggressive, loud, or rude behavior) that made discussing or using the test more difficult (e.g., “They laughed at me…”).

Table 4.

Problems with Home Test use due to substances.

Did alcohol cause any problems when initially talking to a partner about the test?	44
No		38 (86%)
Yes, minor problems		5 (11%)
Yes, major problems		1 (2%)
Did alcohol cause any problems when you actually used the test?	44
No		40 (91%)
Yes, minor problems		4 (9%)
Yes, major problems		0 (0%)
Did recreational drugs cause any problems when initially talking to a partner about the test?	28
No		25 (89%)
Yes, minor problems		2 (7%)
Yes, major problems		1 (4%)
Did recreational drugs cause any problems when you actually used the test?	28
No		27 (96%)
Yes, minor problems		1 (4%)
Yes, major problems		0 (0%)

Discussion

Substance use was very common among this sample of MSM and TGW. Yet, only two of fifty-three participants who used either alcohol or drugs in the context of partner testing reported that substance use resulted in major problems when initially talking to a partner about the test or using it. Only a minority of participants indicated that substance use did occasionally interfere with HIVST use. For some, it impaired their ability to discuss HIVST use with a potential partner; some did not even bring it up because they were too drunk/high. Also, 11–15% said it was fairly hard or very hard to use the test when they or a partner was under the influence. Therefore, overall, the use of HIVST prior to sexual intercourse appears to be a feasible although limited harm-reduction tool, and the potential for alcohol or drug use to complicate HIV testing should be part of any educational/intervention program designed to promote HIVST use. In high-risk populations such as this, sexual partners are often met in contexts where alcohol or drugs are being used, so interventions should take into account this reality. Previous work has considered the role of substance use alongside partner-based HIV prevention strategies (e.g., condom use). Work from substance use/condom use programs including one by Zellner et. al. 2015 (e.g., study tested a culturally-tailored intervention to provide substance use/HIV education to reduce perceived stress young Black Americans) (20), and another by Kurtz et. al. 2013 (e.g., a grouped series of sexual and substance use risk reduction interventions based on empowerment theory) (21) could provide useful starting points for future educational efforts.

Importantly, our results indicate that people seem willing to modify their behavior in the interest of proper HIVST use. On their own, nearly half avoided or delayed drinking and over a quarter avoided or delayed using drugs in anticipation of using the tests. Given the possibility that substance use might interfere with using the tests to screen sexual partners, educational programs could include this as a strategy for the most effective HIVST protection. Participants might find it more acceptable if presented in terms of delaying or limiting substance use until after testing, rather than avoiding substance use altogether. However, it was also clear that it was often the partner’s substance use that caused problems. Activities such as role-playing could be used to prepare participants for how to discuss HIVST in situations where a partner may be under the influence. This could be an especially useful component of future HIVST interventions, since role playing is a commonly used tactic in other studies (22) to practice HIV prevention behaviors that can be challenging in a real world settings.

Limitations

This study had some important limitations. First, all data were collected via self-report, so declared levels of substance use among participants and their partners could have been impacted by social desirability or recall bias. Second, given our sample size, we were unable to identify potential associations between substances and HIVST use variables. In addition, substance use was not an inclusion criterion for this study so nearly half of the intervention participants did not attempt to use HIVST in the context of substance use. A study involving those with more alcohol/drug use/abuse issues may result in different findings concerning use of HIVST. Finally, MSM and TGW were not analyzed separately as there were only 13 TGW. The relationship between substance use and HIVST may be different for these two groups and a larger sample of TGW would have allowed that to be explored.

Conclusions

In conclusion, although substance use was common in this sample, most participants did not experience problems when attempting to use HIVST with potential sexual partners. Some participants took measures to delay or forego substance use on some occasions when they anticipated that it could be a barrier to HIVST use. These findings demonstrate the feasibility of HIVST use as an additional HIV prevention intervention and harm reduction tool even among those who use substances, and can be used to guide future interventions.