The sexual partnerships of people with serious mental illness.

The transmission of Human Immunodeficiency Virus (HIV) and Acquired Immune Deficiency Syndrome (AIDS) among people with serious mental illness has garnered attention in the public health community over the past 20 years (National Institute of Mental Health, 2002). Seroprevalence rates among people with serious mental illness are typically higher than in the general population and have most recently been estimated at between 1.0% and 5.4% (Essock et al., 2003; Rosenberg et al., 2001).

Because of this elevated risk, clinicians and public health officials recognize an eminent need for research on patterns of sexual behavior in this population (Buckley, Robben, Friedman, & Hyde, 1999; Lukoff, Sullivan, & Goisman, 1992). Existing studies are beset by sampling and methodological constraints and are informed by an individual-focused approach that limits our understanding of the sexuality of people with serious mental illness (Buckley et al., 1999). Our analysis employed a systematic, venue-based, random sample of mental health clients drawn from various treatment settings. Placing emphasis on the role of social context in shaping individuals' sexual behavior within relationships (Laumann, Gagnon, Michael, & Michaels, 1994), we compared the primary partnerships of people with serious mental illness and the general population.

In the past 10 to 20 years, there have been over a dozen studies of client sexual behavior (see Buckley et al., 1999, and McKinnon & Rosner, 2000, for reviews). The prevalence of sexual activity (anal or vaginal intercourse) among people with serious mental illness has been estimated at between 44% and 80% (Buckley et al.; Cournos, Guido, & Coomaraswamy, 1994; Rozensky & Berman, 1984). Importantly, multiple studies have identified an increased tendency to engage in unprotected, high-risk sexual behaviors (Baker & Mossman, 1991; Buckley et al.; Carey et al., 2001, 2004; Chuang & Atkinson, 1996; Cournos, McKinnon, & Meyer-Bahlburg, 1994; DiClemente & Ponton, 1993; Kalichman, Kelly, Johnson, & Buero, 1994; Kelly et al., 1992; Levan & Johnson, 2002; McKinnon & Rosner, 2000; Susser, Valencia, Miller, & Tsai, 1995; Tubman, Gil, Wagner, & Artigues, 2003). In particular, rates of homosexual activity are high, as is the likelihood of having multiple concurrent sexual partners (Carey et al., 2001; Coverdale, Turbott, & Roberts, 1987; Cournos et al., 1993, 1994; DiClemente & Ponton; Kalichman et al., 1994; Kelley et al., 1992; Lavan & Johnson; McDermott, Sautter, Winstead, & Quirk, 1994; Rozensky & Berman; Susser et al.).

Although the risk-avoidance literature has contributed a great deal to our understanding of the levels and types of sexual risk behaviors in which people with serious mental illness engage, cultural and clinical attitudes characteristic of this research are a limitation. Rarely is sex conceptualized as "an act of communication, expression of emotion, or an attempt at interpersonal bonding" (Lilleleht & Leiblum, 1993, p. 252). Rather, many clinicians believe that their clients are asexual, that their sexuality is inappropriate, or that they are incapable of successful romantic relationships (Buckley et al., 1999; Hogan, 1980; Lilleleht & Leiblum; Nesteros, Lehman, & Ban, 1981; Sadow & Corman, 1983; Thomas, 1989; Vandereycken, 1993). In short, sexual behavior is seen as another symptom to be treated or as a possible risk factor for slowing or reversing recovery (Pinderhughes, Barrabee, & Reyna, 1972).

These attitudes likely contribute to the tendency to ignore the relationship context of sexual behavior in research involving people with serious mental illnesses. Instead of examining how social characteristics such as relationship dynamics, social network pressures, and stigmatizing public attitudes influence sexual risk behavior, clinicians frequently assume that their patients act independently of external social pressures and constraints. Risk behavior is described as a consequence of individual factors, such as severity of psychiatric illness, diagnosis, lack of social skills, or an inadequate understanding of how to protect oneself. As Laumann and colleagues (1994) argued, an exclusive focus on individual-level behaviors may limit our understanding of mechanisms underlying sexual risk behavior, as well as the consequences of such behavior.

In addition to these theoretical limitations, several methodological problems contribute to the need for additional research in this area. Most notably, nearly all of the studies reviewed above are characterized by small (fewer than 100 participants) convenience samples (Buckley et al., 1999). Further, very few studies include inpatient mental health clients in their sample, so the effects of institutionalization on client sexuality are largely unknown (Buckley et al.). Most importantly, few studies have included a comparison group from the general population. Instead, researchers assess their findings about people with serious mental illness by contrasting their patterns of sexual behavior to a subjective sense of what is "normal" (Howell et al., 1987; see Dickerson et al., 2004, for a recent exception).

To fill the gap left by these conceptual and methodological limitations, we addressed the following research questions: first, does the level of sexual activity differ between the general population and people with serious mental illness? Second, what variables predict sexual activity in these populations? Finally, among those who are sexually active, are there systematic differences in the sexual partnerships of people with serious mental illness and the general population?

DATA AND METHOD

The data used in these analyses are from two sources: the 1992 National Health and Social Life Survey (NHSLS) and the Indiana Mental Health Services and HIV-Risk Study (IMHSHRS). The NHSLS, conducted in 1992 by a research team at the University of Chicago (Principal Investigators Edward O. Laumann, John H. Gagnon, Robert T. Michael, and Stuart Michaels), is the most comprehensive representative survey of sexual behavior in the U.S. general population. The NHSLS sample is comprised of 3,432 non-institutionalized American men and women between the ages of 18 and 60. Respondents were chosen for participation using a multi-stage area probability sample. The response rate was 79%. The data were collected via in-depth, face-to-face interviews and self-administered questionnaires.

The second data set used in these analyses is from the Indiana Mental Health Services and HIV Risk Study (Wright, 1999), which explores the impact of HIV on the mental health treatment system. A total of 401 mental health clients from two state hospitals and three community mental health centers took part in the IMHSHRS. After a case-wise deletion of missing data (32 cases dropped), the target sample was reduced to 369. All respondents were diagnosed with a serious mental disorder (major depression, bipolar disorder, schizophrenia, schizoaffective disorder, or psychosis NOS) and were between the ages of 18 and 60 years old. Respondents receiving treatment at two of the community mental health centers were recruited via a random sampling of clients, while all clients who met the sampling criteria at the two hospital sites and one CMHC were invited to participate. The response rate was 74%. Data were obtained via in-depth, face-to-face interviews, and data collection for each participating site occurred sequentially during intensive field periods ending in 2001. A significant proportion of the interview schedule, including all questions pertaining to relationship characteristics and sexual activity included in these analyses, was taken directly from the NHSLS.

Although the NHSLS sample likely contains some individuals with serious mental illness, the omission of institutionalized individuals from the sampling frame reduced overlap between the NHSLS and IMHSHRS. To the degree that this overlap exists, it would moderate differences between the two samples rather than exaggerate them. Also, because people aged 18 to 60 are more likely to be sexually active than younger teenagers and older adults, limiting the sample to this age range may overestimate differences between these two populations. Finally, because respondents in the NHSLS were asked about their primary partnership in the past year, and respondents with serious mental illness were asked about partnerships in the past three months, people in the IMHSHRS sample had fewer relationships to choose from. This could exaggerate observed differences in partnerships.

Measures

We created and coded standard demographic variables according to convention (see Table 1), including a dummy variable Female and categorical variables for Race, Education, and Marital Status. We also created a continuous variable for Annual Household Income.

We also constructed variables pertaining to lifestyle and global sexual characteristics. These include the dummy variables Sexually Active (1 if sexually active in the three months prior to the interview, 0 otherwise) and Very or Extremely Happy with Personal Life (1 if Extremely Happy or Very Happy Most of the Time when asked, "How happy have you been with your personal life lately?" and 0 otherwise). When creating this and other dummy variables from categorical items, we attempted to use cut-off points that resulted in the most even distribution of respondents in each cell. The variable Attendance at Religious Services is the number of times per month respondents attended a church, mosque, or synagogue. Finally, we created the variable Number of Symptoms of Sexual Dysfunction by summing items asking whether respondents had experienced any of the following in the past three months: "Lacked interest in having sex," "Were unable to orgasm," "Had trouble lubricating or getting wet (women)/getting an erection (men)," "Came to a climax too soon," "Had physical pain during intercourse," "Did not find sex pleasurable," and "Were anxious about performing sexually."

We also created variables describing respondents' sexual partnerships. We classified primary partnerships according to duration for the variable Relationship Type. Three categories, Short-term, Longer-term, and Long-term, were created according to coding used in The Social Organization of Sexuality (Laumann et al., 1994). These categories refer to relationships that endured less than one month, one month or more without cohabitation or marriage, and relationships involving either cohabitation or marriage. We also created three dummy variables measuring homogeneity (Partner is Same Gender, Partner is Same Race, and Partner Has Same Degree as the respondent), and dummy variables for Knew Partner One Month or More Before Having Sex, Partner Got to Know Friends, and Partner Got to Know Family. Finally, the categorical variable Who Introduced Partner is equal to 1 for Respondent or Partner Introduced Himself/Herself, 2 for Family Introduced Partner, and 3 for Friend Introduced Partner.

We created several variables measuring the quality of primary partnerships. Respondents in both studies were asked to rate the physical and emotional quality of their partnerships (Relationship is Extremely Physically Pleasurable and Relationship is Extremely Emotionally Satisfying). Similarly, both the NHSLS and the IMHSHRS asked, "How often do you orgasm during sex?" and "How often does your partner orgasm during sex?" The variables Respondent Always Orgasms and Partner Always Orgasms were coded 1 if the respondent answered Always, and 0 otherwise. About half of respondents in the NHSLS indicated that they or their partners always orgasm, hence the use of this high cut-off point.

We created a series of variables measuring the frequency of sexual activity. These include the dummy variable Have Sex 3-6 Times/Week or More and six additional variables (Always Have Vaginal Sex, Always Use Condom During Vaginal Sex, Always or Usually Give Oral Sex, Always or Usually Receive Oral Sex, Ever Have Anal Sex, Always Use Condom During Anal Sex). These were created using different response categories and criteria in order to parallel their appearance in Lauman, et al.'s (1994) original study. Because the sex frequency variables all have different distributions, dummy variables were coded so at least 20% of the sample appeared in each category. Further, while NHSLS respondents were asked how often (i.e., Always, Usually, Sometimes, Rarely, or Never) they engaged in these sexual activities, IMHSHRS respondents were prompted for the number of times they actually engaged in each behavior during the past three months. Consequently, we collapsed the responses from both samples into a series of dummy variables. Thus, for the sample of people with serious mental illness, Always indicates that they engaged in a certain type of sex or used condoms 100% of the time (e.g., had vaginal sex 10 times in 10 sexual episodes), Always or Usually means that they did so between 51% and 100% of the time (e.g., had oral sex 6 times in 10 sexual episodes), and Ever means that they did so at least once (e.g., had anal sex once in 10 sexual episodes). For example, the dummy variable Always Have Vaginal Sex is equal to 1 if NHSLS respondents answered Always to the frequency question, and for IMHSHRS respondents, if the total number of times respondents engaged in sexual activity equals the total number of times they had vaginal sex. This variable is equal to 0 if NHSLS respondents had vaginal sex Usually, Sometimes, Rarely, or Never and if IMHSHRS respondents had vaginal sex fewer times than they engaged in sexual activity.

Analyses

We used frequency distributions on variables in these analyses to compare the NHSLS sample to the IMHSHRS sample. We used the Kolmogorov-Smirnov Z test, a two-sample nonparametric test, to determine whether cumulative frequency distributions from the NHSLS and IMHSHRS were significantly different. We chose this test because it is sensitive to differences in central tendency and in shape of distribution, which need not be identical for the two populations being tested (Miller & Wilson, 1983).

We used the Area Under the ROC Curve (AUC) to determine effect size. The AUC represents the extent to which the two populations can be discriminated by a given dependent variable. An AUC of 50% represents a case with no discrimination at all, while an AUC of 100% represents perfect discrimination. Although odds ratios, risk ratios, and relative risk reduction are popular among researchers, these effect size indices have two major limitations that the AUC does not (Kraemer et al., 2003). First, the magnitude of these measures may approach infinity even if the association is near random. Second, these measures vary widely with the choice of cut-off point for binary variables. A further advantage of the AUC is its comparability to Cohen's d. According to Kraemer et al., an AUC of [greater than or equal to] 76% is similar to a d of [greater than or equal to] 1.00 (very large effect), an AUC of 71% to a d of 0.80 (large effect), an AUC of 64% to a d of 0.50 (medium effect), and an AUC of 56% to a d of 0.20 (small effect).

We also employed binary logistic regression to identify predictors of sexual activity and to examine unstandardized regression coefficients (b) and their corresponding p-values in order to determine the direction and magnitude of these effects.

RESULTS

About 56% of NHSLS respondents and 43% of IMHSHRS respondents were female, and 24% and 35%, respectively, were non-White (see Table 1). NHSLS respondents ranged in age from 18 to 59 years (mean = 37), while IMHSHRS respondents ranged from 18 to 59 years (mean = 38). About 56% of the NHSLS sample had some college or vocational training, while only 26% of the IMHSHRS sample did. Finally, most NHSLS respondents were currently married or cohabitating (61%), while most IMHSHRS respondents were currently single and not cohabitating (89%). In sum, the sample of people with serious mental illness was significantly older (Z = 2.7, p [less than or equal to] .001) and less educated (Z = 5.3, p [less than or equal to] .001), and contained significantly more men (Z = 2.7, p [less than or equal to] .001), racial minorities (Z = 2.1, p [less than or equal to] .001), and single people (Z = 8.1, p [less than or equal to].001).

We also found a significant difference (Z = 7.9, p [less than or equal to] .001) in the number of sex partners the sample of people with serious mental illness and the general population sample had over a three-month period (see Table 2). A majority of people with serious mental illness had no partners in three months (67%), while only 24% of people in the general population had no sex partners (AUC = 71.8%). Further, the majority of NHSLS respondents had one partner (71%) in three months, but only 21% of IMHSHRS respondents had one (AUC = 75.1%).

The results of the binary regression of the variable Sexually Active on demographic and other independent variables are presented in Table 3. For NHSLS respondents, aging significantly decreased the likelihood of being sexually active (b = -0.04, p [less than or equal to].001). Also, marital status affected sexual activity such that people in the general population who were neither married nor cohabitating and individuals who were separated, widowed, or divorced and not cohabitating were significantly less likely to be sexually active than married people (b = -2.72, p [less than or equal to] .001; b = -2.07, p [less than or equal to] .001). Also, the higher a respondent's household income, the more likely he was to be sexually active (b = 0.01, p [less than or equal to] .001), while attendance at religious services significantly decreased the probability of being sexually active (b = -0.14, p [less than or equal to] .001). Finally, individuals in the general population who were very or extremely happy with their personal life were significantly more likely to be sexually active than those who were not (b = 0.22, p [less than or equal to] .05).

People in the general population were somewhat similar to people with serious mental illness with respect to factors that influenced sexual activity (see Table 3). Increasing age also decreased the probability of being sexually active in the population with serious mental illness (b = -0.05, p [less than or equal to] .001). Additionally, never being married and not cohabitating decreased the likelihood of being sexually active in both populations (b = -2.14, p [less than or equal to] .001). However, unlike the general population, individuals with serious mental illness who were non-White were more likely to be sexually active than Whites (b = 0.86, p [less than or equal to] .001). Also, individuals with serious mental illness who experienced more symptoms of sexual dysfunction were more likely to be sexually active (b = 0.20, p [less than or equal to].01).

Table 4 compares the quality and characteristics of the primary partnerships of the NHSLS and IMHSHRS respondents. People with serious mental illness were more likely to have short-term relationships (13%) that lasted less than one month than people in the general population (3%; Z = 5.7, p [less than or equal to] .001; AUC = 54.8%). Also, more individuals with serious mental illness (65%) than people in the general population (21%) had longer-term relationships (those lasting one month or more) that did not culminate in cohabitation or marriage, and the primary partnerships of the general population (75%) were significantly more likely to progress to marriage or cohabitation than those of people with serious mental illness (23%). The effects of these differences were large (AUC = 72.1% and 76.3%, respectively). Additionally, people with serious mental illness were significantly more likely to have sexual relationships concurrent to their primary partnerships (Z = 2.4, p [less than or equal to] .001). About 28% of the IMHSHRS sample had one or more concurrent partners, while only 6% of the NHSLS sample did (AUC = 61.0%). Further, people in the general population were more likely to wait one month or more before having sex than individuals with serious mental illness (Z = 2.7, p [less than or equal to] .001). While 79% of NHSLS respondents waited, only 54% of IMHSHRS respondents did (AUC = 62.5%). Finally, more people in the general population were the same race as their primary partners than people with serious mental illness (86% and 72%, respectively; Z = 1.4, p [less than or equal to] .05). However, this effect was small (AUC = 56.8%).

According to these analyses (see Table 4), the general population and people with serious mental illness also differed with respect to how they met primary partners (Z = 3.5, p [less than or equal to] .001). About 43% of NHSLS respondents met their primary partners because either they or the partner introduced themselves. Significantly more IMHSHRS respondents met partners in this manner (78%). A family member introduced the couple 13% of the time for NHSLS respondents and 9% of the time for IMHSHRS respondents. Lastly, 42% of the NHSLS sample was introduced to their primary partner by a friend, while only 13% of the IMHSHRS sample was. We found a medium effect of serious mental illness on the likelihood of introduction by the respondent or partner (AUC = 67.4%) and introduction by friends (AUC = 64.7%), but a small effect on introduction by family members (AUC = 51.7%).

These data also indicated that people in the general population had primary partnerships that were more satisfying, on average, than those of individuals with serious mental illness (see Table 4). However, most of these effects were small. About 43% of the general population had partnerships they described as extremely physically pleasurable, while only 33% of people with serious mental illness did (Z = 1.3, p [less than or equal to] .05; AUC = 55.2%). Furthermore, 50% of the NHSLS sample always had an orgasm during sexual encounters with their primary partner, and 63% said their partners always had an orgasm. Only 41% of IMHSHRS respondents always had an orgasm during sex, and 56% of their partners did (Z = 1.3, p [less than or equal to] .05; AUC = 54.7% and Z = 1.4, p [less than or equal to] .05; AUC = 53.4%, respectively). Table 4 also suggests people in the general population had sex more frequently (Z = 2.6, p [less than or equal to] .001). Only 7% of the IMHSHRS sample had sex with their primary partner three to six times per week or more, on average. A significantly higher proportion of NHSLS respondents did (31%), constituting a medium effect (AUG = 62.0%).

We also compared the types of sexual activities in which NHSLS and IMHSHRS respondents typically engaged (see Table 4). In general, there were few significant differences. One notable exception is that people with serious mental illness were more likely always to use condoms during vaginal sex than people in the general population (54% versus 12%; Z = 2.7, p [less than or equal to] .001), and this effect is large (AUC = 71.6%).

Information about whether partners got to know family and friends is also presented in Table 4. Primary partners of people in the general population were more likely to have gotten to know friends than were primary partners of people with serious mental illness (87% versus 73%; Z = 1.5, p [less than or equal to] .05). However, this is a small effect (AUC = 57.4%).

DISCUSSION

These results suggest important differences between the partnerships of people with serious mental illness and the general population. While most people with serious mental illness had no partner, the majority of people in the general population had one. In both populations, respondents who were younger, less educated, and married were more likely to be sexually active. However, sexual dysfunction was associated with a higher likelihood of being sexually active among people with serious mental illness.

Because patterns of high- and low-risk sexual behavior are due in part to differences in the nature of relationships themselves and in the ways in which they are socially structured (Laumann et al., 1994), we also examined sexual partnerships. We found that people with serious mental illness were more likely to have short-term relationships and longer-term relationships that did not culminate in a more serious commitment like cohabitation or marriage. People with serious mental illness were also more likely to have one or more concurrent sexual partnerships and to be in biracial relationships. Additionally, people with serious mental illness engaged in sexual activity with new partners sooner, and they were more likely to use condoms consistently (people typically use condoms in the initial period of uncertainty about sexual and STD history before they have achieved an elevated level of trust and commitment [Laumann et al., 1994]). Furthermore, consistent with existing research (Deegan, 1999; Friedman & Harrison, 1984; Howell et al., 1987; Lyketsos, Sakka, & Mailis, 1983), people with serious mental illness reported being less physically satisfied with their partnerships and orgasmed less frequently than individuals in the general population. Finally, people with serious mental illness had partners that were less integrated into their social networks.

These patterns suggest relationships that lack intimacy and commitment. Unfortunately, these data cannot answer whether these partnerships have the potential to reach a high level of commitment. Certain features of mental illness, particularly schizophrenia spectrum disorders, as well as some side effects of psychiatric medication, may impair social functioning, thereby limiting the development of intimacy (Badger, McNiece, Bonham, & Jacobson, 2003; Cassano & Favo, 2004; Grant, Addington, Addington, & Konnert, 2001; Harris et al., 2005).

In addition, external social and institutional factors associated with serious mental illness may complicate efforts to develop and maintain committed long-term partnerships. For instance, people with serious mental illness often have limited access to the appropriate support, information, and professional help needed to deal with sexual and relationship problems effectively. Moreover, the organization of mental health treatment systems and the generally negative sexual culture of treatment facilities, particularly institutional settings, can constrain behavior and opportunity for sexual interaction and can shape sexual activity when it does occur (Wright, 2001; Wright & Gayman, 2005; Wright & Martin, 2003). Finally, the pervasive poverty, unemployment, low educational attainment, homelessness, and stigmatizing public attitudes associated with serious mental illness both limit access to sexual and romantic partners and reduce the value of this population in the sexual marketplace (Dowdall & Goldstein, 1979; Estroff, 1981; Laumann et al., 1994; Rosenfield, 1991).

These social and institutional constraints play a vital role in shaping the sexual opportunities and partnerships of people with serious mental illness. Consequently, as our data suggest, people with serious mental illness are often forced to try to meet their sexual needs or forge a relationship with other "social undesirables." In short, they seem to take what they can get in terms of where, when, and with whom they have sex. This not only results in relationships that are less satisfying, less intimate, and much shorter-lived, but also increases HIV risk by concentrating sexual activity within high-risk populations like IV drug users, sex workers, and others with serious mental illness. When these individuals do manage to find a viable partner, the social and interpersonal challenges associated with mental illness, coupled with the lack of support for sexual expression from treatment providers and social networks, may undermine the relationship's success.

Our results point to a need for a paradigm shift in the way clinicians, clinical researchers, and public health officials think about and study the sexual lives of people with serious mental illness. Our research demonstrates that many people with serious mental illness are sexually active, and the recognition that mental health clients have normal sexual desire and expression is an important quality of life issue. Further, worry and anxiety about one's sex life and relationships could exacerbate symptoms of mental illness or trigger relapse, and sexual fears may impose real barriers to community reintegration (Lukoff et al., 1992). Finally, healthy sexual partnerships can promote the development and maintenance of new relationships and contribute to social integration, which can increase quality of life and decrease the risk of relapse (Lukoff et al., 1992; Miller & Ingham, 1976; Vaughn & Left, 1976).

Interventions to improve the likelihood of relationship success and satisfaction for people with serious mental illness are an important clinical component of this paradigm shift toward attending to social and relationship context (Lukoff et al., 1986, 1992). Psychosocial rehabilitation maximizes self-sufficiency by giving people with serious mental illness the skills they need to interact in the community and to have successful platonic and romantic relationships (Anthony, 1993; Barton, 1999; Brekke, Ansel, Long, Slade, & Weinstein, 1999). These services are particularly important now that effective psychotropic medications and the continuation of the deinstitutionalization movement have allowed more patients to live in community care settings.

A shift away from an individualized, pathology-focused approach to client sexuality could also reduce the HIV risk of people with serious mental illness. Our findings confirm that some features of the relationships of people with serious mental illness put them at higher risk of contracting HIV and other sexually transmitted diseases. Although about half of people with serious mental illness use condoms consistently, they also tend to have sex sooner with new partners than the general population and are more likely to have concurrent sexual relationships. Addressing sexuality issues more frequently in the context of sexual partnerships and helping clients to cope with the social and institutional factors affecting their sexual behavior would introduce opportunities to provide STD education to mental health clients, and would also support them in engaging in risk-reduction strategies (Lukhoff et al., 1992).

This research was supported by grants to Eric R. Wright from the National Institute of Mental Health (R01 MH59717) and from the Indiana University Research and University Graduate School.

Manuscript accepted November 8, 2005

REFERENCES

Anthony, W. (1993). Recovery from mental illness: The guiding vision of the mental health service system in the 1990s. Psychosocial Rehabilitation Journal, 16, 11-23.

Badger, T. A., McNiece, C., Bonham, E., Jacobson, J., & Gelenberg, A. J. (2003). Health outcomes for people with serious mental illness: A case study. Perspectives in Psychiatric Care, 39, 23-32.

Baker, D. G., & Mossman, D. (1991). Potential HIV exposure in psychiatrically ill hospitalized adolescent girls. American Journal of Psychiatry, 148, 528-530.

Barton, R. (1999). Psychosocial rehabilitation services in community support systems: A review of outcomes and policy recommendations. Psychiatric Services, 50, 525-534.

Brekke, J. S., Ansel, M., Long, J., Slade, E., & Weinstein, M. (1999). Intensity and continuity of services and functional outcomes in the rehabilitation of persons with schizophrenia. Psychiatric Services, 50, 248-256.

Buckley, P. E, Robben, T., Friedman, L., & Hyde, J. (1999). Sexual behavior in persons with serious mental illness: Patterns and clinical correlates. In P. F. Buckley (Ed.), Sexuality and serious mental illness (pp. 1-21). Amsterdam: Harwood Academic Press.

Carey, M. P., Carey, K. B., Maisto, S. A., Gordon, C. M., & Vanable, P. A. (2001). Prevalence and correlates of sexual activity and HIV-related risk behavior among psychiatric outpatients. Journal of Consulting and Clinical Psychology, 69, 846-850.

Carey, M. E, Carey, K. B., Maisto, S. A., Schroder, K. E., Vanable, P. A., & Gordon, C. M. (2004). HIV risk behavior among psychiatric outpatients: Association with psychiatric disorder, substance use disorder, and gender. Journal of Nervous Mental Disease, 192, 289-296.

Cassano, E, & Favo, M. (2004). Tolerability issues during long-term treatment with antidepressants. Annals of Clinical Psychiatry, 16, 15-25.

Chuang, H. T., & Atkinson, M. (1996). AIDS knowledge and high-risk behavior in the chronically mentally ill. Canadian Journal of Psychiatry, 41,269-272.

Cournos, F., McKinnon, K., & Meyer-Bahlburg, H. (1993). HIV risk activity among persons with severe mental illness: Preliminary findings. Hospital and Community Psychiatry, 44, 1,104-1,106.

Cournos, E, Guido, J. R., & Coomaraswamy, S. (1994). Sexual activity and risk of HIV infection among patients with schizophrenia. American Journal of Psychiatry, 151,228-232.

Coverdale, J. H., Turbott, S. H., & Roberts, H. (1997). Family planning needs and STD risk behaviors of female psychiatric outpatients. British Journal of Psychiatry, 171, 69-72.

Deegan, P. E. (1999). Human sexuality and mental illness: Consumer viewpoints and recovery principles. In P.F. Buckley (Ed.), Sexuality and serious mental illness (pp. 21-33). Amsterdam: Harwood Academic Press.

Dickerson, F. B., Brown, C. H., Kreyenbuhl, J., Goldberg, R. W., Fang, L. J., & Dixon, L. B. (2004). Sexual and reproductive behaviors among people with mental illness. Psychiatric Services, 55, 1,299-1,301.

DiClemente, R. J., & Ponton, L. E. (1993). HIV-related risk behaviors among psychiatrically hospitalized adolescents and school-based adolescents. American Journal of Psychiatry, 150, 324-325.

Dowdall, G. W., & Goldstein, M. S. (1979). Occupational problems, unemployment, and mental illness: How a community mental health center deals with job troubles. Presented at the Annual Meetings of the Association for Humanist Sociology, Johnstown, PA.

Essock, S. M., Dowden, S., Constantine, N. T., Katz, L., Swartz, M. S., Meador, K. G., et al. (2003). Risk factors for HIV, hepatitis B, and hepatitis C among persons with severe mental illness. Psychiatric Services, 54, 836-841.

Estroff, S. E. (1981). Making it crazy: An ethnography of psychiatric clients in and American community. Berkley, CA: University of California Press.

Friedman, S., & Harrison, G. (1984). Sexual histories, attitudes, and behavior of schizophrenic and normal women. Archives of Sexual Behavior, 13, 555-567.

Grant, C., Addington, J., Addington, D., & Konnert, C. (2001). Social functioning in first and multiepisode schizophrenia. Canadian Journal of Psychiatry, 46, 746-749.

Harris, A., Brennan, J., Anderson, J., Taylor, A., Sanbrook, M., Fitzgerald, D., et al. (2005). Clinical profiles, scope, and general findings of the Western Sydney first episode psychosis project. Australian and New Zealand Journal of Psychiatry, 39, 36-43.

Hogan, R. (1980). Human sexuality: A nursing perspective. New York: Appleton-Century-Crofts.

Howell, J. R., Reynolds, C. F., Thase, M. E., Frank, E., Jennings, J. R., Houck, P. R., et al. (1987). Assessment of sexual function, interest, and activity in depressed men. Journal of Affective Disorders, 13, 61-66.

Kalichman, S. C., Kelly, J. A., Johnson, J. R., & Buero, M. (1994). Factors associated with risk for HIV infection among chronic mentally ill adults. American Journal of Psychiatry, 151, 221-227.

Kelly, J. A., Murphy, D. A., Bahr, G. R., Brasfield, T. L., Davis, D. R., & Hauth, H. C. (1992). AIDS/HIV risk behavior among the chronic mentally ill. American Journal of Psychiatry, 149, 886-889.

Kraemer, H. C., Morgan, G. A., Leech, N. A., Gliner, J. A., Vaske, J. J., & Harmon, R. J. (2003). Measures of clinical significance. Journal of the American Academy of Child and Adolescent Psychiatry, 42, 1,524-1,529.

Laumann, E. O., Gagnon, J. H., Michael, R. T., & Michaels, S. (1994). The social organization of sexuality: Sexual practices in the United States. Chicago: The University of Chicago Press.

Lavan, H., & Johnson, J. G. (2002). The association between axis I and II psychiatric symptoms and high-risk sexual behavior during adolescence. Journal of Personality Disorders, 16, 73-94.

Lilleleht, E., & Leiblum, S. (1993). Schizophrenia and sexuality: A critical review of the literature. Annual Review of Sex Research, 4, 247-276.

Lukoff, D., Gioia-Hasick, D., Sullivan, G., Golden, J. S., & Nuechterlein, K. H. (1986). Sex education and rehabilitation with schizophrenic male outpatients. Schizophrenia Bulletin, 12, 669-577.

Lukoff, D., Sullivan, G., & Goisman, R. M. (1992). Sex and AIDS education. In R. P. Liberman (Ed.), Handbook of psychiatric rehabilitation (pp. 171-182). New York: MacMillan Publishing Co.

Lyketsos, G. C., Sakka, D., & Mailis, A. (1983). The sexual adjustment of chronic schizophrenics: A preliminary study. British Journal of Psychiatry, 143, 376-382.

McDermott, B. E., Sautter, F. J., Winstead, D. K., & Quirk, T. (1994). Diagnosis, health beliefs, and risk of HIV infection in psychiatric patients. Hospital and Community Psychiatry, 45,580-585.

McKinnon, K., & Rosner, J. (2000). Severe mental illness and HIV-AIDS. In F. Cournos & M. Forstein (Eds.), New directions for mental health services: What mental health practitioners need to know about HIV and AIDS (pp. 69-76). San Francisco, CA: Jossey-Bass.

Miller, A. D., & Wilson, V. J. (1983). Vestibular-induced vomiting after vestibulocerebellar lesions. Brain, Behavior, and Evolution, 23, 26-31.

Miller, P., & Ingham, J. G. (1976). Friends, confidants, and symptoms. Social Psychiatry, 11, 51-59.

National Institute of Mental Health. (2002). HIV/AIDS research at the National Institute of Mental Health. Article published at www.mental-health-matters.com.

Nesteros, J. N., Lehman, H. E., & Ban, T. A. (1981). Sexual behavior of the male schizophrenic: The impact of illness and medications. Archives of Sexual Behavior, 10, 421-442.

Pinderhughes, C. A., Barrabee, E., & Reyna, L. J. (1972). Psychiatric disorders and sexual functioning. American Journal of Psychiatry, 128, 1,276-1,283.

Rosenberg, S. D., Goodman, L. A., Osher, F. C., Swartz, M. S., Essock, S. M., Butterfield, M. I., et al. (2001). Prevalence of HIV, hepatitis B, and hepatitis C in people with severe mental illness. American Journal of Public Health, 91, 31-36.

Rosenfield, S. (1991). Homelessness and rehospitalization: The importance of housing for the chronically mentally ill. Journal of Health and Social Behavior, 25, 14-23.

Rozensky, R. H., & Berman, C. (1984). Sexual knowledge, attitudes, and experiences of chronic psychiatric patients. Psychosocial Rehabilitation Journal, 8, 21-27.

Sadow, D., & Corman, A. (1983). Teaching a human sexuality course to psychiatric patients. Sex and Disability, 6, 47-53.

Susser, E., Valencia, E., Miller, M., & Tsai, W. Y. (1995). Sexual behavior of homeless mentally iii men at risk for HIV. American Journal of Psychiatry, 152, 583-587.

Thomas, B. (1989). Asexual patients. Nursing Times, 85, 49-51.

Tubman, J. G., Gil, A. G., Wagner, E. F., & Artigues, H. (2003). Patterns of sexual risk behaviors and psychiatric disorders in a community sample of young adults. Journal of Behavioral Medicine, 26, 473-500.

Vandereycken, W. (1993). Shrinking sexuality: The half-known sex life of psychiatric patients. Therapeutic Communities, 14, 143-149.

Vaughn, C. E., & Left, J. P. (1976). The influence of family and social factors on the course of psychiatric illness: A comparison of schizophrenic and depressed neurotic patients. British Journal of Psychiatry, 129, 125-134.

Wright, E. R. (1999). HIV risk of people with SMI across service sites. National Institute of Mental Health Grant Proposal, R01 MH59717. Indianapolis, IN: Department of Sociology, Indiana University-Purdue University Indianapolis.

Wright, E. R. (2001). The sexual culture of psychiatric treatment facilities and mental illness clients' HIV risk: A theoretical framework and pilot study. In S. Hartwell & R. K. Schutt (Eds.), The organizational response to social problems (pp. 275-309). Amsterdam: JAI, Elsevier Science.

Wright, E. R., & Gayman, M. (2005). The sexual networks and HIV risk of people with severe mental illness in institutional and community-based care. AIDS and Behavior, 9, 341-353.

Wright, E. R., & Martin, T. N. (2003). The social organization of HIV/AIDS care in treatment programs for adults with serious mental illness. AIDS Care, 15, 763-773.

Brea L. Perry

Indiana University

Eric R. Wright

Indiana University-Purdue University Indianapolis

Address correspondence to Brea L. Perry, Indiana University, Department of Sociology, Ballantine Hall 744, 1020 E. Kirkwood Ave, Bloomington, IN, 47405-7103; e-mail: [e-mail].

Table 1. Comparison of Sample Demographic
Characteristics for the National Health and Social
Life Survey (NHSLS, N = 3,432) and the
Indiana Mental Health Services and HIV-Risk
Study (IMHSHRS, N = 369)

                                NHSLS          IMHSHRS
                                                             Z-value
                                 N       %      N      %       (1)

Female                          1921      56    159   43.1   2.7 ***
Race                                                         2.1 ***
  Asian/Pacific Islander          67       2      5    1.4
  African American               552    16.1    104   28.4
  Hispanic                       162     4.7      2    0.5
  White                         2604    75.9    239   65.3
  Other                           45     1.3     16    4.4
Age                                                          2.7 ***
  18 to 25 years                 644    19.1     46   12.5
  26 to 35 years                1096    32.5     88     24
  36 to 45 years                 918    27.2    141   38.4
  46 to 55 years                 512    15.2     92   25.1
  56 or more years               207     6.1      0    0.0
Education                                                    5.3 ***
  Less Than High School          506    14.8    113   31.7
  High School Diploma or
    Equivalent                  1007    29.5    150   42.1
  Some College/Vocational        1115    32.7     72   20.2
  College Degree                 541    15.9     13    3.7
  Master's Degree or Higher      239       7      8    2.2
Marital Status                                               8.1 ***
  Currently Married             1809    53.4     35    9.5
  Never Married,
    Not Cohabitating             813      24    232   63.2
  Never Married, Cohabitating    150     4.4      1    0.3
  Separated/Widowed/Divorced,
    Not Cohabitating             514    15.2     96   26.2
  Separated/Widowed/Divorced,
    Cohabitating                 103       3      3    0.8

(1) Kolmogorov-Smirnov Z test of independent samples.

*** p [less than or equal to] .001

Table 2. Number of Sex Partners in a Three-Month Period
for the General Population (NHSLS, N = 3,432)
and Individuals with Severe Mental Illness
(IMHSHRS, N = 369)

                                NHSLS          IMHSHRS

                                 N       %      N      %

# of Sex Partners in
Past 3 Months
  None                           812    23.7    248   67.2
  1                             2450    71.4     89   21.2
  2                              122     3.6     22    6.0
  3 or More                       48     1.4     10    2.7

                                   AUC (1)     Z-value (2)

# of Sex Partners in
Past 3 Months                                    7.9 ***
  None                              71.8%
  1                                 75.1%
  2                                 51.2%
  3 or More                         50.7%

(1) AUC = area under the ROC curve.

(2) Kolmogorov-Smirnov Z test of independent samples.

*** p [less than or equal to] .001

Table 3. Binary Logistic Regression of Sexually
Active on Independent Variables from the NHSLS
(N = 3,432) and the IMHSHRS (N = 369)

                                     NHSLS       IMHSHRS

                                      b            b
                                    (S.E.)       (S.E.)

Female                              -0.18         0.34
                                    (0.12)       (0.27)
Nonwhite                             0.15         0.86 ***
                                    (0.14)       (0.27)
Age (Years)                         -0.04 ***    -0.05 ***
                                    (0.01)       (0.02)
Years of Schooling                  -0.05        -0.07
                                    (0.03)       (0.05)
Marital Status1
  Never Married, Not Cohabitating   -2.72 ***    -2.14 ***
                                    (0.17)       (0.45)
  Never Married, Cohabitating       (0.29)       19.72
                                    (0.42)       (4.02x[10.sup.4])
  Separated/Widowed/Divorced, Not
    Cohabitating                    -2.07 ***    -0.80
                                    (0.16)       (0.44)
  Separated/Widowed/Divorced,
    Cohabitating                     1.16        21.09
                                    (0.73)       (2.29x[10.sup.4])
Annual Household Income
  (Thousands of Dollars)             0.01 ***     0.03
                                    (0.01)       (0.02)
Attendance at Religious Services
  (Days/Month)                      -0.14 ***    -0.06
                                    (0.03)       (0.04)
Number of Symptoms of Sexual
  Dysfunction (2)                    0.01         0.20 **
                                    (0.04)       (0.07)
Very or Extremely Happy with
  Personal Life                      0.22 *       0.20
                                    (0.12)       (0.28)
Constant                             4.68 ***     2.46 **
                                    (0.51)       (0.99)
[R.sup.2]                            0.33         0.25

(1) Omitted variable is "Currently Married."

(2) Symptoms include diminished interest in sex, trouble reaching
orgasm, premature orgasm, physical pain during sex, performance
anxiety, erectile dysfunction, and not finding sex pleasurable.

* p [less than or equal to] .05

** p [less than or equal to] .01

*** p [less than or equal to] .001

Table 4. Comparison of Primary Partnerships among the General
Population (NHSLS, N = 2,988) and Individuals with Severe
Mental Illness (IMHSHRS, N = 120)

                                         NHSLS     IMHSHRS

                                       N      %    N      %

Relationship Type
  Short-term (<1 Month)                89    3.0   15   12.6
  Longer-term (>1 Month, Not
    Cohabitating/Married)             612   20.5   77   64.7
  Long-term (Cohabitating/Married)   2248   75.2   27   22.7
Has Concurrent Relationship           168    5.6   33   27.5
Partner is Same Gender                 30    1.0    8    6.6
Partner is Same Race                 2554   85.5   87   71.9
Partner Has Same Degree              1214   40.6   36   29.8
Who Introduced Respondent/Partner
  Respondent or Partner              1111   43.1   84   77.8
  Family                              377   12.6   10    9.3
  Friend                             1092   42.3   14   13.0
Knew Partner 1 Month or
  More Before Having Sex             2334   78.8   64   53.8
Relationship is Extremely
  Physically Pleasurable             1269   43.0   36   32.7
Relationship is Extremely
  Emotionally Satisfying             1181   39.9   35   32.1
Respondent Always Orgasms            1471   50.1   44   40.7
Partner Always Orgasms               1842   63.0   58   56.3
Have Sex 3-6 Times/Week or More       752   30.7    8    6.7
Always Have Vaginal Sex              2290   79.3   78   72.2
Always Use Condoms (Vaginal Sex)      335   11.8   22   55.0
Always/Usually Give Oral Sex          465   15.9   28   23.9
Always/Usually Receive Oral Sex       473   16.2   21   17.8
Ever Have Anal Sex                    271    9.3   11    9.3
Always Use Condoms (Anal Sex)          48   17.8    2   18.2
Partner Got to Know Friends           939   87.3   79   72.5
Partner Got to Know Family            860   79.9   74   67.9

                                     AUC (1)   Z-value (1)

Relationship Type                              5.7 ***
  Short-term (<1 Month)              54.8%
  Longer-term (>1 Month, Not
    Cohabitating/Married)            72.1%
  Long-term (Cohabitating/Married)   76.3%
Has Concurrent Relationship          61.0%     2.4 ***
Partner is Same Gender               52.8%     0.6
Partner is Same Race                 56.8%     1.4 *
Partner Has Same Degree              55.4%     1.1
Who Introduced Respondent/Partner              3.5 ***
  Respondent or Partner              67.4%
  Family                             51.7%
  Friend                             64.7%
Knew Partner 1 Month or
  More Before Having Sex             62.5%     2.7 ***
Relationship is Extremely
  Physically Pleasurable             55.2%     l.3 *
Relationship is Extremely
  Emotionally Satisfying             53.9%     1.1
Respondent Always Orgasms            54.7%     1.3 *
Partner Always Orgasms               53.4%     1.4 *
Have Sex 3-6 Times/Week or More      62.0%     2.6 ***
Always Have Vaginal Sex              53.6%     0.7
Always Use Condoms (Vaginal Sex)     71.6%     2.7 ***
Always/Usually Give Oral Sex         54.0%     0.9
Always/Usually Receive Oral Sex      50.8%     0.2
Ever Have Anal Sex                   50.0%     0.1
Always Use Condoms (Anal Sex)        50.2%     0.1
Partner Got to Know Friends          57.4%     1.5 *
Partner Got to Know Family           56.0%     1.2 *

(1) AUC = area under the ROC curve.

(2) Kolmogorov-Smirnov Z test of independent samples.

* p [less than or equal to] .05

*** p [less than or equal to] .001