Correlates of sexual outcome expectations and risk of sexually transmitted infections ( STIs ) among male inmates in the United States

This study sought to identify variables that independently correlate with sexual outcome expectations among a sample of 187 convicted felons in Georgia. Trained interviewers collected data from study participants after explaining the purpose of the study and obtaining informed consent. Relationships between individual sexual outcome expectation variables and other demographics were examined using multiple linear regression models. Models specified self-reported sexual outcome expectation regarding a given infectious disease risk as an outcome with selected variables. Respondents’ age was a predictor of using condoms each time they had sex (Beta = 0.15, P < 0.04) and limiting the number of sex partners (Beta = 0.19, P < 0.008). Incarceration history (first time versus prior incarceration) was significantly associated with using and/or requesting that their partners use condoms each time they had sex (p < 0.03), and being able to ask partners about their sexual history (p < 0.03). Our findings imply that interventions to prevent infectious diseases among inmates should address education and illicit substances specifically in sexual situations so as to better understand the nature of risk practices and sexual behavior among this group.


INTRODUCTION
Infectious diseases and sexually transmitted infections (STIs) are both expensive and principal public health problems in the United States and are even more so among incarcerated populations (Lang and Belenko, 2001;Harris et al., 2002;Narevic et al., 2006;Stephens et al., 2002).Nationally, rates of human immunodeficiency virus (HIV), tuberculosis (TB) and heppitatus C (HCV) are greater for incarcerated populations than the general community (Bruneau et al., 2001;Mullings et al., 2003;Dolan et al., 2005).Specifically, rates of STIs are substantially higher in prison populations than in the public worldwide (Massad et al., 1999;Clarke et al., 2001).
Many of these studies that examine inmate populations and psychosocial and behavioral constructs tend to look at overt behaviors.Lang and Belenko (2001) found among other factors, that age, physical abuse history, anxiety, family problems and poly-substance use were predictive of sex-related HIV risk.Incarcerated populations tend to have multiple HIV risk behaviors.
These include having a relatively high lifetime number of sexual partners, limited condom use and having sex under the influence of drugs and alcohol (Bruneau et al., 2001;Belenko et al., 2005;Stephens et al., 2004;Johnson et al., 2004;Narevic et al., 2006).

Theoretical frame work
Rooted in social cognitive theory, the concept of outcome expectation is an important component of health-related behavioral change.According to this theory, outcome expectations may be shaped by social norms, knowledge, and self-efficacy (Bandura, 1977;DiIorio et al., 2006).Previous studies of HIV risk behaviors and sexual health have consistently shown that low outcome expectations are strong predictors of behaviors that increase risk for infectious disease among adults with mental illness (Kelly et al., 1995), heterosexual adult female STD patients (Thompson et al., 1997) and heterosexual adult males (Bengel et al., 1996).
Little is known about outcome expectations related to sexual practices among inmates and how they may differ by sociodemographic factors unique to inmate populations.In addition, few data on sexual outcome expectation among inmate populations exist; yet, this population is at particularly high risk for infectious diseases, including STIs.What is documented is that sexual self expectation is a predictor of HIV risk among inmate populations and can be impacted positively with peer-based interventions (Braithwaite et al., 2005).Thus, the objectives of this study were to estimate independent correlates of sexual outcome expectations with respect to infectious disease risk behaviors of a sample of male inmates in Georgia and to identify inmate characteristics associated with these psychosocial outcomes.

METHODOLOGY Study site and sample
This study collected data from adult male inmates in Georgia, USA.Baseline level data was collected after the identification of a pool of eligible participants by the Department of Correction's Personnel.Inmates had to be between 60 and 90 days prior to release from the facility, returning to the metropolitan Atlanta area (this criterion would make follow up more feasible), 18 years of age or older, and male to be eligible for inclusion.The participants were recruited from a population housed at three medium security correctional institutions for men located in middle Georgia and a transitional center located in a major southeastern city.

Data collection
Trained interviewers collected data from study participants after explaining the purpose of the study and obtaining informed consent.The interviewer provided an overview of the major areas to be covered with the data collection instrument to each participant to make certain that they understood the response format.The instrument was written on a fourth grade reading level.
Participants were 187 adult male inmates recruited from three Stephens and Braithwaite 93 medium security prisons and one Area Transition Center in Georgia.The ages of participants ranged from 18 to 59 years (M = 35.3,SD = 8.9) with the majority of the participants (126 = 67.4%)being African American.Although nearly half (87 = 46.28%)had been incarcerated prior to the current incarceration, the mean total lifetime years incarcerated was 9.17 (SD = 15.4).

Analysis
Relationships between individual sexual outcome expectation variables and other demographics were examined using multiple linear regression models.All models specified self-reported sexual expectation regarding a given infectious disease risk as an outcome with selected demographic variables (race, age, education, income prior to incarceration, length of incarceration, first time incarceration and total years incarcerated) as predictor variables.All measures were entered independently into the estimated equations.

Measures
Participants were asked to provide descriptive and scale information on several variables using standard response formats.Socio-demographic variables on the data collection instrument included: ethnicity, marital status, educational level, income prior to incarceration, and incarceration history.Several individual items represented incarceration status.Length of incarceration was an open ended item that asked participants to write the total years served during their recent incarceration.Prior or first time incarceration was measured with a dichotomous response item worded: "Was this your first time being incarcerated?"A measure of sexual outcome expectations was obtained by asking individuals eight items about "how confident are you that you will be able to do the following activities when you are released from prison?" Example items included "Using a condom or request that your partner use them every time you have sex"; "limiting the number of sexual partners you have"; "engaging in sexual activity after using marijuana"; and "asking your partner whether they have a sexually transmitted infection (that is, herpes, syphilis)."Responses were measured on a 100-point Likert-type scale from (0) "not at all confident" to (100) "very confident".Cronbach's alpha coefficient for the eight items was 78.

RESULTS
Table 1 presents results of the linear regression analysis that examined whether the selected demographic variables and attributes of incarceration were associated with self-reported sexual outcome expectation among inmates prior to their current incarceration.As shown, race, age, and first time incarceration were significantly predicted and associated with inmate's ability to use and/or request that they use a condom every time during sex.Total years' incarceration was almost significantly associated with variable (p < 0.06).Race was not significantly correlated with any of the other independent sexual selfexpectation outcomes; however, it was almost associated with inmate's propensity to engage in sexually activity ex post facto drinking (p < 0.08).
Respondents' age was a predictor of using condoms each time they had sex (Beta = 0.15, P < 0.04) and limiting the number of sex partners (Beta = 0.19, P < 0.008).Age was nearly predictive of inmates being able to discuss safe sex practices with their partners (p < 0.06) and when engaging in sexual activity after using marijuana (p < 0.08).In addition, participant education approached significance and was inversely associated with inmates being able to discuss safe sex practices with their partner(s) (p < 0.08).Incarceration history (first time versus prior incarceration) was significantly associated with using and/or requesting that their partners use condoms each time they had sex (p < 0.03), and being able to ask partners about their sexual history (p < 0.03).This variable approached significance for discussing safer sex practices with their partners ( p < 0.07) and asking their partners if they have or have had an STI (p < 0.09).

DISCUSSION
In this cohort of inmate's, several variables including race, age and incarceration history were associated with several individual outcome expectations related to sexual risk practices.Of note was that education, albeit not a significant correlate of any of the individual outcome expectations for this sample changed the direction of the sample distribution below the median with the exception of "engaging in sexual activity after drinking."Thus, the more education the more inmates believed they could perform the aforementioned activities unless it was regarding behaviors under the influence of alcohol.This is not unusual since education has been considered a protective factor for a host of health risks related to infectious disease (Kral et al., 2001;Stephens et al., 2005).Income was also observed to be inversely associated with multiple outcome expectations with the exception of "asking one's partner about their history of STIs.
It is difficult to compare study findings to prior research on the construct in particular with respect to incarcerated populations.What was observed from the research, especially in the area of dietary behavior change and physical activity, was that outcome expectations can be instrumental in producing selfefficacious health behaviors (Clark and Dodge, 1999;Tercyak and Tyc, 2006;Baranowski et al., 1999).In addition, the impact of education on HIV risk as documented in this study has been presented in other studies related to predictors of HIV risk (Kral et al., 2001;Moore et al., 1999;Srinivas et al., 2000).
Thus, true primary prevention with regards to this population in terms of sexual outcome expectations and their association with HIV risk practices would entail targeting education, specifically keeping potential inmates from ever entering the prison system.This means that focusing on education, staying in school and overall academic performance can be a resilient factor that reduces both the likelihood of being incarcerated, as well as enhancing individuals ability to reduce risk based on risk outcome expectations related to sex.
One limitation of this study is the possible bias associated with asking an inmate to project future behaviors when released from prison as well as the reliability of self-reported data.Future research should incorporate other measures of sexual outcome expectations such as specific negotiation skills under the influence of other elicit substances such as ecstasy and heroin.Nonetheless, our position is that sexual outcome expectations can provide valuable insight into how well inmate populations adapt and are amenable to infectious disease risk reduction interventions.
This study was also limited by the selectivity of the sample (inmates from three prisons in middle Georgia and one area transition center), with the majority of the participants being African American.Another major problem with the study was that sexual outcome expectation measures were derived from measures used with adolescent and college samples and based on selfreport.Finally, it is possible that social desirability response biases exerted an influence on the findings.Despite these limitations, this study is an important first step to developing a reliable understanding regarding outcome expectations related to STI risk practices among inmate populations.Strengthening outcome expectations related to STI risk reduction and thereby increasing selfefficacious health practices can improve overall health and ultimately the quality of life for inmates, since they will eventually be returned to the community from which they came.
In conclusion, this study employed linear regression models to examine relationships between sociodemographic variables and sexual outcome expectations among a sample of male inmates.The findings of this study imply that interventions to prevent STIs among inmates should address their use of illicit substances, specifically in sexual situations, so as to better understand the nature of risk practices and sexual behavior among this group.

Table 1 .
Relations between sexual outcome expectations and selected variables.