Peer-based treatments are gaining popularity in traditional service venues for persons with severe mental illness (1). Several authorities have suggested that the success of these services may be largely attributable to peer providers' ability to readily forge favorable relationships with clients (2,3). Recent evidence supports these suggestions. In a previous study comparing clients of peer providers and of traditional providers, the former reported that they perceived higher levels of validating qualities such as empathy; moreover, clients who perceived that their providers had such qualities were more likely to report motivation for and engagement in community-based treatments (4). Such findings are consistent with those of traditional psychotherapy research, in which a therapist's warmth, understanding, and acceptance are shown to foster strong working alliances and positive therapeutic outcomes (5).
Interpersonal transactions between health care providers and clients are complex, however, typically reaching beyond the provision of validation. Nevertheless, it seems a strongly held belief that the positive impacts of peer-based treatments are attributable to clients' experiences of generic validation through such qualities as provider warmth and understanding. However, this conceptualization is incomplete and conflicts with documented and successful practices in other peer-based services, such as therapeutic communities, where potentially invalidating confrontation may be enacted within the spirit of therapeutic benefit (6,7). For instance, a peer provider may choose to express dissatisfaction with a client based on the client's recent drug use or may refuse to understand or may cut short a client's self-denigrating statements. Although such practices may not typify peer-based treatment, the prevalence and impact of clients' experiences of invalidation from providers are underresearched and, to the best of our knowledge, have yet to be examined in the context of peer-based services within traditional service venues for persons with severe mental illness.
This article reports results of a randomized controlled trial that compared the occurrence and outcomes of clients' experiences of invalidation in peer-based and traditional intensive case management services for persons with severe mental illness, the majority of whom had co-occurring drug use problems. We hypothesized that compared with clients in traditional intensive case management, those who received services from peer providers would perceive that provider communications were more validating than invalidating as reflected by levels of positive regard, empathy, and unconditional acceptance. We also hypothesized that perceived invalidation from peer providers but not from traditional providers would be significantly associated with favorable client outcomes, including enhanced quality of life and fewer perceived obstacles to recovery.
This project ran from July 2001 to June 2003. The Human Investigations Committee at the Yale School of Medicine approved the investigative protocol before data collection began. Participants included 84 men and 53 women aged 20 to 63, with a mean±SD age of 41±9. Participants described their racial background as follows: 89 (65%) Caucasian, 39 (28%) African American, and nine (7%) "other." Nine participants (7%) self-identified as Hispanic. Diagnostic data were unavailable for five participants (4%); all others had a diagnosis of severe mental illness; 70% had a co-occurring substance use disorder. As shown in Table 1, the distribution of diagnoses was similar across the peer-based and traditional treatment group.
We randomly assigned 69 participants to a control condition in which they received traditional intensive case management services but not from staff serving as peer providers. Sixty-eight participants were randomly assigned to an experimental condition in which they received one year of services from intensive case management teams that included peer providers as primary contacts.
The Barrett-Lennard Relationship Inventory (BLRI) (8) is a 64-item client self-report questionnaire designed to gauge dimensions of the client-provider relationship relevant to favorable therapeutic change. Respondents rate agreement with items on a 6-point scale, ranging from 1, definitely false, to 6, definitely true. For the purposes of this investigation, we included items pertaining to six subscales, specifically across the positive-negative dimensions of client-perceived regard, empathy, and unconditional acceptance. An example of a subscale item that reflects empathy is "She [the provider] usually senses or realizes what I am feeling." An example of an item that reflects lack of empathy is "She looks at what I do from her own point of view." The BLRI has been successfully employed in several investigations—particularly of traditional psychotherapy—and has been shown to be a valid and reliable measure of therapeutic alliance (9). As in previous research (4), after pilot administrations we made slight modifications to ten BLRI items to render them more comprehensible to the target population. Two items were deleted because they were shown in the pilot testing to be difficult for participants to understand.
The Addiction Severity Index (ASI) (10) is a structured interview that measures obstacles to effective treatment and recovery in areas notably affected by alcohol and drug use, including alcohol or drug use behavior, medical concerns, and psychiatric problems. Composite subscales gauge severity of difficulties over the past 30 days, with higher scores indicating more difficulties. The ASI has been rigorously assessed in numerous investigations, and its subscales have been proven valid and reliable for measuring alcohol and drug use (11) as well as psychiatric and medical problems (12).
The Quality of Life Inventory-Brief Version (QOLI-B) is an interview assessment based on Lehman's (13) instrument. It is designed to gauge respondents' quality of life across subscale domains, including general satisfaction, living situation, work or school functioning, familial relationships, and health. QOLI-B question formats vary and include use of a respondent's brief description, frequency ratings (for example, "In the past six months, how often did you get together with a member of your family?"), and satisfaction ratings (for example, "How do you feel about your physical condition?"). Higher scores reflect better life quality. Several empirical investigations attest to the validity and reliability of the QOLI-B in populations with severe mental illnesses (14,15).
Criteria for participation included a diagnosis of severe mental illness (that is, a major mood or psychotic disorder) and a clinician rating on the Level of Care Utilization System (16) that reflected disengagement with treatment. Potential participants were identified by mental health centers. Investigators invited those who were interested to take part in an interview; during the interview they were given a complete description of the project, including the potential benefits and risks of participating. If they were still interested, participants read and signed an informed consent form. Investigators randomly assigned participants to either a control condition (traditional treatment) or an experimental condition (peer-based treatment).
Peer staff were individuals who had chosen to publicly disclose their history of mental illness (and for some, co-occurring substance use problems) and subsequent recovery, with the intention of using these experiences in concert with their clinical talents and skills to assist clients who were currently dealing with active psychiatric problems—and often co-occurring substance use problems. All peer staff participated in broad-based didactic, experiential, and practical training in applying personal experiences to working effectively with clients. Further details about the training have been published elsewhere (4).
After training, peer staff worked as providers within the Peer Engagement Specialist Project, a statewide initiative spanning four Connecticut towns and based at agencies that provide intensive case management services—three public health centers and one nonprofit organization. At all sites, peer providers carried an average caseload of ten to 12 clients and received guidance from clinical supervisors. Traditional providers who participated in the project worked in tandem with peer providers on treatment teams and typically carried twice the peer provider caseload.
Statistical analyses. We employed mixed analysis of variance (ANOVA) to test the between-group hypothesis that clients in peer-based treatment would perceive that their providers' communications were more validating than invalidating compared with clients in traditional treatment, as reflected by client-perceived positive regard, empathy, and unconditional acceptance. This analyses progressed according to a 2×2×3 design, with two levels of the between-group variable of condition (peer and traditional provider), two levels of the within-group variable of relationship quality level (negative and positive), and three levels of the within-group variable of relationship element (regard, empathy, and unconditionality).
To examine hypotheses regarding the link between invalidating communications and favorable outcomes, we followed closely the recommendations of Baron and Kenny (17), adopting a multiple linear regression approach to determine whether the predictive effect of client-perceived invalidation by the provider was moderated by provider type (peer or traditional) upon the ASI and QOLI-B subscales. We included an interaction term reflecting our hypothesis that provider type moderates the association between perceptions of invalidation and favorable outcomes. Each moderation analysis included the following steps: assessing demographic variables, including race, gender, and age as potential covariates; developing final regression equations with a full set of outcome predictors, including the outcome variable at baseline, intervention conditions, invalidation, invalidation-by-intervention condition, and baseline scores by intervention; assessing whether interactions between baseline scores and intervention terms were significant, and if not, then removing them from the model; and assessing whether final regression equation results yielded significant interactions for the invalidation-by-intervention terms, analyzing simple effects per intervention condition. The squared semipartial correlation coefficient (sr2) was used as an effect size measure, reflecting the amount of variance in outcomes explained by predictors after the analysis controlled for baseline levels of outcome variables.
Results of mixed ANOVA revealed a robust main effect for relationship quality (F=96.54, df=1 and 102, p<.05; η2=.49); clients in both peer-based and traditional treatment perceived that interactions with and communications from their providers were significantly more validating than invalidating. More pertinent to the hypotheses, results also showed a significant valence-by-condition interaction (F=9.54, df=1 and 102, p<.05; η2=.09), in which participants in peer-based treatment perceived communications as significantly more validating than invalidating compared with participants in the control condition. There were no main effects for condition.
Table 2 presents six-month means and standard deviations for BLRI scores, which indicate relationship quality, by condition, valence (positive and negative), and relationship element (regard, empathy, and unconditionality). At the six-month follow-up, 23% of the overall sample did not complete the BLRI questionnaire, either because of study drop-out or incomplete interviews, which is not uncommon in studies that include persons with co-occurring psychiatric and substance use problems (18). Chi square analyses confirmed that the difference in attrition between groups was not significant.
Results revealed significant moderator effects of invalidation-by-study condition across ASI and QOLI-B subscales at the six-month follow-up but not at the 12-month follow-up. Specifically, negative regard and a composite negative overall index (mean score for negative empathy, negative regard, and negative unconditionality) were found to have significant predictive effects on outcomes. Table 3 presents statistical data for all significant moderating analyses, and Table 4 shows results of follow-up simple-effects tests.
As shown in Table 3, after controlling for ASI medical composite baseline scores, regression analyses revealed that among those in peer-based treatment, an increase in the negative overall index at six-month follow-up predicted a decrease in the ASI medical composite score at six-month follow-up significantly more so than for those in traditional treatment. Simple-effects findings, which are shown in Table 4, specified that when the analyses controlled for baseline levels, the negative overall index at six months significantly predicted a decrease in the ASI medical composite score at six months (indicating improvement) among participants in peer-based treatment but not in the control condition.
When the analyses controlled for baseline scores on the ASI psychiatric composite, an increase in the negative overall index at the six-month follow-up predicted a decrease in ASI psychiatric composite score at six months significantly more so than for those in traditional treatment. Simple-effects analyses showed that the negative overall index significantly predicted a decrease in the ASI psychiatric composite score for those in peer-based treatment but not for those in traditional treatment. No other ASI composite subscale (for example, drug use) showed a significant moderation pattern.
After the analyses controlled for baseline scores on the QOLI-B health subscale, an increase in the negative overall index at the six-month follow-up among those in peer-based treatment predicted an increase in the QOLI-B health subscale at six months significantly more so than for those in traditional treatment. Simple-effects analyses showed that the negative overall index significantly predicted increases in the QOLI-B health score for those in peer-based treatment but not for those in traditional treatment.
When the analyses controlled for baseline scores on the QOLI-B family subscale at six-month follow-up, an increase in perceived negative regard among those in peer-based treatment at six months predicted an increase in the QOLI-B family score significantly more so than for those in traditional treatment. Treatment condition also moderated the relation between baseline scores on the QOLI-B family subscale and the same subscale scores at six months. Simple-effects analyses showed that negative regard significantly predicted increases in QOLI-B family scores for those in peer-based treatment but not for those in traditional treatment. No other QOLI subscale scores showed a significant moderation pattern.
No other significant results were found at 12 months when we used the same analyses.
Results of this study support the hypothesis that participants who received peer-based intensive case management services would perceive that their providers communicated in ways that were more validating than invalidating compared with participants who received services from traditional providers. In addition, results support our hypothesis that provider type (peer or traditional) would moderate the association between the perception of invalidation from the provider and favorable client outcomes.
These findings are consistent with the notion that peer providers may serve a validating role for clients with severe mental illness. However, they also suggest that early in the course of treatment, peer providers' invalidating communications predict greater client benefits than the invalidating communications of traditional providers in terms of fewer psychiatric and physical health problems, better perceived medical health, and better relationships with family members. It should be noted, by contrast, that traditional providers' invalidating communications did not appear to lead to unfavorable client outcomes across these analyses. Moreover, the differential effects we found for provider type held true only early in the treatment process, at six months but not at 12 months. Thus it may be that the relative positive effects of invalidation in peer-based services operate chiefly within the early, engagement phase of treatment.
These findings speak to the potential benefits of employing peer providers in traditional service venues serving persons with severe mental illnesses. The role of these providers, although typically conceived within a constellation of validating qualities such as ally and advocate (19), appears to possess favorable transformative power through invalidating channels as well. Although the precise nature of these channels is unclear, it may share a conceptual foundation with what White (19) described as the "truth-teller" role of peer-based treatment relationships, through which a peer provider may offer measured invalidating communication to a client about her or his potentially deleterious behaviors, beliefs, and so on. These communications may not differ substantially from those received from traditional providers, but within the context of the peer provider's disclosed experiential background and greater client-perceived validation, clients may more readily appreciate invalidating communications and use them as an impetus for improving their lives. It is up to future research to specify how and when such invalidating "truths" are offered to clients in peer-based and traditional treatment and, furthermore, to empirically link those instances to clients' experiences of invalidation by providers and client outcomes.
This research had several limitations. First, although the study examined differential impacts of provider type on client process and outcome, the work did not specifically differentiate treatment practices among traditional and peer providers as has been done by others in this area (20,21). As a result, the study was not able to distinguish specific validation practice, and thus it remains unclear to what extent validation was differentially offered by peer and traditional providers to address personal, behavioral, or other considerations. Furthermore, it is possible that the programmatic context of this investigation, in which peer and traditional providers worked together on the same teams, may have compromised internal validity through contamination of distinct traditional and peer-based treatment practices. Moreover, the intensity of client services in the peer-provider condition, in which peer and traditional providers often work side by side, may have accounted for observed differences between conditions. Nevertheless, in traditional service agencies that employ peer providers, it is not uncommon to have them work with traditional staff (22); although these practices may pose challenges to investigative internal validity, the service context of this investigation does represent the actual circumstances under which such treatments are often delivered and thus serves external validity.
Second, it is possible that the results obtained in this investigation could be attributed to peer providers' smaller caseloads, insofar as traditional case managers may have had neither adequate time nor energy to work with clients with the same level of demonstrated investment as peers. It is important for future research to better control for client caseload and also to determine suitable caseload levels for optimal service provision for both traditional and peer-based intensive case management services.
Third, data in this study were based primarily on client self-report. Provider judgments on measures such as the BLRI have been shown by investigators to predict client outcomes (23); clinician ratings have also been shown to be linked to aspects of recovery such as hope (24). Although client self-ratings appear to be good predictors of outcome in psychotherapy investigations, future research should determine the best predictors of outcome in peer-based treatment. In addition, it is possible that changes made to the BLRI for the purposes of this study compromised its validity.
Fourth, study participants represented a convenience sample whose reports may have diverged from those who declined participation; although the refusal rate was not tracked systematically, staff estimations made after the study was completed suggest that it was below 5%. In addition, the study sample was small, and missing data with respect to the BLRI at six months was 23%, which may limit statistical validity and study generalizability. The statistical tests indicated robust differences, and therefore statistical validity was not of primary concern. Future research on this topic, however, would do well to include larger samples to maximize generalizability. This is particularly important in studies of persons with co-occurring psychiatric and substance use disorders, who are known to show high levels of attrition in community-based services research. Nevertheless, chi square analyses showed that attrition did not differ between peer-based and traditional treatments.
Findings from this investigation of clients with severe mental illness in intensive case management suggest that early in the course of treatment, clients with peer providers perceive that their providers' communications are more validating compared with clients with traditional providers. Moreover, peer providers' invalidating communications appeared to be linked to clients' subsequent improvements in social relationships and health, whereas no corresponding associations were found for clients of traditional providers. It may be that peer providers' greater tendency toward client validation renders them more credible and effective invalidating truth-tellers about clients' need for change and improvement. Future research should examine links between therapeutic outcomes and specific instances of validating and invalidating dyadic processes in traditional and peer-based treatment practices and clients' perceptions of validation and invalidation in the treatment relationship.
This research was supported by funding from the Yale Institution for Social and Policy Studies. The peer-based treatment option was sponsored by the Connecticut Department of Mental Health and Addiction Services.
The authors report no competing interests.