Peer support among persons with severe mental illnesses: a review of evidence and experience

Abstract

Peer support is largely considered to represent a recent advance in community mental health, introduced in the 1990s as part of the mental health service user movement. Actually, peer support has its roots in the moral treatment era inaugurated by Pussin and Pinel in France at the end of the 18th century, and has re-emerged at different times throughout the history of psychiatry. In its more recent form, peer support is rapidly expanding in a number of countries and, as a result, has become the focus of considerable research. Thus far, there is evidence that peer staff providing conventional mental health services can be effective in engaging people into care, reducing the use of emergency rooms and hospitals, and reducing substance use among persons with co-occurring substance use disorders. When providing peer support that involves positive self-disclosure, role modeling, and conditional regard, peer staff have also been found to increase participants’ sense of hope, control, and ability to effect changes in their lives; increase their self-care, sense of community belonging, and satisfaction with various life domains; and decrease participants’ level of depression and psychosis.

“As much as possible, all servants are chosen from the category of mental patients. They are at any rate better suited to this demanding work because they are usually more gentle, honest, and humane” – Jean Baptiste Pussin, in a 1793 letter to Philippe Pinel 1.

Peer support among persons with severe mental illnesses has been largely considered a recent phenomenon, with the first published account of this presumably “new” form of service delivery dating to 1991 2 and attributed to the mental health service user movement that began in the 1970s 3. As the passage quoted above suggests, however, the idea that persons in recovery may be especially well-suited to helping others suffering from a severe mental illness has a longer, if unacknowledged, history.

At the time of his writing the letter from which the above passage was taken, Pussin was serving as the governor (i.e., superintendent) of the Bicêtre Hospital in Paris, where he had himself been a patient. Pinel had been assigned to become the chief physician there, and had asked Pussin to describe how the hospital was functioning prior to his arrival. As noted above, one of Pussin’s key management strategies was to hire as many staff for the hospital as possible from among recovered patients. In addition to being “gentle, honest, and humane”, Pinel found these former pa- tients recruited by Pussin to be “averse from active cruelty” (which was a common management strategy in the asy- lums of the day) and “disposed to kindness” 4 toward the patients in their care. It was then to a significant degree through the hiring and deployment of such staff that Pinel, and Pussin, were able to do away with shackles and abuse, and institute what has since come to be called the “moral treatment” era 5.

Recognition of the value of peer support among persons with severe mental illnesses thus goes back centuries, with the practice showing up periodically and with apparently good effect throughout the history of psychiatry. For example, Harry Stack Sullivan used a similar strategy of hiring patients who had recovered from psychotic episodes to staff his inpatient unit in the US in the 1920s 5.

Over the last twenty years, the practice of peer support has virtually exploded around the globe, with many more recovering persons being hired to provide peer support than ever before. Estimates place the number of peer support staff currently to be over ten thousand in the US alone 6, and this number continues to rise at an astonishing rate despite the global recession and high unemployment rates. It is thus timely to step back from what has become something of a juggernaut within contemporary mental health policy and practice to review what is known thus far about this particular strategy within the context of community-based practice. This paper will do so by, first, reviewing the existing evidence base and, second, by describing some of the concerns that have emerged as more peers have been hired, along with some of the strategies that have been found useful in addressing and overcoming these concerns.

EVIDENCE

“Yeah, it’s nice to know… it’s like having someone that you can confide in, you feel like you’re kind of in the same boat… She was depressed, homeless, with a drug problem. And that’s where I was. And I’m newer to it. She’s got a car, she’s got her apartment, and I’m building those things, and it’s just… you know, somebody who really knows” – Person with severe mental illness describing experiences with a peer provider

Much of the research conducted on peer support to date can be conceptualized as falling into one of three categories lying along a linear continuum 7,8. The first stage of research involved feasibility studies, in which the main aim was to demonstrate that it was in fact possible to train and hire persons with histories of severe mental illnesses to serve as mental health staff. Given the history of stigma and discrimination against persons with severe mental illnesses, it was necessary first to show that such people could perform the tasks involved. At this initial stage, the roles for peer staff were conceptualized primarily as ancillary to and supportive of conventional staff as case manager aides or companions; roles for which few specific skills or competencies were required. Four randomized controlled trials conducted during the 1990s demonstrated consistently that peer staff were able to function adequately in these roles and to produce outcomes at least on a par with those produced by non-peer staff 9,10,11,12, with two studies showing slightly better outcomes for those receiving peer support in addition to usual care as compared to those receiving usual care only 9,11.

The second stage of research involved studies comparing peer staff and non-peer staff, with both functioning in conventional roles such as case managers, rehabilitation staff, and outreach workers. In these studies of conventional services provided by peers, most studies again found that peer staff functioned at least as well in these roles as non-peer staff, with comparable outcomes 13,14,15,16. Several studies in this second stage of research began to detect consistent differences between these two conditions, however, with peer-delivered services generating superior outcomes in terms of engagement of “difficult-to-reach” clients, reduced rates of hospitalization and days spent as inpatient, and decreased substance use among persons with co-occurring substance use disorders 17,18,19,20. These promising findings led investigators in this area to emphasize the need for the next generation of research to specify and begin to evaluate the ways in which peer staff may perform their roles differently from non-peer staff, based on their first-hand experiences of disability, stigma, and recovery, and whether new roles can be created in the mental health system in which these life experiences can be used most effectively to promote the recovery of others 8,21,22,23,24.

In other words, insofar as the second stage of research evaluated the functioning of peers in providing conventional services within conventional roles, these studies stopped short of investigating whether or not peers could perform new functions in new roles that were unique to them because they were derived specifically from their own first-hand experiences of illness, recovery, and service use – experiences that were not shared by non-peer staff.

It has thus required a third generation of studies to begin to answer the following questions: a) Do interventions provided by peers differ in any significant way from the same interventions provided by non-peers? b) Are there any interventions that cannot be provided by people who do not have their own first-hand experience of mental illness, which thereby make peer support a unique form of service delivery? c) If so, what are the active ingredients of these aspects of peer support, and what outcomes can they produce?

Thus far, the literature has suggested three basic contributions of peer support that would seem to be unique to, or at least especially well-suited for, peer staff. The first is the instillation of hope through positive self-disclosure, demonstrating to the service recipient that it is possible to go from being controlled by the illness to gaining some control over the illness, from being a victim to being the hero of one’s own life journey 23,25. The second expands this role modeling function to include self-care of one’s illness and exploring new ways of using experiential knowledge, or “street smarts”, in negotiating day-to-day life, not only with the illness but also with having little to no income, with being unstably housed, with overcoming stigma, discrimination, and other trauma, all the while trying to negotiate the complex maze of social and human service systems 23,26.

The third aspect of peer support focuses on the nature of the relationship between peer provider and recipient, which is thought to be essential for the first two components to be effective. This relationship is characterized by trust, acceptance, understanding, and the use of empathy; empathy which in this case is paired with “conditional regard” – otherwise described as a peer provider’s ability to “read” a client based on having been in the same shoes he or she is in now. Their ability to empathize directly and immediately with their clients can be used in this particular way by peer providers because they may have higher expectations and may place more demands on their clients, knowing that it is possible to recover, but also that it takes hard work to do so (e.g., “I know how hopeless you feel now, but I also know that you can work toward a better life”) 26,27. These expectations may at times lead to conflict, but also are just as likely, if not more so, to lead to encouragement and inspiration 26,27,28,29.

While this third stage of peer research – which focuses on any potential unique qualities that alternative peer-provided services may have – is only just getting underway at this time, a couple of recently completed studies are suggestive of what may be in store. For example, our research team completed a study, funded by the National Institute of Mental Health, of culturally-responsive, person-centered care for psychosis among adults of African and/or Hispanic origin in which peer staff played two new roles that were developed specifically for that study. Using the evidence-based practice of illness management and recovery (IMR) as our comparison condition 30, we trained peer staff to provide one of two sets of interventions. The first set involved acting as an advocate to facilitate person-centered care planning for participants to engage them more actively in directing their own treatment and recovery. The second set involved acting as a “community connector” to support participants in pursuing the community activities and roles they had identified in their care plan. A total of 290 adults with serious mental illness were randomly assigned to one of three conditions that built on each other in a graduated way: a) usual care plus the invitation to participate in the evidence-based practice of IMR; b) usual care plus IMR plus a peer-facilitated person-centered planning process (PCP); and c) usual care plus IMR and PCP with the addition of the peer-run community connector program.

In this study, we found benefits to both forms of peer support as compared to usual care plus IMR. In particular, the addition of a peer-facilitated person-centered care planning process increased the degree to which participants felt their care was responsive and inclusive of non-treatment issues (such as housing and employment), and increased their sense of control and ability to bring about changes in their lives. The peer-run community connector program increased their sense of hope and degree of engagement in managing their illness, degree of satisfaction with family life, positive feelings about themselves and their lives, social support, and sense of community belonging. Finally, and perhaps most unexpectedly, the peer community integration program decreased participants’ level of psychotic symptoms, while increasing the amount of distress they experienced due to these symptoms. Qualitative data suggested that this increase in distress may have been due to the fact that participants were encouraged to do more with their lives and perceived their remaining symptoms as barriers to pursuing activities that interested them, while in the past these same symptoms, though more prominent, were not viewed in the same way as barriers to a fuller life 31.

The second study built on earlier findings, mentioned above, which suggested that peer support might be useful in decreasing rates of hospitalization and days spent in the hospital for persons with histories of multiple hospitalizations. For this study, we trained and deployed peer staff to serve as “recovery mentors” (the name they chose for themselves), a broader role that integrated the interventions of both the PCP advocate and community connector of the earlier study. Feedback from both participants and peer staff in the previous study indicated a strong preference for having both of the functions of advocacy and community integration performed by one person in a continuous manner, rather than requiring the participant to develop trust in two different people. As a result of this feedback, we developed a model of recovery mentoring that absorbs these and other related functions into the role of one peer provider who, most importantly, was trained in how to use his or her own life experiences to the benefit of his or her clients. In addition to the positive uses of self-disclosure, peer staff were trained in developing empathic relationships, using conditional regard, and role modeling self-care.

For this study, participants were randomly assigned either to usual care or usual care plus a peer recovery mentor. To be eligible, patients had to have experienced two or more hospitalizations during the 18-month period prior to the current admission and have a documented history of a severe mental illness. Data were collected at baseline (during index hospitalization) and again at 3 and 9-month post-discharge. The main outcomes were the number of hospitalizations and hospital days during the 9-month study period, measured through a combination of medical records, administrative databases, and self-report.

We used an intention-to-treat analysis including a total of 74 participants. Primary statistical analyses utilized a univariate analysis of covariance to assess differences between the conditions in inpatient admissions and total number of days in the hospital, both at the end of 9-month participation and controlling for pre-18-month baseline levels. Partial eta squared (?2) served as an estimate of between-condition effect size. For the primary outcome analyses, we tested if the readmission experiences (events and days) reflected statistically significant changes at the end of the study between the conditions (recovery mentors or control) as the between-subjects independent variable. Unlike generalized linear models, linear mixed models – which are commonly employed in community-based research – examine variation within individuals, at the same estimating levels of correlation with other key factors, and are capable of interpolating values for uneven numbers of repeated measurements. We set the significance criterion at p≤0.05, and, in the case of the hospitalization experience, used a one tailed test based on our hypothesis that having a recovery mentor would be associated with less use of the hospital.

There were statistically significant main findings for the number of hospitalizations and the number of days spent in the hospital, with participants assigned recovery mentors doing significantly better than those without a recovery mentor on both number of admission events (0.89±1.35 vs. 1.53±1.53 events, F=3.07, df=1, p<0.042, one tailed) and number of hospital days (10.1±17.31 days vs. 19.1±21.6 days, F=3.63, df=1, p<0.03, one tailed). In addition, we analyzed a range of measures tapping into possible intervening variables that might reflect the therapeutic mechanisms of peer support, and these findings supported the general hypothesis that the assignment of a recovery mentor also had other beneficial effects. Consistent with previous studies, there was a significant decrease in substance use for people receiving recovery mentors. New findings, however, included a decrease in depression and increases in hope, self-care, and sense of well-being 32 – all important domains of recovery that are consistent with the model of peer support described above 33,34,35.

Currently, we are pursuing the next step in this line of research, conducting a randomized controlled trial that controls for frequency and intensity of contact and compares the effectiveness of peer recovery mentors to peer case managers, on the one hand, and non-peer recovery mentors, on the other, to attempt to tease out the most crucial elements of peer support (i.e., the third of our three questions above). While obviously much work remains to be done in understanding both the nature and effectiveness of peer support, much work has been done already, and much more is being carried out currently, to bring this central practice of the 18th century moral treatment era into the mainstream of 21st century community-based care.

EXPERIENCE

Since the hiring of peer staff both stimulates and requires significant changes in the culture of traditional mental health settings, as much, if not more, has been written about challenges involved in implementation, and strategies for overcoming these challenges, as about research on outcomes. We review both the challenges and strategies below, drawing from our twenty years of experience in recruiting, training, deploying, and retaining peer staff and from the experiences of others involved in this work 27,36,37,38,39,40,41,42,43,44. We begin with the five most common questions asked by staff and administrators in mental health settings, followed by brief answers to each. We then present a series of recommendations for implementation.

Common practitioner concerns

Aren’t peer staff too “fragile” to handle the stress of the job?

No. Jobs in mental health settings are stressful for everyone, not just for peer staff. As a result, self-care is an important area of focus for all mental health staff, not just peer staff. It is true, however, that peer staff are asked to take on the additional burdens of disclosing some of their most personal experiences and putting these experiences to good use in helping others as well as bearing the additional scrutiny of having to represent all peers (i.e., if they do not do well in the job, it may factor in whether or not the agency will continue to value peer services). Managing these processes are important foci for supervision. With respect to “fragility”, though, people in recovery should be considered to have shown a considerable amount of persistence and resilience, as opposed to fragility, in battling back against the illness. Recovery, after all, is hard, taxing, and ongoing work.

Despite the considerable efforts peer staff have had to make to overcome the more debilitating aspects of the illness, many administrators continue to worry about potential relapses or setbacks and have looked for various indicators of stability in their hiring practices. These have taken various forms, such as stipulating at least a year since last hospital admission or two years since using substances. In addition to lacking reliability in terms of their predictive value, instituting such requirements in hiring would constitute discrimination under many countries’ disability rights legislation. As long as the person can perform the essential functions of the job, consideration of his or her psychiatric history in terms of the use of arbitrary criteria of functioning is no longer acceptable practice. This is one of the many significant changes that are introduced when mental health practitioners shift from viewing an individual with a mental illness as a patient to viewing him or her as an employee.

Don’t peer staff relapse?

All employees, including peer staff, take off time because of illness. Many staff who are not identified as peers take off time because of mental health issues. The same expectations for sick time and accommodations for illness should be applied for all employees, including peer staff. Even when facing adversity or not feeling well, peer staff can still serve as role models in showing the kind of determination, resilience, and persistence it takes to come back to work following a difficult period. In addition, the stress of working may be considered less onerous in comparison to the stress of prolonged involuntary unemployment, poverty, and isolation.

Can peer staff handle the administrative demands of the job?

While some people might not have worked for a prolonged period before joining the peer workforce, and others might have had limited educational opportunities, many peer staff are equally if not more competent at administrative tasks than other staff members. For those who do struggle with these tasks, peer staff can be shown how to manage the administrative details of their jobs and, when needed, provided with supports to enable them to do so. As just one example, for people with cognitive or linguistic impairments, such strategies as speaking into a recorder and having one’s notes transcribed may be useful in meeting documentation requirements.

Won’t peer staff cause harm to clients by breaking confidentiality or by saying the “wrong” things?

Peer staff, like all other employees, are expected to conform to policies and regulations regarding confidentiality and privacy. Training and supervision support this, and peer staff are employees who are just as responsible as any other staff for keeping client information confidential. There is no reason to believe that this will be any more difficult for peer staff than for anyone else. In fact, given the sensitivity peer staff have to privacy issues, based on their own experiences as a service user, it is more likely that they will guard their clients’ confidentiality even more so than non-peer staff. Also, there is no reason to believe that peer staff who have been trained and are supervised will be more likely than other staff to say the “wrong” things. If “wrong things” include demeaning and disrespectful treatment, then peer staff may indeed be less likely to engage in this kind of behavior, as initially witnessed and remarked upon by Pussin.

Won’t peer staff make my job harder rather than easier?

Peer support provides an important and useful complement to existing mental health services. Peer staff can be especially effective in engaging people into care and acting as a bridge between clients and other staff. When well-trained and supervised, peer staff can serve rather to lessen the load carried by other practitioners, enriching consumers’ lives while allowing other staff to concentrate on their respective roles.

Effective strategies for implementing peer services

The following are some of the strategies that have been found effective in introducing peer support into conventional mental health settings:

A clear job description and role clarification – fully endorsed by key stakeholders (including program administrators, supervisors, and potential coworkers) – with relevant competencies, and a clear policy for evaluating competencies and job performance.

Involving non-peer staff and organizational leaders, as well as people in recovery, along with organizational leaders early and throughout the process of creating peer positions, including in formulating job descriptions and making hiring decisions.

Identifying and valuing the unique contributions that peers can make to the programs and settings where they will work. For example, the optimal benefits of hiring peers as part of case management teams will not be realized if the peer staff are only trained and expected to carry out traditional case management tasks. In other words, it is important that peers have jobs in which they can use the skills they have acquired through their life experiences and training, rather than being assigned tasks that other staff are simply too busy to perform (such as filing or providing transportation).

Starting with at least two peer staff within any program, team, or work unit to facilitate their transition to this new role and giving them the opportunity to share job experiences and provide mutual support to each other.

Having a senior administrator take on the role of peer staff “champion” who can address issues that arise on a systemic – as opposed to individual – level and who keeps the development of peer services a priority for the agency.

Providing training for peer staff that covers the specific skills and tasks required by their roles, such as using their recovery story to the benefit of the people they work with, effective listening skills, creating positive relationships, goal identification and setting, what to do in an emergent situation, agency documentation requirements and how to fulfill them, ethics and confidentiality, boundaries, self-care, and ways of resolving conflicts in the workplace, including how to talk openly about issues of power and hierarchy within the organization.

Providing supervision for peer staff that concentrates on job skills, performance, and support rather than on the person’s clinical status, and which establishes expectations of peer staff that are equivalent to the agency’s expectations of other employees.

Providing training and education for non-peer staff that covers relevant disability and discrimination legislation and its implications for hiring and the provision of reasonable accommodations, expectations of peer staff, ethics, boundaries, adopting person-first language and a respectful attitude toward all coworkers, and ways of resolving conflicts in the workplace, including how to talk openly about issues of power and hierarchy within the organization.

Dissemination of success stories that inspire hope and persistence in all parties.

Conclusions

As suggested by the list of strategies above, implementing peer support services in mental health settings is messy and complicated work that brings about significant culture change in these institutions. Bringing about such change has been one of the driving forces behind the dissemination of peer support since the beginning, however, as we saw in the case of Pinel and moral treatment. While the need for such change may be less obvious today – in that people with severe mental illnesses are no longer in shackles and subjected to ongoing cruelty and abuse in most societies – there remains a pressing need to restore to such persons their basic human rights as full citizens of their communities. The forms of deprivation and discrimination experienced today may be different, but they require changes of the same order of magnitude as those introduced through moral treatment.

The transformation from being a service recipient to being a service provider represented by peer support is one concrete manifestation of this order of magnitude, and contributes to the needed changes in many more ways than simply by improving individual outcomes as evidenced by controlled research trials.