Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Soc Sci Med. Author manuscript; available in PMC 2008 November 1.
Published in final edited form as:
PMCID: PMC2098875

Disappearing acts: The social networks of formerly homeless individuals with co-occurring disorders


Studies of the social lives of men and women living with co-occurring disorders (substance abuse and serious mental illness) suggest that social networks critically influence recovery. In this paper, we examine some of the reasons that the social networks of individuals with co-occurring disorders are small, and the impact of small networks for this population. Using a social capital framework with cross-case analysis, we analyze 72 in-depth qualitative interviews with 39 formerly homeless mentally ill men and women who were substance abusers. All were participants in the New York Services Study (HYSS), a federally funded study of mentally ill adults in New York City. The patterns suggest that networks shrunk because 1) social network members died prematurely, 2) study participants withdrew or pushed others away, and 3) friends and family members faced so many obstacles of their own that they could not provide resources for the study participants. We suggest that as networks diminished, some participants responded by attempting to rebuild their networks, even if the networks provided negative social capital, and others isolated themselves socially to escape the pressures and disappointments of interaction.

Keywords: USA, New York Services Study (NYSS), social capital, social networks, co-occurring disorders, mental illness, homeless, substance abuse

The role of family and friends for individuals with co-occurring disorders (substance abuse and serious mental illness) in recovery is complex and often ambiguous. While close friends and family members can forge ties that produce social capital in the form of significant financial, emotional, and in-kind support to individuals with co-occurring disorders, they can also pose specific problems. Studies of the social lives of co-morbid individuals suggest that social support from family and friends can decrease symptomatology (Albert, Becker, McCrone, & Thornicroft, 1998; Erickson, Beiser, & Iacono, 1998), help individuals reduce substance use (Mueser, Bellack, & Blanchard, 1992; Alverson, Alverson, & Drake, 2000; Laudet, Magura, Vogel, & Knight, 2000), and keep them active in the rhythms of daily life (Bellack & Muesser, 1986). At the same time, however, close relatives and friends can produce a negative form of social capital by creating conflict, making demands, influencing relapses (O'Farrell, Hooley, Fals-Stewart, & Cutter, 1998), being abusive, and functioning as reminders of past emotional trauma (Savage & Russell, 2005).

Mental illness and substance abuse each create certain social obstacles. The stress and challenges of negotiating relationships may be a part of the reason that individuals living with co-occurring disorders have small social networks, as some may break ties with or alienate those close to them with behaviors brought on by mental illness such as psychotic symptoms, fear, difficulty with stress, and lack of motivation (Albert et al., 1998; Trumbetta, Mueser, Quimby, Bebout, & Teague, 1999; Drake, Brunette, & Mueser, 1998; Blankertz & Cnaan, 1994). While it is certainly not always the case, some individuals may engage in desperate or alienating behaviors as a result of substance addiction (Perese & Wolf, 2005; Albert et al., 1998; Drake et al., 1998). At times, the combination of substance abuse and mental illness can mean that family members endure such a high level of stress that they react by rejecting the individual (Copello, 2003).

Finally, many homeless co-morbid individuals come from impoverished backgrounds (Caton et al., 1994) and are likely to have social networks populated with people who may also be low-income, substance abusers, and mentally ill. This may compound the difficulties of engaging in social relationships. As a result, the social networks of individuals living with co-occurring disorders tend to be small (Albert et al., 1998; Trumbetta, et al., 1999; Drake, et al., 1998; Blankertz & Cnaan, 1994). Those who have been homeless are especially likely to have very limited social capital. In this paper, we analyze in-depth interviews to better understand the role of social relationships in the lives of 39 formerly homeless persons with co-occurring disorders. We use a social capital framework (Bourdieu, 1977; 1985; Coleman, 1990; Putnam, 2000; Szreter & Woolcock, 2004) to highlight that it is not just the presence of supportive others that encourages well-being and recovery, but that it is also a matter of the type of social capital characterizing the bonds (in short, the potential of others to connect co-morbid individuals to economic, social, medical, occupational, or other resources). This study describes what can happen to the social networks of individuals with co-occurring disorders, and begins to explore the possibility of social interventions.

Social Capital as a Conceptual Model for Understanding Social Networks

We employ a social capital framework to better understand the social ties of our sample. It is important to distinguish among several different concepts. While there is some inconsistency in the literature, social networks, social support, and social capital are quite distinct (Putnam, 2000; Lin, 2000; Briggs, 1998; Portes, 1998). Cleak and Howe (2003), for example, define a social network as “all those people with whom we have ongoing relationships and through which individual people are linked into groups and society” (p.21). Tracy and Abell (1994) describe social networks as “the number and structure of relationships with others,” while they define social support as “the amount and types of supportive exchanges that take place among network relationships” (p.56). Others have framed social support similarly, with Findler (2000) defining it as the extent to which an individual receives assistance and help (e.g. emotional encouragement, advice, information, guidance, concrete aid, etc.) from others. Finally, some researchers combine the two concepts to describe a “social support network” (Peek & O'Neill, 2001; Jankowski, Videka-Sherman, & Laquidara-Dickinson, 1996), making little or no distinction between the terms.

Many researchers draw on Coleman's (1988) conceptualization of social capital as a by-product of social networks and social support systems that exist within a community, both inside and outside of the family. According to Coleman, within the family, social capital refers to the cooperative relationships among family members as well as the time and attention each member pays the others. Outside of the family, social capital encompasses the social relationships that exist among individuals, families, and institutions that create opportunities for consistent positive expectations and experiences.

Debate continues, however, about how social capital should be defined. Some view the concept as existing only as a by product of membership in a larger community, not as an individual resource (Lochner, Kawachi, & Kennedy, 1999; Kawachi, 1999; Carpiano, 2005). Others view the concept in broader terms, as an element that operates both at the individual and family level as well as within communities (Portes, 1998; 2000; Coleman, 1990; Poortinga, 2005). Each perspective integrates the individuals, families, and surrounding community institutions that comprise social support networks, role models, and information channels (Portes, 1998; 2000; Putnam, 2000). While recognizing that definitions of social capital vary, in this paper we are primarily concerned with social capital as an individual resource.

Important to this study is the ongoing debate about whether social capital is solely a positive element or if it has both positive and negative factors. Studies that attempt to measure the amount of social capital in someone's life, for example, assume that the presence of social capital alone is positive (Kawachi, Kennedy, Lochner, & Prothrow-Stith, 1997; Kawachi, 1999; Lochner et al., 1999). Putnam (2000), Coleman (1990), Portes (2000), and others contend that social capital also can have a negative component, preventing individuals from accessing needed resources, depleting limited time and energy, and introducing destructive behaviors (Corcoran & Adams, 1997; Roschelle, 1997; Antonucci, Akiyama, & Lansford, 1998). Savage and Russell (2005), for example, studied homeless co-morbid women who suffered traumas, and found that a “small but definite” portion of the women's support networks also abused them, caused trauma, and facilitated their drug use.

Other researchers have moved beyond the positive vs. negative debate to further refine the concept. One useful model was conceptualized by Gitell (1998) and expanded by Szreter and Woolcock (2004). These scholars identified three kinds of social capital: bonding, bridging, and linking. Bonding social capital refers to relationships among members of a group or network who see themselves as relatively equal, for example, neighbors or schoolmates. Bridging social capital refers to relationships among people and groups of people who are fundamentally different such as age, socio-economic status, race/ethnicity, or education (Szreter & Woolcock, 2004). Linking social capital represents the extent to which individuals build relationships with the institutions and people who have relative power over them (e.g., to provide access to services or jobs) (Szreter & Woolcock, 2004; Woolcock, 2001).

How Social Networks and Social Support Systems Matter

If social capital is a by-product of social networks and social support, the size and quality of networks matter. Individuals with co-occurring disorders tend to have smaller and weaker social support networks than individuals with a single disorder (Trumbetta et al., 1999); and the limited data on the homeless suggest their networks are smaller still. In a study comparing two samples of co-morbid homeless women, respondents in New York City had an average of only 1.96 individuals in their networks, while those from California's Central Valley had an average of 5.65. Additionally, 11% of the New York women reported no network members at all and 20% reported having no close ties (Savage & Russell, 2005). In another study of homeless individuals with co-occurring disorders, 51% reported having no friends at all, 18% had only one friend, and the remaining 31% had more than one friend (Blankertz & Cnaan, 1994).

Social support networks are usually made up of family members, friends, peers, and coworkers (Mlynarczyk, 2004). Research suggests families are especially important as a social support system to individuals with co-occurring disorders. When they are accessible, family members commonly become a foundation for survival, providing material necessities like a place to live, food, transportation to appointments and activities, access to leisure activities, and/or financial support (Clark, 1996; 2001; Clark & Drake, 1994). However, for homeless individuals with co-occurring disorders, this type of support either was never available or the ties have been broken. Where the bonds remain intact, families may not be able to spare any helpful resources. In a large, controlled study of people with schizophrenia using public mental health services, adequate family support was the single most important factor that distinguished individuals who were housed from those who were homeless (Caton et al., 1994; Caton et al., 1995). The authors suggest that, at least among men, family history may explain how family support helped meet housing needs. The men with schizophrenia who were homeless were significantly more likely to come from families that provided inconsistent nurturing, had unstable housing, had insufficient income, and relied on public assistance. These men were also more likely to have parents with a history of involvement with the criminal justice system, mental illness, and substance abuse (Caton et al., 1994). Many of these variables suggest that the families were poor and similarly socially situated to the respondents. If they wanted to help with material support, they may not have been able to do so.

In addition to providing for basic needs like housing, social support can alleviate psychological symptoms for people with co-occurring disorders. For example, support from friends and other non-relatives improved recovery from schizophrenia at both 18 months and 60 months after diagnosis (Erickson et al., 1998), and people with psychosis had a higher quality of life when their social networks included confidantes with whom they shared positive and frequent interactions (Albert et al., 1998).

However, the presence of large numbers of family members, drug users, and individuals unaware of previous traumas detract from the ability of the network to alleviate symptoms (Savage & Russell, 2005). Even supportive family and friends may simultaneously be abusive, cause trauma or stress, or function as reminders of earlier abuse (Savage & Russell, 2005). In addition, their presence may unwittingly create pressure to conform or “fit in” (Whitley & McKenzie, 2005), their emotional reactions and criticism of the co-morbid individual may lead to relapse (O'Farrell et al., 1998), or they might become burdens in other, as-yet-unexplored, ways (Whitley & McKenzie, 2005).

In general, individuals with co-occurring disorders may experience high levels of conflict with their families. Their families may in turn try to minimize conflict by withdrawing or choosing to tolerate or ignore substance abuse (Copello, 2003). Under any of these conditions, co-morbid individuals have an increased risk of relapse (Mueser et al., 1992; Alverson et al., 2000). In fact, interpersonal conflict and social pressure are generally identified as two of the biggest antecedents of substance abuse relapse (Copello, 2003). In addition, network members who are drug users themselves may continue to encourage drug use even after someone has undergone treatment (Savage & Russell, 2005).

Co-morbid individuals with family members who are not users often distance themselves from their families when they begin substance use (Test, Wallish, Allness, & Ripp, 1989; Drake et al., 1998). In their stead, mutual substance use can form a basis for new friendships, especially for those who might otherwise have trouble socializing due to mental illness symptoms (Drake et al., 1998). But while substance use may make it easier to build relationships, abstinence and reduction involve the challenging task of disconnecting from this network and ideally replacing it with an entirely different group of non-using people, usually while undergoing treatment (Drake et al., 1998). This is a challenging task for anyone, but particularly for the individual who feels stigmatized, experiences symptoms of mental illness, or not actively engaged in social settings such as work or school.



The sample consists of formerly homeless mentally ill individuals who participated in the New York Services Study (NYSS), a federally-funded study of homeless mentally ill adults in New York City. We used purposive sampling to select participants from a group of 225 respondents who participated in a longitudinal study of homeless mentally ill individuals from 1998-2002. That longitudinal study, the New York Housing Study (NYHS), assessed differences in “housing first” and “treatment first” approaches to service for homeless mentally ill individuals, and the 225 participants were randomly assigned to one of the two conditions (Tsemberis, Gulcur, & Nakae, 2004). For our sample selection, several participants were removed from the list of 225 because of unknown whereabouts (n=19), death (n=17), they had moved out of the greater New York area (n=16), or they had not signed a release for contact for participation in further studies after the NYHS (n=4). The remaining 169 were rated on several dimensions by two members of our study team who previously had been interviewers with the NYHS. Each participant was rated as having “positive,” “neutral,” or “negative” outcomes at the end of the NYHS on such qualities as housing stability, psychiatric rehabilitation, controlling substance use, employment, and involvement with the criminal justice system. The inter-rater agreement rate was 78%. Through further discussion and case review with a larger group, the list was reduced to four groups (“housing first” positive, “housing first” negative, “treatment first” positive, and “treatment first” negative) and the individuals in each group were prioritized based on expected likelihood of making contact with the participant, potential richness of interview, and gender (in order to oversample females). Thirty-nine were recruited and available for full participation. They included 21 who had been originally in the experimental housing and 18 in the control group, roughly evenly split between positive and negative outcomes. All 39 consented to participate.


The study included two life history interviews occurring about a month apart, each lasting approximately two hours. Each interview was conducted at a setting chosen by the participant, including their own residence, a quiet public space, and the study's offices. All participants gave informed consent and received $30 for each interview. Participants understood that they were part of a study to better understand how formerly homeless individuals navigated service use, including mental health, substance abuse, and any other services they had used. Of the 39 participants, 6 did not complete a second interview because they were unavailable, impaired, or the research team felt the first interview had achieved saturation. In total, 72 interviews were conducted by four trained interviewers, all of whom had previous experience conducting qualitative research interviews with New York City social service consumers. During interview #1, participants were asked to tell their life stories, paying particular attention to major events and their significance. Interviewers probed for turning points; experiences with mental illness, substance abuse, and homelessness; service use; social relationships; and life philosophies. The second interview was individually tailored for each respondent. It was designed to fill in gaps in the life history, verify events, clarify uncertainties, and to follow-up in great detail the experiences with each service used for mental health, substance abuse, and housing. Interviews were tape recorded and transcribed verbatim. Researchers participated in written and oral debriefings following every interview. The team compiled a case summary for each participant and created a visual life trajectory charting the timing and sequencing of key life events.


Multiple coders read several transcripts and developed a systematic coding scheme for the data. Separately, they each analyzed the transcripts for preliminary themes using open coding, and developed a list of categories and concepts (Strauss & Corbin, 1998). They arrived at consensus on a list of codes through group discussions. Coders then developed primary and secondary codes through axial coding, including a focus on social networks, with sub-codes for family, friends, service providers, positive and negative support. Several themes and patterns were then identified through selective coding. ATLAS-ti software was used to support the coding, annotating, analyzing, and integrating of interview and observational data. Social capital was not considered in the original study design, but the theme was identified by the first author of this study after reading several original transcripts.


The sample was 67% male, 33% female, ranging in age from 23 to 62 (mean age 48). The racial/ethnic distribution was 41% African American, 41% Caucasian, 15% Hispanic, and 3% Arab. Slightly more than half (56%) were living in supportive housing, while 23% were living in long-term transitional housing, 8% in residential substance abuse treatment facilities, 5% in apartments with Section 8 vouchers (United States government subsidies for low income housing), 5% in adult care facilities, and 2% in homeless shelters. Eighty-five percent reported long-term substance abuse. The primary psychiatric diagnoses were schizophrenia (56%), bipolar disorder (22%), and major depression (22%).

We identified several themes related to the nature of the social networks, how these networks function, and what effect they had on study participants. Study participants identified an average of two ties each, although the closeness of these ties was not always clear. Nine participants had no ties at all. The research team identified three patterns in the data that address why many of the participants' networks are limited: 1) former members of the network died; 2) study participants pushed away network members or removed themselves from their networks; and 3) network members had problems of their own and were unable to engage with the study participants.

Deaths within the Social Network

Almost all participants lost social network members through death. Often, participants' parents died while the participants were young. Siblings, children, and close friends died too, leaving a social support void that was not refilled. Participants described several types of deaths: accidents, drug use, violence, and poor health.

Some deaths left an immediate void, especially when there were multiple deaths in a brief span of time and when they occurred while the participants were young. One participant, age 59, lost a brother, an aunt, and both parents by age 19.

  • Every now and again, I would take a trip back to Pittsburgh to see my family. And, uh… my brother… He had passed away…My aunt. I had one aunt…She committed suicide…what really had an impact on me, was when my father passed away, because I thought he was rather young and he had high blood pressure and a bad heart.

Some participants witnessed the deaths at young ages. At age five, Mario (now 57) saw his older brother decapitated in a motorcycle accident. During childhood, Patrice saw her mother die and was then raised by her older sister. By the time Patrice was 23, her older sister and two of her sister's children died as well.

Maria, whose mother had thrown her out of her house at age 12, returned home at 21 to reconcile with her mother, only to learn that her mother had passed away. A few years earlier, Maria lost her fiancé in a shootout at a strip club where she worked. Her fiancé's death left her alone to raise a baby. She explains,

  • So, I opened the door to his jeep. He dropped in my arms out of the car seat and fell in the street. He was shot up. He was shot. I'm screaming, and somebody called the ambulance. By the time the ambulance came he was already dead. The last words he said to me was, “Take care of our baby.”

Even when the deaths were not violent, many were sudden. Karla, for example, experienced the early death of her husband.

  • In ‘72 I was married. And my husband passed away in 1976. That was traumatic for me as well…I came home from work and I found him … dead. I didn't know that he was dead. I thought he was still alive because his vital organs were still functioning. But he had had a massive brain hemorrhage.

Perhaps the combination of homelessness, mental illness, and addiction involves more individuals living in environments of high stress, risk of disease, poor health care, and poverty, all of which can lead to early death. Such environmental factors appear to be even more likely culprits when we consider how members of the network passed away: through violence, substance overdose, suicide, and health problems.

The fact that deaths figured so prominently in life narratives has several implications. First, when people died, the support void was usually not filled by other people, either personal or professional (such as service providers). The absences remained pivotal even many years later. Second, the continuing traumatic impact of long-ago deaths may suggest a particular vulnerability to loss. This is perhaps a symptom of mental illness for some, and it may make it more difficult to want to become close with others again.

Pushing and Pulling Away

Study participants described behavior motivated by addiction, mental illness, or both, that they believed damaged their relationships. At times, they were demanding, violent, or wore out their welcome, and at other times they withdrew in order to avoid difficult situations or negative influence from substance users.

Richard, 44, for example, alienated his family by physically assaulting both his parents and stealing from his aunt to pay for drugs. While his aunt may have been forgiving, Richard never felt the same closeness.

  • Yeah. I stole a ring from one of my aunts…I was in the bathroom and I saw on the side of the sink a ring. I didn't know it was gold…So I picked it up and I put it in my pocket…I sold it for crack, but got a half ounce of grass…And my aunt said…Remember when you went to the bathroom.” I said, “Yes.” “That ring that you took. That was one of my favorite rings. You should not have taken that ring.” I was like, “Uh, Uh, Uh…” I was gonna lie, but then I said no. I said, “I'm sorry I took your ring.” She goes, “It's OK. I understand”… But ever since then I felt that I was never invited back to her house again.

Other participants similarly felt that their behavior distanced people. One participant, a woman in her early 40s, pointed to substance abuse as a problem in her relationship with her family. “I had started building up my drug use. I had family problems…I got caught up in the drug use and started crime and my family problems really escalated.”

Pete also encountered family problems with his former wife and family members. He described finding reasons to leave the house to escape the tension and conflict that he felt his presence generated, and pushing his family to the point of rejection:

  • I used to walk out. And I'd go over to the corner store…and we used to sit there and drink beer and stuff…So I'd get back home and hear it again, and again, and again … It was just the same thing everyday. Or I'd get into a fight with (my wife)…The whole family. I'm fightin’ the whole family. I mean, fightin’-not just arguments. It was fightin’. They're like, “…you gotta get out. You have to leave.”

Pushes and pulls are often hard to disentangle, but in most cases the participants viewed them in more of an agentic way, acknowledging their own role in the separation. For example, one participant described feeling pushed away by peer group members because of his mental illness, but his description also includes his own pulling back.

  • And my peers they made me do that cause they were saying things about my original self, that made me want to change into something I'm not…‘I don't like you, cause you this way or that way. Or your opinion just sounds stupid.’ Or ‘you're weird.’ Or ‘you, you, you sound like you have low-low self-esteem about yourself, self-confidence, so, we don't want you with our crowd. We don't want you to fit in with us. We don't like your kind around here.’

Problems of Their Own

Because many of the individuals had a family history of poverty, they had relatives and friends who faced challenges of their own. Some participants' family members and friends faced mental illness as well as physical illnesses, had served time in jail or were currently incarcerated, or were limited by their own substance abuse. Others felt burdened by stresses of managing families and low income jobs.

One Latino male, age 28, discussed how his parents served as a role model for his own drug use as a child.

  • My mother was doing weed and my father was shooting up dope, he's in the gang, my mother's in the gang, a bunch of drug war stuff. That's why I started using drugs…

It was not uncommon for social network members, especially family members, to have mental illness or substance addiction as well, sometimes diagnosed and sometimes not.

  • I was living in Bronx when I met my husband…He was a drug addict. He was doing it to escape from having to have to live (with a mental illness). And…he was constantly going in and out of mental, men-, mental…he was constantly going out of drug, drug, drug rehabilitation testing.

Another participant visited her boyfriend and family members who were substance abusers while she was in treatment.

  • I had just started sniffing cocaine with my family on the weekends…I'd get a pass to stay out for the weekend and come back.

Dealing with these problems was often very stressful for participants, as this participant explains his father's mental illness.

  • It got to a point where it became overbearing for me…I took my dad for shock treatments; when they come out they don't know where they are. And one time I had to go to the store and I was walking with him and he… was like a little drunk. And I was going to the store to get some cigarettes…And he started to run away from me. I said, “No, dad!” And here I'm running after him on street. And it took everything out of me. And I cried to my brother, cried to my mom. I felt like a rubber band that's being expanded and…and ready to break.

Often, when participants turned to their social network for help, they sought emotional or financial support. While some individuals were helpful, others had their own intense struggles and simply could not spare any extra resources. Jimmy, for example, had a brother who was also mentally ill, and was barely scraping by providing for his two sons.

Bonding, Rebonding, and Social Isolation

If the pre-treatment characteristics of the social networks are described as “bonding,” where the members of the social network shared socio-economic, demographic, and psychological characteristics (Szreter & Woolcock, 2004), then perhaps participants engaged in “rebonding” during and following treatment. Study participants wanted to interact and rebuild a social network, often with people much like themselves: they had a mental illness, were former or current substance abusers, lived in poverty, and were dealing with multiple problems.

One man in his 50s met a woman he helps out from time to time.

  • I got acquaintances that [need] help…Like this girl in Philly…I take to the AA meetings…She has grandchildren. Lives in a single room, and I feel for her…And we both have a friendship. She's there building on me, you know, building up…she's working now like a babysitter. And she's great at it, you know. But I told her, I said, “You just got out of the [psychiatric] hospital, but I hope you told the people that you're working for that you, you know, what's going on.” You understand?

Other study participants found relationships within a treatment program as well. This bonding, however, continues relationships that are similar to the former bonding relationships: they provide emotional support, and are limited in their ability to provide other types of reliable support. One male participant said that he was close friends with one person already in his program and another who was trying to get into the program.

  • He's another person that I would, would talk to. But that's about it, one, two people, because you know, you don't know who you can share this information.

Other participants believed that peer pressure played too large a role in their lives and wanted to break away from that pressure. A 59-year-old woman described the pressure she felt from a group of friends she met at a psychiatric hospital.

  • They [my friends] were drinking…everybody was drinking. They said, “Come on…the guy's buying beer for everybody.” I said, “No, no, no, no. ‘Cause I knew if I started drinking I'd act like a fool. And then, what kind of housing are they gonna get me? I'm there for housing. I'm not there for drink. I'm sorry. “Oh you want to be one of them stuck in the mud kinds of bitches….” So you get a bad name with your peers. But fuck them, you know. You're not there for them.

Many study participants had had their trust betrayed, and spoke of how they no longer trusted very many people. Some wanted to make new friends and build relationships; others, however, preferred to isolate themselves. Although several men isolated themselves, women appeared more likely to do so.

When asked about her friends one female participant said:

  • I don't have no friends…I had one friend. She stabbed me in my back…Not literally stabbed me in my back, no. She said she was my friend, she wasn't my friend.

In contrast, a male participant recalled that after he received treatment, he made significant changes to his social environment.

  • I had to make all new friends. When I came into this fellowship I had to let go of all my friends. I had to stop going to all the old places I used to go. And now I replaced them with new places and new friends that I know now. So I don't go to any of those places any more.

Although it was rare, a few participants created bridging or linking social capital through their service providers. These connections developed beyond the typical case management or counseling relationship into one that seemed more personal. One participant describes such a relationship:

  • I know his (telephone) number, and his address…He became like my adopted father when my mother passed away, he was there through thick and thin. He helped Betty, my mother…He was that kind-hearted. And more respect…he gave more respect for me and my mother than any in my family.

Another participant describes her ongoing friendship with a former provider. She developed a bond with him as he helped her through various life situations, first within a professional setting, but later outside of the institution.

  • He kept up with me…he always had time for me. You know even if I wasn't even affiliated at the hospital…He was my social worker when I was younger. When I was going to day treatment. ‥I'd just go there, hey, talk, you know, whatever. Sometimes I needed some money he would lend me some money. It was like a friendship thing. Not a friendship, but I don't know how to explain it. More like a person I know. A good person I know. Confidant. Yeah. Not a friend, yeah, maybe a confidant, yeah. A friend confidant.

These two examples notwithstanding, bridging and linking relationships were rare. In other research with low-income, non-mentally ill, and non-substance abusing samples, service providers are often viewed by their clients as part of their social/friend network (Dominguez & Watkins, 2003), creating the possibility of generating positive bridging or linking social capital. But, if these relationships existed, they were fleeting among our sample.

Discussion and Implications

The literature indicates that the social networks of individuals with co-occurring disorders are small, but they can be helpful. This study examines why those networks are small, explores the nature of the networks, and illuminates how a group of formerly homeless individuals with co-occurring disorders experience and react to them. The findings may best be illustrated by the conceptual model in Figure 1. We found that members of social networks died at a relatively young age, that study participants withdrew or felt that they could not cope with social relationships and pushed their social network away, and that some network members faced such large obstacles of their own that they could not afford to provide help to the study participants. New people rarely appeared to fill the gaps. These realities led to social isolation at one end of a spectrum, and the rebuilding of a social network that looked very similar to their original networks at the other. In managing this process, some participants “rebonded” with old, tenuous ties, while others chose or settled for social isolation. The social capital is limited and consists almost solely of bonding relationships. Those who “rebonded” found themselves in situations that looked very similar to a point earlier in their lives.

Either end of this spectrum can have negative outcomes for people with co-occurring orders. Social isolation can limit possibilities for treatment and recovery into a more fulfilling life. A rebuilt social network, however, can lack socio-economic power, influence, and connections, and risks negative social capital. However, both results can have a positive valence as well, as in some cases a rebuilt network is rich in emotional support, or isolation provides comfort and safety. Of course, these situations represent extremes, and much occurs in between. Only a few participants were completely isolated, and none were completely re-connected to their old social network. Few options existed in study participants' lives for building social capital or for benefiting from the bridging and linking types of social capital that are important to making social and economic advances. In fact, with few exceptions, bridging and linking social capital are entirely absent from these life stories. Some participants had meaningful relationships with service providers, but for a service provider relationship to develop beyond a standard professional connection was extremely rare.

This research provides implications for public policy. Service providers may expect that their clients would have small networks because of the client's psychological state or history of substance abuse. Knowing exactly why a social network is small helps providers better understand how a client handles relationships and interactions with others. Not only would a deeper assessment help providers better understand their client's social interactions, but it provides insight into the client's personality, social abilities, and ways of re-integrating into the community and building new relationships or reconstructing those that could produce positive social capital.

Service providers can consider three areas with respect to the social networks of clients with co-occurring disorders. In general, they can:

  1. Consider not only who is in the network, but who has disappeared from the network and why;
  2. Work with clients to better handle the positive and negative elements of social ties; and
  3. Help provide opportunities for creating bridging and linking, as well as enhanced bonding.

For some co-morbid individuals, social skills training can be explored. In social skills training, individuals learn coping, cognitive, behavioral, interpersonal, and instrumental skills that they may have not fully developed due to mental illness (Liberman, 1998). These new skills might help in building personal and professional relationships. Providers can also look for other ways in which clients can develop better social skills, such as by creating social and professional situations that may help bridging and linking opportunities develop naturally. One approach might be a form of Supported Employment (SE), where individuals with disabilities are integrated into competitive work environments (Substance Abuse and Mental Health Services Administration, 2003; Shankar & Collyer, 2002). SE programs help clients find and keep employment while maintaining treatment and other services. Currently, SE and related employment programs exist in the United States for people with mental illnesses, and have been relatively successful in enhancing income, integration into a community, and quality of life. However, to date, they have limited implementation, and social service programs have paid little attention to their impact among people with co-occurring disorders (Cook et al., 2001; Latimer, 2001). Supported Employment, coupled with social skills training and relationship building, could potentially enhance not only employability but also the ability to create bridging and linking social capital among individuals with co-occurring disorders.

Limitations of the Study

Generalizability in this study is limited by our small sample size and purposive sampling method. As in most qualitative studies, we did not intend to produce a representative sample of the entire formerly homeless co-morbid population. Instead, the sample size of 39 allowed for an in-depth analysis to raise new questions and generate theories (Strauss & Corbin, 1998; Feagin, Orum, & Sjoberg, 1991), but we cannot say how far these patterns extend. Further, while this paper addresses social capital and social networks, the original study design did not focus on these topics, suggesting that at least some important details may have been omitted or overlooked by interviewers or study participants. This study also omits the role that faith communities might play in recovery. While few study participants mentioned church or religion specifically, there are studies that find that religion and spirituality are central dimensions to self-understanding and recovery, especially among African Americans with serious mental illness (Fallot, 1998).

In addition, memories are fallible, and relationships are fragile, and undoubtedly, if we were to interview social network members, we would hear a different version of events. But, what matters is the meaning that participants have made of events, rather than the factual truth of the events. We believe it is important to see events as the individuals see them, in order to be better able to understand and to work with them.


While previous studies have suggested that the social support networks of individuals with co-occurring disorders are small, they usually only speculate about why this might be the case. In this paper, we examined what happens to these social networks and discovered a multi-level process. Using a social capital framework, we found that the networks shrunk when members of social networks died, the participants pushed and pulled away, or network members become caught up in deep-rooted problems of their own related to poverty, mental illness, or substance abuse. Later, a process of rebonding sometimes occurred, mostly among men, while a larger proportion of the women tended to isolate themselves socially. Each of these situations could be positive or negative.

To our knowledge, no other study has examined the functioning of social networks and social capital with this population in their own words and perspectives. Researchers and practitioners working with the homeless co-morbid population should look beyond the size of the network and consider the dynamics of the relationships between an individual and his/her family and friends. Understanding the network, its voids, how it functions or fails to function, and its meaning in a person's life provides a window into someone's life that may help explain their motivations and decisions, and also assists them in moving toward recovery. Future research can further examine the complexity of these social networks, emphasizing how they function and how they are understood by the co-morbid client. And more attention can be paid to intervention models that include opportunities for clients to build social capital.


This project was supported by grant #R01-MH69865 from the National Institute of Mental Health, Deborah Padgett, PI. The authors are grateful to the study participants for sharing their lives with us. We would like to thank Deborah Padgett for her comments on earlier drafts and for her guidance and generosity in assisting us with this study. We also wish to thank Deesha Narichania for her help in preparing this paper and our NYU colleagues Dan Gardner, Caroline Rosenthal Gelman, Allison Werner-Lin, and Ellen Tuchman for their comments on earlier versions of this paper. We would also like to thank New York Services Study interviewers Andrew Davis, Ana Stefancic, and Maureen Hayes.


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


  • Albert M, Becker T, McCrone P, Thornicroft G. Social networks and mental health service utilization—a literature review. International Journal of Social Psychiatry. 1998;44(4):248–266. [PubMed]
  • Alverson H, Alverson M, Drake RE. An ethnographic study of the longitudinal course of substance abuse among people with a severe mental illness. Community Journal of Mental Health. 2000;36:557–569. [PubMed]
  • Antonucci TC, Akiyama H, Lansford JE. Negative effects of close social networks. Thousand Oaks CA: Sage Publications; 1998.
  • Bellack A, Muesser K. A comprehensive treatment program for schizophrenia and chronic mental illness. Community Mental Health Journal. 1986;22(3):175–189. [PubMed]
  • Blankertz LE, Cnaan RA. Assessing the impact of two residential programs for dually diagnosed homeless individuals. Social Service Review. 1994;68:536–560.
  • Bond G, Becker D, Drake R, Rapp C, Meisler N, Lehman A, Bell M, Blyler C. Implementing supportive employment as an evidence-based practice. Psychiatric Services. 2001;52(3):313–322. [PubMed]
  • Bourdieu P. Cultural reproduction and social reproduction. In: Karabel J, Balsey AH, editors. Power and ideology in education. New York: Oxford University Press; 1977. pp. 487–511.
  • Bourdieu P. The forms of capital. In: Richardson JG, editor. Handbook of theory and research for the sociology of education. New York: Greenwood; 1985. pp. 241–258.
  • Carpiano R. Toward a neighborhood resource-based theory of social capital for health: Can Bourdieu and sociology help? Social Science & Medicine. 2005;62:165–175. [PubMed]
  • Caton CLM, Shrout PE, Dominguez B, Eagle PF, Opler LA, Cournos F. Risk factors for homelessness among women with schizophrenia. American Journal of Public Health. 1995;85(8):1153–1156. [PubMed]
  • Caton CLM, Shrout PE, Eagle PF, Opler LA, Felix A, Dominguez B. Risk factors for homelessness among schizophrenic men: A case-control study. American Journal of Public Health. 1994;84(2):265–270. [PubMed]
  • Clark RE. Family support for persons with dual disorders. New Directions for Mental Health Services. 1996;70:65–77. [PubMed]
  • Clark RE. Family support and substance use outcomes for persons with mental illness and substance use disorders. Schizophrenia Bulletin. 2001;27(1):93–101. [PubMed]
  • Clark RE, Drake RE. Expenditures of time and money by families of people with severe mental illness and substance use disorder. Community Mental Health Journal. 1994;30:145–163. [PubMed]
  • Cleak H, Howe JL. Social networks and use of social supports of minority elders in East Harlem. Social Work in Health Care. 2003;38(1):19–38. [PubMed]
  • Coleman JS. Social capital in the creation of human capital. American Journal of Sociology. 1988 94:S95–S120.
  • Coleman JS. Foundations of social theory. Harvard University Press; Cambridge MA: 1990.
  • Cook JA, Pickett-Schenk SA, Grey D, Banghart M, Rosenheck RA, Randolph F. Vocational outcomes among formerly homeless persons with severe mental illness in the ACCESS program. Psychiatric Services. 2001;5(8):1075–1080. [PubMed]
  • Copello A. Substance misuse and psychosis in context: The influences of families and social networks. In: Graham HL, Copello A, Birchwood MJ, Mueser KT, editors. Substance Misuse in Psychosis: Approaches to Treatment and Service Delivery. NJ: John Wiley and Sons; 2003. pp. 43–60.
  • Corcoran M, Adams T. Race, sex, and the intergenerational transmission of poverty. In: Brooks-Gunn GJ, Brooks-Gunn J, editors. Consequences of growing up poor. New York: Russell Sage Foundation; 1997. pp. 461–517.
  • Dominguez S, Watkins C. Creating networks for survival and mobility: Social capital among African-American and Latin-American low-income mothers. Social Problems. 2003;50(1):111–135.
  • Drake RE, Brunette MF, Mueser KT. Substance use disorders and social functioning in schizophrenia. In: Mueser KT, Tarrier N, editors. Handbook of Social Functioning. Boston: Allyn & Bacon; 1998. pp. 280–289.
  • Erickson DH, Beiser M, Iacono WG. Social support predicts 5-year outcome in first-episode schizophrenia. Journal of Abnormal Psychology. 1998;107:681–685. [PubMed]
  • Fallot RD. The place of spirituality and religion in mental health services. In: Fallot RD, editor. New Directions for Mental Health Services. Vol. 80. 1998. Winter. pp. 3–12. Spirituality and Religion in Recovery from Mental Illness [Special issue] [PubMed]
  • Feagin J, Orum A, Sjoberg G. A case for case study. Chapel Hill NC: University of North Carolina; 1991.
  • Findler LS. The role of grandparents in the social support system of mothers of children with a physical disability. Families in Society; The Journal of Contemporary Human Services. 2000;81(4):370–381.
  • Finfgeld-Connett D. Empowerment of individuals with enduring mental health problems results from concept analyses and qualitative investigations. Advances in Nursing Science. 2004;27(1):44–52. [PubMed]
  • Gitell RV. Community organizing: Building social capital as a development strategy. Newbury Park: Sage Publications; 1998.
  • Jankowski S, Videka-Sherman L, Laquidara-Dickinson K. Social support networks of confidants to people with aids. Social Work. 1996;41(2):206–213. [PubMed]
  • Kawachi I. Social capital and community effects on population and individual health. Annals of the New York Academy of Sciences. 1999;896:120–130. [PubMed]
  • Kawachi I, Kennedy B, Lochner K, Prothrow-Stith D. Social capital, income inequality and mortality. American Journal of Public Health. 1997;87:1491–1498. [PubMed]
  • Latimer EA. Economic impacts of supported employment for persons with severe mental illness. Canadian Journal of Psychiatry. 2001;46:496–505. [PubMed]
  • Laudet AB, Cleland CM, Magura S, Vogel HS, Knight EL. Social support mediates the effects of dual-focus mutual aid groups on abstinence from substance use. American Journal of Community Psychology. 2004;34(34):175–185. [PMC free article] [PubMed]
  • Laudet AB, Magura S, Vogel HS, Knight EL. Support, mutual aid and recovery from dual diagnosis. Community Mental Health Journal. 2000;36(5):457–476. [PMC free article] [PubMed]
  • Liberman RP. International perspective on skills training for the mentally disabled. International Review of Psychiatry. 1998;10:5–8.
  • Lochner K, Kawachi I, Kennedy BP. Social capital: a guide to its measurement. Health & Place. 1999;5:259–270. [PubMed]
  • Mlynarczyk S. Social support. In: Miller J, Lerner RM, Schiamberg LB, Anderson PM, editors. The encyclopedia of human ecology. Vol. 2. Denver, CO: ABC-CLIO, Inc; 2004. pp. 655–658.
  • Mueser KT, Bellack AS, Blanchard JJ. Comorbidity of schizophrenia and substance abuse: Implications for treatment. Journal of Consulting and Clinical Psychology. 1992;60:845–856. [PubMed]
  • O'Farrell TJ, Hooley J, Fals-Stewart W, Cutter HSG. Expressed emotion and relapse in alcoholic patients. Journal of Consulting and Clinical Psychology. 1998;67:744–752. [PubMed]
  • Peek MK, O'Neill GS. Networks in later life: An examination of race differences in social support networks. The International Journal of Aging and Human Development. 2001;52(3):201–229. [PubMed]
  • Perese EF, Wolf M. Combating loneliness among persons with severe mental illness: Social network interventions' characteristics, effectiveness, and applicability. Issues in Mental Health Nursing. 2005;26:591–609. [PubMed]
  • Poortinga W. Social capital: An individual or collective resource for health? Social Science & Medicine. 2005;62:292–302. [PubMed]
  • Portes A. The two meanings of social capital. Sociological Forum. 2000;15(1):1–12.
  • Portes A. Social capital: Its origins and applications in modern sociology. Annual Review of Sociology. 1998;24(1):1–24.
  • Putnam RD. Bowling Alone: The collapse and revival of American community. New York: Simon & Schuster; 2000.
  • Roschelle AR. No more kin: Exploring race, class, and gender in family relations. Family Relations. 1997;47(4):379–384.
  • Savage A, Russell LA. Tangled in a web of affiliation: Social support networks of dually diagnosed women who are trauma survivors. Journal of Behavioral Health Services and Research. 2005;32(2):199–214. [PubMed]
  • Shankar J, Collyer F. Support needs of people with mental illness in vocational rehabilitation programs -the role of the social network. International Journal of Psychosocial Rehabilitation. 2002;7:15–28.
  • Strauss AL, Corbin J. Basics of qualitative research: Techniques and procedures for developing grounded theory. 2nd. Thousand Oaks CA: Sage; 1998.
  • Substance Abuse and Mental Health Services Administration. Supported employment: A guide for mental health planning & advisory councils. Rockville MD: U.S. Department of Health and Human Services; 2003. pdf file. Available from SAMHSA's National Mental Health Information Center Web site,
  • Szreter S, Woolcock Health by association? Social capital, social theory, and the political economy of public health. International Journal of Epidemiology. 2004;33(4):650–667. [PubMed]
  • Test MA, Wallish L, Allness D, Ripp K. Substance use in young adults with schizophrenic disorders. Schizophrenia Bulletin. 1989;15:465–476. [PubMed]
  • Tracy EM, Abell N. Social network map: Some further refinements on administration. Social Work Research. 1994;18(1):56–60. [PubMed]
  • Trumbetta SL, Mueser KT, Quimby E, Bebout R, Teague GB. Social networks and clinical outcomes of dually diagnosed homeless persons. Behavior Therapy. 1999;30:407–430.
  • Tsemberis S, Gulcur L, Nakae M. Housing First, consumer choice, and harm reduction for individuals who are homeless with dual diagnosis: A 24-month follow up. American Journal of Public Health. 2004;94(4):651–656. [PubMed]
  • Whitley R, McKenzie K. Social capital and psychiatry: Review of the literature. Harvard Review of Psychiatry. 2005;13(2):71–84. [PubMed]
  • Woolcock MM. The place of social capital in understanding social and economic outcomes. Isuma: Canadian Journal of Policy Research. 2001;2(1):11–17.