Stigma toward psychiatric patients remains a challenge for our field despite efforts at reduction (1, 2). In psychiatric education, stigma plays a significant role in shaping medical students’ views of individuals living with psychiatric disorders (3, 4). Stigma also impedes recruitment of talented students to join our field (5, 6).

The “contact hypothesis” proposes that experiences characterized by interpersonal contact can reduce stigma, particularly when participants with relatively equal status interact around cooperative tasks. This hypothesis is based on observations made in efforts to change prejudicial attitudes across a variety of populations, including racial and ethnic groups (7–9). Even brief interventions addressing stigma have shown relatively long-lasting impact (10). Yet efforts to reduce stigma via intergroup interaction can be complex, with unintended consequences, such as exacerbating social distancing when individuals become self-conscious about being perceived as prejudiced (11, 12). Based upon these principles, our present study describes an intervention designed to reduce preclinical medical students’ stigmatization of people with chronic mental illness (CMI).

Context

We utilized an off-site facility housed within a state psychiatric hospital–the Living Museum (LM)–that functions as an art studio and display space. The artwork is produced by patients; many, currently inpatients, some previously hospitalized but currently living in the community. The often vibrant paintings, drawings, sculptures, and photographs cover a wide range of styles, content, and emotions: from abstract to realistic, landscapes to portraits, intensely personal to social, and serious or unsettling to light-hearted.

Participants

Two consecutive classes of second-year medical students enrolled in Psychiatric Medicine, a required two-semester (Spring and then Fall) preclinical course sequence at Columbia University's College of Physicians and Surgeons, were potential participants (combined total: 314). Visiting the LM was required for all students. There were 120 participants in 2008 (76% of the class) and 102 participants in 2009 (65% of the class), for a combined response rate of 71% (222 students). We obtained Institutional Review Board (IRB) approval.

Materials

Our self-assessment instrument (available upon request) included 24 items eliciting participants’ ratings of their attitudes and feelings toward people with CMI. These questions were developed de novo, using an evaluation process model methodology (13) that generated a set of observable behaviors from existing literature and previous feedback about LM visits. Given that students’ empathy levels may influence their attitudes (14), we included an existing Empathic Concern subscale (15). We also obtained information about the LM visit experience and students’ previous exposure to people with CMI.

Procedures

The usual course-administration procedure was followed regarding assignment of the students to 1 of 24 preceptor-led groups of 13 persons. Specifically, the course director’s administrative assistant shuffled, and then randomly sorted, the students’ individual photos into the designated number of groups. These groups were then randomly assigned a date to visit; in both cohort-years, half of the groups visited during the Spring semester, and half visited the following Fall. During each student group visit, patient-artists provided tours, showed their work, and discussed creating art. Students had the opportunity to ask questions or speak with the artists if they chose to do so. All students completed the questionnaire at the beginning of the Fall semester–whether or not they had attended the LM (thus constituting the Visit or No-Visit groups)–during a psychiatry lecture that all second-year students were expected to attend. No personal identification was elicited, ensuring student anonymity.

Analysis

The final valid sample for analysis consisted of 174 students, of which 64 students had visited the LM and 110 students had not yet visited. Of those who had visited, 44 students reported having spoken individually with a patient-artist. Surveys were included in the analysis only if they were sufficiently complete and, for students who had visited, if patient-artists were present at the LM. Comparisons were planned between 1) the Visit/No-Visit groups, and 2) within the Visit group, between those students who reported having spoken individually with a patient-artist, and those who did not. Background analyses indicated that there were no group differences in students’ previous exposure to people with CMI in the students’ personal lives or in a work or volunteer setting, with one exception: those students who reported previous exposure to people with CMI in a work or volunteer setting were more likely to report speaking individually with a patient-artist during their LM visits (p=0.005). The possible implications of this group difference will be addressed in the discussion. Background analysis also found no group differences in baseline empathy.

An exploratory factor analysis of the 24 questionnaire items that we developed to survey students’ attitudes and feelings toward people with CMI yielded the following six factors, which included 18 questions and accounted for 60.5% of the variance: 1) Contentment of people with CMI (16.5%); 2) Negative Feelings toward people with CMI (12.3%); 3) Positive Feelings toward people with CMI (9.9%); 4) Discomfort in Interactions with people with CMI (8.4%); 5) Positive Personality Traits of people with CMI (7.1%); and 6) Hopelessness of people with CMI (6.3%). (Details of the factor analysis, for example, eigenvalues and principal components, are available upon request). These factors served as the basis for subsequent analyses. In particular, an analysis of covariance (ANCOVA) tested for significant differences in attitudes and feelings toward people with CMI through attendance at the LM and intensity of the experience (i.e., whether students reported having spoken individually with a patient-artist during their visits). Students’ empathy scores were included as a covariate for each analysis, to control statistically for individual differences in empathy. Finally, student comments routinely collected as part of the ongoing evaluation of the curriculum were searched for mention of the LM and content-analyzed for relevance.

We present a sample of students’ written reflections initially to illustrate the complicated impact of the LM experience. They were, in many ways, positive—for example: “This man defied my preconceived notion of a patient hospitalized for psychiatric illness; he made me realize that mental illness is only one part of the patient’s story, and should not necessarily define him or her as a person;” but, with a down-side as well: “… though we have heard delusions shared from psychiatric patients before, it has been in our space—a medical space dominated by white coats and an environment controlled by us. Here, we were solidly in Mr. A’s space. The discomfort among our group as Mr. A’s delusions accelerated was palpable.” Nevertheless, the students seemed to vary in their capacity to handle the setting and understand how it may influence their learning: “It suddenly became so real to me, the first time I had ever talked to someone who freely admitted to having a psychotic disorder. I think the trip to the Living Museum really brought me out of the comfort zone of class . . . which will certainly better my understanding of the reality of psychiatric illness.”

In our quantitative data, significant differences were found for two of the three positive-focused dependent measures in comparing students who visited the LM to students who had not yet visited (Table 1). Specifically, students who visited reported more Positive Feelings toward and more Positive Personality Traits of people with CMI. The covariate, Empathy, was also significant for these two factors. Among students who visited the LM, two additional significant findings emerged (Table 2). First, students who reported speaking individually with a patient-artist endorsed less Hopelessness in people with CMI than those who had visited yet indicated not speaking with a patient-artist. Second, students who reported speaking individually with a patient-artist endorsed less Positive Feelings toward people with CMI than those who had visited yet indicated not speaking with an artist.

This latter finding was counter to our expectations that more intense engagement with artists would add to the benefit of the intervention. Therefore we conducted three additional post-hoc analyses, with a Bonferroni-adjusted significance level of 0.016, to further investigate the relationship among the three positive factors. Using a repeated-measures ANCOVA, with Positive Feelings and Positive Personality Traits as the dependent measures and Speaking Individually as the independent measure, the interaction was significant (F=8.5; p=0.005). Interestingly, the group of students who indicated not speaking individually with an artist indicated more positive feelings toward people with CMI. No other significant interactions were observed in the remaining comparisons. The covariate, Empathy, was not significant in any of these analyses.

Finally, for exploratory purposes, we compared Empathy scores with each of our six dependent measures. As coded by a simple median split and group-mean analysis of variance, high-Empathy students had more Positive Feelings (p=0.002) toward and perceived more Positive Personality Traits (p=0.013) in people with CMI than did low-Empathy students. These findings lend support to our original decision to include the Empathy variable as a covariate measure to isolate as much as possible the effect of visiting the LM.

Preclinical medical students’ visits to the Living Museum (LM) had a positive impact on their self-reported attitudes toward people with chronic mental illness (CMI), as compared with students who had not yet visited the facility, even when controlling for empathy level. This finding is consistent with an extensive “stigma” literature, and it confirms our initial hypothesis that exposure to patients in a setting focusing on their functioning as individuals (i.e., artists) rather than their psychiatric symptoms, would contribute to medical students’ less-stigmatized view of people with CMI. This study is the first to extend the stigma literature and the implications of the contact hypothesis into the area of medical student education in psychiatry with a systematically planned and assessed intervention in which students were assigned to a patient Contact or No-Contact condition (16).

In light of our findings, however, it is also the case that those students who reported themselves as having spoken individually with artists indicated more hopefulness about—but also less positive feelings toward—people with CMI, when compared with those students who had visited the LM yet reported not speaking individually with an artist. Follow-up analyses corroborated our finding that students who visited and did not speak individually with an artist reported a higher level of agreement with statements denoting positive feelings toward people with CMI than all other comparisons within the interaction. It is the Positive Feelings Toward factor that appears to account for these differences; other comparisons were not significant.

This finding is contrary to our initial hypothesis, in that we expected that the degree to which students’ feelings and attitudes toward people with CMI improved would be related to the intensity of interaction, with more interaction and individual contact predicting more improvement. On the other hand, this finding is consistent with previous work indicating that interactions with stigmatized groups—specifically people with CMI—can be unappealing (17–22) as well as anxiety-producing, and even stressful (14, 23–25). Furthermore, the contact-hypothesis literature suggests that the nature of the contact is critical for the desired result, and intergroup contact may either change or confirm existing stereotypes, depending on the nature of the experience (11, 26). As one student’s comments indicate, leaving the “comfort zone” of the medical setting is not easy. For example, recognizing the shared humanity in disorganized, delusional patients can make us all aware of our own vulnerabilities and can bring to the surface the fear that we ourselves may be at risk of disorganization and dysfunction. Stigmatization is a ubiquitous and entrenched phenomenon: it should not be surprising that diminishing the stigmatization of people with CMI would defy simple solutions.

The medical education literature has tended to focus on empathy as a target of change, yet the factors contributing to its change in medical students and health professionals remain a matter of debate (27, 28). In the present study, students’ Empathy scores (when included as a covariate) accounted for much of the variance in our significant findings. From the outset, we did not expect empathy to change as a result of this one brief intervention. Therefore, different from other studies in medical education in which interest lies in increasing empathy (or at least preventing it from diminishing), we controlled for it statistically to focus on the effect of the intervention itself.

This study illustrates that nontraditional settings can be a valuable adjunct to more-traditional clinical settings to reduce stigma when introducing medical students to psychiatry. Our unexpected finding that students who reported themselves as speaking individually with the artists indicated less-positive feelings toward people with CMI suggests that interventions involving students’ interactions with people with CMI can have complex and sometimes unintended consequences. Thus, educators should proceed cautiously and monitor the nature of the contact carefully during any such interventions.

Our ability to generalize from our results to our entire medical school class is limited by the response rate of slightly less than three-quarters of the students invited to participate. This study is also limited by the nature of the cross-sectional design and self-report questionnaires. In particular, our design prevents confirmation of any causal inferences, since we are unable to compare the same students’ attitudes and feelings before and after the intervention, and are not able to comment on the long-term stability of any changes in attitudes. Also, the questionnaire item that asked whether a student “spoke individually” with a patient-artist was not further defined and was therefore subject to varying interpretation. Students may have identified an interaction with an artist as “speaking individually” in any number of ways, including one-on-one questioning, brief comments, and/or conversing within a group discussion. Thus, our data only indicate that students perceived themselves as having spoken individually with patient-artists. Perhaps the response to the unspecific question indirectly sorted the students according to some particular characteristic of the students, rather than a genuine difference in their levels of interaction while at the LM. On the other hand, our finding that students with previous exposure to people with CMI were overrepresented among the students who reported speaking individually with artists suggests that students with previous exposure may have been more likely to seek out such one-on-one interactions (16). In fact, their tendency to seek out such interactions may have preceded—and even led to—their previous exposure as well, particularly since the association with previous exposure was only observed in volunteer and work settings, and was not observed in students’ personal lives.

The contents of this paper have been presented in part at the Annual Meeting of the American Psychiatric Association (APA) in San Francisco, CA, May 16–21, 2009, and the Annual Meeting of the Association of Directors of Medical Student Education in Psychiatry (ADMSEP) in Portsmouth, NH, June 18–20, 2009.

The authors thank the artists and staff at Creedmoor Psychiatric Center’s Living Museum and acknowledge the assistance of Dr. Janos Marton, the Director of the Living Museum.

This work was supported by Innovative Teaching Grants from the American Psychiatric Association in 2008 and 2009.