The morbidity for standardized patients (SPs) associated with participation in medical education activities has been identified as an area in need of research.1 Hanson et al2 hypothesized that suicide contagion could be associated with simulation of suicidality by adolescent standardized patients (ASPs) with a history of suicidality. Suicide contagion refers to the link between adolescents’ exposure to a suicide stimulus and subsequent rise in suicide attempts. It most likely occurs with vulnerable adolescents within a two-week period after suicidality stimulus.3 It is important to evaluate the presence of evidence for this theoretical risk because ASPs regularly participate in simulations of depression and/or suicidality.4,5
Our group has previously investigated ASP effects in high-stress psychosocial conditions.2,6 Although negative effects were not prevalent, it is important to look beyond explicit effects that adolescents verbally express because (a) they may be hesitant to provide open responses about emotionally evocative issues, and (b) they may not be aware of all impacts of simulating stressful situations, as research suggests that humans maintain an adaptive unconscious, a mental system that guides our emotions and behavior yet is unavailable to introspection.7,8
This paper evaluates whether ASP simulation of depression/suicidal ideation is associated with increased risk of psychological harm, using both explicit (i.e., self-report) and implicit (i.e., behavioral) measures. To do so, we (a) used case-specific self-report measures and (b) adapted behavioral measures thought to be related to depression, including success with word puzzles and walking speed after participation in SP training.8–10 Unlike preceding work,6 ASPs identified as vulnerable continued with their training, but with oversight by mental health professionals.
The SP program recruited adolescents aged 14 to 17. Adolescents and parents attended SP recruitment evenings to introduce the study. Consenting procedures involved describing the study as an examination of the emotional impact of acting as a standardized patient. Parental consent and adolescent assent was obtained for adolescents under age 16. Adolescent consent was obtained for those over age 16. Recruitment and training activities were completed by the SP trainer of the Clinical Learning Centre, McMaster University. The study received ethics approval from the Hamilton Health Sciences Research Ethics Board.
Adolescents completed a screening procedure the evening of recruitment. The screen comprised two self-report questionnaires regarding the outcomes of interest: suicidal ideation and depression. The Suicidal Ideation Questionnaire (SIQ and SIQ-Jr) is a paper-and-pencil self-report scale. It employs a seven-point response format with which respondents indicate the frequency of various thoughts during the past month. The SIQ has demonstrated reliability and validity as a measure of adolescent suicidal ideation.11 Critical items and endorsement patterns were considered as suggested in the professional manual, and the liberal cutoff score of 30 or above (for the SIQ, 23 for the SIQ-Jr) was used to identify participants in need of follow-up interviews; 10% to 17% of participants were expected to be in this range according to normative data.11 We also employed the Reynolds Adolescent Depression Scale-2 (RADS-2), a 30-item self-report scale that evaluates depressive symptoms. RADS-2 is a reliable measure with demonstrated internal consistency, test–retest reliability, and construct validity.12 Critical items and endorsement patterns were considered as suggested in the professional manual, and the liberal cutoff score of 60 or above was used to identify participants in need of follow-up interviews; 12% to 16% of participants were expected to be in this range according to normative data.12 Both instruments were administered in a group-based format. Adolescents scoring in the clinical range on either instrument participated in an interview with an adolescent mental health specialist, but they remained eligible for study participation.
ASPs were randomly assigned across conditions (depression/suicidality or cough), with stratification to ensure that ASPs with clinical scores were spread across condition. Training took place in two phases of four hours’ total duration, with four to six ASPs per group. During the first training session, ASPs were presented case-specific educational materials, and they discussed the condition with the SP trainer and adolescent mental health specialist. Those in the depression/suicidal ideation condition received specific instruction in stress-relief techniques. The second training session was approximately two weeks later, and ASPs practiced their role with feedback from the SP trainer and adolescent mental health specialist. Implicit behavioral measures were administered at the conclusion of each training session.
Implicit behavioral measures
After the first SP training session, ASPs were told that the SP trainer had to leave the room to determine whether the research assistant was ready to run the participants through the rest of the study protocol. To pass the time, each ASP was provided a series of 30 anagrams and instructed to see how many they could solve before the trainer returned. After five minutes, the trainer reentered the room, collected the anagrams, and instructed participants to walk down the hallway to the computer lab. The number of anagrams accurately solved was calculated, and the amount of time it took each ASP to walk down the hall and enter the computer lab was recorded. Once in the lab, ASPs were instructed on how to perform a computer-based Implicit Association Test (IAT) that was modeled after those available at (https://implicit.harvard.edu/implicit), but created locally for use in previous research. To perform the task, participants were shown two response categories (e.g., funerals versus birthday parties), one on each side of the computer screen, and asked to categorize each of a series of 28 words (e.g., cake, hearse) into a category as quickly as possible by pressing the appropriate response key. Priming ASPs with words related to depression or age has previously been found to slow participants down. Five such tasks were completed (one pair of categories directly contrasting positive and negative emotions—funerals versus birthdays, the others contrasting good and bad things—e.g., health versus illness). We recorded (a) the proportion of correct responses, (b) time to record each response, and (c) time to complete the entire task. Finally, ASPs were provided with a pair of stories and instructed to rate (using a seven-point scale), the extent to which each of 20 personal qualities (e.g., thoughtfulness, cynicism) was demonstrated by the story’s main characters. They were also instructed to rate their perceptions of acting as an SP using 11 questions such as whether “training to be a SP was boring.” The amount of time required to complete this pair of tasks was recorded.
At the conclusion of the second training phase ASPs completed a similar series of tasks. The differences were that (1) when the SP trainer left the room for five minutes, participants were shown the words simulated patient and instructed to write down as many words as they could think of, using the letters in that phrase instead of completing the word scramble, and (2) instead of rating characters in a pair of stories, participants were simply shown the list of 20 adjectives and asked to rate how positive each characteristic was, using a seven-point scale.
Approximately two weeks after study completion, all ASPs were convened to readminister the RADS-2 and SIQ and to administer a questionnaire regarding ASPs’ overall perceptions of their experience.
Analyses of variance were used, treating training condition as a grouping factor and time as a repeated measure (when repeated measures were available) to assess differences in performance/ perceptions across group.
Twenty-eight adolescents attended recruitment meetings and 24 (17 female) participated in the study. Four were excluded because they were under 14 years of age (N = 3) or had a scheduling conflict (N = 1). The average age was 15.5 years (SD = 1.14). According to self-report, ASPs’ grade point average was 76.4% (range 55%–90%), and 88% planned to attend university. ASPs assigned to the depression/suicidality condition did not differ from those assigned to the cough condition on the basis of any demographic variable collected except grade point average (means = 80.0% and 72.8%, respectively) t = 2.2, P < .05).
RADS-2 and SIQ
RADS-2 scores did not differ across groups (54.3 versus 52.7, respectively) or across time (54.6 at T-1 versus 52.5 at T-2); there was no interaction between these two variables (P > .05 in all cases). The test–retest reliability of this measure was 0.90. Similarly, SIQ scores did not differ across groups (9.9 versus 9.3) or across time (10.3 versus 8.6), nor did time and group interact (P > .05 in all cases). The test–retest reliability of this measure was 0.84. Three of 24 ASPs had clinically elevated scores at T-1 (via the RADS-2, with one ASP also scoring clinically on the SIQ). Two of these three ASPs, including the SP with a clinical SIQ, were randomly assigned to the depression/suicidal ideation role. At T-2, two of these vulnerable ASPs continued to have clinically elevated RADS-2 scores but no elevated SIQ scores. None of these vulnerable ASPs were assessed as suicidal during either T-1/T-2 mental health interviews. At T-1 interviews, referrals to community mental health services were discussed with all three ASPs. At T-2 interviews, a service referral for one ASP was arranged. No ASP with a T-1 RADS-2 or SIQ score in the normal range had a T-2 clinical score.
Implicit behavioral measures
Mean scores on the implicit behavioral measures (and accompanying P values) are illustrated in Table 1 in the chronological order of their collection. ASPs in the depression/suicidality condition revealed behaviors consistent with having experienced a negative reaction to portraying this scenario: they were less successful in completing the word puzzles, walked more slowly from one room to another, and took longer to complete the IAT in the computer lab. Within the IAT, a ceiling effect was found with regard to accuracy, the proportion correct being 90% and 89% for the depression/suicidality and cough groups, respectively, P > .85. However, ASPs in the depression/suicidality condition took longer to categorize words into funeral versus birthday party categories (i.e., the only IAT aimed directly at positive versus negative emotions)—the difference was significant (P < .05) when both administrations of the measure were analyzed together. Interestingly, the magnitude of differences between groups declined as a function of the order in which the data were collected.
Perceptions of simulating patient scenarios
No differences were observed between groups across the 11 questions pertaining to participants’ perceptions of being an ASP. The mean ratings were 5.2 and 5.4 in the depression/suicidality and cough groups, respectively, with 7.0 indicating the positive end of the scale and 1.0 indicating the negative end (P > .6). Similarly, the groups did not differ in their perceptions of the strength of their own performance (means = 8.3 and 8.2, respectively, P > .7). Finally, 2/24 ASPs (neither of whom were preidentified as vulnerable ASPs) reported, in response to open-ended questions, some depression after participation; both ASPs participated in the depression/suicidality group. Fisher exact test reveals the difference in proportion (2/12 versus 0/12) to be nonsignificant (P > .2).
The risk of negative reactions to ASP participation in a suicidality scenario was the primary focus of this study. Follow-up self-report measures of depression and suicidal ideation and mental health interviews were completed within the one-month period after ASP participation to encompass the suicide contagion risk period. These self-report measures did not reveal a significant effect of group pre or post participation. Furthermore, there was no deterioration in mental status following participation, even for vulnerable ASPs.
Given the low base rate of suicide contagion, it is impossible to conclude from this study that contagion specifically is a nonissue, but it is still worth noting that SP training safeguards were built into this study that may lessen the chance of suicide contagion effects. Safeguards included: (1) adolescent mental health specialist participation in SP training, (2) case-writing techniques, and (3) ASP stress-relief methods. These safeguards were informed by (a) school-based suicide prevention programming which suggests that adoption of a mental illness model of suicidal behavior may limit suicide contagion,13 and (b) the Centers for Disease Control and Prevention guidelines for limiting suicide contagion.14
Despite these safeguards, a transient simulation effect was identified. The implicit behavioral measures suggested induction of a depressive reaction with the depression/suicidality simulation. Performance on the word-scramble and word-generation tasks, as well as walking time and IAT completion time, were significantly hampered for this condition. Differences on these implicit measures dissipated with time. These combined results suggest a transient effect, but task order was not counterbalanced across ASPs, making it difficult to conclude that with certainty. Nonetheless, identification of this simulation effect supports ASP monitoring and debriefing strategies.15 In our original study, peer pressure may have prevented full verbal expression of a potential adverse effect.6 Use of implicit behavioral methods has, therefore, proven to be a valuable research methodology for ASP effects evaluation, in part because they help overcome the influence of social desirability/faking good that may hamper more explicit measures of simulation effects.
Screening with measures of depression and suicidal ideation provided further evidence that vulnerable ASPs populate SP programs. However, this method was not sufficiently sensitive to identify the two ASPs who described depressed feelings (on a self-report form) with the simulation of depression/suicidality. We previously recommended psychological screening of ASPs,2,6 for simulations of psychiatric conditions, and we agree with Adamo’s16 identification of the need to clarify the use of SP screening. The utility of psychometric screening strategies to match ASPs to specific simulations, however, awaits further investigation.
To summarize, this small-scale study, the first to enroll ASPs with clinical indicators of depression and suicidal ideation, revealed that ASPs, irrespective of these histories, safely completed a depression/suicidality simulation. There did, however, seem to be sufficient impact of simulating depression/ suicidality as to warrant continued vigilance. The generalizability of these findings must be tested further, given the small number of ASP participants and the study SP training safeguards used. Furthermore, similar work should be performed on adult populations of SPs because the results are consistent with the work of Baumeister et al,17 who have noted that emotionally draining situations can deplete an individual’s self-control. Simulation effects research across multiple psychiatric conditions is necessary to elucidate the specific or generic nature of the transient implicit behavioral reactions identified.
The authors wish to thank Ms. Betty Howey for assistance with data collection and anonymous reviewers for their insightful commentary on this manuscript. Financial assistance was provided by an AFP Research and Educational Development Infrastructure Grant provided by the Department of Psychiatry and Behavioural Neurosciences, Faculty of Health Sciences, McMaster University.
1Spence J, Dales J. Meeting the needs of simulated patients and caring for the person behind them? Med Educ. 2006;40:3–5.
2Hanson M, Tiberius R, Hodges B, et al. Implications of suicide contagion for the selection of adolescent standardized patients. Acad Med. 2002;77(10 suppl):S100–S102.
3Shaffer D, Hicks R. Suicide. In: Pless IB, ed. Epidemiology of Childhood Disorders. New York, NY: Oxford University Press; 1994:339–365.
4Loschen EL. Using the objective structured clinical examination in a psychiatry residency. Acad Psychiatry. 1993;17:95–104.
5Hardoff D, Schonmann S. Training physicians in communication skills with adolescents using teenage actors as simulated patients. Med Educ. 2001;35:206–210.
6Hanson M, Tiberius R, Hodges B, et al. Adolescent standardized patients: method of selection and assessment of benefits and risks. Teach Learn Med. 2002;14:104–113.
7Wilson TD. Strangers to Ourselves: Discovering the Adaptive Unconscious. Cambridge, Mass: The Belknap Press of Harvard University Press; 2002.
8Bargh JA, Chartrand TL. The unbearable automaticity of being. Am Psychol. 1999;54:462–479.
9Guatieri CT, Johnson LG, Benedict KB. Neurocognition in depression: patients on and off medication versus health comparison subjects. J Neuropsychiatry Clin Neurosci. 2006;18:217–225.
10Greenwald AG, McGhee DE, Schwartz JLK. Measuring individual differences in implicit cognition: the Implicit Association Test. J Pers Soc Psychol. 1998;74:1464–1480.
11Reynolds WM. Suicidal Ideation Questionnaire (SIQ): Professional Manual. Odessa, Fla: Psychological Assessment Resources Inc.; 1987.
12Reynolds WM. Reynolds Adolescent Depression Scale–2nd Edition (RADS-2): Professional Manual. Odessa, Fla: Psychological Assessment Resources Inc.; 2002.
13Garland A, Shaffer D, Whittle B. A national survey school-based adolescent suicide prevention programs. J Am Acad Child Adolesc Psychiatr. 1989;28:931–934.
14Centers for Disease Control and Prevention (CDC). Programs for the prevention of suicide among adolescents and young adults. Suicide contagion and the reporting of suicide: recommendations from a national workshop. MMWR Morb Mortal Wkly Rep. 1994;43.
15Woodward CA, Gliva-McConvey G. Children as standardized patients: initial assessment of effects. Teach Learn Med. 1995;7:188–191.
16Adamo G. Simulated and standardized patients in OSCEs: achievements and challenges 1992–2003. Med Teach. 2003;25:262–270.
17Baumeister RF, Bratslavsky E, Muraven M, Tice DM. Ego depletion: is the active self a limited resource? J Person Soc Psychol. 1998;74:1252–1265.