TIME SINCE RELEASE FROM INCARCERATION AND HIV RISK BEHAVIORS AMONG WOMEN: THE POTENTIAL PROTECTIVE ROLE OF COMMITTED PARTNERS DURING RE ENTRY By Lauren E. Hearn A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2014
2014 LAUREN E. HEARN
To my family
4 ACKNOWLEDGMENTS I would like to thank my mentor, Dr. Nicole Ennis Whitehead, for her support and guidance on this project. I would also like to thank Dr. William Latimer for the invitation to collaborate on the NEURO HIV study and Dr. Maria Khan for her contributions to this project. Further, I would li ke to recognize the members of my supervisory committee: Dr. Deidre Pereira, Dr. Catherine Price, and Dr. Brenda Wiens. This study was supported by Award Number R01DA14498 from the National Institute on Drug Abuse (William W. Latimer, Ph.D., M.P.H., PI).
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST OF ABBREVIATIONS ................................ ................................ ............................. 9 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 12 Re entry Stressors ................................ ................................ ................................ .. 13 The Stress Buffering Model of Social Support ................................ ........................ 15 Social Support in the Context of Female Offende rs ................................ ................ 15 The Stress Buffering Model of Social Support ................................ ........................ 17 2 METHODS ................................ ................................ ................................ .............. 20 Introduction ................................ ................................ ................................ ............. 20 Study Population ................................ ................................ ................................ ..... 20 Measures ................................ ................................ ................................ ................ 21 HIV risk behavior interview ................................ ................................ ............... 21 Measure of incarceration ................................ ................................ .................. 21 Measure of partnership status ................................ ................................ .......... 22 Measure of past month drug use ................................ ................................ ...... 22 Measure of past month sexual risk behaviors ................................ .................. 23 Sociodemographic and other potential confounding variables ......................... 23 Statistical Analysis ................................ ................................ ................................ .. 24 3 RESULTS ................................ ................................ ................................ ............... 25 Sociodemographics ................................ ................................ ................................ 25 Incarceration History ................................ ................................ ............................... 25 Associations between incarcerat ion history and past month drug use .................... 25 Associations between incarceration history and past month sexual risk behavior .. 26 Interaction with partnership status ................................ ................................ .......... 26 4 DISCUSSION ................................ ................................ ................................ ......... 30 LIST OF REFERENCES ................................ ................................ ............................... 36
6 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 44
7 LIST OF TABLES Table page 3 1 Demographic, drug use, and sexual risk behavior comparisons by incarceration history ................................ ................................ ........................... 27 3 2 arceration history & HIV related drug use in past month, by committed partnership status ............... 28
8 LIST OF FIGURES Figure page 1 1 Stress buffering mode l of social ties and mental health. ................................ .. 19
9 LIST OF ABBREVIATIONS AOR Adjusted Odds Ratio CI Confidence Interval GED General Educational Development OR Odds Ratio SNAP Supplemental Nutrition Assistance Program TANF Te mporary Assistance for Needy Families
10 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science TIME SINCE RELEASE FROM INCARCERATI ON AND HIV RISK BEHAVIORS AMONG WOMEN: THE POTENTIAL PROTECTIVE ROLE OF COMMITTED PARTNERS DURING RE ENTRY By Lauren E. Hearn May 2014 Chair: Nicole Ennis Whitehead Major: Psychology After release from incarceration, former female inmates face considera ble stressors, which may influence drug use and other risk behaviors that increase risk for HIV infection. Involvement in a committed partnership may protect women against the re entry stressors that may lead to risky behaviors. The present study investiga ted the association between time since release from incarceration and HIV risk behaviors and explored whether these associations differed among women with versus without committed partners. Multivariable logistic regression measured adjusted odds ratios (A ORs) for the associations between time since release from inca rceration (1 6 months ago, and greater than 6 months ago versus never incarcerated (the referent)) and HIV related drug use (past month binge drinking, crack cocaine smoking, injection drug use) and sexual risk behaviors (condom use, transactional sex, drug and alcohol use before or during sex). Women released within the past 6 months were significantly more likely to have smoked crack cocaine, used injection drugs and engaged in transactional se x in the past month compared to never incarcerated women and women released more distally. Stratified analyses indicated that incarceration within the past 6 months was
11 associated with crack cocaine smoking, injection drug use and transactional sex among w omen without a committed partner yet unassociated with these risk behaviors among those with a committed partner. While the first months following release from incarceration represent a high risk period for HIV related risk behaviors, involvement in a comm itted partnership may protect against these behaviors following release.
12 CHAPTER 1 INTRODUCTION Women with an incarceration history represent a vulnerable population for HIV infection, with HIV rates five to fifteen times that of the general female popula tion [ 1 ]. Drug incarceration [ 2, 3 ]. In addition to driving incarceration rates among women, drug use and sexual risk behaviors are strong determinants of HIV infection [ 4 7 ]. Among individuals with an incarceration history, most HIV infections ar e acquired in the community rather than in jail and prison settings [ 8 ]. Given that many incarcerated women are in jail or prison as a result of drug use and/or drug trade, it is not surprising that high levels of drug use have been observed among former female inmates following release from incarceration. Extant literature shows that half of all female inmates report drug use or alcohol intoxicati on within ten months of release [ 9 ]. Additionally, high rates of repeated arrests have been reported among women who engage in transactional sex [ 10 incarceration are well documented, the HIV risk behaviors of former female inmates in the community remain poorly understood. While there is evidence that time since release from incarceration may influence HIV risk behaviors, these studies have been primarily conducted in men. Some studies suggest that the first weeks and months in the community are characte rized by high levels of substance use and risky sexual behavior followed by a decline in prevalence, while other studies suggest that the prevalence of these behaviors may remain steady or increase in the months after release [ 9, 11, 12 ]. There is a need for additional research on the time frame when risk of engaging in HIV related risk behaviors is
13 greatest after release from incarceration and the degree to which incarceration itself may increase risk of these beha viors among women. Re entry Stressors Upon returning to the community, former female inmates face considerable stressors that may contribute to HIV related drug use and sexual risk behaviors. Following release from incarceration, former inmates have immedi ate needs for food, clothing, safe housing and medical care [ 13 ]. While some states h ave begun implementing evidence based transition programs for inmates leaving prison, release from jail is generally very unpredictable, often precluding appropriate planning and transition services [ 14, 15 ]. Among women who used drugs prior to incarceration, few women enter or maintain participation in treatment programs after release [ 16, 17 ]. In addition to the need for substance use treatment, estimates suggest that 60% to 95% of female inmates have experienced prior physical or sexual abuse or a traumatic event, and approximately one third of female inmates suffer from a serious mental illness [ 18 21 ]. Despite recognizing need fo r treatment, individuals re entering the community unsurprisingly often prioritize their subsistence needs and financial needs above seeking substance use or mental health treatment [ 15 ]. Achieving stable employme nt is viewed as critical to not only allow individuals to meet basic needs and pay for legal expenses, but also to increase self efficacy and to form new, prosocial relationships [ 22, 23 ]. However job prospects of former female inmates are often diminished by limited work experience and low educational attainment; slightly more than half of female inmates have completed high school or the General Education Development (GED) testing equivalent [ 24, 25 ]. In addition, employer reluctance to hire individuals with a criminal record may create particular
14 challenges for those with drug related offenses, as former female inmates with substance abuse problems face greater challenges to finding employment than former inmates c onvicted of other offenses . Federal assistance programs such as Temporary Assistance for Needy Families (TANF) and Supplemental Nutrition Assistance Program (SNAP) restrict states from providing assistance to individuals convicted of a drug related fel ony, though states may modify or scale back this ban [ 26 ]. Women who still qualify for federal benefits may need to re apply for their benefits or request to have the agencies restart their benefits upon their release; both of these processes can take months [ 27, 28 ]. Resuming and repairing ties with family members and frie nds represent s an additional priority of women exiting the correctional system. Approximately 60% to 70% of female inmates have at least one child under age 18, and childcare and custody issues represent a significant stressor [ 29, 30 ]. In addition, family and friends may pull 31 ]. Women who use illicit drugs often receive low levels of financial and emotional support from friends and family; therefore, women may return to drugs to help meet financial challenges and to cope with the stressors of re entry [ 9 ]. In addition, drug users may encounter substance using acquaintances and the locations where they once purchased or used drugs, making them more likely to use [ 17, 32, 33 ]. Women who engage in transactional sex also generally receive little to no support from family members, with nearly 50% leaving home permanently by age 16 [ 34 ]. Incarceration also exposes individuals to high risk networks that may increase drug use and risky sexual be havior upon return to the community [ 35, 36 ]. Women exiting jail and prison must
15 navigate multiple and simultaneous challenges often without institutional support to facilitate a successful transition back to the community. The Stress Buffering Model of Social Support Social support broadly refers to processes that promote health through the provision of emotional, informational or instrumental (i.e., tangible) resources [ 37 ]. Beginning in the 1950s, studies began to emerge demonstrating that individuals who received regular contact with family and friends during the experience of incarceration were less likely to recidivate than individuals with few social ties [ 38, 39 ]. Men and women who successfully reintegrate after incarceration often attribute their success to supportive friends and family that provide housing and lend financial and emotional support [ 40, 41 ]. The stress buffering model of Cohen and Wills proposes that social support redu ces negative emotional responses to stressful events and promotes more positive and adaptive behavioral responses [ 42 ]. Receipt of actual emotional, informational or tangible support can also improve health outcomes through helping resolve the sourc e of stress [ 37 ]. Notably, the perception of the availability of support is hypothesized to exert an equal or even greater influence than actual received support on emotional and behavioral responses to stress. Perceived social support is y to cope, thereby reducing negative emotional and physiological responses to the stressor (Figure 1 1) [ 37, 43 ]. Social Support in the Context of Female Offenders Social support needs for former female inmates may be higher than that of men; women in general tend to have larger social network s and report stronger emotional bonds with their social ties than men [ 44
16 and social c Termed the Relational Model, Miller proposed that seeking connection with others represents a primary motivation for women and that women suffer negative emotional consequences from disrupted social ties [ 45 ]. Subsequent research has found that in response to stressful events, women tend to cope by seeking out emotional support from those in their social n etwork, and women may be more susceptible to negative health outcomes, such as depression, in response to low levels of social support [ 46 48 ]. Despite the overall low levels of social support documented for women convicted of drug and transactional sex offenses, male romantic partners appear to represent a prominent source of social support in this population In a sample of female jail inmates, most women reported low perceived social support in general and from friends and family. However, three fourths of this sample reported emotional support and comfort from a significant other [ 13 ]. Despite filling a potentially important source of emotional support, a significant body of literature demonstrates that the presence of a male partner may also exert negati arly in contexts where the male partner is abusive, controlling or attaches emotional intimacy to risky behaviors such as needle sharing [ 49 52 ]. Male partners in general appear particularly substance use and sexual risk behaviors. earlier work, explains that women often turn to substance use to initiate or maintain connections with others, and to cope with feelings of stress and isolation [ 53 ]. Women may use drugs to connect with drug using partners, to deal with pain in their
17 relationships or to cope w ith the dissolution of a relationship [ 54, 55 ]. For some women, involvement in a partnership during re entry may increase drug risk. There is empirical evidence to suggest marriage or cohabitation with a male partn er increases the likelihood of illicit drug use [ 56, 57 ]. A partner may also increase risk of sexual in order to maintain harmony in the relationship [ 58 ]. However, low social support has been associated with more frequent post release drug use [ 40 ]. Because women in particular may experience negative consequences from re lationship strain and dissolution, absence of a partner, potentially due to incarceration related disruption of partnerships, may lead to lower tangible and emotional support among former inmates and to increased risk taking [ 54 ]. The Stress Buffering Model of Social Support While few studies have examined the relationship between time since releas e from incarceration and HIV risk behaviors among recently released female offenders, none to our knowledge have examined the potential for involvement in a committed partnership to influence this relationship. Therefore, to better understand risk and prot ective factors in the post release period, this study explored HIV related drug use and sexual risk behaviors after release from incarceration and the potential moderating effect of a committed partner on post release risk behaviors. We used data from fema le respondents of NEURO HIV Epidemiologic Study, a study of non injection and injection drug users in Baltimore, MD to investigate associations between time since release from incarceration and past month: binge drinking, crack cocaine smoking, injection d rug use, transactional sex, condom use, alcohol use before or during sex, and drug use before or during sex. Additionally, we explored whether involvement in a committed
18 partnership, a relationship lasting at least three months and where the partners cohab itate or are married, affected these associations. We hypothesized that women released from incarceration within the past 6 months would exhibit significantly higher rates of drug use and sexual risk behaviors compared to never incarcerated women and women released more distally. Further, we anticipated a lower prevalence of HIV risk behaviors among recently released women who reported a committed partner compared to unpartnered women.
19 Figure 1 1. Stress buffering model of social ties and mental he alth. Adapted from Cohen S, Underwood LG, Gottlieb BH, 2000.
20 CHAPTER 2 METHODS Introduction The current study used the baseline data female participants included in the NEURO HIV Epidemiologic Study. The study was approved by the Univers ity of HIV Epidemiologic Study is a longitudinal epidemiological examination of neuropsychological, social and behavioral risk factors of HIV, hepatitis B, hepatitis C, and se xually transmitted infections among injection and non injection drug users. Study participants were recruited in Baltimore, MD using a variety of community based outreach strategies, including local newspaper advertisements, street recruitment, and referra l; participants were enrolled if they were 18 years of age or older and reported injection or non injection drug use in the past six months. Trained research staff provided detailed information about the study and obtained informed consent before administe ring the baseline assessment in interview format. Blood samples were collected by a phlebotomist at the baseline assessment, and HIV antibody testing was performed using standard ELISA screening and confirmatory Western Blots. Participants were subsequentl y notified of their HIV status and were referred to social services for counseling. Study Population The present analyses were conducted on a subset of the 479 female participants who completed the baseline assessment. The final sample size for these anal yses ( n = 450) was reached after excluding participants who were missing data on their
21 incarceration history ( n = 15) or their partnership status ( n = 3). Additionally, participants that reported a release date within the past month were excluded ( n = 11) to ensure that the behaviors of interest occurred in the community following release from a correctional setting. Sample character istics are presented in Table 3 1. Measures HIV risk behavior interview The HIV risk behavior interview included questions on sociodemographics, medical, educational, and neurodevelopmental histories, and behavioral information about illicit drug use and sexual practices. Drug use questions assessed for drug type, route of administration and frequency of usage over the 30 days pr eceding the assessment. Measure of incarceration Participants were asked if they had ever in their lifetime spent time in juvenile detention, jail, or a correctional facility and were asked the date of their last release. The current study limited the def inition of incarceration to individuals who reported serving time in jail or a correctional facility. For the purposes of this study, the experience of jail and the experience of prison were both regarded as disruptive and stressful life events that destab social ties [ 59, 60 ]. Participants who reported no lifetime history in either institution were assigned to the Never incarcerated group. Ba sed on the reported last release date, participants with an incarceration history were assigned to Released 1 6 months ago or Released >6 months ago. Previous studies have defined recent incarceration as occurring within the past 6 months [ 40, 61 ].
22 Measure of partnership status In the baseline assessment, participants were asked their marital status, who they live with, if they had a steady male sexual partner within the past 3 months, and if so, how long that partn ership had lasted. The current study defined a Committed partnership as a relationship with a main partner lasting at least 3 months and one in which the couple either cohabitates or is married. Participants were assigned to No committed partner if their p resent partnership did not meet the committed criteria or if they did not report a steady sexual partner in the past 3 months. The required duration of the partnership (3 months) and the additional requirement of marriage or cohabitation were applied to ca pture a context where both individuals have expressed commitment to, and potentially support for, the other. A 3 month duration is an accepted length of steady sexual partnerships in previous studies of risk behaviors within the context of a steady sexual partner [ 62, 63 ]. Measure of past month drug use We examined using the following drugs in the past month (yes/no): binge drinking, smoked crack cocaine, and injection of any drug. These drugs were included because they have been associated with heightened risk of acquiring HIV. Alcohol reduces inhibitions, impairs judgment and has been associated with unprotected sexual encounters and multiple sex partners [ 64, 65 ]. In this study, binge drinking was defined according to the National Institute of Alcohol Abuse and Alcoholism guidelines as four or greater alcoholic beverages per occasion for women [ 66 ]. Crack cocaine has also been associated with impulsivity and risky sexual behaviors, while injection drug use directly facilitates vi rus transmission [ 67 69 ]. Additionally, we examined sharing injection
23 equipment in the past month (yes/no), which encompassed sharing needles, cookers, cottons, and/or rinse water. Measure of past month sexual risk behaviors We examined report of the following behaviors in the past month (yes/no): transactional sex, alcohol use before/during sex, drug use before/during sex, and condom use at last sex. Alcohol and drug use before or during sex have been associated wi th unprotected sexual encounters and risky sexual partners [ 15, 70, 71 ]. Condom use at last sex is considered a valid indicator of typical condom use behaviors over longer periods of time [ 72 ]. Soc iodemographic and other potential confounding variables Characteristics that were examined as potential confounders in the analyses included race (dichotomized as Black or non Black), self reported history of psychopathology (an emotional/ behavioral cond ition that was treated by a psychologist/ psychiatrist), educational level (dichotomized to those with less than a high school diploma or GED equivalent, and those with at least a high school diploma or equivalent), lifetime number of years incarcerated, p ast six month employment history ( employment at either a regular or temporary job ) and biologically confirmed HIV infection. Individuals that tested positive for HIV were retained in the analyses in order to better appreciate the behaviors of former femal e inmates within the community. Previous studies of risk factors for HIV and sexually transmitted infection s have included individuals that self reported or tested positive for the infection(s) of interest [32, 61, 73]. The majority of the individuals assi gned to the non Black racial group identified as White (96%), with the remainder identifying as Hispanic/ Latina (2%), Native American (1%), and mixed race (1%).
24 Statistical Analysis The primary goals of this study were to: investigate the main effect of incarceration history on past month HIV risk behaviors, and to explore the potential moderating effect of partnership status on this relationship. Preliminary chi square analyses were conducted to determine whether significant differences in HIV risk behav iors were found between HIV positive and HIV negative. In the overall sample, we first examined associations between the sociodemographic and other potentially confounding variables with incarceration history and past month drug use using the chi square te st. The odds ratio and 95% confidence interval (CI) for the main effect of incarceration history on past month risk behaviors was then calculated for each of the drug outcomes using binary logistic regression. This regression was re run, stratifying by par tnership status to better appreciate the effect of a committed partner on these relationships. To obtain the adjusted odds ratios, the confounding variables that were associated with incarceration history ( p < .05) were entered in the first step of the mod el and incarceration history was entered in the second step. Covariates included: age, history of psychopathology, and educational attainment.
25 CHAPTER 3 RESULTS Sociodemographics Majority of the sample identified as Black (71%), and no significant racial differences were found across groups based on incarceration history. Socio demographic variables that were significantly associated with incarceration history include d : age, educational attainment, and history of psychopat hology. Specifically, never incarc erated women were younger than women who reported an incarceration history and reported a lower prevalence of past mental illness. Additionally, a significantly higher percentage of never incarcerated women reported obtaining a high school degree or GED eq uivalent. Women that tested positive for HIV were retained in the sample, a s preliminary analyses did not find significant differences in HIV risk behaviors, with the exception of condom use, based on HIV status. There were no significant differences in HI V prevalence based on incarceration history. Incarceration History Majority of the sample had spent time in either jail or prison, with 17% of the sample released 1 6 months ago (Table 3 1 ). Fifty seven percent of the sample was released greater than six m onths ago. The two groups of previously incarcerated women did not differ in lifetime total years spent incarcerated, t (329) = .02, p = .90. Associations between incarceration history and past month drug use Recent incarceration (i.e., within the past 1 6 months) was strongly associated with past month crack cocaine smoking (odds ratio (OR) = 2.86, 95% confidence interval (CI): 1.56 5.23). There was also a trend towards significance for the association between recent inc arceration and binge drinking ( p = .0 8) and for the association
26 between recent incarceration and injection drug use ( p = .082). After adjusting for confounding variables, the pattern of significance remained unchanged for crack cocaine smoking. Sharing injection equipment had a very low preva lence in the sample (2.4%) and thus was excluded from analyses. Associations between incarceration history and past month sexual risk behavior Recent incarceration was associated with past month transactional sex (OR = 11.30, 95% CI: 4.05 31.5). However th ese analyses are considered exploratory due to the low prevalence of this behavior reported in the sample. The other sexual risk behaviors (condom use at last sex, alcohol use before/during sex, and drug use before/during sex) did not differ significantly based on incarceration history. Interaction with partnership status The sample was stratified by committed partnership status, and the adjusted odd s ratios are presented in Table 3 2. Recently released women without a committed partner were more likely to smoke crack cocaine (AOR = 2.55, 95% CI: 1.22 5.32 ) and engage in injection drug use (AOR = 2.66, 95% CI: 1.18 6.01). Additionally, they were more likely to engage in transactional sex (AOR = 11.3, 95% CI: 3.65 34.9), though this result should be interpret ed with caution due to the low overall prevalence of the behavior in the sample. Among those with a committed partner, no significant differences were observed between never and recently incarcerated women for smoked crack cocaine, injection drug use, and transactional sex.
27 a N may vary slightly according to missing data Table 3 1. Demographic, drug use, and sexual risk behavior comparisons by incarceration history Variable Entire Sample Never Incarcerated Released 1 6 months ago Released >6 months ago Test statistics M or N S.D. or % M or N S.D. o r % M or N S.D. or % M or N S.D. or % p value N 450 120 26.7 75 16.7 255 56.7 Age 35.89 9.31 32.43 10.96 35.72 7.99 37.56 8.37 13.04 < 0 .001 Race/ethnicity Black 319 70.9 82 68.3 46 61.3 191 74.9 9.2 0 .056 Non Black 131 29.1 38 31.7 29 38.7 64 25.1 Education 13.56 0 .001 < High School 212 47.1 39 32.8 39 52.0 134 52.5 High school or GED 237 52.7 80 67.2 36 48.0 121 47.5 Partnership status 10.61 0 .005 Committed partner 167 37.1 36 30.0 20 26.7 111 43.5 Lifetime years incarcerated 1.62 3.55 2.27 4.14 2.35 4.06 .02 0 .90 History of psychopathology 290 64.4 67 55.8 53 70.7 170 67.2 5.99 0 .05 Employed past 6 months 153 34.0 50 41.7 25 33.3 78 30.7 4.38 0 .11 HIV+ 36 8.0 9 7.5 2 3.2 25 13.0 5.21 0 .074 Past 30 day sex risk behavior s Transactional sex 46 10.2 6 5.0 24 32.0 16 6.3 46.66 < 0 .001 Condom used at last sex 150 33.3 43 36.4 25 35.7 82 32.3 0.6 0 .74 Used alcohol before/during sex 91 20.2 22 18.3 19 25.3 50 19.6 1.54 0 .46 Used drugs before/during sex 158 34. 0 45 37.5 34 45.3 79 31.0 5.65 0 .059 Past 30 day drug use Binge drinking 110 24.4 28 23.5 26 35.1 56 22.5 5.04 0 .08 Smoked crack cocaine 160 35.6 33 27.5 39 52.0 88 34.5 12.37 0 .002 Injection drug use 123 27.3 32 26.7 29 38.7 62 24.3 6. 05 0 .049 Risky injection practice 11 2.4 2 1.7 0 0 9 3.5
28 Table 3 2 related drug use in past month, by committed part nership status Entire Sample Entire Sample No Committed Partner Risk Behavior; % Reporting Behavior Unadjusted OR a (CI b ) p Adjusted OR c (CI b ) p Unadjusted OR a (CI b ) p Smoked Crack Cocaine Never Incarcerated (N = 122) 28% Referent Referent Referent Released > 6 months ago (N = 255) 35% 1.39 (.86 2.23) 0 .18 1.14 (.69 1.91) 0 .59 1.19 (.67 2.11) 0 .56 Released 1 6 months ago (N = 75) 52% 2.86 (1.56 5.23) 0 .001 2.61 (1.39 4.87) 0 .003 2.68 (1.32 5.41) 0 .006 Injection Drug Use N ever Incarcerated (N = 122) 27% Referent Referent Referent Released > 6 months ago (N = 255) 34% .77 (.46 1.28) 0 .62 1.02 (.60 1.74) 0 .07 .91 (.46 1.79) 0 .78 Released 1 6 months ago (N = 75) 39% 1.62 (.87 3.01) 0 .08 1.9 (.99 3.63) 0 .052 2.08 (.9 6 4.49) 0 .022 Transactional Sex Never Incarcerated (N = 122) 5% Referent Referent Referent Released > 6 months ago (N = 255) 6% 1.27 (.49 3.39) 0 .65 1.68 (.60 4.71) 0 .32 1.18 (.39 3.57) 0 .77 Released 1 6 months ago (N = 75) 32% 8.94 ( 3.45 23.2) < 0 .001 11.30 (4.05 31.5) < 0 .001 10.53 (3.67 30.18) < 0 .001 a Odds ratio. b 95% confidence interval. c Adjusted for age, high school diploma or equivalent, and history of psychopathology
29 Table 3 2. Continued No Committ ed Partner Committed partner Committed partner Risk Behavior; % Reporting Behavior Adjusted OR c (CI b ) p Unadjusted OR a (CI b ) p Adjusted OR c (CI b ) p Smoked Crack Cocaine Never Incarcerated (N = 122) 28% Referent Referent Referent Releas ed > 6 months ago (N = 255) 35% 0.97 (0.52 1.80) 0 .93 1.96 (.78 4.93) 0.07 1.74 (0.65 4.54) 0.35 Released 1 6 months ago (N = 75) 52% 2.55 (1.22 5.32) < 0 .001 2.7 (.79 9.29) 0.09 2.56 (0.73 8.92) 0.14 Injection Drug Use Never Incarcerated ( N = 122) 27% Referent Referent Referent Released > 6 months ago (N = 255) 34% 1.08 (0.74 2.28) 0 .83 .58 (.26 1.28) 0.65 0.71 (0.31 1.63) 0.41 Released 1 6 months ago (N = 75) 39% 2.66 (1.18 6.01) 0 .02 1.05 (.34 3.20) 0.93 1. 12 (0.35 3.60) 0.8 5 Transactional Sex Never Incarcerated (N = 122) 5% Referent Referent Referent Released > 6 months ago (N = 255) 6% 1.45 (0.44 4.74) 0 .54 2.00 (.23 17.19) 0.53 3.65 (0.39 34.65) 0.26 Released 1 6 months ago (N = 75) 32% 11.3 (3.65 34.9) < 0 .001 3.89 (.33 45.82) 0.28 7.66 (0.54 102.1) 0.12 a Odds ratio. b 95% confidence interval. c Adjusted for age, high school diploma or equivalent, and history of psychopathology
30 CHAPTER 4 DISCUSSION Consistent with previous r esearch, this study highlights the months following release from incarceration as a high risk period for drug use and sexual risk behaviors [ 11, 73 76 ]. Women in this sample released from incarceration 1 6 months a go were more likely to have engaged in illicit drug use and transactional sex than women who were never incarcerated or released more distally. Additionally, there was a trend towards significantly higher prevalence of binge drinking among those released w ithin the past 6 months. In order to evaluate the incarceration as an independent risk factor for HIV related drug use and sexual risk behaviors, potentially confounding variables were identified and controlled for. However in reality, mental illness and e ducational attainment have strong links to incarceration as well as to illicit drug use and sexual risk behaviors [ 19, 25, 77 ]. In contrast to prior work, this study did not find that Black women were more likely t han other races to report an incarceration history [ 2, 3 ]. However, the majority of this sample identified as Black, which may have contributed to this lack of significant findings. This elevated rates of drug use and transactional s ex following release from incarceration appeared to diminish with time, as women with more distal exposure to incarceration reported past month behavior patterns that did not differ significantly from women who were never incarcerated. Longitudinal studies of released jail inmates and prisoners have found a trend of reduced drug use and risk behaviors at one year post release [ 36, 76, 78 ]. This trend may reflect a gradual shift towards successful reintegration and s ocial stability with time. Further, emotional reactions to the stress of
31 reintegration may have eased, resulting in a greater ability for some women to resist drug use. Community re entry is a period characterized by instability and stressors that may trig ger substance use and high risk sexual behaviors. Former inmates often face unstable housing, homelessness and unemployment and may resort to drugs to cope with the emotional distress from a disrupted life [ 79, 80 ] They may also engage in transactional sex to meet financial and material needs [ 81, 82 ]. Mental health and substance treatment needs often remain unaddressed, and women may continue to engage in transactional sex to provide for a drug dependency or due to untreated mental illness [ 77, 83, 84 ]. At the same time, many women experience a diminished social network as they reenter the community due to strained relationships wit h friends and family. Additionally women may attempt to form new social ties outside of the risky social networks that they may have belonged to prior to incarceration [ 41, 85 ]. Consistent and positive social supp ort has been found to reduce many re entry challenges. It has also been associated with less drug use and risky sexual behavior in the first six months following prison release among males [ 40 ]. Among males, those who resumed behaviors characteristic of those prior to their arrest, such as illicit drug use and multiple sex partners, were less likely to have a steady sexual partner than those who more successfully reintegrated. In this sample, the moderating effect o f a committed partner on recently released increase in drug use was not observed in the first six months following release from incarceration. Women without a committed par tner had a higher use prevalence of
32 crack cocaine smoking and injection drug use in the first six months. The low prevalence of transactional sex in the sample does not allow us to draw any conclusions about this behavior. Ten events per variable are recom mended when performing logistic regression to prevent confidence intervals from exceeding 95% coverage [ 86 ]. Too few women reported transactional sex in this sample to me et the recommended number of events per variable. When examining the general pattern of results for transactional sex, recently released women without a committed partner appeared more likely to engage in transactional sex, while a corresponding increase w as not found among recently released women that reported a partner. All except one recently released woman with a committed partner reported that her partnership was greater than six months in duration. Therefore, it is likely that partnered women release d from incarceration within the past six months were in a relationship with their partner at the time of their release. The presence of a committed partner may have served as a source of emotional or instrumental support and protected against the stressors of re entry that can lead to substance use and transactional sex. Partnered women may have perceived the availability of support from her partner regardless of whether or not support was offered or given; women who perceived available support may have vie wed reentry challenges and her ability to cope with them in a more positive manner [ 42 ]. Alternatively, women with committed partners may have been less likely to engage in HIV related risk behaviors than unpartnered women. However, past month drug use did not differ significantly between women with and without committed partners in the overall sample, though there were significant differences in sexual risk behaviors in the overall sample based on partnership status.
33 Those with a partner were less l ikely to engage in transactional sex and condom use, though they were more likely to use alcohol or drugs before or during sex. These findings related to sexual risk behaviors are consistent with previous studies showing that some risky sexual behaviors, s uch as inconsistent condom use, may be more likely to occur in the context of a steady sexual partnership [ 87, 88 ]. Previous research suggests that a romantic partner provides an important source of social support for female inmates. Despite low support from friends and family, a significant portion of jail inmates report that a significant other provides emotional support and comfort [ 13 health, sense of well being, and ability to cope with adversity rely heavily on relationships [ 45 ]. While some suggest women may use drugs and engage in sexual risk taking to maintain connections, loneliness or an ended relationship can also trigger substance abuse to cope with strong feelings of distress [ 55 89 91 ]. This study did not obtain information on the pa behaviors. However romantic partners, even substance abusing partners, may provide critical support when friends or other family members pull away due to frustration or 31 ]. In a study of former female inmates in drug treatment, 63% named a partner as a main source of support. Over half of the partners actually tried to help the women stop using, though 40% supported the woman whi le also enabling her drug use. The women described family and friends as much less likely to help them stop using [ 92 ]. In another study of former female inmates, negative partner influences were not significantly a ssociated
34 with drug use [ 93 ]. The relationship complexities of female former inmates who abuse substances or engage in transactional sex are considerable. This represents an important area for research in order to situate interve relationships [ 94 ]. It should be pointed out that this sample does not capture women who were re arrested, and the reduced prevalence of risk behaviors observed am ong women released greater than six months ago may reflect those who avoided risk behaviors and more successfully reintegrated. Recent estimates suggest that one third of female jail inmates are rearrested within a year of their release [ 36 ]. The HIV risk behavior interview did not collect information on number of children. Children often exert a cipate in drug treatment, and it will be release from incarceration [24, 31]. Additional limitations include the cross sectional nature of the data, limiting the ability to draw a causal connection between the presence of a committed partner and reduced drug use. This study did not verify self reported incarceration, partner status or past month drug use information. It is possible that recall bias or social desirability bias created inaccuracies in the data. Despite these limitations, the prevalence of risk behaviors and incarceration experiences were consistent with previously reported figures [ 61, 95, 96 ]. While acknowledging these patterns of drug and alcohol use and sexual risk behaviors after release and highlights the potential protective influence of a committed partner on drug use during the risky community re entry period. With high rates of relationship dissolution during
35 incarceration, the potential protective influence of an intact relationship upon release is particularly noteworthy. Future studies should investigate the qualities of relationships that may help women re sist substance use following release. Additionally, factors that allowed the relationship to r emain intact should be explored.
36 LIST OF REFERENCES 1. Chen NE. Advances in the prevention of heterosexual transmission of HIV/AIDS among women in the United S tates. Infectious disease reports. 2011;3(1):e6. 2. Greenfeld L, Snell T. Bureau of Justice Statistics Special Report: Women Offenders. Washington, DC: U.S. Department of Justice, Bureau of Justice Statistics. 2000. 3. Snyder HN. Arrest in the United Sta tes, 1990 2012. Washington, DC: US Department of Justice, Bureau of Justice Statistics. 2012. 4. Hoffman JA, Klein H, Eber M, Crosby H. Frequency and intensity of crack use as predictors of women's involvement in HIV related sexual risk behaviors. Drug Al cohol Depen d 2000;58(3):227 236. 5. Mathers BM, Degenhardt L, Phillips B, et al. Global epidemiology of injecting drug use and HIV among people who inject drugs: a systematic review. Lancet. 2008;372(9651):1733 1745. 6. Strathdee SA, Stockman JK. Epidemi ology of HIV among injecting and non injecting drug users: current trends and implications for interventions. Curr HIV/AIDS Rep. 2010;7(2):99 106. 7. Weller SC D BK. Condom effectiveness in reducing heterosexual HIV transmission: Cochrane Database Syst Re v; 2002; Art. No.: CD003255. 8. Spaulding A, Stephenson B, Macalino G, Ruby W, Clarke JG, Flanigan TP. Human Immunodeficiency Virus in Correctional Facilities: A Review. Clin Infect Dis. 2002;35(3):305 312. 9. Mallik Kane K VC. Health and prisoner reentr y: How physical, mental, and substance abuse conditions shape the process of reintegration. Washington, DC: The Urban Institute; 2008. 10. D.C. AfaSD. Move along: Policing sex work in Washington, D.C.: A report by the Alliance for a Safe & Diverse D.C. Wa shington DC: Different Avenues; 2008. 11. Binswanger IA, Stern MF, Deyo RA, et al. Release from Prison A High Risk of Death for Former Inmates. New Engl J Med. 2007;356(2):157 165. 12. Farrell M, Marsden J. Acute risk of drug related death among newly released prisoners in England and Wales. Addiction. 2008;103(2):251 255. 13. Singer MI, Bussey J, Song LY, Lunghofer L. The psychosocial issues of women serving time in jail. Soc Work. 1995;40(1):103 113.
37 14. Jannetta J NS, Davies E, Horvath A. Transitio n from prison to community initiative: Process evaluation final report: Urban Institute Justice Policy Center; 2012. 15. Wilson AB. How people with serious mental illness seek help after leaving jail. Qual Health Res. 2013;23(12):1575 1590 16. Proctor SL. Substance use disorder prevalence among female state prison inmates. Am J Drug Alcohol Abuse. 2012;38(4):278 85. 17. Taxman FS, Perdoni ML, Harrison LD. Drug treatment services for adult offenders: the state of the state. J Subst Abuse Treat. 2007;32( 3):239 54. 18. Lynch SM, Fritch A, Heath NM. Looking beneath the surface: The nature of incarcerated women's experiences of interpersonal violence, treatment needs, and mental health. Fem Criminol. 2012;7(4):381 400. 19. Lynch SM DD, Belknap J, Green BJ. Women's pathways to jail: examining mental health, trauma, and substance use. Washington, D.C.: U.S. Department of Justice; 2013. 20. Green BL, Miranda J, Daroowalla A, Siddique J. Trauma exposure, mental health functioning, and program needs of women in jail. Crime Delinq. 2005;51(1):133 151. 21. Derkzen D, Booth L, Taylor K, McConnell A. Mental health needs of federal female offenders. Psychol Serv. 2013;10(1):24 36. 22. McLean RL TM. Repaying debts. New York: US Department of Justice, Bureau of Justi ce Assistance; 2007. 23. Warr M. Life course transitions and desistance from crime. Criminology. 1998;36(2):183 216. 24. Covington S. A woman's journey home: challenges for female offenders and their children. From Prison to Home Conference. Bethesda, MD ; 2002. 25. Harlow CW. Education and correctional populations: US Department of Justice, Bureau of Justice Statistics; 2003. 26. McCarty M, Aussenberg RA, Falk G, Carpenter DH. Drug testing and crime related restrictions in TANF, SNAP, and housing assist ance. Washington, DC: Congressional Research Service; 2012.
38 27. Bloom B OB, Covington S. Gender responsive strategies: Research, practice and guiding principles for women offenders. Washington, D.C.: Department of Justice, National Institute of Correction s; 2003. 28. Koyanagi C. Arrested? What happens to your benefits if you go to jail or prison? Bazelon Center for Mental Health Law. Washington DC; 2006. 29. El Bassel N, Gilbert L. Correlates of crack abuse among drug using incarcerated women: Psychologi cal trauma, social support, and coping behavior. Am J Drug Alcohol Abuse. 1996;22(1):41 56. 30. Glaze LE ML. Parents in prison and their minor children. Washington, DC: U.S. Department of Justice, Bureau of Justice Statistics; 2008. 31. Harp KL, Oser C, Leukefeld C. Social support and crack/cocaine use among incarcerated mothers and nonmothers. Subst Use Misuse. 2012;47(6):686 694. 32. Belenko S, Langley S, Crimmins S, Chaple M. HIV risk behaviors, knowledge, and prevention education among offenders unde r community supervision: a hidden risk group. AIDS Educ Prev. 2004;16(4):367 385. 33. Binswanger IA, Nowels C, Corsi KF, et al. Return to drug use and overdose after release from prison: a qualitative study of risk and protective factors. Addict Sci Clin Pract. 2012;7(1):3. 34. Norton Hawk MA The counterproductivity of incarcerating female street prostitutes. Deviant Behav. 2001;22(5):403 417. 35. Dumont DM, Brockmann B, Dickman S, Alexander N, Rich JD. Public health and the epidemic of incarceration. A nnu Rev Publ Health. 2012;33:325 339. 36. Freudenberg N. Jails, prisons, and the health of urban populations: a review of the impact of the correctional system on community health. J Urban Health. 2001;78(2):214 235. 37. Cohen S UL, Gottlieb B, ed. Socia l support measurement and interventions: A guide for health and social scientists. New York: Oxford; 2000. 38. Ohlin LE. The stability and validity of parole experience tables: Sociology, University of Chicago; 1954. 39. Laub JH, Nagin DS, Sampson RJ. Tr ajectories of change in criminal offending: Good marriages and the desistance process. Am Sociol Rev. 1998;63(2):225 238.
39 40. Seal DW, Eldrige GD, Kacanek D, Binson D, MacGowan RJ, Grp PSS. A longitudinal, qualitative analysis of the context of substance use and sexual behavior among 18 to 29 year old men after their release from prison. Soc Sci Med. 2007; 65(11):2394 2406. 41. Parsons ML W RC. Factors that support women's successful transition to the community following jail/prison. Health Care for Women International. 2002;23(1):6 18. 42. Cohen S, Wills TA. Stress, social support, and the buffering hypothesis. Psychol Bull. 1985;98(2):310 357. 43. Thoits PA. Social support as coping assistance. J Consult Clin Psychol. 1986;54(4):416 423. 44. Belle D. Gender differences in the social moderators of stress. In: Barnett RC, Biener L, Baruch GK, ed. Gender and Stress. New York: Free Press; 1987:257 277. 45. Miller JB. Toward a new psychology of women. Boston, MA: Beacon Press; 1976. 46. Ptacek JT, Smith RE, Zanas J. Gender, appraisal, and coping a longitudinal a nalys is. J Pers. 1992;60(4):747 770. 47. Elliott M. Gender differenc es in causes of depression. Women Health. 2001;33(3 4):163 177. 48. Olstad R, Sexton H, Sogaard AJ. The Finnmark Study. A prospective population study of the social support buffer hypothesis, specific stressors and mental distress. Soc Psych Psych Epid. 2001;36 (12):582 589. 49. Noar SM, Webb E, Van Stee S, et al. Sexual Partnerships, risk behaviors, and condom use among low income heterosexual African Americans: A qualitative study. Arch Sex Behav. 2012;41(4):959 970. 50. Tortu S, Beardsley M, Deren S, Davis WR. The risk of HIV infection in a national sample of women with injection drug using partners. Am J Public Health. 1994;84(8):1243 1249. 51. Knudsen HK, Leukefeld C, Havens JR, et al. Partner relationships and HIV risk behaviors among women offenders. J Psychoactive Drugs. 2008;40(4):471 481. 52. Kenrick DT Neuberg SL, Cialdini RB, ed. Social psychology: Goals in interaction. 5 ed. Boston, MA: Allyn & Bacon; 2010.
40 53. Finkelstein N. Using the relational model as a context for treating pregnant and parenting chemically dependent women. In B. Underhill & D. Finnegan (Eds.), Chemical dependency: Women at risk. New York: Harrington Park Press. 1996:23 44. 54. development: Implications for substance abuse. In S. Wilsnack and R. Wilsna ck (Eds.) Gender and alcohol: Individual and social perspectives. New Brunswick, NJ: Rutgers University Press. 1997:335 351. 55. Staton Tindall M, Duvall JL, Leukefeld C, Oser CB. Health, mental health, substance use, and service utilization among rural and urb an incarcerated women. Womens Health Iss. 2007;17(4):183 192. 56. Alarid LF, Burton VS, Cullen FT. Gender and crime among felony offenders: assessing the generality of social control and differential association theories. J Res Crime Delinq. 2000;37(2):171 99. 57. G riffin ML, Armstrong GS. The effect of local life circumstances on female probationers' offending. Justice Quarterly. 2003;20(2):213 239. 58. El Bassel N, Gilbert L, Rajah V, Foleno A, Frye V. Fear and violence: Raising the HIV stakes. AIDS Educ Prev. 2000;12(2 ):154 170. 59. Western B. The impact of incarceration on wage mobility and inequality. Am Sociol Rev. 2002;67(4):526 546. 60. Fu JJ, Herme M, Wickersham JA, et al. Understanding the revolving door: individual and structural level predictors of recidivism among indiv iduals with HIV leaving jail. AIDS Behav. 2013;17 Suppl 2:S145 155. 61. Epperson MW, Khan MR, Miller DP, Perron BE, El Bassel N, Gilbert L. Assessing criminal justice involvement as an indicator of human immunodeficiency virus risk among women in methadone treatment. J Subst Abuse Treat. 2010; 38(4):75 83. 62. Senn TE, Carey MP, Vanable PA, Coury Doniger P, Urban M. Sexual partner concurrency among STI clinic patients with a steady partner: correlates and associations with condom use. Sex Transm Infect. 2009;85(5):343 347. 63. van Empelen P S cha alma H P Kok G, Jansen MW. Predicting condom use with casual and steady sex partners among drug users. Health Educ Res. 2001;16(3):293 305.
41 64. Davis KC, Maste rs NT, Eakins D, et al. Alcohol intoxication and condom use self efficacy effects on women's condom use intentions. Addict Behav. 2014;39(1):153 158. 65. Weafer J, Fillmore MT. Comparison of alcohol impairment of behavioral and attentional inhibition. Drug Alco hol Depen d 2012;126(1 2):176 182. 66. National Institute of Alcohol Abuse and Alcoholism. NIAAA council approves definition of binge drinking. NIAAA Newsletter. 2004;3:3. 67. Division of HIV/AIDS Prevention, National Center for HIV/AIDS, Viral Hepatitis, Sexual Tran smitted Diseases and Tuberculosis Prevention, Centers for Disease Control and Prevention. HIV and substance use in the United States. 2013. 68. Dehovitz JA, Kelly P, Feldman J, et al. Sexually transmitted diseases, sexual behavior, and cocaine use in inner city women. Am J Epidemiol. 1994;140(12):1125 1134. 69. Lejuez CW, Bornovalova MA, Daughters SB, Curtin JJ. Differences in impulsivity and sexual risk behavior among inner city crack/cocaine users and heroin users. Drug Alcohol Depen d 2005;77(2):169 175. 70. Baliunas D, Rehm J, Irving H, Shuper P. Alcohol consumption and risk of incident human immunodeficiency virus infection: a meta analysis. Int J Public Health. 2010;55(3):159 166. 71. Zablotska IB, Gray RH, Serwadda D, et al. Alcohol use before sex and HIV acquisition: a l ongitudinal study in Rakai, Uganda. AIDS. 2006;20(8):1191 1196. 72. Younge SN, Salazar LF, Crosby RF, DiClemente RJ, Wingood GM, Rose E. Condom use at last sex as a proxy for other measures of condom use: is it good enough? Adolescence. 2008;43(172):927 931. 73. Kha n MR, Wohl DA, Weir SS, et al. Incarceration and risky sexual partnerships in a southern US city. J Urban Health. 2008;85(1):100 113. 74. Morrow KM, Project SSG. HIV, STD, and hepatitis risk behaviors of young men before and after incarceration. AIDS Care. 2009 ;21(2):235 243. 75. Vlahov D, Putnam S. From corrections to communities as an HIV priority. J Urban Health. 2006;83(3):339 348. 76. Wilson ME, Kinlock TW, Gordon MS, O'Grady KE, Schwartz RP. Postprison release HIV risk behaviors in a randomized trial of methadone treatment for prisoners. Am J Addict. 2012;21(5):476 487.
42 77. Graham N, Wi sh ED. Drug Use among Female Arrestees Onset, Patterns, and Relationships to Prostitution. J Drug Issues. 1994;24(1 2):315 329. 78. Adams L, Kendall S, Smith A, Quigle y E, Stuewig J, Tangney J. HIV r isk b ehaviors of m ale and f em ale j ail i nmates p rior to i ncarceration and o ne y ear p ost r elease. AIDS Behav. 2013;17(8):2685 94. 79. McLean RL, Robarge J, Sherman SG. Release from jail: moment of crisis or window of opportunity for female detainees? J Urban Health. 2006;83(3):382 393. 80. Richie BE. Challenges incarcerated women face as they return to their communities: Findings from life history interviews. Crime Delinquency. 2001;47(3):368 389. 81. German D, Latkin CA. Social stability and HIV risk behavior: evaluating the role of accumulated vulnerability. AIDS Behav. 2012;16(1):168 178. 82. Murphy LS Determinants of behavior in women choosing to engage in street level prostitution. Baltimore, MD ; 2007. 83. Farley M Kel ly V. Prostitution: a critical review of the medical and social sciences literature. Women & Criminal Justice. 2000;11(4):29 64. 84. Logan TK, Leukefeld C. Sexual and drug use behaviors among female crack users: a multi site sample. Drug Alcohol Depen d 2000;58(3):23 7 245. 85. Visher CA, Travis J. Transitions from prison to community: Understanding individual pathways. Annu Rev Sociol. 2003;29:89 113. 86. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logisti c regression analysis. J Clin Epidemiol. 1996;49(12):1373 1379. 87. Bryant J, Brener L, Hull P, Treloar C. Needle sharing in regular sexual relationships: an examination of serodiscordance, drug using practices, and the gendered character of injecting. Drug Alc ohol Depen d 2010;107(2 3):182 187. 88. Lescano CM, Vazquez EA, Brown LK, Litvin EB, Pugatch D, Project SSG. Condom use with "casual" and "main" partners: what's in a name? J Adolesc Health. 2006;39(3):443.e1 7. 89. Calhoun S M essina N, Cartier J, Torres S. Implementing gen der responsive treatment for women in prison: Client and staff perspectives. Federal Probation. 2010;74:339 349.
43 90. Leukefeld C, Havens J, Tindall MS, et al. Risky relationships: targeting HIV prevention for women offenders. AIDS Educ Prev. 2012;24(4):339 349. 91. Tucker MB. Social support and coping: Applications for the study of female drug abuse. J Soc Issues. 1982;38(2):117 137. 92. Falkin GP, Strauss SM. Social supporters and drug use enablers: a dilemma for women in recovery. Addict Behav. 2003;28(1):141 155. 93. Stat on Tindall M, Frisman L, Lin HJ, Leukefeld C, Oser C, Havens JR, et al. Relationship influence and health risk behavior among re entering women offenders. Womens Health Issues. 2011;21(3):230 238. 94. El Bassel N, Wechsberg WM, Shaw SA. Dual HIV risk and vulner abilities among women who use or inject drugs: no single prevention strategy is the answer. Curr Opin HIV AIDS. 2012;7(4):326 331. 95. Kuo I, Gre enberg AE, Magnus M, Phillips G 2nd, Rawls A, Peterson J, et al. High prevalence of substance use among heterosexua ls living in communities with high rates of AIDS and poverty in Washington, DC. Drug Alcohol Depen d 2011;117(2 3):139 144. 96. Staton Tindall M, Leukefeld C, Palmer J, et al. Relationships and HIV risk among incarcerated women. The Prison Journal. 2007;87(1):14 3 165.
44 BIOGRAPHICAL SKETCH Lauren Hea rn graduated from Emory University in May 2009 with a Bachelor of Science degree in biology. Lauren is presently pursuing a doctoral degree in clinical and health psychology at the University of Florida. Her academic interests include understanding the uni que factors contributing to disparities in the prevalence, diagnosis and treatment of HIV/AIDS, hepatitis C and sexually transmitted infections among underserved populations.
ORIGINALPAPERTheDifferenceinSelf-ReportedandBiologicalMeasuredHIV Prevalence:ImplicationsforHIVPreventionAlisaE.PedranaMargaretE.HellardRebeccaGuyKimWilsonMarkStoovePublishedonline:29December2011 SpringerScience+BusinessMedia,LLC2011Abstract InAustralia,HIVprevalenceestimatesamong gaymenhavebeenmainlybasedonself-reportedHIV statuscollectedinannualbehaviouralsurveys.WemeasuredbiologicalHIVprevalenceamonggaymeninMelbourne,Australia,usingafacilitybasedsamplingmethod. WecalculatedHIVprevalenceandusedlogisticregression toassesscorrelatesofapositiveHIVtest.Atotalof639 gaymenwererecruitedcompletedasurveyandprovided oraluidforHIVtestingfromsevenvenuesin2008.The medianageoftheparticipantswas35years(range 1875years).OverallbiologicalHIVprevalencewas9.5% (95%CI7.512.0%)comparedwith6.3%(95%CI 4.58.4%)forself-reportedHIVpositivestatus.Wefound asignicantdiscrepancybetweentestdetectedbiological andself-reportHIVstatusinourstudy,with19men (31.1%)unawareoftheirHIVinfection.Theseresults highlighttheimportanceofrepeatablebiologicalestimates toinformandevaluateHIVpreventionstrategies. Keywords HIV BiologicalHIVprevalence Self-report HIVprevalence Surveillance Gaymen Introduction InAustralia,menwhohavesexwithmen(MSM)accountfor morethan65%ofnewlydiagnosedand85%ofnewly acquiredHIVinfectionseachyear[ 1 ].Overthepastdecade therehasbeenasteadyincreaseintheannualnumberofnewly acquiredHIVinfectionsamongMSM,from667in2001to 909casesin2009[ 1 ].Similartrendshavebeenobservedin MelbournethecapitalofVictoria,Australia,ajurisdiction withthesecondlargestgaypopulationinAustralia[ 2 ].The riseinHIVdiagnoseshasbeenattributedtomoreriskysexual behavioursamonggaymen[ 3 ]anddramaticincreasesinother sexuallytransmittedinfections(STIs)includingsyphilis, gonorrhoeaandchlamydia[ 1 4 6 ],thatareknowntoincrease HIVtransmissionrisk[ 7 10 ]. InAustralia,HIVprevalenceestimatesamonggaymenhave beenmainlybasedonself-repo rtedHIVstatuscollectedin annualbehaviouralsurveys;recentprevalenceestimates amongMSMarereportedbetween59%[ 2 9 11 ].However, suchsurveysarelikelytounder-estimateHIVprevalenceas self-reportedHIVstatusisaff ectedbytestingpatternsanda subsetofmenwithHIVinfecti onwhomaynotbeawareoftheir positiveHIVstatusatthetimeofthesurvey.Internationally, studiesamongMSMhavedemonstratedwidediscrepancies betweenself-reportedand biologicalHIVstatus[ 12 13 ]. ThecollectionofbiologicalsamplestoestimateHIV prevalencealongsideunlinkedanonymousbehavioural surveysiswidelyusedasasurveillancetoolthroughout Europe,theUSandAfrica[ 14 16 ].Finger-prickblood [ 15 17 ]andoraluidspecimens[ 14 16 18 ]werethe A.E.Pedrana( & ) M.E.Hellard M.Stoove CentreforPopulationHealth,BurnetInstitute,GPOBox2284, Melbourne,VIC3004,Australia e-mail:email@example.com A.E.Pedrana M.E.Hellard M.Stoove DepartmentofEpidemiologyandPreventiveMedicine, MonashUniversity,Melbourne,VIC,Australia M.E.Hellard TheNossalInstituteforGlobalHealth,TheUniversity ofMelbourne,Melbourne,VIC,Australia R.Guy FacultyofMedicine,NationalCentreinHIVEpidemiology andClinicalResearch,Sydney,NSW,Australia K.Wilson NationalSerologyReferenceLaboratory,Melbourne,Australia123AIDSBehav(2012)16:14541463 DOI10.1007/s10461-011-0116-7
preferredmethodofspecimencollectionamongrecent studies,offeringadvantagesovervenousbloodsamples; beinglessinvasive,lesscostly,easiertohandleand,for thoseofferingrapidtesting,canprovideparticipantswith resultsatthepoint-of-care.Theuseofsuchdevicesin communitysettingsisidealfortheroutineandperiodic estimationofHIVprevalence.Suchmeasurementis importantfordeterminingburdenofdisease,evaluating preventioninitiativesandformodellingepidemictrajectories.Inthispaper,weestimatetestdetectedHIVbiological-prevalenceamonggaymenusingoraluidHIV testingandcomparethistoself-reportedHIVstatus.We alsoexplorecorrelatesofaHIVinfection. Methods Setting Melbourne,Australia,ajurisdictionwiththesecondlargest gaypopulationinAustralia[ 2 ].Melbourne'sgaycommunityvenuesconsistlargelyofgaysocialvenues(barsand clubs)andsex-on-premisesvenues(SOPV)(includinggay saunas),withatleast10SOPVsinoperationin2007[ 19 ]. StudyDesign InJune2008across-sectionalsampleofgaymenwas recruitedthroughsevengaycommunityvenuesininner cityMelbourne;fourSOPVsandthreegaybarsandclubs. Usingafacility-basedsamplingmethod[ 20 ],aconveniencesampleofgaymenwererecruitedonspecicdays andtimesoftheweektomaximiseattendancenumbersand participationrates.Locationsandtimesoftheweek(timelocation-sampling)werechosentomaximiserecruitment (FridayandSaturdaynightsforbars/clubs;Thursday,Friday,SaturdayandSundayeveningsforSOPVs).Recruitmentwasprecededbyasocialmarketingcampaigntoraise awarenessaboutthesurveyandthenovelnatureofspecimencollection. Menwereapproachedatthevenuesbytrainedeld researchersandinvitedtoparticipate.Thestudyinclusion criteriawereanyoneaged18yearsorover,whoself-identiedasgayorhadsexwithanothermaninthepast5yearsand abletoprovideverbalinformedconsent.ConsentingmenselfcompletedaquestionnaireandprovidedanoraluidspecimenonsiteatthevenuesusingtheOraSurecollectionkit (OraSureTechnologies,Inc.,Bethlehem,Pennsylvania, USA)accordingtothemanufacturer'sinstructions.Individual HIVtestresultswerenotprovidedtoparticipantsbecauseoral uidtestingforHIVinfectionisnotregisteredinAustraliafor screeningordiagnosticpurposes.Giventhatparticipantswere notabletobenetfromreceivingtheirtestresultatthepointof-care,studyrecruitmentcardswereprovidedthatincluded thestudywebsiteaddresscontainingdetailsofwheremen couldaccessfreeHIVtestingortelephonecounselling. Researcherrecruitmentlogsrecordedhowmanymenwere approachedtoparticipateandaimedtocapturereasonsfor non-participation. Questionnaire Thequestionnairewasadaptedfromtheinstrumentusedin theMelbourneGayCommunityPeriodSurveys(MGCPS) [ 21 ].TheMGCPSisanannualbehaviouralsurveyamong gaymenwhichcapturesinformationaboutdemographics, sexualself-identity,gaycommunitysocialattachment, sexualrelationships,sexualpracticeswithregularand casualpartners,knowledgeofpartners'HIVstatus,selfreportedperceivedHIVstatusandHIV/STItestinghistory. Thesexualbehaviourquestionsrelatedtothepast 6months,andHIV/STItestingthepastyear.Weincluded additionalquestionsregardingcondenceaboutknowing HIVstatusandacceptabilityoforaluidspecimencollection,whichaskedHowcomfortablehaveyoufound thecollectionofsalivainthisstudy?''andWouldyou participateinfutureMelbourneGayCommunityPeriodic Surveysifasalivaspecimenwascollected?'' HIVTesting OraluidspecimensweretestedforHIVbytheNational SerologyReferenceLaboratory(NRL)usingananti-HIV IgGantibodycaptureELISA(GACELISA)basedonthe methoddevelopedbyParryetal.[ 22 ].AninternalvalidationstudyoftheGACELISAatNRLdemonstrateda sensitivityof100%(95%CI:95.0100.0)andspecicityof 100%(95%CI:95.0100.0%)uponrepeattestingasper thetestingprotocol.TheHIVstatusofallspecimens testingrepeatedlypositivebytheHIV-1EIAtestwere conrmedusinganoraluidbasedwesternblot. StatisticalAnalysis Thequestionnairesandoraluidspecimenswerematched byanumericuniqueidentier.Menwereaskedwhether theyhadsexwithanycasualmalepartner/sinthepast 6months(yes/no)andiftheynever',occasionally'or often'engagedasareceptiveorinsertivepartnerinanal intercoursewithacondom,withoutacondomwithejaculation,andwithoutacondomwithwithdrawalbefore ejaculating,inthelast6months.Fromthisweconstructed avariableofunprotectedanalintercoursewithcasual partners(UAIC)(nocasualpartner,noanalintercourse,no UAIC,anyUAIC).Menwerealsoaskedabouthowmany oftheirfriendsweregayorhomosexualmen(none,afew, AIDSBehav(2012)16:14541463 1455123
some,ormost)andhowmuchoftheirfreetimeisspent withgayorhomosexualmen(none,alittle,some,alot). Fromthese,weconstructedavariableofsocialengagementwithgaymen'(lower,moderate,extensive)basedon amatrixoftheabovetwoquestions. HIVprevalenceestimateswerecalculatedfromthe biologicaltestresultandalsoself-reportedHIVstatus,and 95%condenceintervalswerecalculatedforallestimates usingbinomialstandardformulas.ToassesstheconcordanceinclassicationofHIVstatusbyself-reportandtest detectedmeasures,weusedamatchedMcNemar'stest. Univariableandmultivariablelogisticregressionmodels wereusedtoidentifyfactorsindependentlyassociatedwith bothself-reportedandtestdetectedHIVstatus.Forthe multivariableanalysisabackwardstepwisemethodwas used.DataanalysiswasperformedusingStata10.1 (StataCorp,Texas,USA)[ 23 ].Acutoffof P \ 0.05was usedforallstatisticaltests. EthicsapprovalwasobtainedfromtheVictorian DepartmentofHealthHumanResearchEthicsCommittee andtheMonashUniversityStandingCommitteeonEthics inResearchInvolvingHumans. Results Sample Inthestudy,1,027menwereapproachedtoparticipateand 639men(62.2%)completedaquestionnairethatcouldbe matchedtoanoraluidsample(threequestionnairescould notbematchedtoanoraluidsample).Participantswere recruitedfromthesevengaycommunityvenues(57.0% fromfourSOPVsand43.0%fromthreebarsandclubs). Themedianageoftheparticipantswas35years(range 1875years).Themajorityofparticipants(70.1%)were borninAustralia,89.1%reportedresidinginmetropolitan Melbourne,75.9%wereinfull-timeemploymentand 51.0%hadcompletedtertiaryeducation.Themajorityof menreportedtheirsexualidentityashomosexual(84.2%) orbisexual(11.8%)andwereeitherextensivelyengaged (42.4%)ormoderatelyengagedwiththegaycommunity (38.6%)(Table 1 ).Men'ssexualriskbehaviours,HIV testinghistory,andSTItestinghistoryaresummarisedin Table 2 byself-reportedandtestdetectedHIVstatus. HIVPrevalenceTestDetected Ofthe639men,61providedspecimensthatwereHIV positive(9.5%,95%CI7.512.0%).HIVprevalence increasedsignicantlywithincreasingageupuntilage50 andthenfellslightly;2.6,7.7,17.3and14.6%in1829, 3039,4049and50 ? yearolds,respectively(Table 1 ). Inunivariableanalyses,testdetectedHIVpositivestatus ( n = 61)wassignicantlyassociatedwithbeingrecruited fromSOPVs,olderage,receivingapension/socialsecuritybenets'orbeingunemployed,reportingknowing someonewithHIV,reportinghavingaHIVpositivecurrentregularpartner,highrisksexualpractices(multiple sexualpartners,UAIC,groupsex)andinfrequentHIVand STItesting(Table 3 ).HIVprevalencewashighestamong menreportingareasofresidenceinruralorregionalVictoria(16.7%)comparedtometropolitanMelbourne(9.5%) andother/interstate(6.7%),howeverthisdifferencewas notsignicant(Table 1 ).Multivariablelogisticregression showedindependentassociationsbetweenapositiveHIV resultandbeingolder( C 40years),reportingUAICinthe past6months,reportinggroupsexwithregularand/or casualpartnersinthepast6monthsandreportinghaving HIVpositiveregularpartner,afteradjustingforrecruitment siteandresidentiallocation(Table 3 ). HIVPrevalenceSelfReport Ofthe639men,40menreportedbeingHIVpositive correspondingtoaHIVself-reportprevalenceof6.3% (95%CI4.58.4%),howeverfourreturnedanegativeHIV testresult.HIVprevalenceincreasedsignicantlywith increasingageupuntilage50andthenfellslightly;1.0, 3.3,14.0and12.6%in1829,3039,4049and50 ? year olds,respectively(Table 1 ).Forself-reportedHIVpositive status( n = 40),univariableanalysesshowedsimilarcorrelatestothosefoundfortestdetectedHIV,althoughHIV wasnolongersignicantlyassociatedwithbeingrecruited fromSOPVs(Table 4 ).Multivariablelogisticregression showedindependentassociationsbetweenaself-reported HIVpositivestatusandbeingolder(aged40yearsand over),reportingUAICinthepast6months,reporting havingaHIVpositiveregularpartnerandreportingany STItestinginthepast12months,afteradjustingfor recruitmentsiteandresidentiallocation(Table 4 ). ComparisonsbetweenTestDetectedHIVStatus andSelf-ReportedHIVStatusWhenweexaminedHIVprevalenceaccordingtothetwo outcomemeasures(self-reportedvs.testdetected),the prevalenceofself-reportedHIV(6.3%)was1.5times loweroverallthantestdetectedHIVprevalence(9.5%). AccordingtothematchedMcNemar'stesttherewasa signicantmisclassicationofHIVstatusthroughselfreport( P value \ 0.01).Theoveralllowerprevalenceof HIVbyself-reportcomparedtotestdetectedwasconsistentacrossmostcharacteristics(Table 1 2 ),withafew exceptions.Self-reportedandtestdetectedHIVresults, werediscrepantamongthosewhoreportedbeingborn 1456 AIDSBehav(2012)16:14541463123
outsideAustralia,thosenotidentifyingashomosexualor bisexual,andthosereportingnothavingrecently(inthe past12months)testedforSTIsorevertestedforHIV. Comparingthetwomultivariablelogisticregressionmodels,thetwomodelswerelargelycomparable.However, discrepantfromself-reportedHIVstatus,testdetectedHIV statuswas1)notassociatedwithrecentSTItesting(inthe past12months);2)associatedwithgroupsexwithany partner;and3)lessstronglyassociatedotherUAICand witholderage. Table1 Demographicsand engagementwithgaymenby self-reportHIVstatusandtest detectedHIVresult( n = 639) Unknownsexcludedfromtable CI condenceinterval, OR unadjustedoddratioP value \ 0.05aUnemployedincludes: unemployed,studentsandotherbOtherincludes:heterosexual, openminded,sexual,undened/ unlabelledcSocialengagementisa summaryvariablemadeupof matrixoftwoquestions,namely Proportionsoffriendsthatare gay'andTimeisspentwith gayorhomosexualmen' DemographicsHIVstatus Self-reportTestdetected HIV ? /Total(%)HIV ? /Total(%) 40/6396.361/6399.5 Recruitmentsite Bars/Clubs14/2575.519/2756.9 SOPVs26/3607.242/36411.5 Agegroup(years) 18292/1931.05/1952.6 30396/1803.314/1817.7 404921/15014.026/15017.3 50 ? 11/8712.613/8914.6 Medianage45/3644/36 Countryofbirth Australia32/4487.143/4529.5 Other8/1624.915/1629.3 Residentiallocation MetropolitanMelbourne35/5206.750/5249.5 Rural&RegionalVictoria2/1811.13/1816.7 Interstate/Other2/454.43/456.7 Ethnicity Anglo-Australian23/3366.931/3399.2 Other17/2746.227/2769.8 Education Secondaryorless11/1766.319/17710.7 Further/Vocational9/1068.512/10711.2 Degree/postgraduate19/3245.926/3268.0 Employment Employed(full/parttime)27/5085.345/5128.8 Unemployeda5/776.55/776.5 Pensioner/socialsecuritybenets7/2231.87/2231.8 Sexualidentity Homosexual/Bisexual39/5946.658/6129.5 Otherb1/234.33/2611.5 SocialengagementwithgaymencLow/Moderate18/3565.134/3649.3 Extensive21/2578.226/2689.70 KnowingsomeonewithHIV No8/2203.612/2215.4 Yes32/3259.942/32812.8 Acceptabilityoforaluidtesting Uncomfortable/Veryuncomfortable4/429.54/439.3 Comfortable/verycomfortable35/5616.253/5659.4 AIDSBehav(2012)16:14541463 1457123
CharacteristicsofUndiagnosedHIVPositiveCases Ofthe61mentestingHIVpositive,36menself-reported asHIVpositive(fourofthemenself-reportingasHIV positivereturnedaHIV-negativetestresults),19selfreportedasHIV-negative,withsixmendidnotreporttheir HIVstatusatall.Ofthese19undiagnosedHIVpositive men,six(31.6%)reportednoHIVtestinghistory,six (31.6%)reportedtheirlastHIVtestasmorethan 12monthsagoandseven(36.8%)reportedaHIVtestin thepast12months.Almostathirdofmen(31.6%)with undiagnosedHIVreportedmorethan10sexpartnersand overhalf(52.9%)reportedunprotectedanalintercourse (UAI)withcasualpartnersinpast6months.Themajority (80.0%)reportedgroupsexinthepast6months.Almost two-thirds(63.1%)ofundiagnosedHIVpositivemen reportedbeingverycondent'orcondent'inknowing theirHIVstatus. Discussion ThisistherststudytoassessbiologicalHIVprevalence amonggaymeninsocialandsexvenuesinMelbourne, Australia.TestdetectedbiologicalHIVprevalencewas 9.5%(95%CI7.512.0),whileself-reportedHIVprevalencewas6.3%(95%CI4.58.4%).Wefoundasignicant Table2 Reportedsexualriskbehaviours,HIVtestinghistory,STI testinghistorybyself-reportHIVstatusandtestdetectedHIVresult ( n = 639) HIVStatus Self-reportedTestdetected HIV ? / Total (%)HIV ? / Total (%) 40/63961/639 No.sexpartners(inpast6months) 10orless16/4223.832/4367.3 Morethan1023/18812.228/19214.6 Lookforsexontheinternet Never11/1517.312/1527.9 Occasionally/often24/3646.635/3679.5 Sexwithregularpartnerinpast6monthsaYes23/3267.034/34010.0 UAIwithregularpartnera(inpast6months) Never7/1265.615/13211.4 Occasionally/Often15/1858.118/1929.4 Groupsexwithregularpartner(inpast6months) No5/1463.412/1537.8 Yes14/8616.315/9016.7 Relationshipwithcurrentregularpartner Monogamousrelationship1/611.61/611.6 Openrelationshipb8/1996.715/12512.0 Severalregularmale partnersc3/2611.54/2615.4 HIVstatusofcurrentregularpartner Negative3/1422.111/1437.7 Positive9/1850.08/1844.4 Don'tknow1/571.82/573.5 Sexwithcasualpartnerinpast6monthsaYes35/4967.152/51110.2 UAIwithcasualpartnera(inpast6months) Never10/3053.320/3166.3 Occasionally/Often25/16814.932/17218.6 Groupsexwithcasualpartner(inpast6months) No6/2192.713/2225.9 Yes29/25611.338/25914.7 DiscloseofHIVstatustocasualpartners None8/2273.515/2316.5 Some19/15712.123/15814.6 All8/839.611/8313.3 DiscloseofHIVstatusbycasualpartners None9/2393.818/2437.4 Some/All25/23010.932/23213.8 40/63961/639 ReceivedPEPinthepast6months No37/5516.752/5559.4 Yes2/296.93/2910.3 Table2 continued 40/63961/639 AnySTItestinpast12months No1/1400.77/1405.0 Yes36/4468.148/44910.7 EverhadaHIVTest No0/1420.06/1424.2 Yes40/4758.453/47511.2 IfYestoeverhadaHIVtest40/4758.453/47511.2 HIVtestinghistorydInthepast12months28/3208.832/32010.0 Between14yearsago3/744.16/748.1 Morethan4yearsago7/3619.49/3625.0 Unknownsexcludedfromtable CI condenceinterval, OR unadjusted oddratioP value \ 0.05aNotmutuallyexclusivebOpenrelationshipreferstoeither/bothmypartner&Ihavecasual sexwithothermen)cSeveralmaleregularpartnersreferstomorethanoneregularmale partnerdBasedonlastHIVantibodytest,excludethosenevertested 1458 AIDSBehav(2012)16:14541463123
discrepancybetweentestdetectedbiologicalandselfreportHIVstatusinourstudy,with19men(31.1%ofall mentestingpositiveforHIV)unawareoftheirHIV infection.MenwithundiagnosedHIVcommonlyreported highrisksexualbehaviourforthetransmissionofHIVand infrequentHIVtestingpatterns.Althoughnumberswere small,wenotedthatthediscrepancyinprevalencewas morepronouncedamongparticularparticipantsandindependentcorrelatesofHIVvariedacrossself-reportedand testdetectedHIVstatus.Whileoursamplemaynotbe representativeallgaymen[ 2 ]andmaybeconsideredhigh riskgiventheproportionofSOPVrecruitedmen,sample demographicsandsexualriskbehavioursarelargely comparablewithotherAustralianstudies[ 9 11 24 ]. Recruitmentprotocolswerealsolikelytohaveprovideda samplerepresentativeofgaymenmostatriskofHIV[ 25 ] andbetterplacedtomeetthestudyaims. ThetestdetectedbiologicalHIVprevalenceof9.5%is similartothatreportedintheonlyothercommunity-based HIVprevalencestudyconductedamonggaymeninAustralia(8.8%)[ 11 ]andintheUK(9.1%)[ 26 ],though slightlylowerthanstudiesintheUS(12.119%)[ 16 27 ]. Table3 CorrelatesofHIV Infection:Univariableand multivariablelogisticregression modeloffactorsrelatedtotest detectedHIVstatus( n = 639) Unknownsexcludedfromtable CI condenceinterval, OR unadjustedoddratioP value \ 0.05aAdjustedORforrecruitment siteandareaofresidencebOtherincludeRural/Regional Victoria,InterstateandothercUnemployedincludes: unemployed,studentsandotherdGroupsexwithregularand/or casualpartnerseMenwithoutaregularpartner wereclassiedasNegative/ Don'tknow, X Hosmer Lemeshowgoodness-of-ttest Self-reportsurveydataUnivariableMultivariable P value OR(95%CI)AdjustedORa(95%CI) Recruitmentsite Bars/clubs1.01.0 SOPVs1.8(1.03.1)1.3(0.62.6)0.49 Residentiallocation MetropolitanMelbourne1.01.0 Otherb1.0(0.42.4)0.9(0.32.4)0.82 Agegroup \ 40years1.01.0 [ =40years3.7(2.16.5)2.7(1.45.2)\ 0.01 Employment Employed(full/parttime)1.0 Unemployedc0.7(2.81.9)Pensioner/socialsecuritybenets4.8(1.912.5)KnowingsomeonewithHIV No1.0 Yes2.6(1.35.0)No.sexpartners(inpast6months) 10orless1.0 Morethan102.2(1.33.7)UAIwithcasualpartner(inpast6months) No1.01.0 Yes3.6(2.16.3)3.0(1.65.6)\ 0.01 GroupSexwithanypartnerd(inpast6months) No1.01.0 Yes2.0(1.65.1)2.1(1.04.2)0.04 HIVstatusofcurrentregularpartnereNegative/Don'tknow1.01.0 Positive11.7(5.126.7)10.0(3.726.7)\ 0.01 HIVtestinghistory Inthepast12months1.0 [ 12monthsago1.4(0.72.6) Nevertested0.4(0.20.9)AnySTItestinpast12months No1.0 Yes2.9(1.17.4)Goodness-oft = 0.6760 X AIDSBehav(2012)16:14541463 1459123
Theself-reportprevalenceof6.3%inourstudyisconsistentwithpreviousHIVprevalenceestimatesinAustralia basedonself-reportsurveillancedatawhichvarybetween 510%[ 9 28 29 ].Despiteahighproportionofmen accuratelyself-reportingtheirHIVstatus,31%(19/61)of HIVpositiveparticipantsinthisstudywereunawareof theirinfection.ImportantlyundiagnosedHIVinfectionsare knowntocontributedisproportionatelytonewtransmissions[ 9 ]andposeaseriousthreattotheeffectivenessof sero-sortingstrategiesforreducingHIVtransmission[ 30 ]. However,giventherelativelysmallnumbersof undiagnosedinfectionsinthisstudy( n = 19),results shouldbeinterpretedcautiously.Thisstudyprovidedusefuldataontheprevalenceandcharacteristicsofundiagnosedinfectionthatcanonlybecapturedbytheadditionof abiologicaloutcometobehaviouralsurveillancedata. Suchdatacanbeusedtotargetmenwhomaybemore likelytohaveundiagnosedHIVwithinitiativestopromote regulartestingandpreventsecondarytransmissions. InthecontextofaHIVprevalenceestimateapproaching 10%inthissampleandcontinuinghighratesofotherSTIs inthispopulation[ 1 ],furtherconsiderationofthe Table4 CorrelatesofHIV Infection:Univariableand multivariablelogisticregression modeloffactorsrelatedtoselfreportedHIVStatus( n = 639)P value \ 0.05aAdjustedORforrecruitment siteandareaofresidence, X HosmerLemeshow goodness-of-ttestbOtherincludeRural/Regional Victoria,InterstateandothercGroupsexwithregularand/or casualpartnersdMenwithoutaregularpartner wereclassiedasNegative/ Don'tknow. CI condence interval,ORunadjustedodd ratio Self-reportsurveydataUnivariableMultivariable P value OR(95%CI)AdjustedORa(95%CI) Recruitmentsite Bars/clubs1.01.0 SOPVs1.3(0.72.6)1.3(0.53.3)0.58 Residentiallocation MetropolitanMelbourne1.01.0 Otherb1.0(0.32.7)0.8(0.23.7)0.94 Agegroup \ 40years1.01.0 [ =40years7.1(3.215.7)8.0(3.021.1)\ 0.01 Employment Employed(full/parttime)1.0 Unemployedb0.7(0.31.9) Pensioner/socialsecuritybenets4.8(1.912.5)KnowingsomeonewithHIV No1.0 Yes2.9(1.36.4)No.sexpartners(inpast6months) 10orless1.0 Morethan103.5(1.86.9)UAIwithcasualpartner(inpast6months) No1.01.0 Yes5.1(2.610.1)6.5(2.815.3)\ 0.01 Groupsexwithanypartnerc(inpast6months) No1.0 Yes4.0(1.88.5)HIVstatusofcurrentregularpartnercNegative/Don'tknow1.01.0 Positive20.9(8.849.4)26.7(7.595.8)\ 0.01 HIVtestinghistory Inthepast12months1.0 [ 12monthsago1.1(0.52.4) Nevertested AnySTItestinpast12months No1.01.0 Yes12.2(1.789.8)10.8(1.483.7)0.02 Goodness-oft = 0.9993 X 1460 AIDSBehav(2012)16:14541463123
effectivenessofcurrentHIVtestingasprevention approachesinAustraliaiswarranted.CurrentAustralian testingguidelinesrecommendHIVtestingannuallyfor sexuallyactivegaymenandmorefrequenttesting (36monthly)formenathighrisk'[ 31 ].Despitethese recommendedtestingfrequenciesapplyingtoalargeproportionofourstudysample,mostmenreportedmuch lowerthantherecommendedtestingrateswithonly52% reportingrecent(withinthepast12months)HIVtesting. Althoughpreviousself-reportedannualHIVtestingrates amonggaymeninAustraliaarehigh(6070%)[ 24 32 33 ]comparedtoothercountries(e.g., 40.0%intheUK [ 34 ]),ourndingssuggestthatcurrentHIVtestingrates maybeinsufcienttolimittheimpactofundiagnosedHIV ontransmissionsamongAustraliangaymen. Arangeofstrategiesshouldbeconsideredtoaddressselfperceivedandstructuralbarriers[ 35 ]totestingandincrease thefrequencyofHIVtestingamonggaymeninAustralia. Greaterawarenessoftheneedformorefrequenttesting throughhealthpromotionshouldbeconsidered,including enhancingawarenessofpersonalriskprolesindrivingtestingfrequency.Optimisingclinicalsystemsshouldalsobe exploredsuchasutilisingnursesorpeer-educatorsfortesting orusingrecallsystemsandelectronicpromptstoencourage morefrequenttestingpatterns.Theuseofnewtechnologies, includingtextmessaging[ 36 ]andcomputerbased-technology[ 37 ]haveshownpositiveresultsinimprovingclinic attendanceandshort-termbehaviouraloutcomes.Formen whohavenevertested,itmaybeworthexploringothertesting modelssuchascommunity-basedsites[ 38 ]andHIVrapid testing[ 39 ].Unlikemanyothercountries[ 38 40 41 ],AustraliadoesnotofferrapidtestingaspartofHIVscreening despiteincreasingevidenceofstrongconsumerandprovider supportforthisformoftesting[ 42 43 ]. Thecollectionoforaluidspecimensalongside behaviouralsurveyswaswellreceivedwithaparticipation rateofover60%,similartotheannualGayCommunity PeriodicSurveyswhichdonotincludeoraluidtesting [ 21 ],demonstratingahighdegreeofacceptabilityof community-basedbiologicalprevalencetestinginthis population.Nearlyallparticipantsreportedtheywere comfortable'orverycomfortable'withhavinganoral uidsamplecollectedforHIVtestingandover90%of participantsreportedthattheywouldparticipateinfuture MCGPSiforaluidspecimenswerecollected.These ndingssuggestthatitwouldbefeasibletoincorporate oraluidspecimencollectioninfuturebehaviouralsurveys orsimilarcommunity-recruitedstudies.HIVtestingusing nger-prickandoraluidspecimencollectionforestimatingHIVprevalenceamongMSMhasbeenwidely adoptedintheU.S[ 16 ]andthroughoutEurope[ 14 ],Africa [ 44 ]andAsia[ 45 ],toenhancesurveillancesystemsandthe accuracyofHIVestimates. Thendingsinthisstudyaresubjecttoseverallimitations.Therecruitmentstrategyandsamplingmethodmay haveresultedinselectionbias,thuslimitingtherepresentativenessofthissampleandthegeneralisabilityoftheresults. Ourrecruitmentprotocolreplicatedonlythevenue-based recruitmentfortheMGCPS.Wedidnotrecruitatthegayfair dayorMidsummaCarnival',whichconstitutesapproximatelytwo-thirdsofMGCPSrespondents.Thiscontributed tothehigherproportionofmenrecruitedatSOPVscomparedtogaybars/clubsandthussomewhatlimitsdirect comparisonsbetweenthetwosamples;althoughdemographicallythesamplesareverysimilar[ 21 33 ].Butin relationtotheaimsofthisstudy,giventheplacesgaymen reportmeetingsexpartners[ 33, 35 ],socialvenuerecruitmentislikelytobetterrepresentthosemostatriskofHIV. Reportingbiasmayhaveaffecteddatacollection;asmall numberofparticipant's( n = 4)self-reportedtheywereHIV positivebutreturnedanegativeHIVtest.Thereasonsforthis arenotclearbutcouldrelatetoincorrectformcompletion, misinterpretationofthequestion,abeliefthattheyreally wereHIVpositiveorthetestresultwasafalsenegative (althoughunlikelygiventhewelldocumentedhighperformanceofthetest)[ 22 46 ].Responderbiasmayhavealso haveaffecteddatacollectionhowever,giventhatindividual testresultswerenotprovidedbacktoparticipantsand,as writtenresearcherlogsandverbalfeedbackfromrecruiters suggest,onlyaverysmallproportionofmenapproached chosenottoparticipatebecauseoforaluidspecimencollection,webelievethiswasminimised.Asthepaper-based surveyswereself-administered,thereisnowaytodenitivelyverifymen'sself-reportedHIVstatus.Thisstudydid providesomeadvantagesovercurrentself-reportHIV prevalenceestimatesinAustralia;oraluidtestingprovided non-invasivespecimencollectionallowingthemeasurement ofthetrueHIVprevalenceamongthispopulationinan acceptablemanner. Conclusions HIVprevalenceinAustralianMSMremainshigh.SelfreportHIVprevalenceunderestimatesactualprevalence, withover30%ofHIVpositiveMSMinthisstudybeing unawareoftheirstatus.Thesedataaddgreatlytoour understandingoftheHIVepidemicinAustralia,byprovidinginformationaboutself-reportedversustestdetected HIVprevalenceandfactorsassociatedwithHIVinfection. TheseresultsalsohighlighttheimportanceofongoingHIV preventionprogramstoencourageregulartestingtoreduce hightransmissionrates,andalsotoensurethatHIVbiologicaltestingisundertakeninconjunctionwithcommunity-basedsurveysaboutHIVtoimproveHIVsurveillance inAustralia. AIDSBehav(2012)16:14541463 1461123
Acknowledgments Thefollowingco-authorshavecontributedto thework,DrKimWilsoninstudydesign,laboratoryworkand manuscriptconception;DrRebeccaGuyinstudydesign,manuscript preparationandmanuscriptreview;DrMargaretHellardinstudy design,manuscriptconceptionandreviewandDrMarkStoovein studydesign,manuscriptconception,manuscriptpreparationand review.WeacknowledgetheassistanceofSueBestfromtheNational SerologyReferenceLaboratorywithplanningandprovidingassistancewiththetestingprotocol.Wealsowouldliketoacknowledge thehelpoftheprojectadvisorycommittee,particularlythatofGarrett Prestageforguidingtheplanningandimplementationofthestudy. AlisaPedranareceivesfundingfromtheAustraliaGovernment throughaNationalHealthandMedicalResearchCouncil (NH&MRC)PublicHealthPostgraduateScholarshipandtheSidney MyerHealthScholarship.MargaretHellardreceivesfundingfromthe NH&MRCasaseniorresearchfellow.RebeccaGuyreceives fundingfromtheNH&MRCasapostdoctoralfellow.Theauthors gratefullyacknowledgeVictorianDepartmentofHealthforproviding fundstosupportthisstudyandthecontributiontothisworkofthe VictorianOperationalInfrastructureSupportProgram.Thisproject wasfundedbytheVictorianDepartmentofHealth. Conictofinterests Nocompetingintereststoreport.References1.NationalCentreinHIVEpidemiologyandClinicalResearch. HIV/AIDS,viralhepatitisandsexuallytransmissibleinfectionsin AustraliaAnnualSurveillanceReport2010.Sydney:National CentreinHIVEpidemiologyandClinicalResearch,TheUniversityofNewSouthWales;2010. 2.PrestageG,FerrisJ,GriersonJ,ThorpeR,ZablotskaI,ImrieJ, etal.HomosexualmeninAustralia:population,distributionand HIVprevalence.SexHealth.2008;5(2):97102. 3.ZablotskaIB,PrestageG,MiddletonM,WilsonD,GrulichAE. ContemporaryHIVdiagnosestrendsinAustraliacanbepredicted bytrendsinunprotectedanalintercourseamonggaymen.AIDS. 2010;24(12):19558. 4.MiddletonMG,GrulichAE,McDonaldAM,DonovanB, HockingJS,KaldorJM.Couldsexuallytransmissibleinfections becontributingtotheincreaseinHIVinfectionsamongmenwho havesexwithmeninAustralia?SexHealth.2008;5(2):13140. 5.TeagueR,MijchA,FairleyCK,SidatM,WatsonK,BoydK, etal.Testingratesforsexuallytransmittedinfectionsamong HIV-infectedmenwhohavesexwithmenattendingtwodifferent HIVservices.IntJSTDAIDS.2008;19(3):2002. 6.NationalCentreinHIVEpidemiology,ClinicalResearch.HIV/ AIDS,viralhepatitis,sexuallytransmissibleinfectionsinAustraliaAnnualSurveillanceReport2008.Sydney:NationalCentre inHIVEpidemiologyandClinicalResearch,TheUniversityof NewSouthWales;2009. 7.BuchaczK,PatelP,TaylorM,KerndtPR,ByersRH,Holmberg SD,etal.SyphilisincreasesHIVviralloadanddecreasesCD4 cellcountsinHIV-infectedpatientswithnewsyphilisinfections. AIDS.2004;18(15):20759. 8.WasserheitJN.Epidemiologicalsynergy.Interrelationships betweenhumanimmunodeciencyvirusinfectionandother sexuallytransmitteddiseases.SexTransmDis.1992;19(2): 6177. 9.WilsonDP,HoareA,ReganDG,LawMG.ImportanceofpromotingHIVtestingforpreventingsecondarytransmissions: modellingtheAustralianHIVepidemicamongmenwhohave sexwithmen.SexHealth.2009;6(1):1933. 10.deWitJ,TreloarC,WilsonH.HIV/AIDS,hepatitisandsexually transmissibleinfectionsinAustralia:Annualreportoftrendsin behaviour2009.Sydney:NationalCentreinHIVSocial Research,TheUniversityofNewSouthWales;2009. 11.BirrellF,StauntonS,DebattistaJ,RoudenkoN,RutkinW,Davis C.Pilotofnon-invasive(oraluid)testingforHIVwithina communitysetting.SexHealth.2010;7(1):116. 12.SifakisF,FlynnCP,MetschL,LaLotaM,MurrillC,KoblinBA, etal.HIVprevalence,unrecognizedinfection,andHIVtesting amongmenwhohavesexwithmen-FiveUScities,June2004 April2005.MorbMortalWklyRepSurveillSumm.2005;54(24): 597601. 13.WilliamsonLM,DoddsJP,MerceyDE,HartGJ,JohnsonAM. SexualriskbehaviourandknowledgeofHIVstatusamong communitysamplesofgaymenintheUK.AIDS.2008;22(9): 106370. 14.MirandolaM,FolchTodaC,KrampacI,NitaI,StanekovaD, StehlikovaD,etal.HIVbio-behaviouralsurveyamongmenwho havesexwithmeninBarcelona,Bratislava,Bucharest,Ljubljana,PragueandVerona,20082009.Eurosurveill.2009;14(48): 229. 15.MishraV,VaessenM,BoermaJT,ArnoldF,WayA,BarrereB, etal.HIVtestinginnationalpopulation-basedsurveys:experiencefromtheDemographicandHealthSurveys.BullWorld HealthOrgan.2006;84(7):53745. 16.SmithA,MilesI,LeB,FinlaysonT,OsterAM,DinennoE. PrevalenceandawarenessofHIVinfectionamongmenwhohave sexwithmen-21cities,US,2008.MorbMortalWklyRepSurveillSumm.2010;59(37):12017. 17.MahfoudZ,AR,RamiaS,KhouryDE,KassakK,BarbirFE, etal.HIV/AIDSamongfemalesexworkers,injectingdrugusers andmenwhohavesexwithmeninLebanon:Resultsoftherst biobehaviorialsurveys.AIDS.2010;24(2):S4554. 18.VanGriensvenF,VarangratA,WimonsateW,TanpradechS, KladsawadK,ChemnasiriT,etal.TrendsinHIVprevalence, estimatedHIVincidence,andriskbehaviouramongmenwho havesexwithmeninBangkok,Thailand,20032007.JAcquir ImmuneDecSyndr.2010;53(2):2349. 19.PhangCW,HockingJ,FairleyCK,BradshawC,HayesP,Chen MY.Morethanjustanalsex:thepotentialforsexuallytransmittedinfectiontransmissionamongmenvisitingsex-on-premisesvenues.SexTransmInfect.2008;84(3):2179. 20.SchwartlanderB,GhysPD,PisaniE,KiesslingS,LazzariS, CaralM,etal.HIVsurveillanceinhard-to-reachpopulations. AIDS.2001;15(3):S13. 21.EvelynL,HoltM,ZablotskaI,PrestageG,McKenzieT,BatrouneyC,etal.GayCommunityPeriodicSurvey:Melbourne 2010.MelbourneNationalCentreinHIVSocialResearch.Sydney:FacultyofArtsandSocialSciences,TheUniversityofNew SouthWales;2010. 22.ParryJV,PerryKR,MortimerPP.Sensitiveassaysforviral antibodiesinsaliva:analternativetotestsonserum.Lancet. 1987;2(8550):725. 23.StataCorp.IntercooledStata.In.9.0ed.CollegeStation:Stata Corp;2004. 24.HoltM,MaoL,PrestageG,ZablotskaI,deWitJ.Gaycommunityperiodicsurveysnationalreport2010.Sydney:National CentreinHIVSocialResearch,NationalCentreinHIVEpidemiologyandClinicalResearch.TheUniversityofNewSouth Wales;2010. 25.DownI,BradleyJ,EllardJ,BrownG,GrulichA,PrestageG. ExperiencesofHIV:theseroconversionstudyannualreport 2010.Sydney:TheKirbyInstitute,TheUniversityofNewSouth Wales,2010. 26.WilliamsonLM,DoddsJP,MerceyDE,HartGJ,JohnsonAM. SexualriskbehaviourandknowledgeofHIVstatusamong 1462 AIDSBehav(2012)16:14541463123
communitysamplesofgaymenintheUK.AIDS.2008;22(9): 106370. 27.SifakisF,HyltonJB,FlynnC,SolomonL,MacKellarDA, ValleroyLA,etal.PrevalenceofHIVinfectionandpriorHIV testingamongyoungmenwhohavesexwithmen.TheBaltimore youngmen'ssurvey.AIDSBehav.2010;14(4):90412. 28.PrestageG,JinF,ZablotskaI,ImrieJ,KaldorJM,GrulichAE. TrendsinHIVprevalenceamonghomosexualandbisexualmen ineasternAustralianstates.SexHealth.2008;5(2):1037. 29.GrulichAE,deVisserRO,SmithAM,RisselCE,RichtersJ.Sex inAustralia:sexuallytransmissibleinfectionandblood-borne virushistoryinarepresentativesampleofadults.AustNZJ PublicHealth.2003;27(2):23441. 30.MontoyaJA,KentCK,RotblattH,McCrightJ,KerndtPR, KlausnerJD.Socialmarketingcampaignsignicantlyassociated withincreasesinsyphilistestingamonggayandbisexualmenin SanFrancisco.SexTransmDis.2005;32(7):3959. 31.BourneC,EdwardsB,ShawM,GowersA,RodgersC,FersonM. Sexuallytransmissibleinfectiontestingguidelinesformenwho havesexwithmen.SexHealth.2008;5(2):18991. 32.JinFY,PrestageG,LawMG,KippaxS,VandeVenP,RawsthorneP,etal.PredictorsofrecentHIVtestinginhomosexual meninAustralia.HIVMed.2002;3(4):2716. 33.LeeE,HoltM,MaoL,McKenzieT,BatrouneyC,KennedyM, etal.GayCommunityPeriodicSurvey:Melbourne2011.Melbourne:NationalCentreinHIVSocialResearch,TheUniversity ofNewSouthWales;2011. 34.HicksonF,BourneA,WeatherburnP,ReidD,JessupK,HammondG.Tacticaldangers:ndingsfromtheUnitedKingdom GayMen'sSexSurvey2008.London:SigmaResearch;2010. 35.PrestageG,McCannP,HurleyM,BradleyJ,DownI,BrownG. Pleasureandsexualhealth:thePASHstudy.Sydney:NCHECR (NationalCentreinHIVEpidemiologyandClinicalResearch); 2009. 36.BourneC,KnightV,GuyR,WandH,LuH,McNultyA.Short messageservicereminderinterventiondoublessexuallytransmittedinfection/HIVre-testingratesamongmenwhohavesex withmen.SexTransmInfect.2011;87:22931. 37.NoarSM,BlackHG,PierceLB.Efcacyofcomputertechnology-basedHIVpreventioninterventions:ameta-analysis.AIDS. 2009;23(1):10715. 38.BowlesKE,ClarkHA,TaiE,SullivanPS,SongB,TsangJ,etal. ImplementingrapidHIVtestinginoutreachandcommunity settings:resultsfromanadvancingHIVpreventiondemonstrationprojectconductedinsevenUScities.PublicHealthRep. 2008;123(3):7885. 39.AguirreD,Mares-DelGrassoA,EmersonC,TsangJ,PincusJ, CalhounC,etal.RapidHIVtestinginoutreachandothercommunitysettingsUnitedStates,20042006(Table).MMWR MorbMortalWklyRep.2007;56(47):12337. 40.SmytheE.Evaluationof3monthAucklandbasedrapidtestpilot. Auckland:NewZealandAIDSFoundationReport:2007. 41.DeLaFuenteL,DelgadoJ,HoyosJ,BelzaMJ,AlvarezJ, GutierrezJ,etal.IncreasingearlydiagnosisofHIVthroughrapid testinginastreetoutreachprograminSpain.AIDSPatientCare STDS.2009;23(8):6259. 42.ChenMY,BilardiJE,LeeD,CummingsR,BushM,FairleyCK. AustralianmenwhohavesexwithmenpreferrapidoralHIV testingoverconventionalbloodtestingforHIV.IntJSTDAIDS. 2010;21(6):42830. 43.LeeD,FairleyC,CummingsR,BushM,ReadT,ChenM.Men whohavesexwithmenpreferrapidtestingforsyphilisandmay testmorefrequentlyusingit.SexTransmDis.2010;37(9):5578. 44.BurrellE,MarkD,GrantR,WoodR,BekkerLG.Sexualrisk behavioursandHIV-1prevalenceamongurbanmenwhohave sexwithmeninCapeTown,SouthAfrica.SexHealth. 2010;7(2):14953. 45.vanGriensvenF,VarangratA,WimonsateW,TapperoJ,etal. HIVprevalenceamongpopulationsofmenwhohavesexwith menThailand,2003and2005.MMWRMorbMortalWklyRep. 2006;55(31):8448. 46.ConnellJA,ParryJV,MortimerPP,DuncanJ.Novelassayfor thedetectionofimmunoglobulinGantihumanimmunodeciency virusinuntreatedsalivaandurine.JMedVirol.1993;41(2): 15964. AIDSBehav(2012)16:14541463 1463123