AIDS Care, 2015 Vol. 27, No. 3, 392–400, http://dx.doi.org/10.1080/09540121.2014.967657
Predictors of condom use in women receiving court-mandated drug and alcohol treatment: implications for intervention Natasha S. DePesaa*, Gloria D. Eldridgeb, Frances Deaversa and Jeffrey E. Cassisia a
Psychology Department, University of Central Florida, Orlando, FL, USA; bPsychology Department, University of Alaska Anchorage, Anchorage, AK, USA (Received 26 February 2014; accepted 15 September 2014) Women who abuse substances are at a high-risk for contracting HIV. Condom use interventions are important in reducing HIV in high-risk populations, but current interventions have small effects. The aim of this study is to examine the relative impact of substance use, personal variables (sexual impulsivity and condom expectancies), and relationship variables (perceptions of relationship commitment and partner risk, perceptions of power within the relationship) on condom use in women in court-mandated substance abuse treatment. Information was collected from 312 sexually active women in an inpatient drug and alcohol treatment facility in the Southeastern US Participants completed questionnaires and were interviewed using the Timeline Follow-back method and provided information about sexual activity in the 30days prior to intake, including type of sexual event, co-occurrence with substance use, condom use, and characteristics of sexual partners and the nature of the relationship. Multilevel logistic modeling revealed that perception of relationship commitment, condom outcome expectancies, and age significantly affected condom use for women in the sample. Specifically, condom use was least likely when women reported that the relationship was committed (odds ratio [OR] = 0.31, 95% confidence interval [CI]: 0.23, 0.43) or when the participant was older (OR = 0.96, 95% CI: 0.94, 0.99), and more likely when women reported more positive condom outcome expectancies (OR = 1.02, 95% CI: 1.00, 1.03). The findings suggest that perceptions of relationship commitment, regardless of perceptions of partner risk, strongly affect condom use among women court-mandated into drug and alcohol treatment. In addition, positive outcome expectancies (e.g., positive self-evaluations and perceived positive partner reactions) are associated with a greater likelihood of condom use. These findings have important implications for condom use interventions, which have failed to produce large or lasting effects within this population. Keywords: condom use; sexual risk behavior; partner commitment; HIV intervention; high-risk women
Individuals who abuse substances often face disadvantages at multiple levels, ranging from societal-level to individuallevel factors (Beyrer, 2007; Pellowski, Kalichman, Matthews, & Adler, 2013). Consequently, they are at greater risk for HIV infection, are less likely to seek and adhere to antiretroviral treatment, and often progress more rapidly to AIDS (Chander, Himelhoch, & Moore, 2006; Kapadia, Vlahov, Donahoe, & Friedland, 2005). Condom use remains one of the primary defenses against HIV transmission. However, meta-analytic reviews of condom use and sexual risk reduction interventions reveal mainly small effect sizes, particularly with individuals who abuse substances (Noar, 2008; Prendergast, Urada, & Podus, 2001). Further, effect sizes have remained relatively stable over time, suggesting limited increases in intervention efficacy (Meader et al., 2013). Focusing on significant moderators of intervention efficacy may enhance the impact of condom use interventions (Peragallo et al., 2005; Scott-Sheldon, Huedo-Medina, Warren, Johnson, & Carey, 2011). Sexual risk reduction and condom use interventions that have been tailored to address the needs of specific *Corresponding author. Email: [email protected] © 2014 Taylor & Francis
groups have shown promising results (Morrison-Beedy et al., 2013; Shoptaw et al., 2005). Women in substance use treatment are at particular risk for HIV (Beyrer, 2007; Brooks et al., 2010), yet condom use interventions in this group have failed to produce large effects (Meader et al., 2013; Prendergast et al., 2001). This suggests that these women may benefit from tailored interventions. Variables that may affect sexual risk behavior in women who abuse substances include factors surrounding the alcohol- or drug-using context, individual differences, and relationship and partner variables. Decreased condom use is clearly linked to the use of alcohol and other substances (Leigh & Stall, 1993). However, when examining individual sexual events, condom use does not appear to vary with concurrent alcohol or substance use (Calsyn, Baldwin, Niu, Crits-Christoph, & Hatch-Maillette, 2011; Hagger-Johnson, Bewick, Conner, O’Connor, & Shickle, 2011; Leigh, Ames, & Stacy, 2008; Scott-Sheldon et al., 2009). As alcohol and substance use do not appear to predict event-level condom use, it seems that other factors may be involved.
AIDS Care Individual factors such as impulsivity and sensation seeking appear to be consistently linked to substance-using behavior, but their link to sexual risk behavior is less straightforward. In one study, impulsivity increased the relationship between alcohol use and risky behavior, while self-regulation abilities had a buffering effect (Neal & Carey, 2007). In another study, both sexual and nonsexual sensation-seeking were related to increased alcohol use before sex and more sexual partners (Hendershot, Stoner, George, & Norris, 2007). Expectancies surrounding the outcomes of condom use have a more straightforward effect. More positive condom outcome expectancies are associated with more consistent condom use among individuals who use injection drugs (Brown et al., 2008; Kapadia et al., 2011). Partner and relationship variables strongly affect condom use in women who abuse substances (Ober et al., 2011; Parks, Hsieh, Collins, & Levonyan-Radloff, 2011). Condom use is more likely with casual or noncommitted partners than with partners perceived to be committed (Ober et al., 2011; Parks et al., 2011). Women who consider themselves to be in committed relationships report lower perceived sexual risk than women who do not (Matson, Chung, Sander, Millstein, & Ellen, 2012; Stein, Nyamathi, Ullman, & Bentler, 2007). However, discordance between perceptions of monogamy and actual partner concurrency are associated with increased rates of STI infection among individuals who misperceive risk (Riehman, Wechsberg, Francis, Moore, & Morgan-Lopez, 2006; Witte, El-Bassel, Gilbert, Wu, & Chang, 2010). Even when risk is perceived, individuals are not likely to use condoms with committed partners (Ober et al., 2011). Finally, among women, more perceived relationship power may be linked to greater consistency of condom use (Pulerwitz et al., 2002). The aim of this paper is to examine substance use, personal variables (sexual impulsivity and condom expectancies), and relationship variables (perceptions of commitment and partner risk, perceptions of power within the relationship) and their impact on condom use in order to synthesize and replicate previous findings. The intent is to inform efforts to tailor HIV prevention efforts for women in substance abuse treatment. The study uses multilevel logistic modeling to examine the simultaneous and relative effects of substance use, individual factors, and perceived relationship and partner characteristics on event-level condom use in women court-ordered into inpatient drug treatment. This type of modeling allows for an integrated and informative view into event-level condom use in a population of women at high risk for HIV-infection. Methods Participants and procedure The sample is comprised of women court-mandated into drug and alcohol treatment at an inpatient chemical
393
dependence unit in a Southeastern state hospital. Data were from a larger study (n = 506) funded by the National Institute on Drug Abuse and designed to test HIV prevention interventions. Women were approached for the study after they completed detoxification and if they were determined to be ready for treatment activities. The project coordinator provided study information to potential participants and obtained informed consent. Participation was voluntary and research protocols were approved by the university and facility Institutional Review Boards. Data were collected prior to intervention, immediately after the intervention and at 3- and 6months follow-up. The present analyses utilized data from the preintervention assessment, during which participants were interviewed about their sexual activity by trained university research staff and provided $10 compensation. Participants were selected for the current analyses if they reported at least one instance of receptive vaginal or anal sex with a male partner in the 30 days prior to entering the drug treatment program. Those who reported being in a controlled environment (e.g., prison, other inpatient treatment) prior to entering treatment reported on the 30-day period prior to being in a controlled environment (see Table 1 for details).
Measures Demographic information Demographic information (e.g., age, race) was obtained using a self-report questionnaire. Addiction Severity Index (ASI) Information about participants’ alcohol and substance use history was obtained using the ASI, a semi-structured interview widely used in drug treatment facilities (Alterman, Cacciola, Habing, & Lynch, 2007; McLellan, Luborsky, Woody, & O’Brien, 1980). Composite scores determined overall severity of alcohol or substance abuse, while individual items provided descriptive information. Comprehensive Timeline Follow-back (TLFB) The TLFB procedure was used to obtain details regarding sexual activity, condom use, and concurrent substance use over the 30 days prior to entering the treatment program or another controlled environment. The TLFB is a calendarbased interview that provides day-by-day information about alcohol and substance use, sexual activity, condom use, perception of relationship commitment with sexual partners, and perceived partner risk (Sobell & Sobell, 1992). Following Morrow, Eldridge, Nealey-Moore, Grinstead, and The Project START Study Group (2007), a partner was defined as “risky” if the participant believed the partner to be HIV+, to have ever injected drugs, to have smoked crack cocaine, or was someone with
394
N.S. DePesa et al.
Table 1. Sample characteristics. n (%) Race Caucasian African-American/Black Other Marital status Single/never married Married Separated/divorced/widowed Education Less than high school High school/GED Beyond high school In a controlled environment prior to treatment M = 9.05 days (SD = 7.00) Ever incarcerated (at least one month) M = 13.47 months (SD = 26.32) Ever charged with prostitution Prior alcohol use treatment M = 3.73 attempts (SD = 3.92) Prior substance use treatment M = 3.18 attempts (SD = 3.39) Prior HIV education Prior HIV testing Positive results Negative results Unsure of results
186 (59.6) 118 (37.8) 8 (2.6) 112 (35.9) 50 (16.0) 150 (48.1) 104 123 85 119
(33.3) (39.4) (27.3) (38.1)
90 (28.8) 10 (3.2) 133 (42.6) 192 (61.5) 77 (24.7) 246 (94.9) 10 (3.2) 222 (71.1) 64 (20.5)
Note: Total n = 312. M indicates mean; SD indicates standard deviation. A controlled environment refers to prison or inpatient treatment other than the current treatment.
whom the participant had exchanged sex for money, drugs, or other things. Previous research has established the TLFB procedure to provide reliable and valid data about sexual behavior and substance use for up to a three month reporting period (Carey et al., 2001; Weinhardt et al., 1998). Sexual Impulsivity Scale (SIS) The SIS is a 15-item self-report scale assessing sexual impulsivity and behavioral outcomes (e.g., “Sometimes I get overexcited sexually and get myself into trouble”) that was developed for this study. Each item was rated on a 5-point scale ranging from 1 (Not true of me) to 5 (Very true of me). Negatively worded items were reverse coded so that a higher scale score represents greater sexual impulsivity. Cronbach’s alpha for the current sample was .81. Relationship Power Scale (RPS) The RPS contains eight self-report items developed for the study that assess the extent to which the participant or her partner controls sexual and relationship decisions (e.g.,
“Who decides whether or not to use a condom?”; “Who makes more decisions in the relationship?”). Responses ranged from 1 (He always does) to 5 (I always do). Cronbach’s alpha was .84. A higher score indicates that the participant makes the majority of the decisions within the relationship while a lower score indicates that her partner makes the majority of the relationship decisions. Condom Outcome Expectancy Scale (COES) The COES is a 29-item scale that reliably assesses attitudes and beliefs about condom use (Dilorio, Maibach, O’Leary, Sanderson, & Celentano, 1997). The scale contains six factors: partner reaction, hedonism, prevention efficacy, social approval, positive self-evaluation, and negative selfevaluation. An example item includes, “Saying we have to use a condom or barrier is like saying ‘I don’t trust you’” (reverse scored). Response choices ranged from 1 (Strongly Disagree) to 5 (Strongly Agree). Negatively worded items were reverse coded so that a higher scale score represents more positive expectancies surrounding the use of condoms.
Analysis Level-1 data were derived from reports on sexual activities with up to three males identified by each participant as her most frequent sex partners in the previous 30 days. Data were organized on the event-level, such that each unit of analysis represents one day where sexual activity (i.e., at least one instance of either vaginal or anal sex) occurred along with relevant contextual information (e.g., condom use, co-occurring substance use, partner and relationship characteristics). If a participant reported sexual activity with more than one partner in a single day, separate lines were included for each partner. If a participant reported sexual activity with more than three male partners over the recall period, event-level data were included in the main analyses for the three most frequent partners and summed data (e.g., total number of instances of protected or unprotected vaginal intercourse) were collected for the remaining partners. Level-2 data were organized on the individual level and consisted of demographic variables (i.e., age and race), total number of partners in the last 30 days, and scale scores derived from self-report measures (i.e., SIS, RPS, and COES). The data were analyzed using random effects multilevel logistic regression modeling. The numbers of sexual events at level-1 were nested within the women who reported them at level-2, allowing for simultaneous testing of individual-level and event-level effects on condom use. Analyses were conducted in several stages using the build-up method described by Bryk and Raudenbush (1992). First, level-1 predictors of condom use were tested individually for their potential to be included in
AIDS Care
395
however, when included in the multivariate model, the impact of perceived partner risk was no longer significant. The interaction of perceived partner risk and relationship commitment was significant; however, the Wald test indicated poor model fit and the interaction was excluded from further analyses. Only perceived relationship commitment was included in the final level-1 model. To illustrate the level-1 effects, Figure 1 highlights the frequency of condom use across partner risk and relationship commitment.
the final level-1 model. Significant univariate predictors were then included in a multivariate model. Predictors that did not affect the likelihood of condom use were excluded and a final level-1 model was fit with the remaining variables. Next, level-2 predictors of condom use were individually added to the final level-1 model to test for their potential inclusion in the final multilevel model. Significant predictors of condom use were then fit into one final multivariate, multilevel model.
Results
Level-2 modeling Age and condom outcome expectancies incrementally affected the likelihood of condom use such that younger age and more positive expectancies were associated with an increase in the likelihood of condom use. Number of partners in the last 30 days, sexual impulsivity, relationship power, and none of the interaction terms were found to affect the likelihood of condom use.
Description of participants and sexual activity The sample was comprised of 312 women (see Table 1 for sample characteristics). Patterns of recent and lifetime substance use, by substance type, across all participants are shown in Table 2. A total of 3146 instances of sexual activity with a male partner (i.e., days with at least one instance of vaginal or anal intercourse) were analyzed. Participants reported between 1 and 30 days with sexual activity in the 30-day reporting period (M = 10.08, SD = 8.98). Fifty women reported more than three male partners in the preceding 30 days, 7 of whom (14.0%) reported 10 or more additional partners. Thirty of the 50 women (60.0%) reported consistent condom use, 6 (12.0%) reported no condom use, and 11 (22.0%) reported occasional condom use with these additional partners. Descriptive information for level-1 and -2 variables can be found in Tables 3 and 4, respectively.
Final integrated model Overall, condom use was relatively unlikely in this sample of women (odds ratio [OR] = 0.56, 95% confidence interval [CI]: 0.43, 0.73). The final multivariate, multilevel model contained perceived relationship commitment (level-1) and age and condom outcome expectancies (level-2). Condom use was less likely in committed relationships (OR = 0.31, 95% CI: 0.23, 0.43) and with older participants (OR = 0.96, 95% CI: 0.94, 0.99) and more likely when women had more positive condom outcome expectancies (OR = 1.02, 95% CI: 1.00, 1.03). An exploratory cross-level interaction between commitment and expectancies was not significant (results not shown), indicating that the impact of condom outcome expectancies did not vary by perceived relationship commitment. All modeling results are displayed in Table 5.
Level-1 modeling Substance use concurrent with sexual activity did not significantly affect the likelihood of condom use. Perceived relationship commitment and perceived partner risk, respectively, significantly decreased and increased the likelihood of condom use in the univariate analyses;
Table 2. Recent and lifetime substance use by substance type. Recent use (days) Substance Alcohol (felt effects) Cannabis Cocaine Opiates/analgesics, heroin, methadone Barbiturates, sedatives/hypnotics/tranquilizers Other substances >1 substance (including alcohol)
n (%) 199 170 197 60 66 59 222
(63.4) (54.1) (62.7) (19.1) (21.0) (18.8) (70.7)
Lifetime use (years)
M (SD) 15.2 12.1 15.3 5.0 6.8 3.6 13.5
(11.2) (11.6) (11.1) (4.1) (6.7) (3.6) (10.9)
n (%) 210 202 217 69 73 86 243
(66.9) (64.3) (69.1) (22.0) (23.2) (27.4) (77.4)
M (SD) 11.2 10.0 5.9 1.9 2.8 1.2 7.7
(7.0) (6.9) (4.8) (1.8) (3.8) (1.3) (5.8)
Note: Information was obtained from the ASI. Recent use refers to use over the 30-day reporting period. Lifetime use refers to years of regular or problematic use. n indicates the number of individuals who reported at least one day of recent substance use (recent use) or the number who reported at least one year of regular or problematic use (lifetime use). M indicates the mean; SD indicates the standard deviation. Alcohol (felt effects) refers to alcohol consumption to the point of some impairment. No significant differences between racial groups were found in the alcohol or drug composites.
396
N.S. DePesa et al.
Table 3. Level-1 descriptive statistics.
Total no. of events Condom used (outcome) Committed partner Risky partner Exchange partner Concurrent substance use
Vaginal sex n (%)
Anal sex n (%)
3144 472 2197 1970 605 1860
22 2 9 20 4 19
(100) (15.0) (69.9) (62.7) (19.2) (59.2)
(100) (9.1) (40.9) (90.9) (18.2) (86.4)
Note: Analyses were conducted across vaginal and anal sexual events. Risky partners are those that participants perceived to be HIV+, to have injected drugs, to have smoked crack cocaine, or was a person with whom sex has been exchanged for money, drugs, or other things. Exchange partners were not included as an individual predictor.
Discussion This study integrates and confirms previous research while providing a vigorous multi-level examination of event-level condom use among women who abuse substances. Given the limited efficacy of condom use interventions in this population, this information is particularly useful for informing future interventions. Results indicated that perceived partner commitment, negative condom outcome expectancies, and older age were all associated with a lower likelihood of condom use. Strikingly, condom use occurred in only 15.0% of sexual events overall. As expected, perceptions of relationship commitment had the largest impact on condom use such that women were less likely to use condoms in a relationship they defined to be “committed,” despite perceptions of partner characteristics deemed “risky” (Ober et al., 2011). During sexual activity with partners perceived to have risky characteristics, condom use occurred 32.2% of the time when the woman defined the relationship as not committed but only 6.9% of the time when the relationship was defined as committed (Figure 1). This dramatic difference may result from inaccurate perceptions of risk or discounted importance of risk (Riehman et al., 2006; Witte et al., 2010).
Table 4. Level-2 descriptive statistics. M (SD) Age Partners in last 30 days Relationship Power Scalea Condom Outcome Expectancies Scaleb Sexual Impulsivity Scalec
32.15 4.2 26.06 91.56 37.87
(7.63) (14.9) (5.81) (15.47) (11.53)
a Range is 8–40; higher scores indicate more relationship power. bRange is 29–145; higher scores indicate more positive expectancies. cRange is 15–75; higher scores indicate greater impulsivity. Mean-centered scale scores were used in the analyses in order to facilitate the interpretation of the resulting odds ratios.
Figure 1. Frequency of condom use during sexual events by partner type. Partners with whom the women identified themselves as being in a committed relationship are labeled as “committed” while others are labeled “noncommitted.” Partners who were perceived to have one or more defined risk characteristics were labeled “risky” while those perceived to have none of the risk characteristics were labeled “nonrisky.” n = number of sexual events with each partner type; the total number of sexual events for all partner types was 3141. The * represents the significant main effect of perceived partner commitment on condom use found in level-1 modeling and included in the final multilevel model.
Results from this and other studies suggest the importance of addressing expectancies surrounding the perceived impact of using condoms in all types of relationships (Brown et al., 2008; Corbett, Dickson-Gómez, Hilario, & Weeks, 2009). The small effect of COES unit increases (OR = 1.02) suggests that small variations in expectancies may not significantly affect outcomes, but that large differences may more noticeably impact behavior. The elevated risk of transmission during anal sexual activities implicates its importance in HIV interventions. Limited data on anal sexual events (n = 22) prevent formal analysis, but it is noteworthy that anal sexual events occurred more frequently with noncommitted and risky partners and involved condoms less frequently compared to vaginal sexual events (Table 3). Finally, though no a-priori hypotheses were made in regards to age, older participants were less likely to use condoms than their younger counterparts, consistent with prior literature (Sormanti & Shibusawa, 2007). As the majority (69.9%) of sexual activity in this study occurred in relationships perceived by the women to be committed, consideration of the nature of the relationship is critical in designing effective condom use interventions. Attempts so far have produced limited success (Meader et al., 2013). A recent randomized trial found that an enhanced sexual risk-reduction and a standard sexual risk-reduction intervention for women who used noninjection drugs resulted in similar
AIDS Care
397
Table 5. Multilevel modeling results.
Level-1 univariate analyses
Level-1 multivariate analyses
Level-1 final model Level-2 univariate analyses
Final fitted multilevel model
Parameter
Wald Chi2
OR
CI
Intercept Substance use Commitment*** Risk* Risk × Commitment Intercept Commitment*** Risk Intercept Commitment*** Sexual impulsivity Relationship Power Expectancies** Impulsivity × Power Impulsivity × Expectancies Power × Expectancies Age** Race Partners in last 30 days Intercept Level-1 Commitment*** Level-2 Expectancies** Age**
1348.54 96.79 131.44 131.85 46.22 10.91 17.42 0.10 250 131.44
253.12
0.25 1.07 0.30 1.44 0.47 0.56 0.33 1.10 0.59 0.30 1.00 1.03 1.02 0.999 0.999 1.00 0.97 1.10 1.04 0.56
[0.20, 0.31] [0.83, 1.38] [0.22, 0.41] [1.04, 1.98] [0.33, 0.66] [0.39, 0.79] [0.24, 0.45] [0.81, 1.51] [0.45, 0.76] [0.22, 0.41] [0.99, 1.02] [1.00, 1.06] [1.00, 1.03] [0.996, 1.00] [0.998, 1.00] [0.999, 1.00] [0.94, 0.99] [0.89, 1.37] [0.98, 1.10] [0.43, 0.73]
133.30
0.31
[0.23, 0.43]
1.02 0.96
[1.00, 1.03] [0.94, 0.99]
* = p < .05, ** = p < .01, *** = p < .001. Note: Commitment indicates perceived relationship commitment. Risk indicates perceived partner risk. Expectancies indicates condom outcome expectancies. Level-2 univariate analyses were conducted along with the final level-1 model. OR, odds ratio. CI, confidence interval.
reductions in unprotected sex (Koblin et al., 2010). Although the enhanced intervention addressed condom negotiation in various relationship contexts, changes in condom use were most likely to occur during sexual activity with casual and exchange partners, providing further evidence that current intervention efforts are more effective for women in these types of relationships than in those they define as committed. Tailoring condom use interventions to address the needs of women who define their relationships as committed necessitates a better understanding of these needs. For some women, perceptions of commitment may be inaccurate or may fluctuate over time, potentially introducing a source of risk (Arriaga, Reed, Goodfriend, & Agnew, 2006; Drumright, Gorbach, & Holmes, 2004; Matson et al., 2012; Witte et al., 2010). Other women may recognize risks but place a higher priority on the emotional needs that are satisfied by the relationship and consider unprotected sex to be “worth the risk” (Corbett et al., 2009). Condoms may be viewed as potential threats to the relationship. Although condom negotiation strategies are part of many HIV-risk reduction interventions, this has not produced consistent behavior
changes among individuals in committed relationships (Meader et al., 2013). One potential limitation of this strategy is that it does not fully address the issue of commitment and trust and assumes that women have the appropriate tools for general communication within a relationship. Showing respect, providing support, and solving problems together, among other behaviors, indicate commitment and predict relationship quality and stability (Weigel, Brown, & O’Riordan, 2011). General and open discussion of relationships, signs of commitment, trust, and respect may provide a context that facilitates adoption of health-positive messages about condoms (Karney et al., 2010). Further, providing women with general communication skills might help them clarify the nature of their relationships (e.g., intentions of exclusivity) and improve their ability to approach difficult topics with partners (e.g., STI testing or condom use; Karney et al., 2010). While the current results provide useful information, a few limitations merit acknowledgment. The use of selfreport data introduces potential bias and inaccuracy. Also, although the TLFB is considered reliable for up to a 3-month recall period, part of the sample was in a
398
N.S. DePesa et al.
controlled environment prior to assessment and reported on a slightly different time period that the other participants. In addition, some participants reported sexual activity with more than three partners in the 30-day reporting period and results might have varied slightly had event-level data concerning additional partners been included in the analyses. Further, scales created for the study proved to be reliable; however, full psychometric analyses were not conducted, limiting the conclusions that can be drawn. These scales are available upon request. Finally, indirect effects and other variables that may impact condom use (e.g., previous STI) were not examined. The present results, along with a review of the existing literature, suggest avenues for strengthening condom use interventions for women in substance use treatment. In particular, increasing condom use within committed relationships appears to be a difficult but important hurdle. Given the consistent lack of success in condom use interventions among this population of women, researchers should examine the efficacy of focusing on risk reduction within committed relationships and enhancing positive condom expectancies along with other factors (e.g., couplelevel interventions; El-Bassel et al., 2010; Karney et al., 2010; Meader et al., 2013). Funding This work was supported by the National Institute on Drug Abuse under [grant number R01 DA10418].
References Alterman, A. I., Cacciola, J. S., Habing, B., & Lynch, K. G. (2007). Addiction severity index recent and lifetime summary indexes based on nonparametric item response theory methods. Psychological Assessment, 19(1), 119–132. doi:10.1037/1040-3590.19.1.119 Arriaga, X. B., Reed, J. T., Goodfriend, W., & Agnew, C. R. (2006) Relationship perceptions and persistence: Do fluctuations in perceived partner commitment undermine dating relationships? Journal of Personality and Social Psychology, 91, 1045–1065. doi:10.1037/0022-3514.91.6.1045 Beyrer, C. (2007). HIV epidemiology update and transmission factors: Risks and risk contexts–16th International AIDS Conference epidemiology plenary. Clinical Infectious Diseases, 44, 981–987. doi:10.1086/512371 Brooks, A., Meade, C. S., Potter, J. S., Lokhnygina, Y., Calsyn, D. A., & Greenfield, S. F. (2010). Gender differences in the rates and correlates of HIV risk behaviors among drug abusers. Substance Use & Misuse, 45, 2444–2469. doi:10.3109/ 10826084.2010.490928 Brown, L. K., DiClemente, R., Crosby, R., Fernandez, M. I., Pugatch, D., Cohn, S., … Schlenger, W. E. (2008). Condom use among high-risk adolescents: Anticipation of partner disapproval and less pleasure associated with not using condoms. Public Health Reports, 123, 601–607. Bryk, A., & Raudenbush, S. (1992). Hierarchical Linear Models. Newbury Park, CA: Sage.
Calsyn, D. A., Baldwin, H., Niu, X., Crits-Christoph, P., & Hatch-Maillette, M. A. (2011). Sexual risk behavior and sex under the influence: An event analysis of men in substance abuse treatment who have sex with women. The American Journal on Addictions, 20, 250–256. doi:10.1111/j.15210391.2011.00123.x Carey, M. P., Carey, K. B., Maisto, S. A., Gordon, C. M., & Weinhardt, L. S. (2001). Assessing sexual risk behaviour with the Timeline Followback (TLFB) approach: Continued development and psychometric evaluation with psychiatric outpatients. International Journal of STDs and AIDS, 12, 365–375. doi:10.1258/0956462011923309 Chander, G., Himelhoch, S., & Moore, R. D. (2006). Substance abuse and psychiatric disorders in HIV-positive patients: Epidemiology and impact on antiretroviral therapy. Drugs, 66, 769–789. doi:10.2165/00003495-200 666060-00004 Corbett, A. M., Dickson-Gómez, J., Hilario, H., & Weeks, M. R. (2009). A little thing called love: Condom use in high-risk primary heterosexual relationships. Perspectives on Sexual and Reproductive Health, 41, 218–224. doi:10.1363/41 21809 Dilorio, C., Maibach, E., O’Leary, A., Sanderson, C. A., & Celentano, D. (1997). Measurement of condom use selfefficacy and outcome expectancies in a geographically diverse group of STD patients. AIDS Education and Prevention, 9(1), 1–13. Drumright, L. N., Gorbach, P. M., & Holmes, K. K. (2004). Do people really know their sex partners? Concurrency, knowledge of partner behavior, and sexually transmitted infections within partnerships. Sexually Transmitted Diseases, 31, 437– 442. doi:10.1097/01.OLQ.0000129949.30114.37 El-Bassel, N., Jemmott, J. B., Landis, J. R., Pequegnat, W., Wingood, G. M., Wyatt, G. E., & Bellamy, S. L. (2010). National Institute of Mental Health multisite Eban HIV/STD prevention intervention for African American HIV serodiscordant couples: A cluster randomized trial. Archives of Internal Medicine, 170, 1594–1601. doi:10.1001/archinternmed.2010.261 Hagger-Johnson, G., Bewick, B. M., Conner, M., O’Connor, D. B., & Shickle, D. (2011). Alcohol, conscientiousness and event-level condom use. British Journal of Health Psychology, 16, 828–845. doi:10.1111/j.2044-8287.20 11.02019.x Hendershot, C. S., Stoner, S. A., George, W. H., & Norris, J. (2007). Alcohol use, expectancies, and sexual sensation seeking as correlates of HIV risk behavior in heterosexual young adults. Psychology of Addictive Behaviors, 21, 365–372. doi:10.1037/0893-164X.21.3.365 Kapadia, F., Latka, M., Wu, Y., Strathdee, S., Mackesy-Amiti, M., Hudson, S., … Garfein, R. (2011). Longitudinal determinants of consistent condom use by partner type among young injection drug users: The role of personal and partner characteristics. AIDS and Behavior, 15, 1309– 1318. doi:10.1007/s10461-009-9569-3 Kapadia, F., Vlahov, D., Donahoe, R. M., & Friedland, G. (2005). The role of substance abuse in HIV disease progression: Reconciling differences from laboratory and epidemiologic investigations. Clinical Infectious Diseases, 41, 1027–1034. doi:10.1086/433175
AIDS Care Karney, B. R., Hops, H., Redding, C. A., Reis, H. T., Rothman, A. J., & Simpson, J. A. (2010). A framework for incorporating dyads in models of HIV-prevention. AIDS Behavior, 14, 189–203. doi:10.1007/s10461-010-9802-0 Koblin, B. A., Bonner, S., Hoover, D. R., Xu, G., Lucy, D., Fortin, P., & Latka, M. H. (2010). A randomized trial of enhanced HIV risk-reduction and vaccine trial education interventions among HIV-negative, high-risk women who use noninjection drugs: The UNITY study. JAIDS Journal of Acquired Immune Deficiency Syndromes, 53, 378–387. doi:10.1097/QAI.0b013e3181b7222e Leigh, B. C., Ames, S. L., & Stacy, A. W. (2008). Alcohol, drugs, and condom use among drug offenders: An eventbased analysis. Drug and Alcohol Dependence, 93(1–2), 38–42. doi:10.1016/j.drugalcdep.2007.08.012 Leigh, B. C., & Stall, R. (1993). Substance use and risky sexual behavior for exposure to HIV: Issues in methodology, interpretation, and prevention. American Psychologist, 48, 1035–1045. doi:10.1037/0003-066X.48.10.1035 Matson, P. A., Chung, S. E., Sander, P., Millstein, S. G., & Ellen, J. M. (2012). The role of feelings of intimacy on perceptions of risk for a sexually transmitted disease and condom use in the sexual relationships of adolescent African-American females. Sexually Transmitted Infections, 88(8), 617–621. doi:10.1136/sextrans-2012050536 Meader, N., Semaan, S., Halton, M., Bhatti, H., Chan, M., Llewellyn, A., & Des Jarlais, D. C. (2013). An international systematic review and meta-analysis of multisession psychosocial interventions compared with educational or minimal interventions on the HIV sex risk behaviors of people who use drugs. AIDS and Behavior, 17(6), 1963–1978. doi:10.1007/s10461-012-0403-y McLellan, A. T., Luborsky, L., Woody, G. E., & O’Brien, C. P. (1980). An improved diagnostic evaluation instrument for substance abuse patients: The Addiction Severity Index. The Journal of Nervous and Mental Disease, 168(1), 26– 33. doi:10.1097/00005053-198001000-00006 Morrison-Beedy, D., Jones, S. H., Xia, Y., Tu, X., Crean, H. F., & Carey, M. P. (2013). Reducing sexual risk behavior in adolescent girls: Results from a randomized controlled trial. Journal of Adolescent Health, 52(3), 314–321. doi:10.1016/j. jadohealth.2012.07.005 Morrow, K. M., Eldridge, G., Nealey-Moore, J., Grinstead, O., & The Project START Study Group. (2007). HIV, STD, and hepatitis risk in the week following release from prison: An event-level analysis. Journal of Correctional Health Care, 13(1), 27–38. doi:10.1177/1078345807299457 Neal, D. J., & Carey, K. B. (2007). Association between alcohol intoxication and alcohol-related problems: An event-level analysis. Psychology of Addictive Behaviors, 21, 194–204. doi:10.1037/0893-164X.21.2.194 Noar, S. M. (2008). Behavioral interventions to reduce HIVrelated sexual risk behavior: Review and synthesis of metaanalytic evidence. AIDS and Behavior, 12, 335–353. doi:10.1007/s10461-007-9313-9 Ober, A., Iguchi, M., Weiss, R., Gorbach, P., Heimer, R., Ouellet, L., … Zule, W. (2011). The relative role of perceived partner risks in promoting condom use in a three-city sample of
399
high-risk, low-income women. AIDS and Behavior, 15, 1347–1358. doi:10.1007/s10461-010-9840-7 Parks, K. A., Hsieh, Y. P., Collins, R. L., & Levonyan-Radloff, K. (2011). Daily assessment of alcohol consumption and condom use with known and casual partners among young female bar drinkers. AIDS and Behavior, 15, 1332–1341. doi:10.1007/s10461-010-9829-2 Pellowski, J. A., Kalichman, S. C., Matthews, K. A., & Adler, N. (2013). A pandemic of the poor. American Psychologist, 68, 197–209. doi:10.1037/a0032694 Peragallo, N., DeForge, B., O’Campo, P., Lee, S. M., Kim, Y. J., Cianelli, R., & Ferrer, L. (2005). A randomized clinical trial of an HIV-risk-reduction intervention among lowincome Latina women. Nursing Research, 54(2), 108–118. doi:10.1097/00006199-200503000-00005 Prendergast, M. L., Urada, D., & Podus, D. (2001). Meta-analysis of HIV risk-reduction interventions within drug abuse treatment programs. Journal of Consulting and Clinical Psychology, 69, 389–405. doi:10.1037/0022-006X.69.3.389 Pulerwitz, J., Amaro, H., De Jong, W., Gortmaker, S. L., & Rudd, R. (2002). Relationship power, condom use and HIV risk among women in the USA. AIDS Care, 14, 789– 800. doi:10.1080/0954012021000031868 Riehman, K. S., Wechsberg, W. M., Francis, S. A., Moore, M., & Morgan-Lopez, A. (2006). Discordance in monogamy beliefs, sexual concurrency, and condom use among young adult substance-involved couples: Implications for risk of sexually transmitted infections. Sexually Transmitted Diseases, 33, 677–682. doi:10.1097/01.olq.0000218882.05426.ef Scott-Sheldon, L. A., Carey, M. P., Vanable, P. A., Senn, T. E., Coury-Doniger, P., & Urban, M. A. (2009). Alcohol consumption, drug use, and condom use among STD clinic patients. Journal of Studies on Alcohol and Drugs, 70, 762–770. Scott-Sheldon, L. A., Huedo-Medina, T. B., Warren, M. R., Johnson, B. T., & Carey, M. P. (2011). Efficacy of behavioral interventions to increase condom use and reduce sexually transmitted infections: A meta-analysis, 1991 to 2010. Journal of Acquired Immune Deficiency Syndromes, 58(5), 489–498. Shoptaw, S., Reback, C. J., Peck, J. A., Yang, X., RotheramFuller, E., Larkins, S., … Hucks-Ortiz, C. (2005). Behavioral treatment approaches for methamphetamine dependence and HIV-related sexual risk behaviors among urban gay and bisexual men. Drug and Alcohol Dependence, 78 (2), 125–134. doi:10.1016/j.drugalcdep.2004.10.004 Sobell, L. C., & Sobell, M. B. (1992). Timeline follow-back: A technique for assessing self-reported alcohol consumption. In R. Z. L. J. P. Allen (Ed.), Measuring alcohol consumption: Psychosocial and biochemical methods (pp. 41–72). Totowa, NJ: Humana Press. Sormanti, M., & Shibusawa, T. (2007). Predictors of condom use and HIV testing among midlife and older women seeking medical services. Journal of Aging and Health, 19, 705–719. doi:10.1177/0898264307301173 Stein, J. A., Nyamathi, A., Ullman, J. B., & Bentler, P. M. (2007). Impact of marriage on HIV/AIDS risk behaviors among impoverished, at-risk couples: A multilevel latent variable approach. AIDS and Behavior, 11(1), 87–98. doi:10.1007/ s10461-005-9058-2
400
N.S. DePesa et al.
Weigel, D. J., Brown, C., & O’Riordan C. K. (2011). Everyday expressions of commitment and relational uncertainty as predictors of relationship quality and stability over time. Communication Reports, 24(1), 38–50. doi:10.1080/ 08934215.2010.511400 Weinhardt, L. S., Carey, M. P., Maisto, S. A., Carey, K. B., Cohen, M. M., & Wickramasinghe, S. M. (1998). Reliability
of the timeline follow-back sexual behavior interview. Annals of Behavioral Medicine, 20(1), 25–30. doi:10.1007/ BF02893805 Witte, S. S., El-Bassel, N., Gilbert, L., Wu, E., & Chang, M. (2010). Lack of awareness of partner STD risk among heterosexual couples. Perspectives on Sexual and Reproductive Health, 42(1), 49–55. doi:10.1363/4204910
Copyright of AIDS Care is the property of Routledge and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
Predictors of condom use in women receiving court-mandated drug and alcohol treatment: implications for intervention Natasha S. DePesaa*, Gloria D. Eldridgeb, Frances Deaversa and Jeffrey E. Cassisia a
Psychology Department, University of Central Florida, Orlando, FL, USA; bPsychology Department, University of Alaska Anchorage, Anchorage, AK, USA (Received 26 February 2014; accepted 15 September 2014) Women who abuse substances are at a high-risk for contracting HIV. Condom use interventions are important in reducing HIV in high-risk populations, but current interventions have small effects. The aim of this study is to examine the relative impact of substance use, personal variables (sexual impulsivity and condom expectancies), and relationship variables (perceptions of relationship commitment and partner risk, perceptions of power within the relationship) on condom use in women in court-mandated substance abuse treatment. Information was collected from 312 sexually active women in an inpatient drug and alcohol treatment facility in the Southeastern US Participants completed questionnaires and were interviewed using the Timeline Follow-back method and provided information about sexual activity in the 30days prior to intake, including type of sexual event, co-occurrence with substance use, condom use, and characteristics of sexual partners and the nature of the relationship. Multilevel logistic modeling revealed that perception of relationship commitment, condom outcome expectancies, and age significantly affected condom use for women in the sample. Specifically, condom use was least likely when women reported that the relationship was committed (odds ratio [OR] = 0.31, 95% confidence interval [CI]: 0.23, 0.43) or when the participant was older (OR = 0.96, 95% CI: 0.94, 0.99), and more likely when women reported more positive condom outcome expectancies (OR = 1.02, 95% CI: 1.00, 1.03). The findings suggest that perceptions of relationship commitment, regardless of perceptions of partner risk, strongly affect condom use among women court-mandated into drug and alcohol treatment. In addition, positive outcome expectancies (e.g., positive self-evaluations and perceived positive partner reactions) are associated with a greater likelihood of condom use. These findings have important implications for condom use interventions, which have failed to produce large or lasting effects within this population. Keywords: condom use; sexual risk behavior; partner commitment; HIV intervention; high-risk women
Individuals who abuse substances often face disadvantages at multiple levels, ranging from societal-level to individuallevel factors (Beyrer, 2007; Pellowski, Kalichman, Matthews, & Adler, 2013). Consequently, they are at greater risk for HIV infection, are less likely to seek and adhere to antiretroviral treatment, and often progress more rapidly to AIDS (Chander, Himelhoch, & Moore, 2006; Kapadia, Vlahov, Donahoe, & Friedland, 2005). Condom use remains one of the primary defenses against HIV transmission. However, meta-analytic reviews of condom use and sexual risk reduction interventions reveal mainly small effect sizes, particularly with individuals who abuse substances (Noar, 2008; Prendergast, Urada, & Podus, 2001). Further, effect sizes have remained relatively stable over time, suggesting limited increases in intervention efficacy (Meader et al., 2013). Focusing on significant moderators of intervention efficacy may enhance the impact of condom use interventions (Peragallo et al., 2005; Scott-Sheldon, Huedo-Medina, Warren, Johnson, & Carey, 2011). Sexual risk reduction and condom use interventions that have been tailored to address the needs of specific *Corresponding author. Email: [email protected] © 2014 Taylor & Francis
groups have shown promising results (Morrison-Beedy et al., 2013; Shoptaw et al., 2005). Women in substance use treatment are at particular risk for HIV (Beyrer, 2007; Brooks et al., 2010), yet condom use interventions in this group have failed to produce large effects (Meader et al., 2013; Prendergast et al., 2001). This suggests that these women may benefit from tailored interventions. Variables that may affect sexual risk behavior in women who abuse substances include factors surrounding the alcohol- or drug-using context, individual differences, and relationship and partner variables. Decreased condom use is clearly linked to the use of alcohol and other substances (Leigh & Stall, 1993). However, when examining individual sexual events, condom use does not appear to vary with concurrent alcohol or substance use (Calsyn, Baldwin, Niu, Crits-Christoph, & Hatch-Maillette, 2011; Hagger-Johnson, Bewick, Conner, O’Connor, & Shickle, 2011; Leigh, Ames, & Stacy, 2008; Scott-Sheldon et al., 2009). As alcohol and substance use do not appear to predict event-level condom use, it seems that other factors may be involved.
AIDS Care Individual factors such as impulsivity and sensation seeking appear to be consistently linked to substance-using behavior, but their link to sexual risk behavior is less straightforward. In one study, impulsivity increased the relationship between alcohol use and risky behavior, while self-regulation abilities had a buffering effect (Neal & Carey, 2007). In another study, both sexual and nonsexual sensation-seeking were related to increased alcohol use before sex and more sexual partners (Hendershot, Stoner, George, & Norris, 2007). Expectancies surrounding the outcomes of condom use have a more straightforward effect. More positive condom outcome expectancies are associated with more consistent condom use among individuals who use injection drugs (Brown et al., 2008; Kapadia et al., 2011). Partner and relationship variables strongly affect condom use in women who abuse substances (Ober et al., 2011; Parks, Hsieh, Collins, & Levonyan-Radloff, 2011). Condom use is more likely with casual or noncommitted partners than with partners perceived to be committed (Ober et al., 2011; Parks et al., 2011). Women who consider themselves to be in committed relationships report lower perceived sexual risk than women who do not (Matson, Chung, Sander, Millstein, & Ellen, 2012; Stein, Nyamathi, Ullman, & Bentler, 2007). However, discordance between perceptions of monogamy and actual partner concurrency are associated with increased rates of STI infection among individuals who misperceive risk (Riehman, Wechsberg, Francis, Moore, & Morgan-Lopez, 2006; Witte, El-Bassel, Gilbert, Wu, & Chang, 2010). Even when risk is perceived, individuals are not likely to use condoms with committed partners (Ober et al., 2011). Finally, among women, more perceived relationship power may be linked to greater consistency of condom use (Pulerwitz et al., 2002). The aim of this paper is to examine substance use, personal variables (sexual impulsivity and condom expectancies), and relationship variables (perceptions of commitment and partner risk, perceptions of power within the relationship) and their impact on condom use in order to synthesize and replicate previous findings. The intent is to inform efforts to tailor HIV prevention efforts for women in substance abuse treatment. The study uses multilevel logistic modeling to examine the simultaneous and relative effects of substance use, individual factors, and perceived relationship and partner characteristics on event-level condom use in women court-ordered into inpatient drug treatment. This type of modeling allows for an integrated and informative view into event-level condom use in a population of women at high risk for HIV-infection. Methods Participants and procedure The sample is comprised of women court-mandated into drug and alcohol treatment at an inpatient chemical
393
dependence unit in a Southeastern state hospital. Data were from a larger study (n = 506) funded by the National Institute on Drug Abuse and designed to test HIV prevention interventions. Women were approached for the study after they completed detoxification and if they were determined to be ready for treatment activities. The project coordinator provided study information to potential participants and obtained informed consent. Participation was voluntary and research protocols were approved by the university and facility Institutional Review Boards. Data were collected prior to intervention, immediately after the intervention and at 3- and 6months follow-up. The present analyses utilized data from the preintervention assessment, during which participants were interviewed about their sexual activity by trained university research staff and provided $10 compensation. Participants were selected for the current analyses if they reported at least one instance of receptive vaginal or anal sex with a male partner in the 30 days prior to entering the drug treatment program. Those who reported being in a controlled environment (e.g., prison, other inpatient treatment) prior to entering treatment reported on the 30-day period prior to being in a controlled environment (see Table 1 for details).
Measures Demographic information Demographic information (e.g., age, race) was obtained using a self-report questionnaire. Addiction Severity Index (ASI) Information about participants’ alcohol and substance use history was obtained using the ASI, a semi-structured interview widely used in drug treatment facilities (Alterman, Cacciola, Habing, & Lynch, 2007; McLellan, Luborsky, Woody, & O’Brien, 1980). Composite scores determined overall severity of alcohol or substance abuse, while individual items provided descriptive information. Comprehensive Timeline Follow-back (TLFB) The TLFB procedure was used to obtain details regarding sexual activity, condom use, and concurrent substance use over the 30 days prior to entering the treatment program or another controlled environment. The TLFB is a calendarbased interview that provides day-by-day information about alcohol and substance use, sexual activity, condom use, perception of relationship commitment with sexual partners, and perceived partner risk (Sobell & Sobell, 1992). Following Morrow, Eldridge, Nealey-Moore, Grinstead, and The Project START Study Group (2007), a partner was defined as “risky” if the participant believed the partner to be HIV+, to have ever injected drugs, to have smoked crack cocaine, or was someone with
394
N.S. DePesa et al.
Table 1. Sample characteristics. n (%) Race Caucasian African-American/Black Other Marital status Single/never married Married Separated/divorced/widowed Education Less than high school High school/GED Beyond high school In a controlled environment prior to treatment M = 9.05 days (SD = 7.00) Ever incarcerated (at least one month) M = 13.47 months (SD = 26.32) Ever charged with prostitution Prior alcohol use treatment M = 3.73 attempts (SD = 3.92) Prior substance use treatment M = 3.18 attempts (SD = 3.39) Prior HIV education Prior HIV testing Positive results Negative results Unsure of results
186 (59.6) 118 (37.8) 8 (2.6) 112 (35.9) 50 (16.0) 150 (48.1) 104 123 85 119
(33.3) (39.4) (27.3) (38.1)
90 (28.8) 10 (3.2) 133 (42.6) 192 (61.5) 77 (24.7) 246 (94.9) 10 (3.2) 222 (71.1) 64 (20.5)
Note: Total n = 312. M indicates mean; SD indicates standard deviation. A controlled environment refers to prison or inpatient treatment other than the current treatment.
whom the participant had exchanged sex for money, drugs, or other things. Previous research has established the TLFB procedure to provide reliable and valid data about sexual behavior and substance use for up to a three month reporting period (Carey et al., 2001; Weinhardt et al., 1998). Sexual Impulsivity Scale (SIS) The SIS is a 15-item self-report scale assessing sexual impulsivity and behavioral outcomes (e.g., “Sometimes I get overexcited sexually and get myself into trouble”) that was developed for this study. Each item was rated on a 5-point scale ranging from 1 (Not true of me) to 5 (Very true of me). Negatively worded items were reverse coded so that a higher scale score represents greater sexual impulsivity. Cronbach’s alpha for the current sample was .81. Relationship Power Scale (RPS) The RPS contains eight self-report items developed for the study that assess the extent to which the participant or her partner controls sexual and relationship decisions (e.g.,
“Who decides whether or not to use a condom?”; “Who makes more decisions in the relationship?”). Responses ranged from 1 (He always does) to 5 (I always do). Cronbach’s alpha was .84. A higher score indicates that the participant makes the majority of the decisions within the relationship while a lower score indicates that her partner makes the majority of the relationship decisions. Condom Outcome Expectancy Scale (COES) The COES is a 29-item scale that reliably assesses attitudes and beliefs about condom use (Dilorio, Maibach, O’Leary, Sanderson, & Celentano, 1997). The scale contains six factors: partner reaction, hedonism, prevention efficacy, social approval, positive self-evaluation, and negative selfevaluation. An example item includes, “Saying we have to use a condom or barrier is like saying ‘I don’t trust you’” (reverse scored). Response choices ranged from 1 (Strongly Disagree) to 5 (Strongly Agree). Negatively worded items were reverse coded so that a higher scale score represents more positive expectancies surrounding the use of condoms.
Analysis Level-1 data were derived from reports on sexual activities with up to three males identified by each participant as her most frequent sex partners in the previous 30 days. Data were organized on the event-level, such that each unit of analysis represents one day where sexual activity (i.e., at least one instance of either vaginal or anal sex) occurred along with relevant contextual information (e.g., condom use, co-occurring substance use, partner and relationship characteristics). If a participant reported sexual activity with more than one partner in a single day, separate lines were included for each partner. If a participant reported sexual activity with more than three male partners over the recall period, event-level data were included in the main analyses for the three most frequent partners and summed data (e.g., total number of instances of protected or unprotected vaginal intercourse) were collected for the remaining partners. Level-2 data were organized on the individual level and consisted of demographic variables (i.e., age and race), total number of partners in the last 30 days, and scale scores derived from self-report measures (i.e., SIS, RPS, and COES). The data were analyzed using random effects multilevel logistic regression modeling. The numbers of sexual events at level-1 were nested within the women who reported them at level-2, allowing for simultaneous testing of individual-level and event-level effects on condom use. Analyses were conducted in several stages using the build-up method described by Bryk and Raudenbush (1992). First, level-1 predictors of condom use were tested individually for their potential to be included in
AIDS Care
395
however, when included in the multivariate model, the impact of perceived partner risk was no longer significant. The interaction of perceived partner risk and relationship commitment was significant; however, the Wald test indicated poor model fit and the interaction was excluded from further analyses. Only perceived relationship commitment was included in the final level-1 model. To illustrate the level-1 effects, Figure 1 highlights the frequency of condom use across partner risk and relationship commitment.
the final level-1 model. Significant univariate predictors were then included in a multivariate model. Predictors that did not affect the likelihood of condom use were excluded and a final level-1 model was fit with the remaining variables. Next, level-2 predictors of condom use were individually added to the final level-1 model to test for their potential inclusion in the final multilevel model. Significant predictors of condom use were then fit into one final multivariate, multilevel model.
Results
Level-2 modeling Age and condom outcome expectancies incrementally affected the likelihood of condom use such that younger age and more positive expectancies were associated with an increase in the likelihood of condom use. Number of partners in the last 30 days, sexual impulsivity, relationship power, and none of the interaction terms were found to affect the likelihood of condom use.
Description of participants and sexual activity The sample was comprised of 312 women (see Table 1 for sample characteristics). Patterns of recent and lifetime substance use, by substance type, across all participants are shown in Table 2. A total of 3146 instances of sexual activity with a male partner (i.e., days with at least one instance of vaginal or anal intercourse) were analyzed. Participants reported between 1 and 30 days with sexual activity in the 30-day reporting period (M = 10.08, SD = 8.98). Fifty women reported more than three male partners in the preceding 30 days, 7 of whom (14.0%) reported 10 or more additional partners. Thirty of the 50 women (60.0%) reported consistent condom use, 6 (12.0%) reported no condom use, and 11 (22.0%) reported occasional condom use with these additional partners. Descriptive information for level-1 and -2 variables can be found in Tables 3 and 4, respectively.
Final integrated model Overall, condom use was relatively unlikely in this sample of women (odds ratio [OR] = 0.56, 95% confidence interval [CI]: 0.43, 0.73). The final multivariate, multilevel model contained perceived relationship commitment (level-1) and age and condom outcome expectancies (level-2). Condom use was less likely in committed relationships (OR = 0.31, 95% CI: 0.23, 0.43) and with older participants (OR = 0.96, 95% CI: 0.94, 0.99) and more likely when women had more positive condom outcome expectancies (OR = 1.02, 95% CI: 1.00, 1.03). An exploratory cross-level interaction between commitment and expectancies was not significant (results not shown), indicating that the impact of condom outcome expectancies did not vary by perceived relationship commitment. All modeling results are displayed in Table 5.
Level-1 modeling Substance use concurrent with sexual activity did not significantly affect the likelihood of condom use. Perceived relationship commitment and perceived partner risk, respectively, significantly decreased and increased the likelihood of condom use in the univariate analyses;
Table 2. Recent and lifetime substance use by substance type. Recent use (days) Substance Alcohol (felt effects) Cannabis Cocaine Opiates/analgesics, heroin, methadone Barbiturates, sedatives/hypnotics/tranquilizers Other substances >1 substance (including alcohol)
n (%) 199 170 197 60 66 59 222
(63.4) (54.1) (62.7) (19.1) (21.0) (18.8) (70.7)
Lifetime use (years)
M (SD) 15.2 12.1 15.3 5.0 6.8 3.6 13.5
(11.2) (11.6) (11.1) (4.1) (6.7) (3.6) (10.9)
n (%) 210 202 217 69 73 86 243
(66.9) (64.3) (69.1) (22.0) (23.2) (27.4) (77.4)
M (SD) 11.2 10.0 5.9 1.9 2.8 1.2 7.7
(7.0) (6.9) (4.8) (1.8) (3.8) (1.3) (5.8)
Note: Information was obtained from the ASI. Recent use refers to use over the 30-day reporting period. Lifetime use refers to years of regular or problematic use. n indicates the number of individuals who reported at least one day of recent substance use (recent use) or the number who reported at least one year of regular or problematic use (lifetime use). M indicates the mean; SD indicates the standard deviation. Alcohol (felt effects) refers to alcohol consumption to the point of some impairment. No significant differences between racial groups were found in the alcohol or drug composites.
396
N.S. DePesa et al.
Table 3. Level-1 descriptive statistics.
Total no. of events Condom used (outcome) Committed partner Risky partner Exchange partner Concurrent substance use
Vaginal sex n (%)
Anal sex n (%)
3144 472 2197 1970 605 1860
22 2 9 20 4 19
(100) (15.0) (69.9) (62.7) (19.2) (59.2)
(100) (9.1) (40.9) (90.9) (18.2) (86.4)
Note: Analyses were conducted across vaginal and anal sexual events. Risky partners are those that participants perceived to be HIV+, to have injected drugs, to have smoked crack cocaine, or was a person with whom sex has been exchanged for money, drugs, or other things. Exchange partners were not included as an individual predictor.
Discussion This study integrates and confirms previous research while providing a vigorous multi-level examination of event-level condom use among women who abuse substances. Given the limited efficacy of condom use interventions in this population, this information is particularly useful for informing future interventions. Results indicated that perceived partner commitment, negative condom outcome expectancies, and older age were all associated with a lower likelihood of condom use. Strikingly, condom use occurred in only 15.0% of sexual events overall. As expected, perceptions of relationship commitment had the largest impact on condom use such that women were less likely to use condoms in a relationship they defined to be “committed,” despite perceptions of partner characteristics deemed “risky” (Ober et al., 2011). During sexual activity with partners perceived to have risky characteristics, condom use occurred 32.2% of the time when the woman defined the relationship as not committed but only 6.9% of the time when the relationship was defined as committed (Figure 1). This dramatic difference may result from inaccurate perceptions of risk or discounted importance of risk (Riehman et al., 2006; Witte et al., 2010).
Table 4. Level-2 descriptive statistics. M (SD) Age Partners in last 30 days Relationship Power Scalea Condom Outcome Expectancies Scaleb Sexual Impulsivity Scalec
32.15 4.2 26.06 91.56 37.87
(7.63) (14.9) (5.81) (15.47) (11.53)
a Range is 8–40; higher scores indicate more relationship power. bRange is 29–145; higher scores indicate more positive expectancies. cRange is 15–75; higher scores indicate greater impulsivity. Mean-centered scale scores were used in the analyses in order to facilitate the interpretation of the resulting odds ratios.
Figure 1. Frequency of condom use during sexual events by partner type. Partners with whom the women identified themselves as being in a committed relationship are labeled as “committed” while others are labeled “noncommitted.” Partners who were perceived to have one or more defined risk characteristics were labeled “risky” while those perceived to have none of the risk characteristics were labeled “nonrisky.” n = number of sexual events with each partner type; the total number of sexual events for all partner types was 3141. The * represents the significant main effect of perceived partner commitment on condom use found in level-1 modeling and included in the final multilevel model.
Results from this and other studies suggest the importance of addressing expectancies surrounding the perceived impact of using condoms in all types of relationships (Brown et al., 2008; Corbett, Dickson-Gómez, Hilario, & Weeks, 2009). The small effect of COES unit increases (OR = 1.02) suggests that small variations in expectancies may not significantly affect outcomes, but that large differences may more noticeably impact behavior. The elevated risk of transmission during anal sexual activities implicates its importance in HIV interventions. Limited data on anal sexual events (n = 22) prevent formal analysis, but it is noteworthy that anal sexual events occurred more frequently with noncommitted and risky partners and involved condoms less frequently compared to vaginal sexual events (Table 3). Finally, though no a-priori hypotheses were made in regards to age, older participants were less likely to use condoms than their younger counterparts, consistent with prior literature (Sormanti & Shibusawa, 2007). As the majority (69.9%) of sexual activity in this study occurred in relationships perceived by the women to be committed, consideration of the nature of the relationship is critical in designing effective condom use interventions. Attempts so far have produced limited success (Meader et al., 2013). A recent randomized trial found that an enhanced sexual risk-reduction and a standard sexual risk-reduction intervention for women who used noninjection drugs resulted in similar
AIDS Care
397
Table 5. Multilevel modeling results.
Level-1 univariate analyses
Level-1 multivariate analyses
Level-1 final model Level-2 univariate analyses
Final fitted multilevel model
Parameter
Wald Chi2
OR
CI
Intercept Substance use Commitment*** Risk* Risk × Commitment Intercept Commitment*** Risk Intercept Commitment*** Sexual impulsivity Relationship Power Expectancies** Impulsivity × Power Impulsivity × Expectancies Power × Expectancies Age** Race Partners in last 30 days Intercept Level-1 Commitment*** Level-2 Expectancies** Age**
1348.54 96.79 131.44 131.85 46.22 10.91 17.42 0.10 250 131.44
253.12
0.25 1.07 0.30 1.44 0.47 0.56 0.33 1.10 0.59 0.30 1.00 1.03 1.02 0.999 0.999 1.00 0.97 1.10 1.04 0.56
[0.20, 0.31] [0.83, 1.38] [0.22, 0.41] [1.04, 1.98] [0.33, 0.66] [0.39, 0.79] [0.24, 0.45] [0.81, 1.51] [0.45, 0.76] [0.22, 0.41] [0.99, 1.02] [1.00, 1.06] [1.00, 1.03] [0.996, 1.00] [0.998, 1.00] [0.999, 1.00] [0.94, 0.99] [0.89, 1.37] [0.98, 1.10] [0.43, 0.73]
133.30
0.31
[0.23, 0.43]
1.02 0.96
[1.00, 1.03] [0.94, 0.99]
* = p < .05, ** = p < .01, *** = p < .001. Note: Commitment indicates perceived relationship commitment. Risk indicates perceived partner risk. Expectancies indicates condom outcome expectancies. Level-2 univariate analyses were conducted along with the final level-1 model. OR, odds ratio. CI, confidence interval.
reductions in unprotected sex (Koblin et al., 2010). Although the enhanced intervention addressed condom negotiation in various relationship contexts, changes in condom use were most likely to occur during sexual activity with casual and exchange partners, providing further evidence that current intervention efforts are more effective for women in these types of relationships than in those they define as committed. Tailoring condom use interventions to address the needs of women who define their relationships as committed necessitates a better understanding of these needs. For some women, perceptions of commitment may be inaccurate or may fluctuate over time, potentially introducing a source of risk (Arriaga, Reed, Goodfriend, & Agnew, 2006; Drumright, Gorbach, & Holmes, 2004; Matson et al., 2012; Witte et al., 2010). Other women may recognize risks but place a higher priority on the emotional needs that are satisfied by the relationship and consider unprotected sex to be “worth the risk” (Corbett et al., 2009). Condoms may be viewed as potential threats to the relationship. Although condom negotiation strategies are part of many HIV-risk reduction interventions, this has not produced consistent behavior
changes among individuals in committed relationships (Meader et al., 2013). One potential limitation of this strategy is that it does not fully address the issue of commitment and trust and assumes that women have the appropriate tools for general communication within a relationship. Showing respect, providing support, and solving problems together, among other behaviors, indicate commitment and predict relationship quality and stability (Weigel, Brown, & O’Riordan, 2011). General and open discussion of relationships, signs of commitment, trust, and respect may provide a context that facilitates adoption of health-positive messages about condoms (Karney et al., 2010). Further, providing women with general communication skills might help them clarify the nature of their relationships (e.g., intentions of exclusivity) and improve their ability to approach difficult topics with partners (e.g., STI testing or condom use; Karney et al., 2010). While the current results provide useful information, a few limitations merit acknowledgment. The use of selfreport data introduces potential bias and inaccuracy. Also, although the TLFB is considered reliable for up to a 3-month recall period, part of the sample was in a
398
N.S. DePesa et al.
controlled environment prior to assessment and reported on a slightly different time period that the other participants. In addition, some participants reported sexual activity with more than three partners in the 30-day reporting period and results might have varied slightly had event-level data concerning additional partners been included in the analyses. Further, scales created for the study proved to be reliable; however, full psychometric analyses were not conducted, limiting the conclusions that can be drawn. These scales are available upon request. Finally, indirect effects and other variables that may impact condom use (e.g., previous STI) were not examined. The present results, along with a review of the existing literature, suggest avenues for strengthening condom use interventions for women in substance use treatment. In particular, increasing condom use within committed relationships appears to be a difficult but important hurdle. Given the consistent lack of success in condom use interventions among this population of women, researchers should examine the efficacy of focusing on risk reduction within committed relationships and enhancing positive condom expectancies along with other factors (e.g., couplelevel interventions; El-Bassel et al., 2010; Karney et al., 2010; Meader et al., 2013). Funding This work was supported by the National Institute on Drug Abuse under [grant number R01 DA10418].
References Alterman, A. I., Cacciola, J. S., Habing, B., & Lynch, K. G. (2007). Addiction severity index recent and lifetime summary indexes based on nonparametric item response theory methods. Psychological Assessment, 19(1), 119–132. doi:10.1037/1040-3590.19.1.119 Arriaga, X. B., Reed, J. T., Goodfriend, W., & Agnew, C. R. (2006) Relationship perceptions and persistence: Do fluctuations in perceived partner commitment undermine dating relationships? Journal of Personality and Social Psychology, 91, 1045–1065. doi:10.1037/0022-3514.91.6.1045 Beyrer, C. (2007). HIV epidemiology update and transmission factors: Risks and risk contexts–16th International AIDS Conference epidemiology plenary. Clinical Infectious Diseases, 44, 981–987. doi:10.1086/512371 Brooks, A., Meade, C. S., Potter, J. S., Lokhnygina, Y., Calsyn, D. A., & Greenfield, S. F. (2010). Gender differences in the rates and correlates of HIV risk behaviors among drug abusers. Substance Use & Misuse, 45, 2444–2469. doi:10.3109/ 10826084.2010.490928 Brown, L. K., DiClemente, R., Crosby, R., Fernandez, M. I., Pugatch, D., Cohn, S., … Schlenger, W. E. (2008). Condom use among high-risk adolescents: Anticipation of partner disapproval and less pleasure associated with not using condoms. Public Health Reports, 123, 601–607. Bryk, A., & Raudenbush, S. (1992). Hierarchical Linear Models. Newbury Park, CA: Sage.
Calsyn, D. A., Baldwin, H., Niu, X., Crits-Christoph, P., & Hatch-Maillette, M. A. (2011). Sexual risk behavior and sex under the influence: An event analysis of men in substance abuse treatment who have sex with women. The American Journal on Addictions, 20, 250–256. doi:10.1111/j.15210391.2011.00123.x Carey, M. P., Carey, K. B., Maisto, S. A., Gordon, C. M., & Weinhardt, L. S. (2001). Assessing sexual risk behaviour with the Timeline Followback (TLFB) approach: Continued development and psychometric evaluation with psychiatric outpatients. International Journal of STDs and AIDS, 12, 365–375. doi:10.1258/0956462011923309 Chander, G., Himelhoch, S., & Moore, R. D. (2006). Substance abuse and psychiatric disorders in HIV-positive patients: Epidemiology and impact on antiretroviral therapy. Drugs, 66, 769–789. doi:10.2165/00003495-200 666060-00004 Corbett, A. M., Dickson-Gómez, J., Hilario, H., & Weeks, M. R. (2009). A little thing called love: Condom use in high-risk primary heterosexual relationships. Perspectives on Sexual and Reproductive Health, 41, 218–224. doi:10.1363/41 21809 Dilorio, C., Maibach, E., O’Leary, A., Sanderson, C. A., & Celentano, D. (1997). Measurement of condom use selfefficacy and outcome expectancies in a geographically diverse group of STD patients. AIDS Education and Prevention, 9(1), 1–13. Drumright, L. N., Gorbach, P. M., & Holmes, K. K. (2004). Do people really know their sex partners? Concurrency, knowledge of partner behavior, and sexually transmitted infections within partnerships. Sexually Transmitted Diseases, 31, 437– 442. doi:10.1097/01.OLQ.0000129949.30114.37 El-Bassel, N., Jemmott, J. B., Landis, J. R., Pequegnat, W., Wingood, G. M., Wyatt, G. E., & Bellamy, S. L. (2010). National Institute of Mental Health multisite Eban HIV/STD prevention intervention for African American HIV serodiscordant couples: A cluster randomized trial. Archives of Internal Medicine, 170, 1594–1601. doi:10.1001/archinternmed.2010.261 Hagger-Johnson, G., Bewick, B. M., Conner, M., O’Connor, D. B., & Shickle, D. (2011). Alcohol, conscientiousness and event-level condom use. British Journal of Health Psychology, 16, 828–845. doi:10.1111/j.2044-8287.20 11.02019.x Hendershot, C. S., Stoner, S. A., George, W. H., & Norris, J. (2007). Alcohol use, expectancies, and sexual sensation seeking as correlates of HIV risk behavior in heterosexual young adults. Psychology of Addictive Behaviors, 21, 365–372. doi:10.1037/0893-164X.21.3.365 Kapadia, F., Latka, M., Wu, Y., Strathdee, S., Mackesy-Amiti, M., Hudson, S., … Garfein, R. (2011). Longitudinal determinants of consistent condom use by partner type among young injection drug users: The role of personal and partner characteristics. AIDS and Behavior, 15, 1309– 1318. doi:10.1007/s10461-009-9569-3 Kapadia, F., Vlahov, D., Donahoe, R. M., & Friedland, G. (2005). The role of substance abuse in HIV disease progression: Reconciling differences from laboratory and epidemiologic investigations. Clinical Infectious Diseases, 41, 1027–1034. doi:10.1086/433175
AIDS Care Karney, B. R., Hops, H., Redding, C. A., Reis, H. T., Rothman, A. J., & Simpson, J. A. (2010). A framework for incorporating dyads in models of HIV-prevention. AIDS Behavior, 14, 189–203. doi:10.1007/s10461-010-9802-0 Koblin, B. A., Bonner, S., Hoover, D. R., Xu, G., Lucy, D., Fortin, P., & Latka, M. H. (2010). A randomized trial of enhanced HIV risk-reduction and vaccine trial education interventions among HIV-negative, high-risk women who use noninjection drugs: The UNITY study. JAIDS Journal of Acquired Immune Deficiency Syndromes, 53, 378–387. doi:10.1097/QAI.0b013e3181b7222e Leigh, B. C., Ames, S. L., & Stacy, A. W. (2008). Alcohol, drugs, and condom use among drug offenders: An eventbased analysis. Drug and Alcohol Dependence, 93(1–2), 38–42. doi:10.1016/j.drugalcdep.2007.08.012 Leigh, B. C., & Stall, R. (1993). Substance use and risky sexual behavior for exposure to HIV: Issues in methodology, interpretation, and prevention. American Psychologist, 48, 1035–1045. doi:10.1037/0003-066X.48.10.1035 Matson, P. A., Chung, S. E., Sander, P., Millstein, S. G., & Ellen, J. M. (2012). The role of feelings of intimacy on perceptions of risk for a sexually transmitted disease and condom use in the sexual relationships of adolescent African-American females. Sexually Transmitted Infections, 88(8), 617–621. doi:10.1136/sextrans-2012050536 Meader, N., Semaan, S., Halton, M., Bhatti, H., Chan, M., Llewellyn, A., & Des Jarlais, D. C. (2013). An international systematic review and meta-analysis of multisession psychosocial interventions compared with educational or minimal interventions on the HIV sex risk behaviors of people who use drugs. AIDS and Behavior, 17(6), 1963–1978. doi:10.1007/s10461-012-0403-y McLellan, A. T., Luborsky, L., Woody, G. E., & O’Brien, C. P. (1980). An improved diagnostic evaluation instrument for substance abuse patients: The Addiction Severity Index. The Journal of Nervous and Mental Disease, 168(1), 26– 33. doi:10.1097/00005053-198001000-00006 Morrison-Beedy, D., Jones, S. H., Xia, Y., Tu, X., Crean, H. F., & Carey, M. P. (2013). Reducing sexual risk behavior in adolescent girls: Results from a randomized controlled trial. Journal of Adolescent Health, 52(3), 314–321. doi:10.1016/j. jadohealth.2012.07.005 Morrow, K. M., Eldridge, G., Nealey-Moore, J., Grinstead, O., & The Project START Study Group. (2007). HIV, STD, and hepatitis risk in the week following release from prison: An event-level analysis. Journal of Correctional Health Care, 13(1), 27–38. doi:10.1177/1078345807299457 Neal, D. J., & Carey, K. B. (2007). Association between alcohol intoxication and alcohol-related problems: An event-level analysis. Psychology of Addictive Behaviors, 21, 194–204. doi:10.1037/0893-164X.21.2.194 Noar, S. M. (2008). Behavioral interventions to reduce HIVrelated sexual risk behavior: Review and synthesis of metaanalytic evidence. AIDS and Behavior, 12, 335–353. doi:10.1007/s10461-007-9313-9 Ober, A., Iguchi, M., Weiss, R., Gorbach, P., Heimer, R., Ouellet, L., … Zule, W. (2011). The relative role of perceived partner risks in promoting condom use in a three-city sample of
399
high-risk, low-income women. AIDS and Behavior, 15, 1347–1358. doi:10.1007/s10461-010-9840-7 Parks, K. A., Hsieh, Y. P., Collins, R. L., & Levonyan-Radloff, K. (2011). Daily assessment of alcohol consumption and condom use with known and casual partners among young female bar drinkers. AIDS and Behavior, 15, 1332–1341. doi:10.1007/s10461-010-9829-2 Pellowski, J. A., Kalichman, S. C., Matthews, K. A., & Adler, N. (2013). A pandemic of the poor. American Psychologist, 68, 197–209. doi:10.1037/a0032694 Peragallo, N., DeForge, B., O’Campo, P., Lee, S. M., Kim, Y. J., Cianelli, R., & Ferrer, L. (2005). A randomized clinical trial of an HIV-risk-reduction intervention among lowincome Latina women. Nursing Research, 54(2), 108–118. doi:10.1097/00006199-200503000-00005 Prendergast, M. L., Urada, D., & Podus, D. (2001). Meta-analysis of HIV risk-reduction interventions within drug abuse treatment programs. Journal of Consulting and Clinical Psychology, 69, 389–405. doi:10.1037/0022-006X.69.3.389 Pulerwitz, J., Amaro, H., De Jong, W., Gortmaker, S. L., & Rudd, R. (2002). Relationship power, condom use and HIV risk among women in the USA. AIDS Care, 14, 789– 800. doi:10.1080/0954012021000031868 Riehman, K. S., Wechsberg, W. M., Francis, S. A., Moore, M., & Morgan-Lopez, A. (2006). Discordance in monogamy beliefs, sexual concurrency, and condom use among young adult substance-involved couples: Implications for risk of sexually transmitted infections. Sexually Transmitted Diseases, 33, 677–682. doi:10.1097/01.olq.0000218882.05426.ef Scott-Sheldon, L. A., Carey, M. P., Vanable, P. A., Senn, T. E., Coury-Doniger, P., & Urban, M. A. (2009). Alcohol consumption, drug use, and condom use among STD clinic patients. Journal of Studies on Alcohol and Drugs, 70, 762–770. Scott-Sheldon, L. A., Huedo-Medina, T. B., Warren, M. R., Johnson, B. T., & Carey, M. P. (2011). Efficacy of behavioral interventions to increase condom use and reduce sexually transmitted infections: A meta-analysis, 1991 to 2010. Journal of Acquired Immune Deficiency Syndromes, 58(5), 489–498. Shoptaw, S., Reback, C. J., Peck, J. A., Yang, X., RotheramFuller, E., Larkins, S., … Hucks-Ortiz, C. (2005). Behavioral treatment approaches for methamphetamine dependence and HIV-related sexual risk behaviors among urban gay and bisexual men. Drug and Alcohol Dependence, 78 (2), 125–134. doi:10.1016/j.drugalcdep.2004.10.004 Sobell, L. C., & Sobell, M. B. (1992). Timeline follow-back: A technique for assessing self-reported alcohol consumption. In R. Z. L. J. P. Allen (Ed.), Measuring alcohol consumption: Psychosocial and biochemical methods (pp. 41–72). Totowa, NJ: Humana Press. Sormanti, M., & Shibusawa, T. (2007). Predictors of condom use and HIV testing among midlife and older women seeking medical services. Journal of Aging and Health, 19, 705–719. doi:10.1177/0898264307301173 Stein, J. A., Nyamathi, A., Ullman, J. B., & Bentler, P. M. (2007). Impact of marriage on HIV/AIDS risk behaviors among impoverished, at-risk couples: A multilevel latent variable approach. AIDS and Behavior, 11(1), 87–98. doi:10.1007/ s10461-005-9058-2
400
N.S. DePesa et al.
Weigel, D. J., Brown, C., & O’Riordan C. K. (2011). Everyday expressions of commitment and relational uncertainty as predictors of relationship quality and stability over time. Communication Reports, 24(1), 38–50. doi:10.1080/ 08934215.2010.511400 Weinhardt, L. S., Carey, M. P., Maisto, S. A., Carey, K. B., Cohen, M. M., & Wickramasinghe, S. M. (1998). Reliability
of the timeline follow-back sexual behavior interview. Annals of Behavioral Medicine, 20(1), 25–30. doi:10.1007/ BF02893805 Witte, S. S., El-Bassel, N., Gilbert, L., Wu, E., & Chang, M. (2010). Lack of awareness of partner STD risk among heterosexual couples. Perspectives on Sexual and Reproductive Health, 42(1), 49–55. doi:10.1363/4204910
Copyright of AIDS Care is the property of Routledge and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
