Relationship Characteristics Associated with Perceptions of Partners’ HIV Testing Behavior Among Male Couples

Abstract

Partnered men who have sex with men are not immune to the risk of HIV. Analyzing dyadic data from 360 male couples recruited from April 2016 to June 2017, we examined how relationship characteristics might influence HIV testing behavior and perceptions of partners’ HIV testing. Increasing levels of mutual trust were associated with lower odds of both partners (versus neither) having been tested in the past year (aOR = 0.91, 95% CI 0.83–0.99), but increasing levels of communal coping were associated with higher odds (aOR = 1.06, 95% CI 1.02–1.10). Only one partner was more likely to be correct about whether or not his partner had been tested in the past year (versus both) if someone had broken their sexual agreement (aOR = 2.60, 95% CI 1.17–5.76). Increasing differences in trust (aOR = 1.10, 95% CI 1.01–1.20) were also associated with higher odds of only one partner being correct. Dyadic HIV prevention efforts should incorporate skills building around negotiating sexual agreements and constructive communication.

Introduction

Epidemiological research conducted over the past decade has drawn attention to the role of male dyads in the national HIV epidemic [1, 2]. One modeling study found that condomless anal sex (CAS) within main partnerships was the source of approximately 39% of HIV infections [1], and another modeling study found that approximately 68% of HIV transmissions were attributed to sex between main partners [2]. Consequently, there has been an increase in the number of published studies investigating how relationship characteristics shape sexual HIV risk behaviors within male couples [3–6]. A recent systematic review noted that positive relationship dynamics such as a desire to establish mutual trust, build emotional intimacy, and demonstrate commitment to each other are associated with greater odds of CAS with main partners [7]. Lack of confirmation of both partners’ HIV serostatus before engaging in CAS, and inadequate viral suppression in serodiscordant partnerships can in turn increase the risk of HIV transmission [8–10]. Therefore, it is important to integrate dyadic concerns into comprehensive HIV prevention interventions for men who have sex with men (MSM), where applicable [11].

Testing for HIV can be considered to be the cornerstone of prevention efforts, as meta-analytic evidence suggests that people who learn they are HIV-positive take steps to reduce the risk of onward transmission [12]. Long-term adherence to antiretroviral treatment ensures that one’s HIV viral load remains suppressed, effectively reducing the risk of HIV transmission among male couples through CAS to zero [13]. Testing for HIV is also the first step in accessing efficacious biomedical prevention options such as pre-exposure prophylaxis (PrEP) among those who are HIV-negative [14]. The Centers for Disease Control and Prevention (CDC) recommends that all sexually active gay, bisexual, and other MSM should be tested for HIV at least once a year (and more frequently if warranted based on an assessment of their individual risk factors, local HIV epidemiology, and local testing policies) [15]. However, screening rates in this subgroup remain suboptimal [16], particularly among partnered MSM. Recent data indicate that MSM in relationships are getting tested for HIV at lower rates compared to the general MSM population, even after being possibly exposed to the virus [17, 18]. Moving beyond individual-level approaches to couples-based approaches might help increase HIV testing levels, and enable the formulation of joint prevention or treatment plans based on dyadic HIV serostatus (i.e. seroconcordant negative, seroconcordant positive, or serodiscordant) [11].

Few published studies have examined how relationship characteristics influence couple-level HIV testing behavior. Motivations to test for HIV among partnered MSM could be event-related (e.g., testing when initiating a new relationship), or partner-related (e.g., testing in response to a partner’s request) [19]. One study involving 275 male couples in the United States (US) found that having a sexual agreement (defined as an explicit understanding between two individuals about the types of sex-related behaviors allowed within and outside their relationship [20, 21]), and having lower levels of confidence in the relationship were positively associated with testing for HIV multiple times in a year [22]. Another study involving 142 male couples in Portland, OR and Seattle, WA found that having a sexual agreement, and being in a relationship where at least one partner had CAS outside the relationship in the past 3 months, were positively associated with testing for HIV during the same time frame [23]. Both studies provided important insights, but there remains a gap in our understanding of how some other relationship characteristics such as trust, communal coping (involving communication, joint decision-making, and working together to reduce a health threat), and agreement regarding sexual health issues potentially affect male couples’ HIV testing behavior.

In-depth interviews conducted with partnered MSM have revealed that trust, intimacy, and commitment to their relationship may impede them from getting regularly tested for HIV [19]. Although the association between these characteristics and engaging in CAS has been well-established [7], there is still a need to explore how levels of mutual trust and differences in trust between partners might directly influence couple-level HIV testing behavior. Communal coping refers to a cooperative problem-solving process, in which partners mutually consider a health threat and formulate a joint plan to reduce their risk [24]. Viewing a sexual health issue from a communal perspective could result in sharing the responsibility of coping, and has been theorized to foster positive health behavior change among couples [25]. For example, both partners could choose to refrain from CAS with outside partners in an attempt to reduce the risk of HIV transmission between them, as opposed to only one partner choosing to do so. Similarly, both partners could formulate a plan to get tested for HIV at regular intervals, rather than only one partner getting tested. We are unable to find published studies evaluating the potential association between communal coping and couple-level HIV testing behavior among male dyads. Understanding this link could help inform future prevention efforts seeking to develop strategies to encourage effective communal coping. Finally, the extent to which both partners communicate and agree with one another about sex and prevention topics has been noted in prior studies with male couples as being an important relationship factor to bolster for HIV prevention [6, 26].

The perceived risk of HIV infection within a relationship is another factor that might influence a male couple’s decision to get tested. Research indicates that partnered MSM often believe they are at lower risk for HIV compared to those who are single, i.e. being in a committed relationship conveys a sense of protection [27]. A recent study with 906 MSM recruited online from across the US revealed that men with a main partner had significantly higher odds of perceiving zero risk of HIV infection, and higher odds of being very confident they will remain HIV-negative [28]. Greater trust in one’s sex partner has been associated with misperceiving him to be HIV-negative when the partner’s self-reported serostatus is positive or unknown [29]. Qualitative research has found that MSM in relationships have fewer discussions about HIV testing, even in the context of sexual agreements that allow for a wide range of sexual activities with outside partners (such as mutual masturbation, rimming, and oral or anal sex with or without condoms) [30]. Data on the annual HIV testing behavior of male couples in the US are available [17, 18], but little is known about each partner’s belief regarding whether or not his partner gets tested for HIV. Incorrect assumptions about a partner’s HIV testing behavior, coupled with an erroneous belief that he is HIV-negative, may result in forgoing condom use during anal sex, thereby increasing the risk of HIV transmission within the relationship. A thorough literature review did not reveal any studies with partnered MSM that have evaluated how relationship characteristics might affect perceptions of their partners’ HIV testing behavior.

To help address these knowledge gaps, we used dyadic data from a national sample of male couples to identify demographic, relationship, and behavioral factors independently associated with annual HIV testing behavior, and perceptions of partners’ HIV testing behavior. This new information can promote an understanding of how constructs such as mutual trust, communal coping (involving communication, joint decision-making, and working together to reduce a health threat), and mutual agreement regarding sexual health issues warrant due consideration when designing HIV and other sexually transmitted infection (STI) prevention interventions geared towards MSM in relationships.

Methods

Study Design

Overview of Study

Data for these analyses come from the baseline survey of a randomized controlled trial (RCT) to evaluate an innovative dyadic HIV testing and counseling intervention for male couples in the US (ClinicalTrials.gov Identifier: NCT02335138). Specifically, the RCT sought to examine the effects of video-based counseling combined with rapid home HIV testing on male couples’ management of HIV risk, formation and adherence to sexual agreements, and sexual risk-taking. Eligible couples first completed a baseline survey programmed in SurveyGizmo, after which they were randomized to receive either rapid home HIV test kits followed by self-reporting of test results via the study’s website (control arm), or rapid home HIV test kits followed by testing together under the facilitation of a remotely-located counselor during a prescheduled video session (experimental arm). Couples were also asked to complete follow-up surveys at 3 months, and at 6 months. For each of the surveys, separate emails containing direct links were sent to both partners at the same time, but participants were requested to take the surveys individually. Additional details regarding the study protocol have been published elsewhere [31]. Approval for the study was obtained from the Institutional Review Board at the University of Michigan (HUM00102906).

Eligibility Criteria

Couples were eligible for the RCT if each partner reported being cis-gender male, 18 years of age or older, in a sexual relationship for at least 6 months, willing to receive rapid home HIV test kits, able to access the Internet at their own or their partner’s home, and willing to participate in video-based counseling along with their partner. Seroconcordant HIV-positive couples and those in whom one or both partners reported experiencing severe intimate partner violence or coercion in the past year were not eligible to participate in the RCT.

Participant Recruitment

Participants were recruited from across the US via targeted advertising on social media platforms (e.g., Facebook, Instagram), and geospatial mobile apps for gay, bisexual, and other MSM (e.g., Grindr, Scruff). Other recruitment methods such as advertising in POZ magazine, and sending emails to MSM who completed the American Men’s Internet Survey were also utilized. Between April 2016 and June 2017, 13,592 individuals accessed the study’s landing page that contained a short description of the RCT including its proposed activities. Of these, 2926 individuals (21.53%) provided electronic informed consent and clicked through to the eligibility screener, of which 862 individuals (29.46%), i.e. 431 couples met the inclusion criteria. Of these, both partners in 424 couples (98.38%) completed individual baseline surveys.

Formulation of Analytic Sample

Data presented in this manuscript are restricted to responses provided in the baseline survey. Three of 424 couples (0.71%) in which participants did not answer questions pertaining to their own or their partners’ past HIV testing behavior were excluded. Additionally, data from 15 serodiscordant couples (3.54%) were not included in the analytic sample because the HIV-positive partners reported being diagnosed more than a year ago, and were currently on antiretroviral treatment. Given our interest in couple-level HIV testing behavior in the past year, we believe this exclusion was appropriate. Finally, an examination of the timestamps recorded in Survey Gizmo indicated that partners in 46 couples (10.85%) completed their surveys within 15 min of each other. Although we requested participants to complete their own survey separately from their partners, it is possible they could have discussed certain questions when responding during a similar timeframe. Both partners in 40 of 46 couples (86.96%) were correct about whether or not their partners had been tested in the past year, which was significantly greater than 252 of 360 couples (70.00%) in which both partners completed their surveys more than 15 min apart (p-value = 0.006). Because of our focus on perceptions of partners’ HIV testing behavior, we decided to exclude data from these 46 couples as a precautionary measure, resulting in a final analytic sample of 360 couples.

Baseline Survey Measures

Demographic, Behavioral and Relationship Characteristics

Demographic information collected from participants included their age, race/ethnicity, sexual orientation, employment status, and highest educational level. Regarding risk behaviors, participants were asked to indicate if they had used any non-prescription drugs in the past 3 months. Those who responded in the affirmative were asked to select one or more options from the following list: marijuana (“pot” or “weed”), amyl nitrite (“poppers”), central nervous system depressants (“downers” such as Valium, Ativan or Xanax), opioid analgesics (such as Oxycontin or Percocet), hallucinogens (such as lysergic acid diethylamide or “acid”), 3,4-methylenedioxy-methamphetamine (“ecstasy” or “molly”), club drugs (such as ketamine or “special K”), non-injection amphetamine (“speed”, “crystal meth” or “crank”), injection amphetamine (“speed”, “crystal meth” or “crank”), non-injection cocaine (smoked or snorted), injection cocaine, non-injection heroin (smoked or snorted), injection heroin, or some other non-prescription drug. Participants were also asked to indicate how often they had consumed ≥ 6 alcoholic drinks on the same occasion in the past 3 months (i.e. never, less than monthly, monthly, weekly, almost daily, or daily). Those who reported consuming ≥ 6 alcoholic drinks on the same occasion weekly, almost daily, or daily were categorized as heavy alcohol users.

The baseline survey also included a series of questions to elucidate specific characteristics of the current relationship, such as legal marital status, relationship duration, whether participants had engaged in CAS with their main partner in the past 3 months, and whether participants had a sexual agreement with their main partner. Participants who reported having a sexual agreement were asked if it was “open” (i.e. sex with outside partners was allowed with or without some restrictions), or “closed” agreement (i.e. sex with outside partners was not allowed), and whether they had ever broken their agreement (i.e. violated the agreed-upon terms of their sexual agreement). Information was also obtained on engaging in CAS outside the relationship in the past 3 months (defined as receptive or insertive anal sex with ≥ 1 outside partners without using a condom the entire time). Survey measures to assess other relationship characteristics of interest (i.e. mutual trust, communal coping, and mutual agreement regarding sexual health issues) are described in detail below.

Mutual Trust

Mutual trust was measured using a modified version of the Dyadic Trust Scale [32]. The original scale is a 7-point Likert tool ranging from “very strongly disagree” to “very strongly agree” on eight items (one’s partner is more interested in his own welfare, there are times when one’s partner cannot be trusted, one’s partner is honest and truthful, one feels he can trust his partner completely, one’s partner is truly sincere in his promises, one feels his partner does not show him enough consideration, one’s partner treats him fairly and justly, one can count on his partner for help), and has been tested and validated among sexual minority populations, including men in same-sex relationships [33]. We used a 5-point Likert tool ranging from “strongly disagree” to “strongly agree” on the same eight items. Data from our sample showed strong internal reliability for this scale (Cronbach’s alpha = 0.87).

Three items were reverse coded, and each participant’s responses were summed to calculate an overall score potentially ranging from 8 to 40 (only for those participants who responded to all items). Trust within the relationship was calculated by averaging the scores for each partner, with higher values indicating greater levels of mutual trust. Differences in trust between partners were calculated by taking the absolute difference between the scores for each partner, which could range from 0 to 32. One or both partners in 19 couples did not respond to all items of the scale, and these cases were not included in our multinomial logistic regression models described below.

Communal Coping

The Communal Coping to Reduce HIV Threat Scale was used to measure the extent to which couples communicate, make joint decisions, and work together to reduce their risk of acquiring HIV. For each of the three domains, data were collected using a 5-point Likert tool ranging from “not to any extent at all” to “to a great extent” on seven items (using condoms when having sex with each other, limiting the number of outside sex partners, having sex outside the relationship, using condoms when having sex outside the relationship, getting regularly tested for HIV and other STIs, being the insertive or receptive partner when having sex with each other, being sexually faithful to each other). This scale was specifically developed and validated among MSM who had been partnered for at least 3 months [34]. Data from our sample showed strong internal reliability for this scale (Cronbach’s alpha = 0.92).

Each participant’s responses to items on the Communal Coping to Reduce HIV Threat Scale were summed to calculate an overall score potentially ranging from 21 to 105 (only for those participants who responded to all items). Communal coping within the relationship was calculated by averaging the scores for each partner, with higher values indicating greater levels of communication, joint decision-making, and working together to reduce their risk of acquiring HIV. Differences in communal coping between partners were calculated by taking the absolute difference between the scores for each partner, which could range from 0 to 84. One or both partners in 57 couples did not respond to all items of the scale, and these cases were not included in our multinomial logistic regression models described below.

Mutual Agreement Regarding Sexual Health Issues

Mutual agreement regarding sexual health issues was measured using the Preferences for Sexual Health Outcomes Scale, a 5-point Likert tool ranging from “we don’t agree at all” to “we completely agree” on the same seven items that comprise the Communal Coping to Reduce HIV Threat Scale. This scale was also developed and validated among men in same-sex relationships of a 3 months or longer duration [34]. Data from our sample showed good internal reliability for this scale (Cronbach’s alpha = 0.72).

Each participant’s responses to items on the Preferences for Sexual Health Outcomes Scale were summed to calculate an overall score potentially ranging from 7 to 35 (only for those participants who responded to all items). Agreement regarding sexual health issues within the relationship was calculated by averaging the scores for each partner, with higher values indicating greater levels of mutual agreement. Differences in agreement between partners were calculated by taking the absolute difference between the scores for each partner, which could range from 0 to 28. One or both partners in 14 couples did not respond to all items of the scale, and these cases were not included in our multinomial logistic regression models described below.

Outcome Measures

Couple-level HIV Testing History

Participants were asked about their past HIV testing behavior using the following questions: (i) “Have you ever been tested for HIV, the virus that causes AIDS? Response options: Yes, No”, and (ii) “When were you last tested for HIV? Response options: Past 6 months, 6 months to 1 year ago, 1 to 2 years ago, 2 to 3 years ago, 3 to 4 years ago, 4 to 5 years ago, More than 5 years ago” (only those who responded Yes to the previous question).

Responses were used to create a dichotomous variable (1 = Yes, 0 = No) reflecting whether or not a participant had been tested in the past year. If a participant selected “Past 6 months”, or “6 months to 1 year ago”, he was assigned a value of 1 for this variable. If a participant selected any other response option, he was assigned a value of 0. Couple-level HIV testing history was derived by combining data from each partner on whether or not he had been tested in the past year (i.e. both partners tested, only one partner tested, neither partner tested).

Couple-Level Perception of Partner’s HIV Testing History

Participants were asked about their partners’ past HIV testing behavior using the following questions: (i) “To your knowledge, has [Partner Name] ever been tested for HIV, the virus that causes AIDS? Response options: Yes, No”, and (ii) “To your knowledge, when was [Partner Name] last tested for HIV? Response options: Past 6 months, 6 months to 1 year ago, 1 to 2 years ago, 2 to 3 years ago, 3 to 4 years ago, 4 to 5 years ago, More than 5 years ago” (only those who responded Yes to the previous question).

Responses were used to create a dichotomous variable (1 = Yes, 0 = No) reflecting whether or not a participant believed their partner had been tested in the past year. If a participant selected “Past 6 months”, or “6 months to 1 year ago”, he was assigned a value of 1 for this variable. If a participant selected any other response option, he was assigned a value of 0. An incorrect or a correct perception of one’s partner’s HIV testing history was quantified by comparing this variable for each participant with the variable reflecting whether or not his partner had been tested in the past year. For example, if the variable reflecting a participant’s belief about his partner’s past year testing behavior was coded 1, but the variable reflecting his partner’s past year testing behavior was coded 0, the participant was deemed to have an incorrect perception of his partner’s HIV testing history. Couple-level perception of partner’s HIV testing history was derived by combining these data for each partner (i.e. neither partner was correct, only one partner was correct, both partners were correct).

Descriptive Analyses and Regression Modeling

Statistical analyses were conducted using SAS 9.4. Descriptive statistics were used to summarize the demographic, relationship, and behavioral characteristics of the sample. Our two main outcomes of interest were: (i) couple-level HIV testing behavior in the past year (i.e. both tested, only one tested, neither tested), and (ii) couple-level perception of partner’s HIV testing behavior in the past year (i.e. neither was correct, only one was correct, both were correct). Both categorical outcomes integrated data from each partner thereby existing at the couple-level, i.e. both partners shared the same value on each of these outcomes [35]. The dyads were indistinguishable, i.e. there was no meaningful way to systematically differentiate between individuals as “Partner 1” versus “Partner 2” across all dyads [36].

Separate multinomial logistic regression models were formulated to identify independent associations of demographic, relationship, and behavioral characteristics with each of these categorical outcomes. Because the dyads were indistinguishable, these models were restricted to couple-level characteristics as recommended in the literature [35]. Both models were true single-level models where the total number of observations used in the analyses of effects equaled the total number of couples who were not missing data on any of the explanatory variables (n = 280). The referent for the first model was “neither partner tested”, and the referent for the second model was “both partners were correct”. Both models were adjusted for age, race/ethnicity, relationship duration, CAS with their main partner, whether their sexual agreement allowed for sex with outside partners, whether someone had ever broken their sexual agreement, CAS outside the relationship, non-prescription drug use, alcohol use, trust, communal coping, and agreement regarding sexual health issues. Results are presented as adjusted odds ratios (aORs) with 95% confidence intervals (CIs).

Results

Table 1 describes the overall composition of our sample. Regarding individual level characteristics, the majority of 720 participants were younger than 29 years (57.50%), and non-Hispanic white (64.72%). Forty-three (5.97%) were non-Hispanic black/African American, 144 (20.00%) were Hispanic, and 67 (9.31%) were of some other race/ethnicity. More than half the participants (53.75%) had completed college or had a higher level of education. Eighty participants (11.11%) were unaware of their HIV serostatus. Regarding dyadic level characteristics, most of the 360 couples (74.44%) had less than a 5-year age difference, and were of the same race/ethnicity (61.94%). However, almost two-thirds of couples (65.00%) were comprised of partners with different educational levels. In the past 3 months, at least one partner in 40.83% of couples had used non-prescription drugs, and at least one partner in 10.28% of couples was a heavy alcohol user. One or both partners were unaware of their HIV serostatus in 67 couples (18.61%).

Table 1.

Individual and dyadic level characteristics of 360 male couples, United States, April 2016–June 2017

Characteristic	N	(%)
Individual level (n = 720)
Age (years)a
18–24	201	(27.92)
25–29	213	(29.58)
30–34	135	(18.75)
≥35	171	(23.75)
Race/ethnicity
Non-Hispanic white	466	(64.72)
Non-Hispanic Black/African American	43	(5.97)
Hispanic	144	(20.00)
Otherb	67	(9.31)
Sexual orientation
Homosexual/gay	653	(90.69)
Bisexual	49	(6.81)
Otherc	18	(2.50)
Employment status
Worked full-time	501	(69.58)
Worked part-time	112	(15.56)
Unemployed	48	(6.67)
Otherd	59	(8.19)
Highest educational level
High school diploma or some high school	107	(14.86)
Associate’s/technical degree or some college	226	(31.39)
Bachelor’s degree	233	(32.36)
Master’s/doctoral degree	154	(21.39)
Non-prescription drug use in the past 3 months
Yese	198	(27.50)
No	522	(72.50)
Heavy alcohol use in the past 3 months
Yesf	41	(5.69)
No	679	(94.31)
Self-reported HIV status
Negative	640	(88.89)
Unknowng	80	(11.11)
Dyadic level (n = 360)
Age (years)
> 5 years apart	92	(25.56)
Same or within 5 years of each other	268	(74.44)
Race/ethnicity
Differenth	137	(38.06)
Samei	223	(61.94)
Sexual orientation
Different	56	(15.56)
Same	304	(84.44)
Employment status
Different	145	(40.28)
Same	215	(59.72)
Highest educational level
Different	234	(65.00)
Same	126	(35.00)
One or both partners used non-prescription drugs in the past 3 months
Yesj	147	(40.83)
No	213	(59.17)
One or both partners were heavy alcohol users in the past 3 months
Yesk	37	(10.28)
No	323	(89.72)
One or both partners were unaware of their HIV serostatus
Yesl	67	(18.61)
No	293	(81.39)

^aAge: mean = 30.48, median = 28, range 18–68

^bIncludes 34 multiracial, 21 Asian, 4 Native American/Alaskan Native, 2 Native Hawaiian/Other Pacific Islander, and 6 other

^cIncludes 15 queer, and 3 questioning/unsure

^dIncludes 38 students, 6 collecting disability, 4 retired and 11 other

^eIncludes 145 who had used marijuana (“pot” or “weed”), 46 who had used amyl nitrite (“poppers”), 25 who had used central nervous system depressants (“downers” such as Valium, Ativan or Xanax), 19 who had used opioid analgesics (such as Oxycontin or Percocet), 11 who had used hallucinogens (such as lysergic acid diethylamide or “acid”), 11 who had used 3,4-methylenedioxy-methamphetamine (“ecstasy” or “molly”), 4 who had used club drugs (such as ketamine or “special K”), 8 who had used non-injection amphetamine (“speed”, “crystal meth” or “crank”), 2 who had used injection amphetamine (“speed”, “crystal meth” or “crank”), 22 who had used non-injection cocaine (smoked or snorted), 1 who had used non-injection heroin (smoked or snorted), and 16 who had used some other non-prescription drug (numbers are not mutually exclusive)

^fDefined as consuming ≥ 6 alcoholic drinks on the same occasion weekly, almost daily, or daily

^gIncludes 75 who had never been tested for HIV, and 5 who did not know their HIV status

^hIncludes 11 couples in which one partner was non-Hispanic white and the other was non-Hispanic black/African American, 65 couples in which one partner was non-Hispanic white and the other was Hispanic, 36 couples in which one partner was non-Hispanic white and the other was of some other race/ethnicity, 7 couples in which one partner was non-Hispanic black/African American and the other was Hispanic, 5 couples in which one partner was non-Hispanic black/African American and the other was of some other race/ethnicity, 8 couples in which one partner was Hispanic and the other was of some other race/ethnicity, and 5 couples in which each partner was of a different other race/ethnicity

ⁱIncludes 177 couples in which both partners were non-Hispanic white, 10 couples in which both partners were non-Hispanic black/African American, 32 couples in which both partners were Hispanic, and 4 couples in which both partners were of the same other race/ethnicity

^jIncludes 96 couples in which one partner had used non-prescription drugs in the past 3 months, and 51 couples in which both partners had used non-prescription drugs in the past 3 months

^kIncludes 33 couples in which one partner was a heavy alcohol user in the past 3 months, and 4 couples in which both partners were heavy alcohol users in the past 3 months

^lIncludes 54 couples in which one partner was unaware of his HIV status, and 13 couples in which both partners were unaware of their HIV status

Table 2 summarizes the relationship and sexual behavioral characteristics of the couples. Nineteen of 185 couples (10.27%) who had been in a relationship for 6 months to 3 years, and 82 of 175 couples (46.86%) who had been in a relationship for 3 or more years were legally married. Of the 261 couples (72.50%) who had formulated sexual agreements, 75 (28.74%) had an “open” agreement (i.e. sex with outside partners was allowed with or without some restrictions), 165 (63.22%) had a “closed” agreement (i.e. sex with outside partners was not allowed), and 21 (8.05%) disagreed about their type of agreement. Seventy of these 261 couples (26.82%) were comprised of partners where at least one had previously broken their sexual agreement (i.e. violated the agreed-upon terms of their agreement). Of these 70 couples, at least one partner in 25 (35.71%) reported engaging in CAS outside the relationship in the past 3 months. Overall, at least one partner in 59 of 360 couples (16.39%) reported engaging in CAS outside the relationship, and of these 59 couples, at least one partner in 27 (45.76%) reported CAS with ≥ 2 outside male sex partners.

Table 2.

Relationship and sexual behavioral characteristics of 360 male couples, United States, April 2016–June 2017

Characteristic	N	(%)
Relationship duration
6 months to 3 years	185	(51.39)
≥ 3 years	175	(48.61)
Had condomless anal sex with their main partner in the past 3 months
Yes	315	(87.50)
No	45	(12.50)
Sexual agreement allowed for sex with outside partners
Yes	75	(20.83)
Noa	285	(79.17)
One or both partners had ever broken their sexual agreement
Yesb	70	(19.44)
Noc	290	(80.56)
One or both partners had condomless anal sex outside the relationship in the past 3 monthsd
Yese	59	(16.39)
No	301	(83.61)

^aIncludes 165 couples whose sexual agreements did not allow for sex with outside partners, 21 couples in which partners disagreed about their type of sexual agreement, and 99 couples who had not formulated sexual agreements

^bIncludes 62 couples in which one partner had ever broken their sexual agreement, and 8 couples in which both partners had ever broken their sexual agreement

^cIncludes 191 couples in which neither partner had ever broken their sexual agreement, and 99 couples who had not formulated sexual agreements

^dDefined as engaging in receptive or insertive anal sex with ≥ 1 outside partners without using a condom the entire time

^eIncludes 46 couples in which one partner had condomless anal sex outside the relationship in the past 3 months, and 13 couples in which both partners had condomless anal sex outside the relationship in the past 3 months

Table 3 presents data on trust, communal coping, and agreement regarding sexual health issues among the couples. The distributions for the averages of each partner’s score on the modified Dyadic Trust Scale (n = 341), the Communal Coping to Reduce HIV Threat Scale (n = 303), and the Preferences for Sexual Health Outcomes Scale (n = 346) were skewed to the left, indicating generally high levels of mutual trust, communal coping, and mutual agreement regarding sexual health issues within the relationships. The absolute differences between each partner’s score on these scales were right-skewed, indicating generally few differences in trust, communal coping, and agreement regarding sexual health issues between partners.

Table 3.

Trust, communal coping, and agreement regarding sexual health issues among 341, 303, and 346 male couples respectively, United States, April 2016–June 2017

Characteristic	Mean	Median	Lower quartile	Upper quartile	Range
Trust within the relationshipa (n = 341)	33.71	34.50	31.50	37.00	20–40
Differences in trust between partnersb (n = 341)	4.77	4.00	2.00	6.00	0–32
Communal coping within the relationshipc (n = 303)	89.02	90.50	82.00	98.50	52–105
Differences in communal coping between partnersd (n = 303)	15.02	12.00	4.00	22.00	0–62
Agreement regarding sexual health issues within the relationshipe (n = 346)	30.93	31.50	29.00	33.50	18–35
Differences in agreement regarding sexual health issues between partnersf (n = 346)	3.43	3.00	1.00	5.00	0–20

^aCalculated by averaging the scores for each partner on the modified 5-point Likert Dyadic Trust Scale (Potential range 8–40). Higher values indicate greater levels of trust within the relationship

^bCalculated by taking the absolute difference between the scores for each partner on the modified 5-point Likert Dyadic Trust Scale (Potential range 0–32). Higher values indicate greater differences in trust between partners

^cCalculated by averaging the scores for each partner on the 5-point Likert Communal Coping to Reduce HIV Threat Scale (Potential range 21–105). Higher values indicate greater levels of communal coping within the relationship

^dCalculated by taking the absolute difference between the scores for each partner on the 5-point Likert Communal Coping to Reduce HIV Threat Scale (Potential range 0–84). Higher values indicate greater differences in communal coping between partners

^eCalculated by averaging the scores for each partner on the 5-point Likert Preferences for Sexual Health Outcomes Scale (Potential range 7–35). Higher values indicate greater levels of agreement within the relationship

^fCalculated by taking the absolute difference between the scores for each partner on the 5-point Likert Preferences for Sexual Health Outcomes Scale (Potential range 0–28). Higher values indicate greater differences in agreement between partners

Regarding HIV testing history, 355 of 720 participants (49.31%) had been tested in the past year (171 tested in the past 6 months, and 184 tested between 6 months to 1 year ago). Both partners in 131 couples (36.39%) had been tested; only one partner in 93 couples (25.83%) had been tested; neither partner in 136 couples (37.78%) had been tested.

Table 4 summarizes the demographic, relationship, and behavioral characteristics associated with couple-level HIV testing behavior in the past year. Both partners were more likely to have been tested if their race/ethnicity differed (aOR = 2.23, 95% CI 1.20–4.15), their relationship was shorter than 3 years (aOR = 3.40, 95% CI 1.80–6.43), their sexual agreement allowed for sex with outside partners (aOR = 3.38, 95% CI 1.44–7.94), and someone had used non-prescription drugs in the past 3 months (aOR = 1.97, 95% CI 1.01–3.85). Increasing levels of mutual trust were associated with lower odds of both partners having been tested (aOR = 0.91, 95% CI 0.83–0.99). However, increasing levels of communal coping were associated with higher odds of both partners having been tested (aOR = 1.06, 95% CI 1.02–1.10). None of the characteristics were significantly associated with only one partner having been tested, except if someone had used non-prescription drugs in the past 3 months (aOR = 2.02, 95% CI 1.01–4.06).

Table 4.

Characteristics associated with couple-level HIV testing history among 280 male couples, United States, April 2016–June 2017

Characteristic	Couple-level HIV testing behavior in the past year
	Both partners tested		Only one partner tested
	aOR	(95% CI)	aOR	(95% CI)
Categorical variables
Age (years)
> 5 years apart	1.17	(0.58–2.36)	1.02	(0.47–2.19)
Same or within 5 years of each other	Ref.		Ref.
Race/ethnicity
Different	2.23	(1.20–4.15)*	1.19	(0.61–2.33)
Same	Ref.		Ref.
Relationship duration
6 months to 3 years	3.40	(1.80–6.43)***	1.57	(0.81–3.02)
≥3 years	Ref.		Ref.
Had condomless anal sex with their main partner in the past 3 months
Yes	0.77	(0.31–1.87)	0.82	(0.31–2.18)
No	Ref.		Ref.
Sexual agreement allowed for sex with outside partners
Yes	3.38	(1.44–7.94)**	1.34	(0.53–3.39)
No	Ref.		Ref.
One or both partners had ever broken their sexual agreement
Yes	1.31	(0.56–3.08)	2.28	(0.97–5.35)
No	Ref.		Ref.
One or both partners had condomless anal sex outside the relationship in the past 3 monthsa
Yes	0.98	(0.34–2.83)	1.43	(0.51–4.00)
No	Ref.		Ref.
One or both partners used non-prescription drugs in the past 3 months
Yes	1.97	(1.01–3.85)*	2.02	(1.01–4.06)*
No	Ref.		Ref.
One or both partners were heavy alcohol users in the past 3 months
Yes	0.51	(0.19–1.39)	0.43	(0.14–1.32)
No	Ref.		Ref.
Continuous variables
Trust within the relationshipb	0.91	(0.83–0.99)*	1.03	(0.93–1.13)
Differences in trust between partnersc	1.04	(0.95–1.13)	1.01	(0.92–1.11)
Communal coping within the relationshipd	1.06	(1.02–1.10)**	1.04	(1.00–1.09)
Differences in communal coping between partnerse	1.00	(0.97–1.03)	1.01	(0.98–1.04)
Agreement regarding sexual health issues within the relationshipf	1.01	(0.88–1.16)	0.95	(0.83–1.10)
Differences in agreement regarding sexual health issues between partnersg	1.00	(0.89–1.12)	0.97	(0.86–1.09)

^*p < 0.05;

^**p < 0.01;

^***p < 0.001

^aDefined as engaging in receptive or insertive anal sex with ≥ 1 outside partners without using a condom the entire time

^bCalculated by averaging the scores for each partner on the modified 5–point Likert Dyadic Trust Scale (Potential range 8–40). Higher values indicate greater levels of trust within the relationship

^cCalculated by taking the absolute difference between the scores for each partner on the modified 5-point Likert Dyadic Trust Scale (Potential range 0–32). Higher values indicate greater differences in trust between partners

^dCalculated by averaging the scores for each partner on the 5-point Likert Communal Coping to Reduce HIV Threat Scale (Potential range 21–105). Higher values indicate greater levels of communal coping within the relationship

^eCalculated by taking the absolute difference between the scores for each partner on the 5-point Likert Communal Coping to Reduce HIV Threat Scale (Potential range 0–84). Higher values indicate greater differences in communal coping between partners

^fCalculated by averaging the scores for each partner on the 5-point Likert Preferences for Sexual Health Outcomes Scale (Potential range 7–35). Higher values indicate greater levels of agreement within the relationship

^gCalculated by taking the absolute difference between the scores for each partner on the 5-point Likert Preferences for Sexual Health Outcomes Scale (Potential range 0–28). Higher values indicate greater differences in agreement between partners

Regarding perception of one’s partner’s HIV testing history, 137 of 720 participants (19.03%) were incorrect about whether or not their partners had been tested in the past year (50 incorrectly believed their partner had been tested, and 87 incorrectly believed their partner had not been tested). Neither partner in 29 couples (8.06%) was correct; only one partner in 79 couples (21.94%) was correct; both partners in 252 couples (70.00%) were correct.

Table 5 summarizes the demographic, relationship, and behavioral characteristics associated with couple-level perception of partner’s HIV testing behavior in the past year. None of the characteristics were significantly associated with neither partner being correct, except mutual agreement regarding sexual health issues (aOR = 0.76, 95% CI 0.62–0.93). Only one partner was more likely to be correct about whether or not his partner had been tested for HIV in the past year if someone had previously broken their sexual agreement (aOR = 2.60, 95% CI 1.17–5.76). Increasing differences in trust between partners were associated with higher odds of only one partner being correct (aOR = 1.10, 95% CI 1.01–1.20).

Table 5.

Characteristics associated with couple-level perception of partner’s HIV testing history among 280 male couples, United States, April 2016–June 2017

Characteristic	Couple-level perception of partner’s HIV testing behavior in the past year
	Neither partner was correct		Only one partner was correct
	aOR	(95% CI)	aOR	(95% CI)
Categorical variables
Age (years)
> 5 years apart	0.96	(0.33–2.82)	0.97	(0.47–1.99)
Same or within 5 years of each other	Ref.		Ref.
Race/ethnicity
Different	0.76	(0.29–1.97)	0.82	(0.44–1.55)
Same	Ref.		Ref.
Relationship duration
6 months to 3 years	1.44	(0.55–3.75)	1.42	(0.75–2.67)
≥ 3 years	Ref.		Ref.
Had condomless anal sex with their main partner in the past 3 months
Yes	1.24	(0.26–5.99)	0.81	(0.33–1.95)
No	Ref.		Ref.
Sexual agreement allowed for sex with outside partners
Yes	0.72	(0.18–2.86)	0.78	(0.32–1.90)
No	Ref.		Ref.
One or both partners had ever broken their sexual agreement
Yes	1.91	(0.53–6.87)	2.60	(1.17–5.76)*
No	Ref.		Ref.
One or both partners had condomless anal sex outside the relationship in the past 3 monthsa
Yes	0.22	(0.02–1.98)	0.89	(0.32–2.45)
No	Ref.		Ref.
One or both partners used non-prescription drugs in the past 3 months
Yes	0.60	(0.22–1.69)	0.77	(0.39–1.50)
No	Ref.		Ref.
One or both partners were heavy alcohol users in the past 3 months
Yes	1.70	(0.39–7.37)	0.53	(0.16–1.74)
No	Ref.		Ref.
Continuous variables
Trust within the relationshipb	1.12	(0.96–1.30)	1.09	(0.99–1.20)
Differences in trust between partnersc	1.03	(0.89–1.18)	1.10	(1.01–1.20)*
Communal coping within the relationshipd	1.07	(1.00–1.14)	1.03	(0.99–1.08)
Differences in communal coping between partnerse	1.02	(0.97–1.07)	1.03	(1.00–1.06)
Agreement regarding sexual health issues within the relationshipf	0.76	(0.62–0.93)**	0.92	(0.80–1.06)
Differences in agreement regarding sexual health issues between partnersg	1.02	(0.87–1.21)	1.02	(0.91–1.13)

^*p < 0.05;

^**p < 0.01

^aDefined as engaging in receptive or insertive anal sex with ≥ 1 outside partners without using a condom the entire time

^bCalculated by averaging the scores for each partner on the modified 5-point Likert Dyadic Trust Scale (Potential range 8–40). Higher values indicate greater levels of trust within the relationship

^cCalculated by taking the absolute difference between the scores for each partner on the modified 5-point Likert Dyadic Trust Scale (Potential range 0–32). Higher values indicate greater differences in trust between partners

^dCalculated by averaging the scores for each partner on the 5-point Likert Communal Coping to Reduce HIV Threat Scale (Potential range 21–105). Higher values indicate greater levels of communal coping within the relationship

^eCalculated by taking the absolute difference between the scores for each partner on the 5-point Likert Communal Coping to Reduce HIV Threat Scale (Potential range 0–84). Higher values indicate greater differences in communal coping between partners

^fCalculated by averaging the scores for each partner on the 5-point Likert Preferences for Sexual Health Outcomes Scale (Potential range 7–35). Higher values indicate greater levels of agreement within the relationship

^gCalculated by taking the absolute difference between the scores for each partner on the 5-point Likert Preferences for Sexual Health Outcomes Scale (Potential range 0–28). Higher values indicate greater differences in agreement between partners

Discussion

Testing for HIV is the gateway to accessing biomedical prevention options such as PrEP among those who test negative, and initiating antiretroviral treatment among those who test positive. Recent data reveal low HIV risk perceptions [27, 28], and inadequate levels of HIV testing among male couples in the US [17, 18]. Although some published studies have evaluated the influence of relationship characteristics on couple-level HIV testing behavior among MSM [22, 23], we are unable to find any that have focused on perceptions of a main partner’s HIV testing behavior. Analyzing dyadic data from the baseline survey of an RCT involving MSM who were in a relationship for at least 6 months, our study provides new information on how relationship characteristics such as trust, communal coping, and agreement regarding sexual health issues might affect perceptions of a partner’s HIV testing behavior. Below we discuss the significance and implications of our findings for advancing HIV prevention efforts with male couples.

Approximately half the participants in our sample had not been tested for HIV in the past year, and neither partner in more than one-third of couples had been tested. These data are consistent with recent observations that partnered MSM are not getting tested at least annually in accordance with CDC recommendations [10, 17, 18], and highlight the exigency of scaling up HIV testing services for MSM in committed relationships. Almost one-fifth of our participants were incorrect about whether or not their partner had been tested for HIV in the past year, and one or both partners in almost one-third of couples were incorrect. Erroneous beliefs about a main partners’ testing behavior, coupled with incorrect assumptions of HIV-negative serostatus, may increase the risk of HIV transmission within male couples [1, 2]. Regular HIV testing is particularly important in the context of sexual risk behaviors such as CAS with outside partners. Overall, one or both partners in approximately 16% of our couples reported engaging in CAS outside their relationship in the past 3 months. Although this proportion is on the lower end of estimates noted in previous studies with male dyads (16% [17], 22% [10], and 25% [18] for CAS with outside partners in the past 3 months), it is concerning that at least one partner in almost half of these couples reported CAS with ≥ 2 outside male sex partners.

Focusing first on the associations between relationship characteristics and perceptions of partners’ HIV testing behavior, we found some noteworthy results. Previous break-age of a sexual agreement by one or both partners was associated with an incorrect perception of a partner’s HIV testing history. Breaking of sexual agreements can induce stress and conflict in the relationship, and potentially increase both partners’ vulnerability to HIV in situations where their agreements prohibit CAS with outside partners [20]. In our sample, increasing differences in trust between partners were associated with a greater likelihood of only one partner being correct about whether or not his partner had been tested in the past year. Additionally, both partners were more likely to be incorrect about their partners’ HIV testing behavior if they were in relationships with lower levels of mutual agreement regarding sexual health issues. These findings add to the evidence that promoting positive relationship dynamics can improve sexual health-enhancing behaviors among male couples, and reduce their risk for HIV [26, 37, 38]. Successes of couples-based efforts will ultimately be contingent upon constructive interactions between partners, emphasizing the need to prioritize interventions that can bolster mutual support for discussing and adopting preferred HIV prevention approaches.

Two evidence-based HIV prevention interventions for male couples supported by the CDC are worth highlighting at this point: Couples HIV testing and counseling (CHTC), and Connect^HIP. CHTC was originally developed for heterosexual couples in sub-Saharan Africa [39], but an adaptation for MSM in relationships [40] is now being disseminated and implemented across the US [41]. It comprises of a structured 30–60 min session between a male couple and a trained HIV tester/counselor, which includes a pre-test discussion of risk, HIV testing of both partners, skills building around negotiating sexual agreements, return of HIV test results to both partners together, and a post-test discussion based on dyadic HIV serostatus [42]. CHTC for male couples is safe and acceptable [43, 44], and can provide the foundation for a comprehensive HIV prevention, treatment, and care package. Besides enabling couples to reduce misperceptions of their partners’ HIV testing behavior and learn their serostatus together, CHTC can facilitate guided discussions around formulating a prevention plan for the future that takes into account potentially high-risk sexual activities with outside partners. Connect^HIP is a three-session intervention that teaches male couples techniques to enhance the quality of their relationship, communication, and shared commitment to HIV risk-reduction behaviors [45]. Testing for HIV and other STIs is part of the intervention. Similar to CHTC, this is an adaptation of an intervention for heterosexual couples that combines content related to HIV prevention with sexual communication and negotiation skills [46]. Connect^HIP was first adapted for black/African American MSM in long-term same-sex relationships where at least one partner used methamphetamine [47], and more recently for Hispanic MSM and their male partners [48].

Turning to focus on couple-level HIV testing behavior, our study found that being in an interracial/ethnic relationship was associated with an increased likelihood of both partners having been tested for HIV in the past year. More than half the interracial/ethnic couples in our sample represented either non-Hispanic white and non-Hispanic black/African American partnerships, or non-Hispanic white and Hispanic partnerships. Four-fifths of the couples in our reference group (partners of the same race/ethnicity) were non-Hispanic white. Some studies show that MSM in the US may have an increased HIV risk perception about having sex with black/African American MSM and Hispanic MSM, as compared to having sex with white or Asian MSM [49, 50]. Previous research has also documented the belief among some black/African American MSM that conversations pertaining to safer sex may lead their partners to mistakenly assume that they are HIV-positive [51]. The possibility that a heightened risk perception might positively influence interracial/ethnic couples’ decisions to test for HIV on a regular basis cannot be ruled out. Future qualitative work comparing motivations and barriers to testing between male partnerships of different and same races/ethnicities could shed light on this issue. Another explanation for our finding could be the significantly higher engagement in preventive behavior, including repeat HIV testing, demonstrated by racial/ethnic minority MSM [52, 53].

In our sample, both partners were more likely to have been tested for HIV in the past year if their relationship was shorter than 3 years, or if their sexual agreement allowed for sex with outside partners. These results are not surprising, and parallel those in the published literature. In-depth interviews with male couples indicate that testing for HIV during the formative stages of a relationship is not uncommon, and this act can help build trust in addition to signifying commitment to one’s partner [19, 30]. Two studies have found that couples in which at least one partner had casual sex outside the relationship were more likely to report getting regularly tested for HIV [22, 23]. In another sample of 314 MSM who were partnered for at least 1 year, those in “open” relationships were more likely to report having been tested in the past 6 months compared to those in “closed” relationships [28]. Given the wide range of sexual activities that might be undertaken with outside partners [54, 55], and the fluid nature of sexual agreements [56], having periodic discussions to identify and adopt mutually acceptable HIV prevention strategies could prove beneficial, regardless of relationship duration.

Similar to recent studies with male couples [57, 58], a substantial proportion of our participants had used one or more non-prescription drugs in the past 3 months. Drug use is of particular relevance to the sexual health of MSM because of its well-established link with sexual risk taking and HIV transmission [59–62]. Our study found that non-prescription drug use by one or both partners was associated with a higher likelihood of having been tested for HIV in the past year. This is an encouraging observation, and possibly reflective of concerted efforts at the national, state and local levels to expand HIV testing to groups at highest risk [16]. Partnered MSM may also regulate each other’s drug use patterns in a manner similar to how they influence other behaviors specific to HIV prevention [63], a hypothesis that has prompted exploratory research on integrating a discussion about the rules or understandings between partners pertaining to drug use into the existing CHTC protocol for male couples [64].

Trust and communal coping were independently associated with HIV testing behavior in our analyses. Both partners were less likely to have been tested for HIV in the past year if they were in relationships with higher levels of mutual trust. Prior research suggests that feelings of trust and commitment may inadvertently suppress HIV testing, especially as the relationship progresses, and that bringing up the need for an HIV test might be viewed as an indication of a broken sexual agreement [19, 30]. In contrast, both partners were more likely to have been tested if they were in relationships with higher levels of communal coping (involving communication, joint decision-making, and working together to reduce a health threat). This is an expected result based on interdependence theory, which posits that behaviors within relationships are interdependent because each partner has a certain amount of control and influence on whether the behavior in question is performed [25].

Collectively, our results underscore the need for designing and evaluating novel relationship-based HIV prevention interventions for partnered MSM. Emphasis needs to be placed on skills building around enhancing mutual communication, jointly formulating sexual agreements, and applying problem-solving techniques to relationship issues such as breaking of agreements and lowering of trust. Dyadic interventions that seek to encourage positive interactions about sexual health issues, including discussions of regular HIV testing as a risk-reduction strategy, have the potential to safeguard partnered MSM against HIV. Ramping up evidence-based prevention strategies that utilize the interplay of couples’ relationship dynamics, such as CHTC for male couples [40, 42] and Connect^HIP [47, 48], remains a priority. The CDC has developed a CHTC training plan for both community-based and clinic-based healthcare providers, and is engaging health departments to support practitioners who are interested in adopting this approach [11]. Finally, longitudinal research guided by relationship-based theories to examine how partners influence each other’s behavior in an effort to reduce their risk for HIV is warranted.

Our study makes an important contribution to the literature, but we recognize its limitations. Using cross-sectional data from a convenience sample precludes us from making causal inferences or generalizing our results to other male couples in the US, who may or may not use social media platforms and geospatial mobile apps. Because we collected personal identifying information, social desirability could have prompted our participants to inaccurately report data on their sexual risk or HIV testing behaviors. Although we requested participants to complete their own survey separately from their partners, it is possible that couples could have discussed certain questions, potentially influencing each other’s responses. We did not include a survey question asking whether partners shared their responses with each other. To minimize the threat to the validity of our results, particularly those pertaining to a participant’s belief regarding whether or not his partner had been tested for HIV in the past year, we excluded data from couples who completed the baseline survey within 15 min of each other. Our survey did not collect detailed information on participants’ outside male sex partners such as whether or not they were on PrEP (if HIV-negative) or virally suppressed (if HIV-positive). We acknowledge these data would have captured risk levels more accurately than just engaging in receptive or insertive CAS. Our survey also did not collect information on outside female sex partners. However, we did not expect many participants to report this behavior based on our eligibility criterion of a cis-gender male being in a sexual relationship with another cis-gender male for at least 6 months (less than 7% of our sample identified as bisexual). Heavy alcohol use was defined as consuming ≥ 6 alcoholic drinks on the same occasion weekly, almost daily, or daily, which might not perfectly correspond to the Substance Abuse and Mental Health Services Administration (SAMHSA) definition, i.e. binge drinking on ≥ 5 days in the past month [65]. Future research that examines testing behaviors among male couples should also assess their history of testing for other STIs.

Despite these limitations, we believe that our analyses further our understanding of how relationship characteristics such as trust, communal coping to reduce HIV risk, and agreement regarding sexual health issues might affect male couples’ HIV testing behavior. Our study also provides new information on how these factors potentially influence perceptions of partners’ HIV testing behavior. Given that partnered MSM are not immune to the threat of HIV, focusing on dyadic prevention efforts that include comprehensive skills building around effective communication and dyadic support for engaging in health-promoting behaviors is important.