How alcohol influences mechanisms of sexual risk behavior change: contributions of alcohol challenge research to the development of HIV prevention interventions

Abstract

This paper examines the contributions of laboratory-based alcohol challenge research (ACR) to the development of HIV prevention interventions. Following a brief overview of HIV prevention interventions and related health behavior change models, we discuss how alcohol may influence mechanisms of behavior change. The paper highlights the value of ACR for: (1) elucidating mechanisms of action through which alcohol affects sexual risk behavior, (2) testing how alcohol may influence mechanisms thought to underlie HIV prevention interventions, (3) clarifying moderators of the causal influences of alcohol, (4) identifying novel intervention targets, and (5) developing strategies to reduce sexual risk among those who consume alcohol. We conclude with a discussion of the importance of using experimental research to identify mechanisms of behavior change that are specific to populations at high risk for HIV and outline some key implications for developing HIV prevention interventions that integrate the role of alcohol.

Introduction

Despite prevention efforts over the past two decades, HIV transmission remains a significant public health concern, especially among minority and men who have sex with men (MSM) populations [1]. As the CDC reports, approximately 65% of new HIV diagnoses in the United States are due to male-to-male sexual contact [2] and incidence rates remain higher in racial and ethnic minority populations (i.e., Hispanic/Latino, African American men) than in White and Asian populations [2]. Condomless intercourse continues to be the major route of transmission, not only for MSM but across populations globally [3,4]. With the development of PrEP there are now pharmacological approaches to prevent HIV that show a high degree of efficacy [5–7]. Unfortunately, PrEP uptake has been slow, especially in populations that are at highest risk for transmission [8]. Thus, to reduce the incidence of HIV, it is critical to identify modifiable risk factors that underlie HIV acquisition and transmission and then develop behavioral HIV prevention strategies accordingly. It is increasingly clear that alcohol use, particularly heavy episodic drinking, represents a central modifiable risk factor that may be relevant to both condomless intercourse [4,9,10] and medication adherence [11,12].

There is growing recognition of the role of heavy drinking in HIV risk behavior [4,13,14]. Despite some exceptions, studies that have examined the association between alcohol use and frequency of unprotected sex support the view that heavy episodic drinking is related to a higher likelihood of condomless intercourse, especially among high-risk populations such as MSM [15]. Event-level studies using timeline follow-back [16], daily diary, and now ecological momentary assessment (EMA) measures have been able to clarify how alcohol intoxication and sexual risk are associated at both the individual and event level [17]. Results from these multilevel modeling studies have pointed to the importance of various individual difference and situational factors that moderate the association between alcohol use and condomless intercourse [18].

However, even with growing evidence of the impact of alcohol consumption on HIV transmission, HIV prevention models and derived interventions have tended to place little emphasis on alcohol. In a meta-analysis of 67 interventions designed to increase condom use from 42 studies, Scott-Sheldon et al. [19] noted that only 10 studies collected data on alcohol use among their participants, and no intervention was identified as specifically focused on reducing consumption. In a more recent meta-analysis of 23 interventions designed to reduce sexual risk behavior in women using alcohol or other substances [20], only three interventions were noted to have specifically targeted alcohol use and collected data on alcohol use outcomes. These meta-analyses clearly illustrate that consideration of alcohol consumption has rarely been central to HIV prevention interventions, despite evidence that drinking, particularly heavy episodic drinking (the consumption of four or more drinks during a single occasion for women and five or more for men) directly increases the risk of condomless intercourse [4].

More recently, investigators have begun to develop intervention approaches that concomitantly address alcohol and risky sexual behavior [21–30]. These studies have targeted diverse populations (i.e., MSM, undergraduate students, emergency room patients, substance-using female sex workers, etc.) commonly through the use of an additive approach which combines existing alcohol and sexual risk reduction strategies: for example, an education and skills-building approach for reducing sexual risk behaviors and a motivational interviewing (MI) [31] approach for reducing alcohol consumption [31,32]. Interventions that include MI-based strategies for alcohol have been shown to be more effective at reducing drinking and risky sexual behavior compared to control interventions, like brief advice [22,27], standard post-HIV test counseling [30], or no intervention [23]. Other interventions have targeted alcohol use with personalized feedback [24,26] and provided education about the association between drinking and risky sexual activity [28].

These recent efforts to include alcohol as a target in HIV prevention interventions have been encouraging. Without diminishing the clear benefit of efforts to reduce sexual risk-taking by using strategies to reduce drinking, we suggest that a more comprehensive approach to address alcohol in HIV prevention interventions is warranted: one that considers how drinking may influence sexual risk-reduction efforts. HIV prevention interventions may be optimized if they are informed by an empirical understanding of how alcohol influences (1) sexual decision-making processes, (2) mechanisms of behavior change (MOBC) related to condom use and other HIV prevention behaviors, and (3) the capacity to use HIV prevention intervention strategies and skills in risky contexts.

Behavioral HIV prevention interventions generally have been developed from models of health behavior change (e.g., the Information-Motivation-Behavioral skills model (IMB skills model) [33] or the Theory of Planned Behavior (TPB) [34,35]) which hypothesize that various sets of psychological processes underlie risk and health protective behavior. Interventions designed to reduce sexual risk do so by targeting theoretically-based MOBC that are specific to sexual risk behavior. An intervention based on the IMB skills model [33], for instance, might include intervention components to increase knowledge about HIV risk, foster more positive attitudes about condoms, and increase condom use skills. Though health behavior change theories have provided an effective basis for intervention design, they do not account for the fact that MOBC that are hypothesized to reduce sexual risk behaviors are not static but change in the presence of alcohol, sometimes markedly. As will be described below, alcohol use, and particularly the consumption of large amounts of alcohol within a single drinking episode, exerts significant and often maladaptive influences on both the psychological processes that underlie sexual decision-making and the capacity to implement self-regulatory strategies to reduce sexual risk. Alcohol challenge research (ACR) has been instrumental in elucidating the nature of these alcohol-induced changes in MOBC and thus offers a particularly valuable approach for developing HIV prevention interventions for those who use alcohol.

Consider what HIV prevention approach might be most useful for an individual who may continue to drink or relapse to heavy episodic drinking. What psychological processes may be most important for reducing sexual risk behavior? Should he or she use the same coping strategies to minimize sexual risk as those who do not engage in heavy episodic drinking? Are there certain types of safer sex skills that may be less susceptible to the deleterious effects of drinking? The fact that MOBC operate differently during periods of sobriety and intoxication has implications for deciding which intervention strategies to use, which MOBC to target, and how interventions should be tailored [36].

Although there are a variety of methods that may contribute to this effort, ACR provides a uniquely valuable approach to address these questions. ACR refers to experimental studies in which the effects of alcohol are studied through the administration of either alcohol, placebo, or control beverage (e.g., water) to participants prior to completing a task or set of tasks that model a behavior of interest (e.g., sexual decision-making). Comparison of responses between these three conditions allows investigators to examine the pharmacological and expectancy effects of drinking on sexual decision-making and MOBC derived from models of HIV-related risk behavior. ACR has provided several methodological and conceptual contributions to understanding how alcohol influences both HIV risk behavior and efforts to modify it. The purpose of the current paper is to review the contributions of ACR to our understanding of how alcohol may influence MOBC related to sexual risk reduction and highlight ways that ACR can contribute to the development of more effective HIV prevention approaches that include a focus on alcohol.

HIV Prevention Interventions

Based on research conducted over the past several decades, there is good evidence that HIV prevention interventions are effective at reducing sexual risk behavior. Large scale reviews of meta-analyses have found significant intervention effects for a number of sexual risk outcomes, including condom use, number of partners, and the incidence of STIs [37,38]. Efficacy of interventions targeting condomless sex has been documented across different demographic groups, including heterosexual men and women [39,40], adolescents [41,42], African Americans [43,44], Latinx individuals [45,46], substance users (including injection drug users) [47], MSM [48–50], and transgender women [51]. Despite indications that many HIV prevention interventions show evidence of efficacy, a closer examination of the evidence suggests that challenges remain. Effect sizes for interventions, for example, are often modest [37], and interventions may not demonstrate efficacy for extended periods of time following completion [48,50]. Moreover, studies on intervention efficacy for specific subpopulations that are both at risk for HIV transmission and engage in heavy episodic drinking are relatively infrequent in the literature and have reported equivocal findings [1,52–54].

Models of HIV prevention

Contemporary psychosocial approaches to HIV prevention have largely focused on the development of behavioral interventions based on one or more theories of health behavior change [37,48,55,56], typically the Health Belief Model (HBM) [57], the IMB skills model [9,33], social-cognitive theory (SCT) [58,59], TPB [34,35], and/or the Transtheoretical Model [60]. Although these theories provide distinct descriptions of how behavior change occurs, they share the idea that health risk behavior (e.g., condomless sex with a partner of unknown HIV status) is a function of similar modifiable psychological variables. Many also identify similar psychological factors as important for change [61,62], including knowledge related to risk, perceptions of the relevance of risk, attitudes, self-efficacy, behavioral skills related to communication and prevention behaviors, and intentions [9,33,55,56]. Behavioral HIV prevention interventions that are based on these theories include specific strategies designed to target those psychological variables that then lead to behavior change such as increased use of condoms. Interventions commonly provide some form of education and skill-building components [30,63–66], but the perceived role of these components may differ as do their interaction with other variables. For example, interventions based on the IMB skills model focus on three broad areas: (1) providing information about HIV infection and transmission, (2) improving motivation to act in ways that protect health by reducing transmission risk, and (3) teaching behavioral skills that have been shown to reduce HIV infection, such as how to properly use condoms and how to discuss condom use with sexual partners [65–70]. Interventions based on TPB conceptualize behavioral intentions (e.g., using a condom) as a consequence of other psychological variables such as perceived behavioral control and condom attitudes, and therefore view intentions as the strongest predictor of behavior [34,35]. In both models, the likelihood of condom use during a sexual situation is hypothesized to be a function of the strength of these targeted MOBC.

Mechanisms of behavior change for HIV prevention interventions

Although research has shown considerable support for their efficacy [37,38] there is less known about whether HIV risk reduction interventions influence behavior by changing psychological processes as hypothesized. In efficacy studies [71], questions of whether theory-based interventions actually change the psychological targets that are specified and whether change in those targets lead to change in behavior often remain unanswered.

The ability to delineate MOBC provides several advantages for developing more efficacious and efficient interventions [72]. Through the measurement of MOBC, investigators can identify whether interventions are indeed influencing the targets of the intervention through processes identified by health behavior change models. This provides insight into the specificity of the intervention and the components of the intervention that are necessary and sufficient to effect change [73]. For example, although a given multicomponent intervention may promote change in sexual risk behavior, its effectiveness may have little to do with changes in HIV knowledge but may rather be a function of increases in self-efficacy from learning skills. Evidence about mechanisms may lead to the refinement of interventions to reduce unnecessary components and bolster elements that serve as mediators of change.

The study of MOBC also helps clarify the dose or gradient of intervention components that are necessary to promote change [74]. Some change processes may be modified with relatively little information (e.g., feedback about drinking norms to influence personal standards for alcohol use), whereas other mechanisms may require multiple presentations and practice (e.g., skills taught during cognitive-behavioral therapy (CBT)). By assessing hypothesized mechanisms of change, the investigator can determine whether the intervention strategy indeed changed the hypothesized process and whether such change occurred in accord with theory.

Finally, better understanding of MOBC provides insight into the individual difference and contextual factors that moderate the efficacy of interventions [73,74]. For example, the demonstration that a hypothesized change in an outcome following a social learning-based intervention for sexual risk is mediated by increases in self-efficacy suggests that those who are lower in self-efficacy at baseline may show particular benefit from the intervention. It also helps to illuminate ways that interventions might be tailored to individual difference factors through the examination of moderated mediation [75], the idea that MOBC may operate differently at different levels of a given individual difference moderator. This provides a greater understanding about what components of an intervention may be most important to emphasize for individual differences in subsequent refinements of the intervention.

MOBC may be studied using a variety of methods, most commonly through the assessment of psychological processes before and after an intervention [73]. In the context of randomized controlled trials, this usually entails assessment of the hypothesized processes derived from the theory that underlie the intervention. Through mediational analyses, one can identify whether there is an association between the intervention and the mediator and whether change in the mediator is linked with change in behavior. To provide stronger evidence of mediation, it is important to demonstrate the temporal precedence of the mediator to outcomes measured and that greater change in the mediator is linked with greater behavioral change. Information about mediators may also be ascertained through dismantling studies in which certain intervention components hypothesized to influence specific mechanisms of behavior change are subtracted from the intervention package to examine the impact of their absence on behavioral outcomes. Comparing outcomes of interventions with and without specific components allows for inferences to be made about how the intervention produces behavioral change.

Reviews of the literature have provided evidence that several hypothesized MOBC are influenced by HIV prevention interventions. Albarracín and colleagues [63], for example, examined potential mediators of intervention effects for 354 intervention conditions administered during the 17 years prior to publication. They found evidence that changes in attitudes, norms, perceptions of the threats of HIV infection, knowledge about HIV, perceptions of control (i.e., self-efficacy), and behavioral skills mediated the relationship between intervention components and changes in condom use. Mediational analyses in more recent HIV prevention studies have also supported the role of self-efficacy, HIV knowledge, motivation to use condoms, and behavioral sex-risk reduction skills as MOBC [76–80]. More research on whether these variables operate as hypothesized in health behavior change theories, however, is needed: while broad conclusions about MOBC can be drawn from existing research, most HIV prevention intervention studies neither collect data relevant to MOBC [71] nor feature mediational analyses that would allow for the estimation of effect sizes. There is also considerable heterogeneity between studies in terms of the MOBC selected for investigation, the type and quantity of data that are collected, and how MOBC identified in meta-analyses change following an intervention. Thus far, the strongest evidence of mediational processes in HIV prevention interventions has been largely derived from questionnaire assessments of psychological processes taken either concurrently with outcome assessment or in lagged assessment at multiple outcome points [63,76–80]. Even though methods of modeling mediation have become increasingly precise, the data that are used are typically gathered from self-report assessments taken in neutral (e.g., laboratory, clinic) settings at time points that are far apart (e.g., baseline, three months, and six months) from one another and from the intervention itself [78–80].

This measurement strategy represents a limitation to understanding MOBC in typical randomized clinical trials as MOBC may operate at various intervals between the intervention and the target risk behavior. In general, the question of how and when specific MOBC may influence a specific sexual decision-making behavior has not been well-specified by the general models of health behavior change that underlie HIV prevention interventions [81]. MOBC that may ultimately influence whether someone engages in sexual risk behavior may occur immediately following an HIV prevention intervention (e.g., knowledge that behavior is risky), in advance of the risk context (e.g., planning to have condoms available), or during the moment that presents risk itself (e.g., activation of coping skills). One challenge for investigators is how to represent MOBC that unfold “in the heat of the moment.” It is unclear how data from self-report questionnaires taken following interventions (e.g., self-efficacy to use condoms) correspond to the processes that occur within a high-risk context. This is not only a question of temporal proximity but also the nature of how MOBC are measured. Identification of intervention mediators depends on valid measures of hypothesized constructs [72] which may be difficult to obtain through self-report alone. For example, self-reported ratings of coping skills do not appear to be significant mediators of CBT treatment for addictions, whereas assessment of skills using in-vivo measurement approaches has supported the role of skills as a mediator [75,82]. Not only is it important to develop alternative methods to accurately assess MOBC, but the salience and impact of MOBC may be significantly changed by the risk contexts themselves [36].

Alcohol alters mechanisms of behavior change

As we describe below, several studies have now shown that alcohol influences risky sexual behavior by changing the way that individuals process information related to sexual risk and condom use. What these studies suggest is that the cognitive and motivational processes that form the basis of HIV prevention intervention targets and strategies may be altered by alcohol. This has implications for how and when interventions may influence sexual risk behavior. To create HIV prevention interventions that fully account for the role of alcohol requires an understanding of how alcohol influences processes that underlie sexual risk and how they might influence MOBC hypothesized to impact interventions effects.

There are a number of methodological approaches that may contribute in this regard. Experience sampling methods are now being used with increasing regularity to assess how sexual risk processes change in the presence of alcohol as assessed in participants’ day-to-day lives [83]. As we outline below, however, experimental approaches, typified by ACR, provide unique contributions to our understanding of how to integrate alcohol into models of HIV risk behavior change and how to improve HIV prevention interventions approaches that include alcohol.

Contributions of ACR for HIV Prevention Interventions

The value of experimental approaches for the development of behavioral health interventions has been well-described [71,84]. Among their many advantages, experiments allow one to manipulate and/or measure hypothesized MOBC under different conditions and then test whether they result in the hypothesized behavior change or a change in a proxy to behavior such as intentions or likelihood to engage in a behavior (e.g., intentions to engage in condomless sex). They also allow investigators to test the impact of various strategies to change MOBC. ACR can be used to test how, when, and for whom MOBC related to sexual decision-making may be affected by alcohol consumption “in the moment” and examine how the efficacy of intervention strategies may be influenced by intoxication [85]. ACR has contributed to our understanding of HIV prevention interventions in a number of ways. Among these contributions, ACR has (1) provided methods to precisely test the causal influences of alcohol on sexual decision-making; (2) identified significant person-level and situational moderators of alcohol effects; (3) contributed to the development of alternative measures of MOBC hypothesized to underlie HIV prevention interventions; (4) tested the causal influences of alcohol on these MOBC, (5) clarified potentially important alternative intervention targets for HIV interventions; and (6) provided a framework to test how intervention strategies may be influenced by intoxication.

Specifying the causal influences of alcohol on sexual decision-making

The main contribution of ACR has been its ability to test the causal influence of alcohol on sexual decision-making. ACR enables investigators to add precision to our understanding of the causal effects of alcohol on sexual risk in terms of both the nature of alcohol effects (e.g., at what point after how much consumption is alcohol most likely to influence sexual risk decision-making processes) and how MOBC are affected. The control afforded by the experimental approach allows one to accurately identify the amount of alcohol consumed and the resulting blood alcohol level (BAL) of a given individual [10]. Alcohol effects on cognitive-motivational processes are a function of both BAL and whether this level occurs while blood alcohol concentration is rising (the ascending limb of the intoxication curve) or falling (the descending limb). The blood alcohol curve is associated with different features of intoxication, including the subjective experiences of stimulation and sedation that occur at various time-points following ingestion. Parametric studies of alcohol doses on sexual decision-making processes [86–88] have allowed for greater insight into how different levels of alcohol consumption may influence psychological mechanisms that underlie sexual decision-making and have helped identify when individuals may be most susceptible to risky sexual behavior both during and following a drinking episode. Through ACR, investigators have been able to model how psychological processes influence sexual decision-making at different doses and at different phases of intoxication [89].

Given that risky sexual behavior cannot be assessed directly in the laboratory, researchers have generally relied on proxy measures of sexual risk behavior such as self-reported intentions to have sex without a condom and likelihood ratings of engaging in condomless sex in the context of experimentally-modeled sexual situations [90–92]. These situations have been modeled through interactions with a confederate [87], written scenarios [90,93,94], video-based sexual scenes [91,95,96] and images of potential partners [97,98]. Across a number of experimental paradigms, ACR studies have shown that consumption of alcohol increases intentions to engage in condomless sex [14,99,100]. In particular, there is evidence across a variety of analogue sexual decision-making tasks that heavy alcohol consumption (achieving a BAL of 0.075% or greater, equivalent to approximately 5 standard drinks in 2 hours for a 170 lb. adult male) leads to stronger intentions to engage in condomless sex [4,10,14,86].

In addition, ACR has shown that subjective sexual arousal (but not physiological measures of sexual arousal, like penile plethysmography or vaginal pulse amplitude [10,101]) can act as an important mediator of alcohol effects on intentions to engage in condomless intercourse. Intoxicating doses of alcohol have been shown to increase sexual arousal across several experimental paradigms. These studies have largely shown that alcohol-induced subjective sexual arousal increases intentions to engage in condomless sex among men and women [10,88,89,91,102–104], though the indirect effect of alcohol on condomless sex intentions via sexual arousal can be moderated by individual difference factors [91,104,105].

Identifying moderators of alcohol effects on sexual decision-making.

Contextual moderators.

In addition to being a mechanism through which alcohol increases sexual risk-taking, arousal also serves as a contextual factor that increases the likelihood that alcohol consumption will lead to risky sexual decision-making [10,14]. Alcohol consumption is most likely to impair sexual decision-making when drinking occurs while individuals are in a state of high sexual arousal. In the context of ACR studies, sexual arousal has been manipulated by random assignment of participants to conditions in which they are exposed to either erotic or non-erotic materials. A number of studies have shown that intoxicated participants are more likely to report intentions to engage in condomless sex under conditions of high arousal that is manipulated experimentally through arousing stimuli [88,102,103,106–110] or as an individual difference [4,91,111,112].

ACR paradigms have also identified how the causal effects of alcohol on sexual decision-making processes can be moderated by partner and relationship factors. Research on partner and relationship factors has been more commonly conducted with heterosexual female samples than with other populations. Broadly, findings suggest that appraisals of the sexual or relationship potential of a partner can override concerns about partner risk [113–115], and that an absence of information about sexual risk is often assumed to represent an absence of actual risk, so that condomless intercourse becomes more likely in situations where women know relatively little about partner risk factors [115]. The cognitive mediation model (CMM) of sexual decision-making [116] proposes that women “enter a potentially sexual situation with goals that provide the basis for evaluating the situation” [113] and that appraisals of various factors, like whether a woman wants to have a relationship with a particular partner, influences her decision-making. Experimental evaluation of the CMM [87,89,94,115,117] has provided support for proposed mechanisms of alcohol-involved sexual risk, including sexual precedence (i.e., whether a woman has had sex with a partner previously) [113], relationship motivation, and partner familiarity [94], such that more positive appraisals of a potential romantic relationship with a prospective partner lead to greater intentions to engage in condomless sex. This work highlights the fact that several partner factors may interact with intoxication to determine whether condoms are likely to be used in a given sexual situation. These are among a variety of contextual factors that have been shown to influence whether alcohol impacts sexual decision-making.

Individual difference moderators.

The association between alcohol use and sexual risk behaviors is moderated by a variety of person-level factors, including gender, sexual orientation, and personality characteristics [118–120]. ACR may be used to clarify whether the causal effects of alcohol interact with an individual difference variable, providing evidence to address questions that are not answered in cross-sectional studies. For example, is heavy episodic drinking more likely to lead to condomless sex for those who are high on the trait of sensation seeking (a moderator effect) or are those who are high sensation seekers simply more likely to engage in more frequent heavy episodic drinking and condomless sex? Studies of individual difference variables may also provide evidence related to theories of behavior change. For example, if condom use self-efficacy is hypothesized to influence condom use, measures of condom use self-efficacy assessed both prior to and after beverage administration can provide information about how that MOBC is associated with behavior while individuals are under the influence of alcohol.

One of the most frequently studied individual difference factors in ACR has been sexual alcohol expectancies. Sexual alcohol expectancies are cognitive schemas that represent the expected results of drinking alcohol on sexual behavior, such as increases in sexual arousal, sexual aggression, and sexual disinhibition [121,122]. Stronger expectancies are associated with a greater likelihood that those outcomes will occur when consuming alcohol [123] and stronger expectations of alcohol’s sexual disinhibitory effects predict a greater likelihood of engaging in risky sexual behavior among young adults [124]. Effects of sexual alcohol expectancies are moderated by gender, such that expectancies are related to greater sexual arousal as well as reduced estimates of risk following alcohol administration among women [105,125], while among men greater alcohol-aggression expectancies are predictive of intentions to engage in sexual assault and other sexually-aggressive actions [121]. In studies of MSM, sexual expectancies have been shown to have a main effect on decisions to engage in condomless sex [4] as well as interaction effects with alcohol intoxication [91] to increase the likelihood of sexual risk-taking.

Another individual difference factor that has been identified as a key moderator of alcohol effects is having a history of sexual abuse or trauma. A large body of research suggests that responses to sexual situations following intoxication differs between women with and without a history of sexual victimization. ACR studies have shown that women with a sexual abuse history have an increased likelihood of engaging in unprotected oral and vaginal intercourse [126]. This appears to occur through a variety of alcohol-involved sexual decision-making processes, including greater self-reported sexual arousal [127], greater tendency to abdicate condom-use decisions to male partners (i.e., letting the man decide whether to use a condom [128,129]), and diminished salience of the negative health consequences that can stem from condomless intercourse following intoxication [130].

Interestingly, there have been some individual difference factors that have been identified in the epidemiological literature as moderators of alcohol effects that have received equivocal support in the ACR literature. Sexual sensation seeking [131,132] and impulsivity have been identified in the literature as moderators of the association between alcohol or substance use with increased intentions to engage in condomless sex [119,132–134]. ACR studies, however, have found that sexual sensation seeking does not interact with the causal influences of alcohol on risky sexual behavior and instead shows direct effects on sexual risk behavior. Both Maisto et al. [135] and Norris et al. [89] found no interaction between alcohol use and sexual sensation seeking in laboratory research with heterosexual women, while Shuper et al. [4] reported the same results in a sample of MSM.

Through assessment of individual differences, ACR has been able to provide insight into who may be most susceptible to alcohol effects on sexual decision-making and the mechanisms through which this heightened risk may be conferred. This is important for both decisions about who may benefit most from HIV prevention approaches that include alcohol and how to tailor such approaches.

Advancing the assessment of MOBC

Another important contribution of ACR has been the development of methods to measure MOBC. First, experimental methods have provided a means to model “in-the-moment sexual decision-making” [90,91] that is sensitive to the contexts in which such decisions occur. This allows for more accurate prediction of behavioral outcomes that take place in those contexts [36]. Second, these experimental studies have helped develop measures of the cognitive processes that are hypothesized to occur during sexual decision-making rather than the resulting “cognitive products” reflected in self-report ratings [136,137]. This distinction has been critical for testing theoretical mechanisms where process measures of skill use demonstrate higher predictive validity than self-reported skill use [36,138,139]. ACR studies have utilized measures of cognitive processes such as skill activation [140], accessibility of benefits and consequences [90,124], and cognitive appraisals [89,93], all of which permit more precise tests of theoretical models regarding how sexual risk decisions are made and how they may be influenced by alcohol consumption. Maisto and colleagues [91,135] for example, developed methods to assess condom negotiation skills through video-based role play tasks. Through a series of video scenarios depicting sexual encounters, participants are presented with several choice points where they are asked to respond to requests to engage in condomless sex from the other partner to demonstrate their capacity to use condom negotiation skills. These responses are then coded to determine skill proficiency.

The experimental methods employed in ACR have led to several contributions to assessment related to HIV prevention. These measures allow for precise testing of the effects of alcohol on sexual risk processes and provide assessment instruments that may have utility as part of the intervention process. As a clinical tool, these measures may be used to assess individual differences in risk or to identify the specific contexts of risk for individuals. This information may be used to determine the need for intervention strategies and to identify potential moderators and mediators of change.

Testing the effects of alcohol on hypothesized MOBC in models of HIV prevention.

In addition to the well-studied impact on intentions, ACR has also shown how intoxication may influence other MOBC that are central to models that underlie HIV prevention interventions.

Risk appraisal.

Perception of risk plays a central role in many health behavior change models. Whether this is risk of an outcome, a perceived vulnerability to the outcome, or negative consequences associated with behavior, a number of behavior change models identify these risk appraisals as important intervention targets [e.g.,33,57,58,60].

ACR studies have shown that intoxication influences a range of risk-related cognitions. For example, consumption of alcohol leads to reduced expectations of negative sexual outcomes [130], lowered estimates of the likelihood of negative consequences [124,141], and increased perception of the benefits associated with sex, all of which likely increase propensity to engage in risky sexual behaviors [124,142]. These effects differ based on the time and amount of drinking. Kruse and Fromme [98], for example, found that men were more likely to appraise sexual risk as high and use mitigation strategies when in the ascending limb of the alcohol absorption curve than when in the descending limb, which resulted in ratings of greater likelihood to engage in condomless sex during the descending portion of the curve.

ACR has shown that not only does alcohol consumption change the nature and magnitude of perceived risk, but it also changes the way that individuals process information related to risk. For example, Norris et al. [89] and Purdie et al. [115] both found that intoxicated participants (in contrast to sober participants) endorsed high levels of sexual potential for high-risk male partners, which in turn predicted a greater likelihood of engaging in condomless sex with those partners despite knowledge of their heightened risk. In those studies, descriptions of potential partners included information indicating higher and lower levels of sexual risk. Other research has shown that intoxicated women are equally likely to have condomless sex with high-risk and low-risk partners [143] and may choose to engage in condomless intercourse if they perceive condomless intercourse as likely to facilitate the development of a relationship [113]. Such changes in risk appraisal following alcohol use may be particularly relevant to the risk of women with a history of sexual abuse [130], who are more likely to engage in condomless intercourse following intoxication than women without such a history [104,114,130].

Consumption of amounts of alcohol that raise BAL to over .075% changes the way that individuals attend to, value, and utilize information in their environments when making sexual decisions. A number of studies have shown that alcohol consumption induces “alcohol myopia” [144], a processing style in which impelling cues about certain behaviors become more salient and inhibiting cues become diminished in their impact. When intoxicated, cues that support risky sexual behavior (e.g., sexual stimuli) are much more salient than those that support condom use. This has now been documented across a number of studies [106,107,142] with findings suggesting that it is the relative strength of impelling and inhibiting cues that is most important in decision-making when intoxicated [145]. The effect of alcohol myopia is associated with a general tendency to focus more on immediate salient rewards (e.g. engaging in sex) rather than long-term goals (e.g., safety, health) when under the influence of alcohol [10].

ACR research thus suggests that alcohol can reduce evaluations of sexual risk, reduce the accessibility of information related to sexual risk, and make it less likely that individuals will attend to sexual risk cues, while at the same time increasing the salience of factors that contribute to condomless sex. To the extent that risk perception, vulnerability to risk, and pros and cons of condoms use are MOBC that operate “in the heat of the moment”, ACR has demonstrated that alcohol can directly interfere with their functioning [106,124,130,142,144,146–149]. The implications of findings that alcohol increases the relative value of information that supports risky sexual behavior and decreases the value of information that may lead to safer sexual behavior has not been integrated into models of behavior change or the interventions derived therefrom. Interventions for HIV prevention should be adapted to account for these findings that risk perception operates differently during periods of intoxication and periods of sobriety.

Attitudes and motivation.

Attitudes toward condoms and condomless sex play a central role in models of health behavior change, including TPB, HBM, and the IMB skills model. For TPB, more favorable attitudes toward condoms contributes to the development of stronger intentions to use condoms [150]. In the IMB skills model, attitudes form a basis for motivation to use condoms and are seen as critical influences on decisions to use condom-promoting behavioral skills [151]. Consistent with the work on risk described above, evidence suggests that acute intoxication can change the influence of attitudes toward condom use during sex. Conner et al. [150] found that alcohol weakened the predictive value of attitudes on intentions to engage in condomless sex among women. Similarly, MacDonald et al. [142] found that even if intoxicated participants expressed negative attitudes toward condomless sex, they were more likely to ignore those attitudes in order to express intentions to engage in condomless sex than those in placebo and control conditions. The influence of alcohol on attitudes interacts with other risk factors as shown in a study by Ebel-Lam and colleagues [102] where sexual arousal increased positive attitudes toward condomless sex for those who were intoxicated but not for those in placebo or control beverage conditions. Direct tests of the impact of alcohol on variables specified by the IMB skills model [151] have shown that intoxication reduces positive attitudes toward condoms without changing social norms or HIV-related knowledge. This was replicated in a study by Gordon et al. [148], who found that greater sexual alcohol expectancies were related to more negative attitudes toward condom use, particularly when intoxicated. Similar to the effects of alcohol on risk appraisal that were discussed above, these studies demonstrate that intoxication can change the way that MOBC are hypothesized to operate: positive attitudes toward condom use when sober cannot be assumed to remain positive when people are intoxicated. These alcohol-influenced changes in attitudes experienced in-the-moment may then result in a lower likelihood of condom use.

Self-efficacy and condom use skills.

All of the major health behavior change models seek to either influence the perception of the ability to implement health-protective behaviors, reflected in self-efficacy [59] or perceived behavioral control [34,35], and/or improve the skills to do so [33]. This is mirrored in the incorporation of various modules to promote skillful use of condoms and partner communication about condom use as seen in HIV prevention interventions across the range of behavioral change models [40,80,152–161]. ACR studies have shown that behavioral skills related to condom use are impaired under an intoxicating dose of alcohol. This is most directly seen in studies of condom use negotiations skills.

Alcohol makes it more likely that women will allow their partner to make decisions about condom use [114]. The effect of intoxication on condom abdication has been shown to be particularly salient for heterosexual women with a history of sexual abuse [128,129] as intoxication increases concern about negative partner reactions to the suggestion of using a condom [128]. Intoxication also appears to impair women’s abilities to insist on condom use in sexual situations [89,94,130,162]. Studies of men have also shown that alcohol may impair condom use negotiation skills [91,148,151,163]. In their test of the IMB model, Gordon et al. [151] found that intoxicated heterosexual men exhibited impaired self-efficacy to initiate condom use. In a follow-up study using a role play task, intoxicated heterosexual men exhibited poorer condom negotiation skills compared to those in placebo and control conditions [148].

Similar impairment in condom negotiation skills has been shown in studies of MSM. In a video role play task which required the participant to respond to prompts, Maisto et al. [91] showed that those who were intoxicated displayed poor communication skills regarding condom use, particularly in sexual contexts that involved a high degree of social pressure. Examination of the mechanisms underlying alcohol effects in this study showed that perceived intoxication served as a mediator between alcohol consumption and skill deficits, specifically among those experiencing high levels of sexual arousal [111]. The capacity to resist condomless sex in this high-pressure social context was also one of the only MOBC to distinguish between alcohol and placebo conditions in an MSM study by Wray and colleagues [92].

Other than the studies mentioned above, the direct effects of alcohol on self-efficacy and perceived control have not received as much direct research, but available evidence tells us that intoxicating doses of alcohol can diminish condom use self-efficacy [151]. Additional evidence regarding the potential impact of intoxication on self-efficacy has come from studies that assessed self-efficacy regarding condom use as a baseline variable (i.e., as part of a questionnaire taken before alcohol administration). Davis et al. [164], for example, showed that the association between self-efficacy ratings and intentions to use condoms was moderated by alcohol consumption. In other words, the association between self-efficacy and intentions to engage in condom use was diminished when individuals consumed alcohol. This exemplifies how failure to consider the role of alcohol may call into question the hypothesized association between high levels of a given MOBC (e.g., high self-efficacy) and health promoting behavior (e.g., condom use) that is a central assumption (so called “target validation” [165]) of theoretically-based interventions. Such findings from ACR research show why it is important to consider that alcohol may change the influence of MOBC targeted by HIV prevention interventions even if it does not change the MOBC themselves.

Identifying novel intervention targets and mechanisms of behavior change

Not only has ACR shown that alcohol has a causal effect on MOBC that are central to models of health behavior change, but it has also (1) expanded our understanding of how these MOBC are altered by alcohol, and (2) identified novel processes through which alcohol exerts its influence on sexual decision-making. Consequently, ACR has provided the basis for the identification of novel intervention targets and theory development.

In particular, investigators have increasingly applied dual process models [166] to the study of health risk behavior [167–169] and sexual risk in particular [170,171]. The basic premise of these models is that the ability to successfully change health risk behaviors depends, in part, on the interaction of automatic, impulsive processes that drive appetitive responses on the one hand, and effortful, reflective processes that support self-control behaviors on the other [71,172,173]. Reflective processes include executive functioning abilities (EF), such as working memory and response inhibition, that vary in strength between individuals and contexts. Those who have high EF abilities (e.g., working memory, response inhibition) will be better able to act according to personal standards and conscious beliefs and attitudes. Conversely, those who have low EF abilities or whose EF capacities are compromised (e.g., by intoxication, emotional state, cognitive load, etc.), are more likely to act in accord with automatic impulsive processes [166].

Alcohol intoxication accentuates the impact of these impulsive processes on behavior in two ways. First, alcohol consumption serves to increase the strength of automatic approach processes toward appetitive cues [174,175]. For example, Simons et al. [109] examined the influence of intoxication on an approach-avoidance task with both sexual imagery (heterosexual couples engaging in oral and vaginal sex) and condom images. The results of this study showed that participants in the alcohol condition exhibited relatively stronger approach biases toward erotic stimuli and away from condom-related stimuli compared to control and placebo participants. Second, alcohol intoxication is associated with deficits in executive functioning [176–178]. Decrements in EF that are induced by alcohol may influence the cognitive capacity to successfully implement risk reduction skills [170]. These dual processes suggest potential alternative approaches (e.g., working memory training [179] and cognitive bias modification [180]) to target these mechanisms.

Developing and refining strategies that will be effective under the influence of alcohol

One of the underutilized benefits of ACR is its ability to provide insight into whether alcohol changes the capacity of HIV prevention strategies to influence MOBC and sexual risk behavior. As characterized by the experimental medicine framework [71,84], laboratory-based experiments can be used in the development of intervention approaches to (1) test whether intervention strategies or microinterventions (the brief administration of a single therapeutic technique predicted to cause measurable change during an experimental session [181]) causally modify MOBC, and (2) examine whether these modifications result in subsequent behavior change. This experimental approach is a low-resource and low-cost method of evaluating the benefits of novel strategies to modify identified targets [182]. The field has generally not made use of this approach to develop intervention strategies for condom interventions that are specifically alcohol-focused but there are illustrative studies in this regard. Hahn and colleagues [180] used the cognitive behavioral modification technique of approach-avoidance training [183] to change approach biases away from alcohol-related stimuli and toward condom-related stimuli during four laboratory-based training sessions. Participants who received active training not only reported greater positive attitudes toward condom use compared to those who had received sham training, they also reported more consistent condom use at a 3-month follow-up session. Such approaches may be integrated with ACR to identify how alcohol influences strategy implementation. Davis and colleagues [85], for example, used ACR to test how two interventions targeted at sexual aggression influenced the hypothesized MOBC of emotion regulation while participants were under the influence of alcohol. The authors found that their cognitive restructuring intervention was associated with lower intentions to coerce sex in an analogue task through the MOBC of emotional modulation. Although there are important considerations for translating experimentally derived strategies into efficacious approaches [184], studies such as Davis et al. [85] may lead to additional novel strategies to address key targets identified from alcohol-administration research and clarify how strategies may be impacted by contextual factors such as alcohol [36].

Limitations of the ACR Approach

Despite the many contributions of ACR, it is important to note limitations of the experimental approach for understanding how alcohol may influence the impact of HIV prevention interventions. The advantages of being able to experimentally control multiple factors and isolate the role of specific processes come at the cost of being able to represent the real-world contexts in which alcohol use, MOBC, and sexual behavior unfold. Moreover, the use of likelihood ratings or intentions as a proxy for sexual behavior is limited given the fact that intentions do not necessarily predict sufficient variance in behavior [71] and may be less-accurate indicators of real-life sexual decisions for certain individuals [185]. Another limitation is that the populations that may be most likely to engage in heavy episodic drinking (e.g., those with severe alcohol use disorders) may be excluded from alcohol administration studies due to ethical concerns about administering alcohol to such individuals. ACR experiments are limited in how much alcohol can be administered which may not provide data on amounts that confer the greatest levels of risk.

Some of the above limitations may be addressed through increasingly sophisticated approaches to modeling sexual risk contexts (like virtual reality) and sexual decision-making responses. However, it is important to keep in mind that the goal of the ACR is to provide one of multiple sources of evidence relevant to our understanding of how alcohol influences MOBC that underlie HIV prevention interventions. It is critical to incorporate ACR into a multi-method approach, including experience sampling methods [83], to corroborate findings and use conflicting results across methods to help develop and refine theory. It also speaks to the importance of having an explicit translational strategy [165,186] to appreciate the reciprocal processes of developing and refining intervention strategies from experimental methods and using data from the implementation of these strategies to develop better experimental paradigms.

Conclusion

As outlined above, ACR has helped to test and expand theory about how alcohol influences HIV risk behaviors and has provided many insights about how MOBC related to sexual risk may be influenced by intoxication. In doing so, it has provided an empirical basis that can be used to further develop and improve HIV prevention interventions for individuals who use alcohol. A central contribution has been to identify how and when alcohol causally influences sexual decision-making processes (e.g., arousal, implicit sexual associations) and specify the processes that may be used to override or change patterns of risk sexual behavior (e.g., condom negotiation skills) [170]. This research has included advances in measurement, improvement of theoretical models of alcohol’s influences on sexual risk behavior, and insights into potentially novel ways to more effectively address “heat of the moment” processes that characterize alcohol-influenced sexual risk behavior. As outlined in Table 1, these findings have several implications for intervention development.

Table 1.

Implications of ACR findings for intervention development

Finding	Intervention implication(s)
Overall, intoxication alters MOBC related to condom use and safer sexual behavior.	• Provide more education on alcohol’s effects during HIV prevention interventions; increase participant awareness of challenges to implementing effective condom use and safer sex negotiations when intoxicated.  • Identify strategies for risk-reduction that will not be disrupted (or will be less disrupted) by alcohol use, including identification of MOBC that are less susceptible to alcohol effects.
Intoxication interferes with the processing of risk and inhibition cues related to condom use and safer sexual behavior.	• Design ways to increase the salience of inhibitory cues. This might include smartphone-enabled just-in-time interventions or physical reminders (e.g., bracelets [145]) that can provide personalized reminders to engage in safer sexual behavior during times of high risk (i.e., when at a bar or club).
Alcohol reduces motivation and creates more negative attitudes toward condoms “in the moment”.	• Develop novel ways to influence attitudes and motivation that will be less susceptible to alcohol effects (e.g., modifying implicit approach/avoid tendencies toward condoms using cognitive bias modification).  • Use state-dependent learning approaches to train participants to consider information related to motivation to change while in a state associated with heightened sexual risk.
Alcohol inhibits the ability to retrieve and utilize condom use skills and decreases perceptions of self-efficacy.	• Focus on increasing planning and use of strategies to increase activation of skills “in the heat of the moment” (e.g., pre- commitment strategies).  • Use alcohol administration or virtual reality to provide state-dependent learning of relevant skills.  • Teach strategies that will be less susceptible to the effects of intoxication, including implementation intentions, simpler strategies with lower complexity and cognitive requirements, and habit training to be able to respond consistently to relevant cues through practice.
Automatic appetitive processes are more salient during periods of intoxication.	• Cognitive bias modification techniques may be a useful adjunct to traditional counselor-led or electronic interventions.
The causal effects of alcohol on sexual decision-making are moderated by individual differences in MOBC.	• Assess person-level risk factors for risky sexual decision-making under the influence of alcohol (e.g., alcohol-related sexual outcome expectancies) and MOBC known to be influenced by alcohol to tailor HIV prevention interventions.
Population-specific influences impact how alcohol and MOBC interact to predict condom use decisions.	• Tailor interventions to consider the relative impact of different MOBC for a given population and how these are influenced by intoxication.

EF = executive functioning; MOBC = mechanism(s) of behavior change.

Alcohol exerts its effects both through experiential changes, such as heightened arousal and perceived intoxication, as well as through changes in cognitive-motivational processes. A number of studies have provided evidence that alcohol can change the experience of sexual arousal [105,111,170], which is associated with a heightened response tendency toward sexual stimuli, increases in the reward salience of sex, and limits in the capacity to consider factors that might improve the likelihood of using condoms during intercourse. This, in conjunction with the tendency for alcohol to increase attention to salient impelling cues through alcohol myopia [107,129,145] and interfere with working memory [187], highlights a change in state that may make deliberate retrieval of skills or the consideration of risks and potential negative consequences of condomless sex more difficult.

The distinctive features of these “hot” states associated with risk behavior have been well-studied [188,189], yet the models of health behavior change that have been applied in the development of most interventions typically have not considered the implications of learning health behavior change strategies in a sober state and then subsequently attempting to implement them when intoxicated. This may interfere with the ability of individuals who complete preventive HIV interventions to activate cognitive strategies in support of behavior change (e.g., evaluating risk) or utilize health-protective behaviors (e.g., condom negotiation skills) during an alcohol-influenced state that includes high arousal and processing biases toward immediate sexual and reward cues. Investigators have characterized these states as those that involve increased influences of automatic, implicit processes on behavior [166,190]. Dual processing models suggest that there may be value to expanding intervention targets and strategies to directly address these impulsive processes as part of efforts to reduce sexual risk-taking. Interventions that attempt to directly modify automatic responses to cues, such as cognitive bias modification or evaluative conditioning [183,191], may provide alternative methods to modify automatic behavioral tendencies to engage in condomless sex. They may also be less susceptible to alcohol-induced changes in executive functioning skills that are necessary to implement deliberate behavioral change strategies. Research on automatic and implicit processes in sexual risk behavior [109] has provided the framework to begin to explore strategies that may address changes in processing that occur in sexual decision-making contexts [109]. Although efforts to utilize this framework to develop interventions are currently at a preliminary stage, they do provide a potentially important new avenue for addressing “heat of the moment” processes.

Another significant contribution of ACR has been to provide a clearer understanding of who may be most susceptible to the effects of alcohol on HIV risk behaviors and the processes through which this heightened risk may occur. The experimental approach has helped clarify how the causal effects of alcohol are shaped by individual difference variables such as alcohol expectancies and sexual abuse history. The study of individual differences in ACR research may be used to identify those who may be at specific risk for alcohol-related sexual risk behaviors and help tailor interventions that address MOBC that are of particular importance for these individuals.

These person-level findings also highlight the importance of conducting ACR with different populations. As indicated in the review above, it is important to keep in mind that the influence of alcohol on MOBC and sexual behavior will vary according to a number of individual difference characteristics [9]. ACR can help illuminate the population-specific ways that alcohol may impact sexual decision-making and health behavior change. It is critical to consider, for example, differences in alcohol and sexual behavior patterns as well as population-specific risk factors that may be associated with condomless sex. Among MSM, for example, stigma and discrimination affect how alcohol is consumed in sexual contexts, particularly for subpopulations (e.g., minority MSM) who experience multiple sources of stigma and discrimination [192,193]. Similarly, alcohol use is linked to sexual behavior patterns and contextual moderators that are specific for MSM [194]. Even the impact of MOBC identified in health behavior models, such as norms regarding condom use [194], will be distinct for different populations [195]. Such findings exemplify the larger point that the relative weight and value of sexual decision-making MOBC vary by population, and consequently so do the effects of alcohol on decisions about when to use condoms

ACR has also elucidated the contextual factors that moderate the effects of alcohol on sexual decision-making [36]. Partner factors have been one of the best-studied contextual moderators of alcohol effects in ACR. Attractiveness of partners and the potential for relationships have been identified as important moderators of alcohol effects on condomless sex. Studies in heterosexual populations have shown that the causal effects of alcohol on sexual decision-making are influenced by partner characteristics. These findings help emphasize that these responses do not occur in a vacuum, as sexual decision-making takes place with an active partner (or partners) through reciprocal processes. This is an important factor to model in experiments and efforts to represent different partner responses have been developed [91]. The extent to which these can represent dynamic responses will advance our understanding considerably.

Using ACR to develop a better understanding of how and when alcohol influences the decision-making process provides insight into what strategies may reduce sexual risk for individuals who continue to consume alcohol. If we know how alcohol modifies behavior change processes in context, we will be able to improve interventions by teaching the use of existing cognitive and behavioral strategies that may be less susceptible to alcohol-related interference and developing strategies that will operate through mechanisms less affected by alcohol consumption. For example, if we know that the execution of condom negotiation skills can be disrupted by alcohol consumption, we might be able to recommend alternative safer sex behavioral change strategies or modify condom negotiation strategies to make them less susceptible to alcohol-related interference (e.g., training on implementation intentions [196,197]). Findings from ACR suggest that there may be ways to better design interventions for those who do not wish to modify their drinking (which will be applicable to many populations) or teach skills in a manner that is less likely to be influenced by alcohol.

To a lesser extent, ACR has also shown that alcohol may also influence the associations between MOBC targets and sexual risk behavior [164]. This suggests that alcohol may interfere with sexual risk reduction efforts not because it alters MOBC such as self-efficacy in context, but rather because it changes the impact of MOBC on sexual risk behavior. This suggests a potentially interesting and important additional consideration for alcohol effects in that they may change the salience, weight, and influence of MOBC on behavior. This represents a change in target validation rather than accessibility of the MOBC.

Although ACR has emphasized the impact of heavy alcohol consumption (i.e., doses to target a BAL of .075% or above) on HIV risk, the use of lower doses of alcohol [89,113,140,162], placebo conditions [92] and manipulation of environmental contexts [198] have identified the importance of cue and expectancy effects on sexual risk MOBC. These effects may be differentially important depending on specific individual difference factors (e.g., for those who have strong alcohol-sexual expectancies). This leads to the possibility that certain individuals may be susceptible to alcohol effects on MOBC relevant to sexual risk at different doses. Cue and pharmacological effects operate through distinct processes that may make them susceptible to distinct change strategies.

Finally, ACR may contribute to efforts to better integrate alcohol into HIV prevention strategies as targets of behavior change. In general, there is little conceptual integration to inform alcohol-condom use interventions, and very little is known about how these sorts of combined interventions might work. As described above, intervention components of HIV prevention interventions that have targeted alcohol have been based on their established efficacy for reducing heavy episodic drinking rather than on an integrated theoretical framework based on an understanding of alcohol’s influences on MOBC. The use of an integrated framework may also provide insight into more efficient ways to address multiple targets. Providing a conceptual rationale for integrating intervention elements has benefits for those receiving the intervention in terms of understanding how and why risk behaviors are associated and what the implications of change in one might be for the other.

In sum, using ACR to understand drinking as a context for sexual decision-making provides for a more refined understanding of MOBC related to sexual risk reduction. This facilitates the optimization of existing interventions, allows for tailored interventions for those who drink heavily, permits more efficient intervention delivery, and provides insight into new strategies to target key self-regulatory mechanisms that predict outcomes. Our ability to develop more efficacious integrated interventions will be facilitated by a more complete understanding of the causal impact of alcohol on processes underlying HIV-related risk: it is in this way that the experimental approach is particularly useful for HIV prevention intervention development. The potential benefits of ACR remain to be fully realized as investigators continue to develop more precise models and measures of “heat of the moment” processes that underlie sexual risk behavior and behavioral change.