Abstract
Urban women with severe mental illness (SMI) are vulnerable to stigma and discrimination related to mental illness and other stigmatized labels. Stigma experiences may increase their risk for negative health outcomes, such as HIV infection. This study tests the relationship between perceived stigma and HIV risk behaviors among women with SMI. The authors interviewed 92 women attending community mental health programs using the Stigma of Psychiatric Illness and Sexuality Among Women Questionnaire. There were significant relationships between personal experiences of mental illness and substance use accompanying sexual intercourse; perceived ethnic stigma and having a riskier partner type; and experiences of discrimination and having a casual or sex-exchange partner. Higher scores on relationship stigma were associated with a greater number of sexual risk behaviors. The findings underscore the importance of exploring how stigma attached to mental illness intersects with other stigmatized labels to produce unique configurations of HIV risk. HIV risk reduction interventions and prevention research should integrate attention to stigmatized identities in the lives of women with SMI.
Keywords: women, severe mental illness, stigma, HIV sexual risk
Many urban women living with severe mental illness (SMI) expose themselves to considerable HIV risk during sexual encounters (Meade & Sikkema, 2005; Wright, Wright, Perry, & Foote-Ardah, 2007). In fact, the prevalence of HIV infection among people with SMI is much higher than in the general population in the United States (McKinnon, Cournos, & Herman, 2002). Among women, in particular, elevated HIV prevalence and risk stem from vulnerability to coercion and victimization, gender dynamics that result in power imbalances, unprotected sex, trading sex for money or other goods, sex with high-risk partners, and substance use (Amaro, 1995; Meade & Sikkema, 2005; Otto-Salaj, Heckman, Stevenson, & Kelly, 1998). Questions remain about the context in which these factors emerge to shape women’s experiences of HIV risk.
Stigma related to mental illness may be one contextual risk factor that shapes health outcomes in the lives of women with SMI; stigma may facilitate some of the sexual risk factors noted above. Contextual risk factors such as stigma affect access to power, money, knowledge, and social connectedness, all of which influence disease risk (Link & Phelan, 1995). Less understood is whether, and through which mechanisms, this stigma affects the sexual lives of women with SMI.
Stigma Processes
Link and Phelan (2001) suggested that stigma occurs when differences are labeled and distinguished; labeled people are linked to stereotypes and distinguished as “other” or “them”; they experience status loss and discrimination. When these components are mobilized by individuals, communities, or governments with greater power, they result in reduced opportunities, rejection, and discrimination for the stigmatized group. Furthermore, laws and institutions that support power relationships maintain these unequal outcomes that stigma produces, thus creating structural discrimination (Parker & Aggleton, 2003).
State-sanctioned sterilization and institutional practices that prioritize reproductive control over HIV prevention provide examples of structural discrimination relevant to women with SMI (P. Y. Collins, 2001; Stern, 2005). A subtler example lies in the diminished opportunities for intimate relationships that can result from being engaged in the mental health care system (Wright et al., 2007). Mental health care providers’ discomfort with sexuality among people with SMI may be reflected in regulations that inadequately address intimate relationships in the lives of women with SMI. Similarly, providers’ sense of duty to protect women from coercion or exploitative sexual encounters can result in regulations that curtail access to relationships.
The discrimination that stigma generates also works through social psychological processes (Link & Phelan, 2001). Modified Labeling Theory provides an illustration. All members of our society—those with mental illness as well as those without—have ideas of what it means to be mentally ill (Link, 1987; Link, Cullen, Struening, & Shrout, 1989). These include beliefs about how employers, relatives, lovers, and others will respond to the mentally ill. When a person is newly diagnosed and labeled as a psychiatric patient, those societal beliefs become particularly important to that individual. Once labeled, people with SMI may internalize stigmatizing beliefs and expect others to reject or discriminate against them. Link and others (Link et al., 1989; Link, Struening, Rahav, Phelan, & Nuttbrock, 1997; Rosenfield, 1997; Wright, Gronfein, & Owens, 2000) have demonstrated that expectations of rejection can potentially lead to reduced confidence and impaired social interactions, constricted social networks, depressive symptoms, unemployment, income loss, and low self-esteem.
Stigma, Sexuality, and HIV Risk
When women internalize the stigma related to mental illness the social and sexual consequences may lead to greater HIV risk. For example, evidence suggests that mental illness stigma and perceived discrimination may contribute to decreased self-esteem (Lai, Hong, & Chee, 2001; Link, Struening, Neese-Todd, Asmussen, & Phelan, 2001; Ritsher & Phelan, 2004; Silverstone, 1991). In turn, low self-esteem has been associated with HIV risk behaviors such as sex with paying partners, more frequent sexual risk-taking, substance use, and reduced self-efficacy to enact risk reduction (Somlai et al., 2000; Sterk, Klein, & Elifson, 2004). Other negative self-perceptions—such as negative body image, less perceived attractiveness, and perceived undesirability—have been linked with behaviors that can increase risk for HIV (e.g., less condom use during sex) and place women in vulnerable situations (e.g., substance use before sex; Littleton, Breitkopf, & Berenson, 2005; Wingood, Diclemente, Harrington, & Davies, 2002).
The relationship between mental illness stigma and decreased self-esteem coupled with the association of low self-esteem and other self-perceptions with increased sexual risk behavior suggest that it is plausible that mental illness stigma could be associated with HIV sexual risk among women with SMI. Qualitative research suggests that some women with SMI perceive themselves as less desirable, have less power in relationships, and may be more likely to engage in sexual encounters of greater risk, involving unprotected sexual intercourse (P. Y. Collins, Unger, & Armbrister, 2008). To our knowledge, these factors have not been examined quantitatively among women with SMI.
Mental Illness and Multiple Stigmatized Statuses
For women of color with SMI, the relationship of stigma to HIV risk requires additional considerations. These women may experience stigma related to mental illness in tandem with discrimination related to race and ethnicity, poverty, and immigration status. Other studies among people with SMI have demonstrated the significance of recognizing multiple stigmatized identities for service provision and mental health (Corrigan et al., 2003; Thompson, Noel, & Campbell, 2004). Link and Phelan (2001, p. 382) noted that many stigmatizing circumstances need to be considered when studying a particular outcome.
The feminist literature on intersectionality adds further support to this approach. Intersectionality assumes that (a) systems of oppression are interlocking and (b) all of these dimensions need to be taken into account because the intersection of these forces produces unique effects on women (P. H. Collins, 1999; Stewart & McDermott, 2004). In terms of HIV risk, an intersectional approach posits that these complex systems of oppression (race, gender, and mental illness) that lead to stigmatization and discrimination interrelate in specific ways that also affect women’s power in the sexual realm. In order to understand the role of stigma in this process, one must simultaneously explore how ethnicity and other oppressed identities affect risk.
P. Y. Collins et al. (2008) showed that the multiple stigmatized statuses of ethnicity, poverty, immigration status, and inability to meet the gender role expectations of their communities contributed to HIV risk among urban Latina women with SMI. Mental illness amplified women’s disadvantage at the intersection of race, gender, and social class. Yet access to financial entitlements linked to their mental illness provided economic empowerment and independence, enabling some women to avoid certain sexual risk situations—such as exchanging sex for money, housing, and other goods—that stem from economic dependence. Quantitative research is now needed to test the extent to which these qualitative findings can be affirmed.
Quantitative Assessment of Stigma Related to Sexuality Among Women
Many women with SMI who reside in urban, impoverished neighborhoods likely carry multiple stigmatized labels (being labeled a psychiatric patient, an ethnic minority, a former drug user, or a woman who does not live up to cultural or gender expectations) that intersect to produce vulnerabilities to HIV infection as well as defenses against infection. In the absence of a measure that assesses stigma as it relates to sexuality in addition to other stigmatized identities, we developed a novel quantitative assessment to further explore the impact of multiple domains of stigma on HIV risk among women with SMI. Building on the work of Link and colleagues (Link, 1987; Link et al., 1989) and qualitative research among women of color with SMI, we measured stigma domains relevant to sexuality in this population using the Stigma in Psychiatric Illness and Sexuality Among Women (SPISEW) questionnaire.
P. Y. Collins and colleagues (2008) found that women described their personal experiences of stigma in terms of being called crazy or loca. Their worries about how revealing their mental illness would adversely affect their romantic and sexual relationships reflected anxieties about devaluation and discrimination in relationships. Some women acknowledged pressures in their daily lives related to dispelling stereotypes about their ethnic group. For others, perceived unattractiveness due to weight gain or other side effects of medications affected their confidence in relationships.
These themes formed the basis of the five domains in the SPISEW (personal experience with mental illness stigma, relationship devaluation and discrimination, ethnic devaluation and discrimination, perceived attractiveness, and experiences of discrimination). Two domains assess women’s agreement with the general public’s perceptions of women with mental illness (relationship devaluation) and other members of their ethnic group (ethnic devaluation). Two domains assess individual discrimination experiences related to mental illness, specifically, and to multiple stigmatized labels. One domain assesses internalized stigma related to perceived attractiveness as a woman with mental illness.
We present preliminary findings on the relationships between five domains of perceived stigma and sexual risk behaviors among urban women with SMI. Our study is informed by an intersectional approach—of particular relevance to women of color with mental illness—which suggests that the effects of multiple domains of stigma contribute to women’s patterns of HIV risk. Thus, experiences of stigma and discrimination may lead a woman to engage in certain risk behaviors and avoid others. We examine the associations between five domains of stigma and these HIV risk behaviors: unprotected sex, multiple partners, substance use before sex, having an exchange sex or casual partner, and having a high-risk partner. Specifically, we explore the following questions: (a) How much do urban women with SMI endorse experiences of stigmatization in any of the five domains? And (b) which, if any, domains of stigma are associated with HIV sexual risk behaviors?
Method
Sample and Data Collection
These data come from a larger study, Project Wisdom, which tested an HIV prevention intervention for women with SMI in a randomized clinical trial. We recruited women age 18 years and older from two residential facilities and four mental health day treatment programs in New York City to participate in Project Wisdom. The research team visited each site and met first with the staff, and subsequently with the women receiving services, to describe the study. Women who were cleared medically and psychiatrically by their clinicians could approach the research team’s designated screening station in each facility. After describing the study to interested women, we obtained written informed consent. Women were eligible to participate in the clinical trial if they had a severe mental illness (any chronic psychotic disorder or major affective disorder), were sexually active in the last 3 months, and had the capacity to give informed consent. HIV serostatus was not assessed. The institutional review boards of the New York State Psychiatric Institute and of Columbia University approved the study.
Eligible women completed a baseline assessment administered by trained, female interviewers. For the purposes of the current analysis, we included all women who met the diagnostic criteria and gave informed consent for the baseline assessment. Of 280 women screened, 121 women met criteria for eligibility; of these, 101 agreed to participate and gave informed consent; and 97 women completed the baseline stigma questionnaire. After exclusion for missing data, the sample included 92 women. Study participants received $30 for completing the baseline assessments.
Measures
Demographic information included participant age (in years), race or ethnicity (White, African American, Latino, or other), marital status (married; divorced, separated, or widowed; or never married), level of education (less than 12th grade, 12th grade, or greater than 12th grade completed), and type of residence (supervised housing, independent apartment, or apartment or house of relatives or friends). For the multiple logistic regression analysis, we created three dichotomous variables for race and ethnicity: Latina (yes or no), other (which included White and mixed race or ethnicity; yes or no), and African American (yes or no). We placed Latina and other in the logistic regression model as covariates.
The Sexual Relationship and Reproductive Support Questionnaire, derived from topics raised in our qualitative research, was used to provide contextual information on women’s sexual and reproductive environments. Participants are asked to provide yes or no responses to nine questions designed to assess institutional (e.g., housing related) restrictions on sexual activity, health care provider and family support of sexual relationships and reproductive health, and participants’ perceived control over sexual relationship and reproductive health concerns. Examples include the following: “Are you free to have sex in the place where you live?” (yes or no); “My family encourages me to pursue romantic relationships” (yes or no); and “I would like to discuss my sexual life with my mental health care provider” (yes or no).
The Structured Clinical Interview for Diagnostical and Statistical Manual of Mental Disorders (4th ed.; DSM–IV) Diagnosis (SCID), a semistructured instrument, was used to assess DSM–IV psychiatric diagnoses of the past 12 months (Williams et al., 1992). The SCID also permits assessment of DSM–IV diagnoses of substance use disorder (SUD) over the course of a participant’s lifetime (Williams et al., 1992). We obtained psychiatric and substance use diagnoses for all but 12 women using the SCID. These variables included schizophrenia-spectrum disorders (schizophrenia, schizoaffective, and psychosis not otherwise specified), depressive disorders, and bipolar disorders. Missing data on SCID diagnoses were replaced with self-reported diagnoses ascertained in the SPISEW. Interviewers trained to administer the SCID conducted these interviews.
For the logistic regression analysis, we created three dichotomous variables for psychiatric diagnosis—schizophrenia spectrum (yes or no), bipolar disorders (yes or no), and depressive disorders (yes or no)—and one dichotomous variable for SUD (yes or no). We included bipolar disorders, depressive disorders, and SUD as covariates in the model.
The Sexual Risk Behavior Assessment Schedule—A—Mentally Ill Sheltered-Female (SERBAS–A–MIS-F) was used to measure sexual risk behavior. This semistructured interview inquires about the number, gender, and type (e.g., casual, steady, and exchange sex) of sexual partners over the previous 3 months; the types of sexual acts performed at each encounter (anal, oral, or vaginal); whether sexual acts were male- or female-condom protected, or whether other nonbarrier methods of contraception were used; whether sex was bought or sold; and the risk behaviors of sexual partners. Researchers have obtained reliable sexuality data using this instrument in a sample of men with SMI (Sohler, Colson, Meyer-Bahlburg, & Susser, 2000).
The SERBAS ascertains the total number of protected and unprotected sexual encounters; sexual encounters preceded by or accompanied by drug or alcohol use; and partners, including main partners, casual and exchange-sex partners and high-risk partners. We used the following variables for our analyses. Unprotected sex was determined by participants’ self-report of engaging in one or more occasions of unprotected vaginal, anal, or oral sex in the last 3 months. Three categories were created from the data for the purpose of conducting the analysis of variance (ANOVA): no sexual activity, no unprotected sex, and any unprotected sex. We created a dichotomous variable (unprotected sex; yes or no) for the subsequent logistic regression analyses, which compared unprotected sex to all other sex. Substance use with sex (yes or no) was defined as use of drugs or alcohol right before or during sex at least once in the past 3 months. Casual/exchange partners was defined as men or women who were not steady partners and with whom the women had sex for fun or exchanged sex for money, drugs, or other goods. Three categories were created for the ANOVA: no partner, main partner, and casual/exchange partner. A dichotomous variable (casual/exchange partner; yes or no) was created for the logistic regression analysis, which compared casual/exchange partners to all others. Risky partners (yes or no) was defined as male partners who injected drugs, had sex with men, had been in jail or prison, were HIV positive, or had recent sexually transmitted infections. Multiple partners was defined as having more than one partner in the past 3 months. Three categories were created for the ANOVA: no partner, one partner, and two or more partners. A dichotomous variable (multiple partners; yes or no) was created for the regression analysis in which having two or more partners was compared to all others. Finally, by summing each of the five dichotomous sexual risk behavior variables, we created a count variable of the total number of risk behaviors participants reported in the past 3 months.
The SPISEW questionnaire was adapted from Link’s Devaluation and Discrimination scale (Link, 1987; Link et al., 1989) and informed by the findings of our previous qualitative study of stigma and sexuality in the lives of women with SMI (P. Y. Collins et al., 2008). The SPISEW encompasses five domains: (a) personal experiences of mental illness stigma, (b) relationship devaluation and discrimination, (c) ethnic devaluation and discrimination, (d) perceived attractiveness, and (e) all current discrimination. A separate score is constructed for each of the five domains of the SPISEW by summing the responses and calculating the mean. The sample mean for each domain is used to test associations with outcome variables (i.e., HIV sexual risk behaviors). In each domain, higher scores reflect greater perceived stigma or discrimination, or unattractiveness. For brevity these will be referred to as mental illness stigma, relationship stigma, ethnic stigma, perceived attractiveness, and discrimination, respectively. We assessed internal reliability with data from the current sample.
Mental illness stigma
The six items in this domain assess personal experiences of discrimination related to mental illness. Items include “Has anyone ever called you ‘crazy,’ ‘loca,’ or ‘nuts’?” and “Have you experienced people trying to take advantage of you because they know that you have a mental illness?” Participants respond either yes or no for all items. Cronbach’s alpha = .86.
Relationship stigma
The five items in this domain assess women’s agreement with the general public’s perceptions of women with mental illness in relationships. Items include “Most women with a mental illness would have difficulty finding a romantic relationship” or “Once people know a woman has a mental illness they would think she was undesirable sexually.” Participants respond on a 4-point Likert-type scale ranging from (1) strongly disagree to (4) strongly agree. Cronbach’s alpha = .79.
Ethnic stigma
The six items in this domain assess women’s agreement with the general public’s perceptions of members of the participant’s ethnic group. Items include “Most people believe that a person from [ethnicity] is just as intelligent as the average person” and “Most people will not take the ideas and opinions of someone from [ethnicity] seriously.” Participants respond on a 4-point Likert-type scale ranging from (1) strongly disagree to (4) strongly agree. Cronbach’s alpha = .73.
Perceived attractiveness
The four items in this domain assess women’s perceptions of their own attractiveness in the context of having a mental illness. Items include “I am an attractive woman” and “Having a mental illness makes me feel less attractive than other women.” Participants respond on a 4-point Likert-type scale ranging from (1) strongly disagree to (4) strongly agree. Cronbach’s alpha = .61.
Discrimination
The eight items in this domain assess the extent to which personal experiences of discrimination related to skin color, language, ethnicity, sexual orientation, method of earning money, drug use, gender, and mental illness block access to opportunities. Items include “Have you ever been kept from being in a relationship, job or an educational opportunity because of your sexual orientation?” Participants respond on a 4-point Likert-type scale ranging from (1) rarely to (4) frequently. Cronbach’s alpha = .80.
Data Analysis
We present analyses based on the domains of mental illness stigma, relationship stigma, ethnic stigma, perceived attractiveness, and discrimination as measured by the SPISEW. Our primary question is whether perceived stigma in these five domains is related to these sexual risk outcomes: unprotected sex, substance use with sex, casual/exchange partners, risky partners, and multiple sexual partners. Except where noted, sexually inactive participants were included in the “nonrisky” group.
We examined bivariate associations between sex risk behaviors and demographic characteristics using chi-square tests for categorical variables and t tests for continuous variables as appropriate. Using a one-way ANOVA, we examined differences in each of the five stigma domains by each of the five categories of HIV risk behaviors. To examine the effect of all five stigma domains on each of the five dichotomous sexual risk variables, we used multiple logistic regression models, including all five of the stigma domains in each mode. We accounted for potential confounding factors by including age, race or ethnicity, psychiatric diagnosis, and substance use disorder as covariates in each of these models. Finally, we conducted an exploratory analysis using Poisson regression to examine the association between the five stigma domains and the total number of all risk behaviors reported.
Results
Descriptive Statistics
The sample characteristics are summarized in Table 1. The majority of women in our sample were members of ethnic minority groups. Most had 12 or fewer years of education, and many lived in supervised housing. More than 60% had a diagnosed psychotic disorder, and more than half had a substance use disorder at some point in their lives. Overall, moderate levels of stigma were reported across all five domains (see Table 1).
Table 1.
Demographic, Stigma, and Relationship Support Characteristics
| Characteristic | Mean (SD) | n (% yes) |
|---|---|---|
| Age | 41.2 (9.3) | |
| Ethnicity | ||
| African American | 60 (65.2) | |
| Latina | 15 (16.3) | |
| White | 11 (12.0) | |
| Other | 6 (6.5) | |
| Education | ||
| Less than 12th grade | 42 (45.6) | |
| 12th grade | 23 (25.0) | |
| Greater than 12th grade | 24 (26.1) | |
| Housing | ||
| Supervised housing | 55 (59.8) | |
| Independent apartment | 30 (32.6) | |
| Apartment/house of relative/friend/other | 7 (7.6) | |
| Marital status | ||
| Never married | 53 (57.6) | |
| Married | 8 (8.7) | |
| Divorced/separated/widowed | 30 (32.5) | |
| Diagnosis | ||
| Schizophrenia/schizoaffective/psychosis NOS | 56 (61) | |
| Depressive disorders | 18 (19.5) | |
| Bipolar disorders | 18 (19.5) | |
| Lifetime substance use disorder | 58 (63.0) | |
| Stigma | ||
| Mental illness stigma (range = 0–6) | 2.65 (2.25) | |
| Relationship stigma (range = 0–4) | 2.53 (0.64) | |
| Ethnic stigma (range = 0–4) | 2.25 (0.49) | |
| Perceived attractiveness (range = 0–4) | 2.22 (0.48) | |
| Discrimination (range = 0–4) | 1.58 (0.64) | |
| Sexual Relationship and reproductive support | ||
| Speak to friends, family, health care providers about sex life | 49 (53.3) | |
| Encouraged by MH providers to discuss sexual relationships | 30 (32.6) | |
| Friends/family encourage to get involved in sexual relationships | 23 (25.0) | |
| Family encourages to pursue romantic relationships | 25 (27.2) | |
| Free to have sex in place of residence | 69 (75.0) | |
| Sometimes has sex with people she doesn’t like | 19 (20.7) | |
| Ever been told you shouldn’t have sex | 26 (28.3) | |
| Anyone ever suggested that you use tubal ligation, hysterectomy, or abortion for birth control? | 26 (28.3) | |
| Women with MI need the emotional support of a partner | 60 (65.2) | |
Note. NOS = not otherwise specified; MH = mental health; MI = mental illness.
A sizable minority (more than 1 out of 3 responding) reported lifetime experiences of discrimination because of their mental illness, ethnicity, and gender. Others reported discrimination because of the use of substances (32%), the way they earn their money (21%), skin color (22%), the way they talk (18%), or sexual orientation (17%).
Three out of 4 women were free to have sexual relationships in their places of residence. Slightly more than half spoke to family, friends, or health care providers about sexual issues, but only one-third were encouraged by mental health care providers to talk about sex. A little more than 1 in 4 were encouraged by family or friends to pursue romantic relationships. Yet the majority believed that women with mental illness need the support of a partner.
Twenty women (21.7%) reported engaging in three or more sexual risk behaviors in the past 3 months, and 19.5% reported engaging in no risk behaviors (range = 0–5 risk behaviors). There was a significant association between engaging in any risk behavior and having a lifetime substance use disorder, χ2(1, N = 91) = 8.48, p = .04; being Latina, χ2(1, N = 91) = 4.36, p = .04; and being younger, t(92) = 2.18, p = .03. Having a casual/exchange partner was significantly associated with being Latina, χ2(1, N = 91) = 5.55, p = .02, and significantly more of those who had unprotected sex were younger, t(92) = 3.15, p = .002.
ANOVA and Regression Analyses
We first examined the relationship between each stigma domain and each of the five sexual risk behaviors using an ANOVA (see Table 2). Of the five stigma domains, mental illness stigma, relationship stigma, and discrimination had significant associations with sexual risk behaviors.
Table 2.
Differences in Stigma by HIV Risk Behaviors in the Past 3 Months Among Women with Severe Mental Illness
| Mental illness stigma |
Relationship stigma |
Ethnic stigma |
Perceived attractiveness |
Discrimination |
||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Sexual risk behavior | M | SD | M | SD | M | SD | M | SD | M | SD |
| Unprotected sex | ||||||||||
| No sexual activity (n = 11) | 1.36 | 1.75 | 2.13 | 0.55 | 2.11 | 0.63 | 2.07 | 0.55 | 1.49 | 0.55 |
| No unprotected sex (n = 28) | 2.93 | 2.34 | 2.51 | 0.53 | 2.21 | 0.42 | 2.25 | 0.58 | 1.53 | 0.60 |
| Any unprotected sex (n = 53) | 2.77 | 2.25 | 2.62 | 0.68 | 2.30 | 0.50 | 2.24 | 0.40 | 1.63 | 0.69 |
| F | 2.13 | 2.73* | 0.84 | 0.63 | 0.34 | |||||
| Substance use with sex | ||||||||||
| No substance use (n = 79) | 2.41 | 2.20 | 2.47 | 0.64 | 2.23 | 0.51 | 2.19 | 0.48 | 1.54 | 0.60 |
| Substance use (n = 13) | 4.15 | 2.08 | 2.86 | 0.59 | 2.36 | 0.39 | 2.38 | 0.42 | 1.85 | 0.84 |
| Fa | 7.17*** | 4.23** | 0.71 | 1.82 | 2.57 | |||||
| Casual/exchange partner | ||||||||||
| No partner (n = 11) | 1.36 | 1.75b | 2.13 | 0.60c | 2.11 | 0.63 | 2.07 | 5.5 | 1.49 | 0.55d |
| Main partner (n = 50) | 2.54 | 2.30 | 2.52 | 0.64 | 2.30 | 0.49 | 2.17 | 0.41 | 1.44 | 0.52 |
| Casual/exchange partner (n = 31) | 3.29 | 2.18b | 2.68 | 0.63c | 2.23 | 0.44 | 2.36 | 0.54 | 1.85 | 0.78d |
| F | 3.25** | 3.10** | 0.71 | 2.35* | 4.22** | |||||
| Risky partner | ||||||||||
| No risky partner (n = 56) | 2.57 | 2.30 | 2.49 | 0.67 | 2.26 | 0.49 | 2.22 | 0.48 | 1.55 | 0.66 |
| Any risky partner (n = 36) | 2.78 | 2.22 | 2.58 | 0.60 | 2.25 | 0.50 | 2.22 | 0.47 | 1.63 | 0.62 |
| Fa | 0.18 | 0.45 | 0.01 | 0.01 | 0.34 | |||||
| Multiple partners | ||||||||||
| No partner (n = 11) | 1.36 | 1.75 | 2.13 | 0.60e | 2.11 | 0.63 | 2.07 | 0.55 | 1.49 | 0.55 |
| One partner (n = 68) | 2.71 | 2.26 | 2.53 | 0.63 | 2.27 | 0.47 | 2.25 | 0.45 | 1.56 | 0.66 |
| Two or more partners (n = 13) | 3.46 | 2.30 | 2.85 | 0.61e | 2.27 | 0.52 | 2.21 | 0.55 | 1.78 | 0.65 |
| F | 2.75* | 3.95** | 0.55 | 0.66 | 0.75 | |||||
Note. N = 92.
Degrees of freedom = 1, 90. For all other F statistics, degrees of freedom = 2, 89.
Means with similar superscripts are significantly different from each other by Scheffé tests at significance levels of 0.05.
p ≤ 0.1.
p ≤ .05.
p ≤ 0.01.
We then examined the role of multiple stigmatized identities on HIV risk. Table 3 presents the association between sexual risk behavior and all five perceived stigma domains, controlling for age, race/ethnicity, diagnosis, and substance use disorder in these models. Intercorrelations between the five stigma domains ranged from r = .28 to r = 0.60. Tests of multicollinearity revealed variance inflation factors ranging from 1.4 to 1.9, suggesting that all stigma domains could be present in the same model. There was no association between unprotected sex and any stigma domain. There was a significant relationship between mental illness stigma and substance use with sex. For each one point increase in score of mental illness stigma, women were 1.74 times more likely report using substances prior to or during sex. Women who ported more ethnic stigma were less likely to report having casual/exchange sex partner. In contrast, women who reported greater discrimination were significantly more likely to having a casual/exchange sex partner. Relationship stigma showed no significant associations.
Table 3.
Effect of Stigma on Risk Behaviors in Past 3 Months
| Unprotected sex (n = 53) |
Substance use with sex (n = 13) |
Casual/exchange partner (n = 31) |
Risky partner (n = 36) |
Multiple partners (n = 13) |
||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Stigma domain | AOR | 95% CI | AOR | 95% CI | AOR | 95% CI | AOR | 95% CI | AOR | 95% CI |
| Mental illness stigma | 1.00 | 0.76–1.33 | 1.74 | 1.04–2.90** | 0.86 | 0.63–1.18 | 0.97 | 0.74–1.27 | 1.09 | 0.73–1.61 |
| Relationship stigma | 2.20 | 0.84–5.75 | 1.77 | 0.45–7.04 | 1.21 | 0.42–3.49 | 1.27 | 0.53–3.07 | 3.15 | 0.89–11.20* |
| Ethnic stigma | 1.27 | 0.40–4.02 | 0.79 | 0.16–3.95 | 0.21 | 0.05–0.85** | 0.77 | 0.26–2.22 | 0.48 | 0.10–2.28 |
| Perceived attractiveness | 0.59 | 0.17–2.11 | 1.13 | 0.20–6.42 | 3.83 | 0.89–16.43* | 0.84 | 0.25–2.77 | 0.41 | 0.08–2.19 |
| Discrimination | 1.27 | 0.50–3.25 | 0.92 | 0.27–3.13 | 4.71 | 1.48–15.00*** | 1.07 | 0.43–2.66 | 1.74 | 0.48–6.29 |
| Model statistics | ||||||||||
| Pseudo R2 | 0.12 | 0.28 | 0.20 | 0.06 | 0.15 | |||||
| χ2 | 15.29 | 19.07 | 23.73 | 7.01 | 11.21 | |||||
| p | 0.12 | 0.02 | 0.01 | 0.73 | 0.34 | |||||
Note. Race/ethnicity (Latina and other vs. African American), age, diagnosis (depressive disorders and bipolar disorders vs. schizophrenia-spectrum disorders), and diagnosis of lifetime substance use disorder are included in all models as covariates. Odds ratios are calculated from five multivariate logistic regressions (one for each HIV sexual risk behavior). N = 92. AOR = adjusted odds ratio; 95% CI = 95% confidence interval.
p < .1.
p < .05.
p < .01.
Finally, we examined the association between the total number of sexual risk behaviors reported and perceived stigma in all five domains by using Poisson regression. Those who endorsed higher scores on relationship stigma were significantly more likely to engage in a greater number of sexual risk behaviors after accounting for the effects of all other domains of stigma (β = 0.30, SE = 0.15, standardized B = 1.96, p = .05).
Discussion
This study used a newly developed instrument to assess stigma and sexuality among urban women with SMI. We posed two research questions: (a) how much do urban women with SMI endorse experiences of stigmatization in any of the five domains? (b) Which, if any, domains of stigma are associated with HIV sexual risk behaviors?
With regard to the first question, as a group, the women in the sample reported low to moderate amounts of mental illness stigma, relationship stigma, ethnic stigma, attractiveness stigma, and discrimination. With regard to the second question, there were significant associations between specific stigma domains (e.g., mental illness stigma, ethnic stigma, and discrimination) and specific sexual risk behaviors (e.g., substance use before or during sex and having a casual/exchange partner type). Consistent with an intersectional approach, endorsement of certain stigma domains was associated with increased sexual risk-taking and endorsement of others with less sexual risk taking. The data imply that specific domains of stigma are differentially associated with sexual risk behaviors among women with SMI.
Our data suggest that women in our sample experience certain aspects of stigma in their interactions with certain partner types. Women with no sexual partners reported the least amount of mental illness stigma, relationship stigma, and discrimination, whereas those in sexual relationships reported more stigma in these domains. In line with Modified Labeling Theory (Link et al., 1989), these findings suggest that women place themselves at increased risk for rejection, criticism, and devaluation when they enter a sexual relationship. These personal experiences of discrimination and mental illness stigma may increase their internalization of negative stereotypes about women with SMI and lead to decreased self-esteem (Link et al., 1989).
This line of thinking is consistent with our multivariate findings that experiences of discrimination were strongly associated with a having a casual/exchange sex partner, even after considering the effect of other stigma domains. Women who experience discrimination from multiple sources may feel disempowered to select a steady partner and maintain such a relationship. With repeated discrimination, women may experience a decline in self-efficacy in relationships and perceive that the power to decide on the nature of the relationship lies in the hands of their partner. Casual and exchange-sex partners, in particular, may be less supportive and more devaluing, creating a cycle of rejection, stigma, and risk. Also consistent is our finding that mental illness stigma–everyday experiences of being called loca or crazy–was associated with substance use before or during sex. Women who have faced this kind of devaluation likely anticipate rejection in sexual encounters. Alcohol and drug use may serve to minimize the stress accompanying sexual encounters or intimate relationships more generally (Stoner, George, Peters, & Norris, 2007).
When participants believed that, in the relationship realm, most people would devalue a woman with mental illness, they tended to engage in more risk behaviors overall. In this situation, poor self-esteem may reduce women’s likelihood of engaging in self-protective behaviors (e.g., negotiating condom use) and increase the likelihood of engaging in sexual risk behaviors (e.g., selecting a riskier partner type or using substances before sex). Although we did not assess self-esteem, our measure of perceived attractiveness may serve as a proxy for women’s self-concept as it relates to sexual desirability and attractiveness.
Unlike the other domains of stigma, greater perception of one’s ethnic group being devalued protected against having a casual or exchange-sex partner. At first glance, this seems to contradict our findings on discrimination, but it draws attention to the complexities of (anticipated) stigma and coping strategies. Studies have shown that recognizing one’s ethnic group as stigmatized does not necessarily produce low self-esteem or negative outcomes (Crocker & Major, 1989). It may reflect a keen awareness of societal attitudes. The women who are critical of racial discrimination are likely to report higher levels of perceived devaluation and discrimination and could also be more politically conscious and better able to protect themselves from abusive or risky relationships. In other contexts, perceiving discrimination as lacking legitimacy has been associated with higher self-esteem and empowerment (Rusch, Lieb, Bohus, & Corrigan, 2006). Empowerment has been positively associated with HIV prevention (Romero et al., 2006).
We found no significant associations between any of the stigma domains and unprotected sex after controlling for potential confounders. One possible explanation is that our measurement of unprotected sex episodes does not capture the relational aspect of sex that our results suggest is important in stigma processes. The experiences of stigma may be a distal risk for unprotected sex, mediated by partner or relationship variables. It is also possible that our sample size had insufficient power to detect a relationship between these stigma domains and unprotected sex. Additional studies using larger samples are required to further examine these relationships.
Implications for Research
Our findings suggest that disregard for women’s experiences of multiple oppressions provides an incomplete picture of sexual risk. Different domains of stigma will have particular relevance for certain communities of women with mental illness. A woman’s ethnicity may influence how significant the stigma related to mental illness will be for her sexual relationships. In our study, women with casual and exchange-sex partners were significantly more likely to be Latina. This may reflect cultural differences in the acceptance of women with mental illness as steady partners or in women’s perceptions of themselves as acceptable partners. Understanding the salience of women’s multiple stigmatized statuses in the realm of sexual relationships and how specific stigma experiences interact to increase or decrease risk requires further study.
It is essential to understand the mechanisms through which stigma domains may affect sexual risk taking. Future research should examine the potential moderating and mediating influences of self-esteem and mastery on the relationship between stigma and HIV risk behavior.
Implications for Interventions in Clinical Settings
Our exploratory data describing women’s sexual and reproductive environments suggest that although women experience stigma and discrimination associated with sexual risk taking, it is unlikely that many will discuss their sexual relationships with their health care providers or family members. Almost half do not. While most are not actively discouraged from having sex, few women are encouraged to be in relationships. Yet the majority of women believe that a woman with a mental illness needs the support of a partner. The mismatch between women’s perceptions of their needs and their perceptions of providers’ and others’ support of relationships raises questions about the nature and impact of structural discrimination in the treatment setting.
In order for effective HIV prevention interventions to occur in clinical settings, providers need to understand how specific ethnic and cultural communities perceive mental illness and what the illness means for women’s perceptions of their own desirability and options for relationships. Immigration status, poverty, and other stigmatized identities should factor into these discussions. Our findings also suggest that clinical interventions around stigma and HIV prevention should target the sexual lives of younger women, who are more likely to engage in risky sexual behaviors. Finally, barriers that reduce the likelihood that sexuality will be routinely and respectfully addressed in the clinical setting must be identified and eliminated.
Limitations
There are several limitations to our study. Stigmatization works on multiple levels—through personal experiences of discrimination, social psychological processes, and structural discrimination (Link & Phelan, 2001). The effects of these processes on health outcomes cannot be fully understood on the basis of self-reported experiences of stigma and discrimination. The SPISEW assesses some of the social psychological components of stigma experience for women, but insights into the relationship between stigma and HIV risk may also be found by examining structural components of the stigma process. The SPISEW requires further psychometric refinement. For example, the Cronbach’s alpha for the domain “perceived attractiveness” is .61. The instrument does not include other variables (e.g., self-esteem, mastery, self-efficacy, and acute psychiatric symptoms) that may moderate the relationship between stigma and sexual behavior. The instrument also assesses lifetime experiences of stigma and discrimination as opposed to current experiences. Our data are cross-sectional; therefore no conclusions about the direction of causality can be drawn. We interviewed a convenience sample of women selected from urban mental health facilities; our findings cannot be generalized to all women with SMI. Finally, because sample size restricted our power, we were unable to examine how specific stigma experiences interact to increase or decrease risk for these women.
Conclusions
To our knowledge, our study is the first to examine, quantitatively, the relationship between stigma, mental illness, and HIV risk among urban, multiethnic women using an intersectional approach. Despite their limitations, our data show that among urban women with SMI, personal experiences of mental illness stigma and experiences of discrimination were related to substance use with sex and having a casual/exchange partner, respectively. Relationship stigma was related to more sexual risk behaviors overall. Conversely, perceived ethnic stigma was associated with less likelihood of selecting a casual/exchange partner. The data underscore the importance of examining multiple sources of stigma among women with multiple stigmatized labels and supporting responses to stigma that lead to reduced risk.
Acknowledgments
This study was supported by grants from the National Institute of Mental Health (NIMH K01MH16091 to Pamela Y. Collins) and from the Robert Wood Johnson Foundation Harold Amos Medical Faculty Development Program. Katherine S. Elkington was supported by Center Grant P30 MH43520 to the HIV Center for Clinical and Behavioral Studies from the National Institute of Mental Health (Anke A. Ehrhardt, principal investigator) and by NRSA Grant T32 MH19139, Behavioral Science Research in HIV Infection (Anke A. Ehrhardt, program director).
Footnotes
We are indebted to Ragini Dutt, Marcela Hoffer, and Bruce G. Link for their input in the development of the SPISEW. We thank the members of the Robert Wood Johnson Health and Society Stigma Working Group for their comments on the manuscript.
Contributor Information
Pamela Y. Collins, Department of Epidemiology, Mailman School of Public Health, and Department of Psychiatry, College of Physicians and Surgeons of Columbia University, and the New York State Psychiatric Institute.
Katherine S. Elkington, HIV Center for Clinical and Behavioral Studies, Department of Psychiatry, and College of Physicians and Surgeons of Columbia University.
Hella von Unger, Social Science Research Center, WZB, Berlin, Germany.
Annika Sweetland, Department of Socio-medical Sciences, Mailman School of Public Health of Columbia University.
Eric R. Wright, Center for Health Policy, School of Public and Environmental Affairs, Indiania University—Purdue University at Indianapolis (IUPUI).
Patricia A. Zybert, Department of Health and Behavior Studies, Teachers College, Columbia University.
References
- Amaro H. Love, sex, and power: Considering women’s realities in HIV prevention. American Psychologist. 1995;50:437–447. doi: 10.1037//0003-066x.50.6.437. [DOI] [PubMed] [Google Scholar]
- Collins PH. Learning from the outsider within: The sociological significance of Black feminist thought. In: Hesse-Biber S, Gilmartin C, Lydenberg R, editors. Feminist approaches to theory and methodology: An interdisciplinary reader. New York: Oxford University Press; 1999. pp. 155–178. [Google Scholar]
- Collins PY. Dual taboos: Sexuality and women with severe mental illness in South Africa: Perceptions of mental health care providers. AIDS & Behavior. 2001;5:151–161. [Google Scholar]
- Collins PY, Unger HV, Armbrister A. Church ladies, good girls, and locas: Stigma and the intersection of gender, ethnicity, mental illness, and sexuality in relation to HIV risk. Social Science & Medicine. 2008;67:389–397. doi: 10.1016/j.socscimed.2008.03.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Corrigan P, Thompson V, Lambert D, Sangster Y, Noel JG, Campbell J. Perceptions of discrimination among persons with serious mental illness. Psychiatric Services. 2003;54:1105–1110. doi: 10.1176/appi.ps.54.8.1105. [DOI] [PubMed] [Google Scholar]
- Crocker J, Major B. Social stigma and self-esteem: The self-protective properties of stigma. Psychological Review. 1989;96:608–630. [Google Scholar]
- Lai Y, Hong C, Chee C. Stigma of mental illness. Singapore Medical Journal. 2001;42:111–114. [PubMed] [Google Scholar]
- Link BG. Understanding labeling effects in the area of mental disorders: An assessment of the effects of expectations of rejection. American Sociological Review. 1987;52:96–112. [Google Scholar]
- Link BG, Cullen FT, Struening EL, Shrout PE. A modified labeling theory approach to mental disorders: An empirical assessment. American Sociological Review. 1989;54:400–423. [Google Scholar]
- Link BG, Phelan JC. Social conditions as fundamental causes of disease. Journal of Health and Social Behavior. 1995;(Suppl):80–94. [PubMed] [Google Scholar]
- Link BG, Phelan JC. Conceptualizing stigma. Annual Review of Sociology. 2001;27:363–385. [Google Scholar]
- Link BG, Struening EL, Neese-Todd S, Asmussen S, Phelan JC. Stigma as a barrier to recovery: The consequences of stigma for the self-esteem of people with mental illness. Psychiatric Services. 2001;52:1621–1626. doi: 10.1176/appi.ps.52.12.1621. [DOI] [PubMed] [Google Scholar]
- Link BG, Struening EL, Rahav M, Phelan JC, Nuttbrock L. On stigma and its consequences: Evidence from a longitudinal study of men with dual diagnoses of mental illness and substance abuse. Journal of Health & Social Behavior. 1997;38:177–190. [PubMed] [Google Scholar]
- Littleton H, Breitkopf CR, Berenson A. Body image and risky sexual behaviors: An investigation in a tri-ethnic sample. Body Image. 2005;2:193–198. doi: 10.1016/j.bodyim.2005.02.003. [DOI] [PubMed] [Google Scholar]
- McKinnon K, Cournos F, Herman R. HIV among people with chronic mental illness. Psychiatric Quarterly. 2002;73:17–31. doi: 10.1023/a:1012888500896. [DOI] [PubMed] [Google Scholar]
- Meade CS, Sikkema KJ. HIV risk behavior among adults with severe mental illness: A systematic review. Clinical Psychology Review. 2005;25:433–457. doi: 10.1016/j.cpr.2005.02.001. [DOI] [PubMed] [Google Scholar]
- Otto-Salaj LL, Heckman TG, Stevenson LY, Kelly JA. Patterns, predictors and gender differences in HIV risk among severely mentally ill men and women. Community Mental Health Journal. 1998;34:175–190. doi: 10.1023/a:1018745119578. [DOI] [PubMed] [Google Scholar]
- Parker R, Aggleton P. HIV and AIDS-related stigma and discrimination: A conceptual framework and implications for action. Social Science & Medicine. 2003;57:13–24. doi: 10.1016/s0277-9536(02)00304-0. [DOI] [PubMed] [Google Scholar]
- Ritsher JB, Phelan JC. Internalized stigma predicts erosion of morale among psychiatric outpatients. Psychiatry Research. 2004;129:257–265. doi: 10.1016/j.psychres.2004.08.003. [DOI] [PubMed] [Google Scholar]
- Romero L, Wallerstein N, Lucero J, Fredine HG, Keefe J, O’Connell J. Woman to woman: Coming together for positive change: Using empowerment and popular education to prevent HIV women. AIDS Education & Prevention. 2006;18:390–405. doi: 10.1521/aeap.2006.18.5.390. [DOI] [PubMed] [Google Scholar]
- Rosenfield S. Labeling mental illness: The effects of received services and perceived stigma on life satisfaction. American Sociological Review. 1997;62:660–672. [Google Scholar]
- Rusch N, Lieb K, Bohus M, Corrigan P. Self-stigma, empowerment, and perceived legitimacy of discrimination among women with mental illness. Psychiatric Services. 2006;57:399–402. doi: 10.1176/appi.ps.57.3.399. [DOI] [PubMed] [Google Scholar]
- Silverstone P. Low self-esteem in different psychiatric conditions. British Journal of Clinical Psychology. 1991;30:185–188. doi: 10.1111/j.2044-8260.1991.tb00936.x. [DOI] [PubMed] [Google Scholar]
- Sohler N, Colson PW, Meyer-Bahlburg HF, Susser E. Reliability of self-reports about sexual risk behavior for HIV among homeless men with severe mental illness. Psychiatric Services. 2000;51:814–816. doi: 10.1176/appi.ps.51.6.814. [DOI] [PubMed] [Google Scholar]
- Somlai A, Kelly J, Heckman T, Hackl K, Runge L, Wright C. Life optimism, substance use, and AIDS-specific attitudes associated with HIV risk behavior among disadvantaged inner city women. Journal of Women’s Health & Gender-Based Medicine. 2000;9:1101–1111. doi: 10.1089/152460900446018. [DOI] [PubMed] [Google Scholar]
- Sterk CE, Klein H, Elifson KW. Self-esteem and “at risk” women: Determinants and relevance to sexual and HIV-related risk behaviors. Women & Health. 2004;40:75–92. doi: 10.1300/j013v40n04_05. [DOI] [PubMed] [Google Scholar]
- Stern AM. Sterilized in the name of public health: Race, immigration, and reproductive control in modern California. American Journal of Public Health. 2005;95:1128–1138. doi: 10.2105/AJPH.2004.041608. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Stewart AJ, McDermott C. Gender in psychology. Annual Review of Psychology. 2004;55:519–544. doi: 10.1146/annurev.psych.55.090902.141537. [DOI] [PubMed] [Google Scholar]
- Stoner SA, George WH, Peters LM, Norris J. Liquid courage: Alcohol fosters risky sexual decision-making in individuals with sexual fears. AIDS & Behavior. 2007;11:227–237. doi: 10.1007/s10461-006-9137-z. [DOI] [PubMed] [Google Scholar]
- Thompson VL, Noel JG, Campbell J. Stigmatization, discrimination, and mental health: The impact of multiple identity status. American Journal of Orthopsychiatry. 2004;74:529–544. doi: 10.1037/0002-9432.74.4.529. [DOI] [PubMed] [Google Scholar]
- Williams JB, Gibbon M, First MB, Spitzer RL, Davies M, Borus J, et al. The structured clinical interview for DSM–III–R (SCID) II: Multi-site test–retest reliability. Archives of General Psychiatry. 1992;49:630–636. doi: 10.1001/archpsyc.1992.01820080038006. [DOI] [PubMed] [Google Scholar]
- Wingood GM, Diclemente RJ, Harrington K, Davies SL. Body image and African American females’ sexual health. Journal of Women’s Health & Gender-Based Medicine. 2002;11:433–439. doi: 10.1089/15246090260137608. [DOI] [PubMed] [Google Scholar]
- Wright ER, Gronfein WP, Owens TJ. Deinstitutionalization, social rejection, and the self-esteem of former mental patients. Journal of Health & Social Behavior. 2000;41:68–90. [PubMed] [Google Scholar]
- Wright ER, Wright DE, Perry BL, Foote-Ardah CE. Stigma and the sexual isolation of people with serious mental illness. Social Problems. 2007;54:78–98. [Google Scholar]