Anti-stigma films and medical students' attitudes towards mental illness and psychiatry: randomised controlled trial

AIMS AND METHOD

To explore the feasibility of a randomised controlled trialof the effects of two anti-stigma films on medical students’attitudes to serious mental illness and psychiatry. Attitudesto serious mental illness, perceived dangerousness, socialdistance and psychiatry, were measured before and after watchingthe films and at 8 weeks.

RESULTS

Intervention films significantly improved general attitudesto serious mental illness and social distance, with a trendtowards reducing perceived dangerousness. These effects appearedto attenuate during the students’ clinical placements,suggesting a possible interaction with their clinical experiences.

CLINICAL IMPLICATIONS

Our results suggest both that it may be possible to conducta substantive trial of the effects of the intervention filmson a larger cohort of medical students and that the films maybe effective in reducing stigmatising attitudes in medicalstudents.

Introduction

TOP

Introduction

The stigma and discrimination associated with mental illnesscontinue to blight the lives of those so diagnosed (Corrigan & Penn, 1999; Hayward, 1997; Crisp et al, 2000), despite extensive attemptsto counter their effects (Estroff et al, 2004; Sartorius & Schulze, 2005).Given that medical professionals are inevitably embedded in the fabric of society it is perhaps unsurprising that theycan also hold stigmatising attitudes (Üçok et al, 2004)and thus contribute to discrimination (Byrne, 1999; Crisp et al, 2000). Indeed, surveys of mental health service users have revealeda relatively high prevalence of stigma and discrimination fromhealthcare professionals (Wahl, 1999). Although several studieshave found that medical students and doctors often regard psychiatric patients as difficult and unrewarding to treat (Nielsen & Eaton, 1981; Lawrie et al, 1998), other research has reported that medicalstudents’ attitudes towards mental illness and psychiatrybecome more positive following undergraduate training (Creed & Goldberg, 1987;Singh et al, 1998), particularly where that training involvesboth patient contact and education about the effects of stigma (Corrigan & Penn, 1999). Given the labour-intensive natureof existing anti-stigma interventions (Pinfold et al, 2003) and the power of audiovisual media to influence societal constructionsof mental illness (Wahl, 1995), researchers have postulatedthat documentary films depicting people diagnosed with mentalhealth problems may offer a more efficient approach to reducing stigma and discrimination among student groups (Penn et al, 2003). Indeed, existing research suggests that anti-stigma films cangarner small and temporary improvements in specific areas suchas social distance (Chung, 2005; Altindag et al, 2006), thoughwith little change in general attitudes (Penn et al, 2003). This latter finding was attributed to the lack of an unambiguous disconfirmation of the mental illness stereotype within thefilm deployed (Penn et al, 2003). We sought to determine theeffects of an anti-stigma intervention based on films producedby mental health service users and combining both educationand stereotype disconfirmation elements, on medical students’attitudes to both serious mental illness and psychiatry.

Method

TOP

Method

Study design
The study was conceived as a pilot project to refine the research methodology in anticipation of a larger scale trial. The twoanti-stigma films used were made in partnership with non-statutorymental health organisations in Nottingham. Participants were4th year medical undergraduates on their psychiatry trainingattachment.

We used a single-masked, randomised controlled trial designto compare a group who watched the intervention films witha group who watched a control film of the same format and length.We tested whether general attitudes to serious mental illnesschanged in the intervention group, whether the films changedspecific attitudes concerning social distance, perceived dangerousness,and psychiatry. After baseline assessment, participants were randomised to either intervention or control groups and werethen reassessed immediately after watching the films, and againat 8 weeks post-intervention. Ethical approval for the studywas granted by the University of Nottingham Medical Schoolresearch ethics committee. Written informed consent was not required as the project was an assessment of an educationalintervention, although students were assured that their responseswould have no influence on their grades or exam scores.

Intervention and control films
The first film, A Human Experience (Smith, 2005), was madein collaboration with service users at Rethink Nottingham.It is 15 min long and adopts a ‘talking head’ documentarystyle approach. Its content revolves around three mental healthprofessionals (a teacher/researcher, a Mental Health Act Commissionerand a psychiatrist) discussing their experiences after beingdiagnosed with a serious mental illness (psychosis, schizophreniaand severe depression - all of which resulted in hospitalisation)and, in particular, their experiences of stigma and discrimination.The film challenges particular stereotyped beliefs, including dangerousness, inability to work and inability to maintainrelationships, and mentions positive aspects of serious mentalillness (such as the importance of the experience of mentaldistress in the forging of personal identity, a sense of overcomingadversity, the celebration of difference, and a formulationof mental distress as a deepening of lived experience). Thesecond film, A Day in the Mind of... (Green, 2005), was madeby service users at Framework Housing Association Nottingham(a non-statutory organisation providing practical and emotional support for people experiencing mental distress and livingin the community). It is 12 min long and adopts a first-personperspective throughout. Its narrative focuses on the subjectiveexperience of psychosis over the course of a typical day. Thefilm attempts to convey to the viewer the first-hand experienceof being diagnosed with a serious mental illness, thereby challengingthe stereotype of psychosis as a condition opaque to understanding.The control film was a 25-min documentary unrelated to mental illness or psychiatry and matched for visual format.

Outcome measures

General attitudes to serious mental illness, as measured bythe Attitudes Toward Serious Mental Illness Scale - AdolescentVersion (Watson et al, 2005), a 21-item validated measureof general attitudes in young people where higher scores indicatemore negative attitudes.
Perceived dangerousness, measuredby the Dangerousness Scale (Link & Cullen, 1986), a 5-itemquestionnaire with goodinternal consistency where ratings aremade on a 5-point Likertscale, with higher scores indicatinggreater perceived dangerousness.
Social distance, measuredby the Social Distance Scale (Penn et al, 1994),a 6-itemquestionnaire with good internal consistency wherehigher scoresare indicative of a tendency to maintain a greatersocial distancefrom people diagnosed with a serious mentalillness.
Attitudesto psychiatry, measured by the Attitudes to PsychiatryScale (Burra et al, 1982), a well-validated 30-item questionnairewhere higher scores indicate a more positive attitude towardspsychiatry.

In addition to the main outcomes we also collected data on students’ previous contact with people diagnosed with a mental illness,affect at the time of assessment, and behavioural intentionstowards such people.

Procedure
At baseline, on the first day of their psychiatric attachment,participants self-completed all four outcome measures. Afterbaseline assessment, they were randomly allocated using a concealedrandomisation method. Those randomised to the interventiongroup watched the two films on the second day of the attachmentand the control group watched the control film. The outcome measures were repeated immediately post-intervention and at8 weeks follow-up, at the end of the psychiatry attachment.Statistical analyses were undertaken by an independent researchermasked to allocation status and all participants were askednot to reveal their group allocation. They were analysed onan intention-to-treat basis, with Mann-Whitney U-tests employedto compare non-parametric data and independent samples t-testsused for parametric data. All results were corrected usingthe Bonferroni method to adjust for multiple comparisons. Therewas evidence that at baseline and post-intervention time pointsthere had been confusion over the polarity of the rating scalefor the Attitudes to Psychiatry questionnaire and scores were discarded for these time points. The rating scale was clarifiedin the final administration of the Attitudes to Psychiatryquestionnaire and the 8 weeks follow-up scores were analysed.

Results

TOP

Results

Although 82 medical students were eligible to participate inthe trial, only 46 took part (56%) and were randomly allocated,23 to each of the study arms (Fig. 1). Almost three-quartersof the sample were female (n=34; 74%) and the mean age was21 years. A great majority (n=37; 80%) were White European;28 (60%) had previous contact with a person diagnosed witha serious mental illness. There were no significant differencesbetween the groups at any of the three assessment time points(Table 1). However, inspection of the data suggested some changeswithin groups and therefore for each group the scores for totalattitudes to mental illness, dangerousness and social distancewere analysed to determine whether they changed significantlyover time, using a Friedman test. Post hoc Wilcoxon matched-pairstest was then used to determine between which time points thesignificant differences were located. Figure 2 shows that,with regard to total score for attitudes to mental illness,there was a significant change in score in the interventiongroup over the three time points (n=23, d.f.=2, P=0.026), withscores demonstrating a significant decrease from baseline topost-intervention (z=–2.614, P=0.009) suggesting thatstudents’ attitudes were less stigmatising after theintervention. There was a trend towards significance for scoresto increase from post-intervention to 8-week follow-up (z=–1.916,P=0.055), suggesting that this anti-stigma effect attenuatedover time. Figure 3 demonstrates that there were significantchanges in perceived dangerousness scores over the three timepoints in the intervention group (n=23, d.f.=2, P=0.062), withscores decreasing significantly from baseline to post-intervention(z=–2.782, P=0.005). There was no significant differencein scores from the post-intervention to 8-week follow-up timepoints, suggesting there was less attenuation than for general attitudes. Figure 4 shows that there was a significant changein social distance scores in the intervention group over thethree time points (n=23, d.f.=2, P<0.0001), with a highlysignificant decrease in social distance from baseline to post-intervention(z=–3.546, P<0.0001). This suggests that the students’inclination to be socially distant from people diagnosed witha serious mental illness lessened after watching the films,though this was not sustained at follow-up, with scores significantlyincreasing (z=–2.169, P=0.03). No significant differencesin the scores for either total attitudes to mental illness,dangerousness, or social distance were found between each timepoint for the control group.

View larger version (17K):
[in this window]
[in a new window]
[as a PowerPoint slide]
Fig. 1. Study profile.

View this table:
[in this window]
[in a new window]

Table 1. Attitudinal and secondary outcome scores

View larger version (15K):
[in this window]
[in a new window]
[as a PowerPoint slide]
Fig. 2. Within-group changes over time in total attitudes to serious mental illness scores for both intervention and control groups. ${blacksquare}$ Total attitudes toward serious mental illness score – baseline; Total attitudes toward serious mental illness score – post-intervention; Total attitudes toward serious mental illness score – 8-week follow-up.

View larger version (14K):
[in this window]
[in a new window]
[as a PowerPoint slide]
Fig. 3. Within-group changes over time in perceived dangerousness scores for both intervention and control groups. ${blacksquare}$ Dangerousness score – baseline; Dangerousness score – post-intervention; Dangerousness score – 8-week follow-up.

View larger version (15K):
[in this window]
[in a new window]
[as a PowerPoint slide]
Fig. 4. Within-group changes over time in social distance scores for both intervention and control groups. ${blacksquare}$ Social distance scale score – baseline; Social distance scale score – post-intervention; Social distance scale score – 8-week follow-up.

Discussion

TOP

Discussion

On the whole, the study was successful, with several methodological problems being highlighted; not least of these being the factthat only 56% of potential participants eventually agreed totake part, thereby making the study vulnerable to type II errorand introducing a selection bias. The effects of this lackof power are most tellingly revealed by the absence of anysignificant between-group differences (with any specific differencesbeing overwhelmed by the much greater variance of individualparticipants’ characteristics in between group comparisons).However, this attrition rate compares favourably with thoseof some recent reports of randomised controlled trials in themainstream psychiatric literature (Wykes et al, 2007). Strengthsof the study were the use of self-report questionnaires anddata analysis masked to group membership which both servedto minimise possible assessment bias.

Comparisons with existing normative data (derived in the mainfrom US studies) suggest that, at baseline, participants inthe present study both held more stigmatising attitudes towardsserious mental illness and were more inclined to maintain socialdistance from people diagnosed with such illness (Watson et al, 2005; Penn et al, 1994) than the general population, but they alsoappeared to view them as less dangerous (Link & Cullen, 1986).Although the possibility of confounding caused by differencesin sociocultural context must be considered when interpreting research results, it is unlikely that cultural differenceswould have wholly accounted for these discrepancies. The participantsmaintained attitudes to psychiatry consonant with previousevaluations of medical students in Nottingham (Singh et al, 1998).The former findings appear to be supported by research examiningthe extent to which mental health professionals stigmatise their patients (Lauber et al, 2006). Interestingly, the latter findingwould appear to be at odds with the same work which found thatmental health professionals are as likely as the general populationto stereotype people diagnosed with mental illness as dangerous.This dissonance may reflect a selection bias, but the factthat this apparently ‘untypical’ group exists maybe important, given that the intervention films were able tofurther reduce perceived dangerousness in a group with relativelybenign extant attitudes. It was considered unfeasible to conductsubgroup analyses (such as by gender) because of the study’slack of power. Any subsequent research can remedy this andcan also seek to examine the impact of the films on participantswith more negative attitudes towards dangerousness.

The within-group results, though highly provisional, are encouraging, suggesting that the intervention films may improve medicalstudents’ attitudes to serious mental illness and decreaseperceived dangerousness and social distance. However, our resultssuggest that the effects on general attitudes and social distancewere attenuated over the 8-week attachment in psychiatry. Althoughno firm conclusions can be drawn from our data, this may havebeen due to the effects of the attachment, and any subsequent investigation may include both quantitative and qualitativeattempts to understand how the films and experience of medicaleducation within a psychiatric service interact. Further researchis needed to investigate strategies to sustain these short-termimprovements in students’ attitudes to people with seriousmental illness, perhaps with an emphasis on the importanceof the patient experience, as has been argued elsewhere (Yang et al, 2007). Such strategies might also include ‘booster’ filmsdesigned to mitigate the corrosive effects of time and experience.A further substantive trial is currently being planned, withthe intention of recruiting sufficient numbers of medical studentsto definitively answer some of the tantalising questions openedup by this pilot work.

Acknowledgments

We thank Mark Smith and Graeme Green for allowing us to employtheir films in the trial, and all the mental health serviceusers and professionals who appeared in the films.

References

TOP