In recent years there has been a movement toward focusing on educational objectives and outcomes in the training of medical students and residents. This is reflected in the emphasis on learning objectives by the Liaison Committee on Medical Education (1) and the adoption by the American Council on Graduate Medical Education of six core competencies, and in the delineation of knowledge, skills, and attitudes related to these competencies by the Residency Review Committee for Psychiatry that are specific to the practice of psychiatry (2). As educators have engaged with the task of teaching toward these competencies and devising accurate modes of assessment of their achievement, various new approaches have been utilized. One innovation that has garnered much positive attention, and also some misgivings, is the use of standardized or simulated patients (SPs)—actors who stand-in for and portray actual patients.
Simulated patients offer numerous advantages in medical education that have been well reviewed in the literature (3, 4). They are often logistically and economically more convenient than real patients, provide for more specified and predictable training experiences, and contain less risk of harm being done by the inexperienced student or resident. They therefore allow the educator and trainee to focus their attention on the aspect of the interaction most relevant to the training need at that time. Because of these advantages SPs have quickly become a significant and valued part of the medical education landscape, used for specific skill training, development of communication abilities, and high stakes examinations such as Objective Structured Clinical Examinations.
Differences in the terminology employed with SPs raise interesting questions about the place of actors in training and assessment. Some papers refer to the actor as a “simulated patient,” others use the term “standardized patient,” and at other points the SP is simply referred to as “the patient.” The terms simulated and standardized have different connotations. Simulation implies we are a step removed from the real thing, with the aim of achieving an accurate facsimile, while standardization suggests that we have achieved something superior to the inconsistencies of real patients. It is important to think carefully about these two aspects of viewing patient-actors in order to tease apart where the use of this method is appropriate to psychiatry, and where it may potentially be misused.
Assessing the appropriateness of SPs in psychiatric education becomes especially complicated when we think about interpersonal aspects of the psychiatrist-patient encounter, such as the quality of empathy. Empathy has been defined as “the action of understanding, being aware of, being sensitive to, and vicariously experiencing the feelings, thoughts, and experience of another of either the past or present without having the feelings, thoughts, and experience fully communicated in an objectively explicit manner” (5). Stedman’s Medical Dictionary defines empathy as “The ability to intellectually and emotionally sense the emotions, feelings, and reactions that another person is experiencing and to effectively communicate that understanding to the individual” (6). This communication may include statements of support, reframing and reflection, and extends to sensitive inquiry which guides exploration. The effectiveness of these interventions is presumed to reflect the accuracy of the interviewer’s apprehension of the patient’s internal emotional experience.
Empathy is essential in building an alliance with disturbed and disturbing patients, and in engaging in the explorations of psychotherapy. To an extent that is unique in psychiatry, clinicians also develop the capacity to use their empathic response to the patient as part of their diagnostic data. For example, the hypomanic patient may make the interviewer feel invigorated, while the depressed patient may inspire fatigue and discouragement; through the use of empathy, experienced clinicians may perceive a patient’s hypomanic euphoria while it is still unknown to the patient. At more advanced levels, clinicians use their own empathic reactions as aids to differentiating specific kinds of character pathology.
Current Use of SPs in Psychiatric Education
The literature on simulated patients in psychiatric education can be divided into three main categories. First, there are reports that focus on the use of SPs to enhance exposure to patient variety and the range of psychopathology. Second, there is an increasingly prominent literature that describes and promotes the use of SPs for the clinical skills evaluation of medical students and residents in psychiatry. These studies are often concerned with improving the reliability and validity of current methods of skills assessment in evaluations of a range of students’ abilities from basic history taking to appraisal of complex interpersonal processes that involve empathic responsiveness to the patient. Finally, there is a small literature that reports piloting the use of SPs in teaching psychotherapy techniques.
Exposure to a Variety of Patients
Simulated patients have been used for the purpose of exposing students to a broader range of psychopathology than would otherwise be consistently available in a relatively short amount of time or on any particular clinical service. One study described using SPs in the third-year medical student clerkship to meet Liaison Committee on Medical Education criteria for exposure to an adequate variety of patients, and reported that students considered the SPs’ display of various psychopathologic disorders to be a highlight of the course (7). Another study reported on students’ use of an SP for round-robin diagnostic interviewing, with stop-action comments by faculty about interviewing technique (8); the students gave high marks to this method of teaching interviewing technique and psychopathology. Using an SP to simulate the phenomenology of schizophrenia in the service of teaching the mental status examination to a large group was similarly well received by students (9). Another study described adding simulated patients to an introductory psychopathology course in which the students reported the advantages of the SPs as being their clear presentation of symptoms, cooperativeness in providing direct answers to the students’ questions, and emphasis on symptoms as opposed to treatment issues (10). Although faculty, when surveyed after the course, agreed that the SPs offered these advantages, they questioned if presentations with such qualities were realistic and felt that actual patients provided a superior educational experience.
Assessment of Skills in Psychiatry
Van Der Vleuten and Swanson (11) state that the purpose of a test is to “draw inferences about the ability of examinees that extend beyond the particular items used to the larger domain from which the items are sampled.” A critical issue in every assessment is whether what is measured in the examination is appropriately representative of the actual skills one wishes to assess (i.e., whether the test is valid) (12). The use of SPs in Objective Structured Clinical Examinations for the purpose of assessing relatively discrete skills, such as remembering to ask specific screening questions about symptoms, seems straightforward; the observed behavior in the exam is probably a good stand-in for the actual clinical skill (13). However, the situation becomes quite complicated when SPs are used to teach or assess empathic responsiveness.
The literature on assessment of empathy in general medical education is instructive of the difficulties that may arise in using SPs instead of actual patients (14–16). For example, using the SP’s assessment of the student’s capacity to sense what the SP was feeling as a measure of empathy can be conceptually problematic. Some studies (15, 16) used a rating of interpersonal skills containing items such as “the doctor usually sensed or realized what I was feeling,” without consideration of what complications may ensue from asking an actor who is simulating an experience to assess whether the trainee sensed what he or she was thinking. Since both know that this is a simulation, the trainee may be identifying what he or she thinks a person ought to feel in the scenario being described, as opposed to perceiving and sensing the actual affect of the actor. This concern about the effects of the examinee knowing it is a simulation is supported by a recent study of internal medicine residents regarding the validity of SP examinations for assessing interpersonal skills (17). The investigators examined the effects of previous experience with SPs on subsequent examination performance. The authors noted that SP exams “may not be valid in their assessment of residents’ interpersonal skills once the resident’s experience with SPs as an exam format reaches a certain level.” They hypothesize that residents may be “learning to approach them as a test and they may ‘go through the motions’ rather than engaging emotionally with the SPs…” (p. 71). As a further complication, the actor’s affective experience, in turn, may or may not be directly related to the scripted role—variables such as acting technique, fatigue, or external preoccupations may all potentially have an impact.
The general medical literature on the use of SPs to assess complex interpersonal skills may also illustrate the challenges involved in validating this method. One study asked SPs to assess whether medical students were empathic or not (14). The authors suggest that since “… the SPs received training in assessing student performance, they might reasonably be expected to be more critical than the ‘real-life’ patient, as the focus of (their) attention is the accurate assessment of student performance, rather than actual medical problems.…Thus SP ratings of student empathy can tentatively be assumed to carry more weight than ratings by actual patients” (p. l0). There is, in my view, a potential contradiction in asserting that an assessment is valid because its results differ (in this case are more stringent) from the actual clinical situation. There is some evidence to support the assumption that SPs’ assessments and real patients’ assessments of physicians do not necessarily correlate, as reported in a recent study of family physicians (18). The authors acknowledge that SPs may become expert connoisseurs of physician communication and perceive failings that a real patient might overlook, but they also note that real patients are able to place the physician’s interventions into the context of a shared history that takes into account how the physician responded in times of great distress, or how individually tailored their psychosocial interventions have been. Simulated patient examinations of interpersonal and communication skills clearly provide interesting and potentially useful data, but this data should not be taken as necessarily reflective of the experience of real patients in the clinical setting.
A growing literature on using SPs in the assessment of complex interpersonal dynamics in psychiatry has also struggled with the question of validation. Grading in such studies is often done using checklists for more straightforward skills and global ratings of interpersonal process factors such as rapport, empathy, and containment of the interviewer’s emotions. Attempts to validate these assessments of complex interpersonal skills have included comparisons of SPs’ assessments with those of faculty (19, 20), with performance on clinical rotations (21), and with prior ranking of students from an interviewing course (22). Standardized patients’ assessments were highly correlated with faculty observer assessments on content checklist items, but poorly correlated for more complex interpersonal process (21, 22). One study that examined the use of an Objective Structured Clinical Examination during psychiatric residency found that clinical and supervisory ratings did not correlate well with examination results (21). Another study yielded more complicated validation results (22). Global ratings by the SPs differentiated medical students and residents, supporting construct validity, while content checklists did not. Faculty supervisors’ ratings of the students correlated with high scores on checklists, but did not correlate with global ratings of the interview by the SPs. Efforts to validate the use of SPs for complex interpersonal skills in psychiatry are, in my opinion, currently inconclusive.
Some studies have found that medical students, residents, and practicing physicians fail to distinguish simulated patients from actual patients in general medical situations, and some authors have suggested that this indirectly supports the validity of the SP examination. A more limited set of studies addresses this issue in psychiatric settings, with conflicting results. One classic study used “pseudopatients” to expose the lack of careful listening and thoughtful diagnosis in the culture of psychiatric hospitals and these SPs went unidentified (23). In a recent study conducted during a behavioral science course in medical school, the students were unable to distinguish the real patients from the actors (24). Interestingly these SPs were not standardized, nor was it clear how much simulation was entailed; they were instructed to create a plausible history, and there is no way to know how much truth may have been involved. Another group, by contrast, reported that medical students were able to accurately differentiate the SPs from the real patients in an introductory psychopathology course (10). In this study the students reported feeling that some of the SPs’ responses seemed shallow, and the students became less emotionally engaged with the interview once they perceived it was a standardized patient.
Standardized Patients in Psychotherapy Courses
The psychiatric residency review committee mandates that residency programs attest to the competence of residency graduates in cognitive-behavioral therapy, psychodynamic psychotherapy, and supportive psychotherapy. Training in the psychotherapies poses special problems for residencies, given the limited possibilities for direct observation of experienced practitioners, and the limited opportunities for faculty to directly witness the resident’s work as a therapist. A particularly intriguing possibility, therefore, is the use of SPs in teaching psychiatric residents to become psychotherapists. One report described using actors to simulate an initial psychotherapy session in a course for beginning psychiatric residents (25). The residents, instructors, and actors all rated the experience positively. Interestingly, in this study the SPs were instructed that they “could give a completely true history or make it up if they felt uncomfortable or did not want to reveal themselves in this manner … there was no effort to standardize the scenarios.” In fact, only one of the five SPs stated that he had “acted” rather than presented his own story. In contrast, another report described using SPs precisely to reduce the “intrinsic ambiguity of psychotherapy situations,” which the authors saw as an impediment to learning (26). In another report, an experienced senior teacher of dynamic psychotherapy described his success using “actor-patients” over many years in an introductory psychotherapy course for psychiatric residents (27). Actors were given a very brief sketch of their identity and were encouraged to rely on their own spontaneous responses. It was noted that the “actor patients frequently explored personal material” (p. 10) and that in order to achieve the goal of teaching exploratory interviewing it was important that the actor patients “not be too scripted in their responses” (p. 10). Thus in the limited efforts to use SPs for psychotherapy training and assessment, it is not clear if these were true SPs or “volunteer” patients.
Inherent Problems in the Use of SPs in Psychiatry
Some have cautioned that the goal of standardization runs counter to the variation in actual clinical work (which includes the personality of the patient, the social and professional context of the encounter, and the specific nuances of this instance of the disease) and that this is central to what is most challenging about being a physician. One review of SP exercises noted, “A disadvantage of SPs is that they may become ideal ‘textbook cases,’ to which real patients with all their idiosyncrasies do not often conform” (4, p. 344). Cox suggests that the use of SPs in Objective Structured Clinical Examinations may result in a preoccupation with eliminating the variability of the clinical task at the expense of the validity of the exam (i.e., that what is being measured in the exam is truly representative of what is intended to be assessed) (28). He argues for assessments based on multiple observations of actual work on the wards, concluding, “The task facing assessment is to foster the study of our judgment of performance on real world responsibilities, not to submerge it within the rigidity of formal structures and a set form of words.” This perspective is echoed by a recent review that suggests that Objective Structured Clinical Examinations may sacrifice validity for reliability (12). Hanna and Fins (29) raise another kind of concern by noting that asymmetrical power relations are intrinsic to doctor-patient encounters (where the doctor has authority and knowledge, and the patient is vulnerable and exposed) and simulations actually invert, rather than re-create, this dynamic.
Psychiatry poses special difficulties for simulation. These difficulties become most apparent when the attempt is made to use SPs for the teaching and assessment of complex interpersonal engagement, such as the use of empathy, and the process of psychotherapy. Some authors, while acknowledging the challenges to assessing and teaching empathy, suggest, “Training must focus on facilitating actors’ ability to convey emotion realistically and therefore evoke empathy in the interviewer” (10, p. 30). In my opinion, it is unlikely that further training for the SPs would suffice because of the intrinsic difficulties discussed below in using simulation to assess or enhance a psychiatrist-in-training’s empathic responsiveness.
Some of these difficulties lie in the nature of empathy in an interpersonal encounter and the confusions that accumulate when we attempt to assess a student or resident’s empathy with a simulated patient. Are we talking about the student acting in a way that we believe would convey empathy if this was a real encounter? In that case, the student is re-creating the behavior that would follow from having an internal experience that was a response to the inner life of the patient. Or are we talking about the student actually having such an internal experience, and thus feeling moved or disturbed by the SP? This situation would be analogous to experiencing strong feelings in response to watching an actor in a movie or a play. Which is the aim in SP encounters—the trainee’s simulation of empathy, or the creation of an actual emotional response to the fictive character? And is there a difference between an empathic response to a dramatic character as opposed to a real person?
To begin to clarify these questions, we might start with examining the difference between an actor and an actual patient. An actor’s intent is to convey to the audience a performance that is meant to evoke an emotional response. The actor on stage explicitly works to make the audience feel amused, heartbroken, frightened, outraged, etc. The patient, on the other hand, does not generally approach the patient-physician encounter with an explicit agenda about the impact of his or her presentation on the physician “audience.” Instead, the desire of real patients to reveal themselves is often conflicted. This is especially true in psychiatric patients, who often feel divided about what aspects of their suffering they wish to be known to their psychiatrist at any given moment, and may even resist knowing themselves what they actually suffer from. Empathic skill in psychiatry involves teasing out ambiguous, conflicted strands of feeling and at times “feeling ahead” of the patient and intuiting what they themselves cannot yet articulate. The actor’s success at intentionally portraying a consistent scripted role limits the usefulness of SPs in psychiatry. This standardization is at the cost of accurate simulation of the situation with a real patient who often may not know the “truth” in advance of the dialogue with the clinician.
A second confusion in assessing a trainee’s encounter with an SP results from the blurring of the two participants’ contributions. If the actor succeeds in evoking an emotional response in the trainee, this may be due more to the efforts of the actor than the empathy of the trainee. Having the SP serve as the one to assess the quality of the student’s empathic response is conceptually problematic, as the SP may be most accurately assessing his or her own capacity to evoke a certain response to a performance. Further, when we standardize the performance and then track the variations in the students, we may be observing something about the range of “audience responsiveness” or the range in the capacity of students to lose themselves in play. One investigator notes that “experiences with standardized patients … may also exert an independent effect through repeated practice of the ‘suspension of disbelief’ necessary for effective standardized patient interaction (30, p. 371). These qualities of audience responsiveness or capacity for suspension of disbelief are not the same as the capacity to empathically feel something with a real, complicated, possibly conflicted, suffering patient. It is true that we may then see differences in the extent to which students can observe their own responses, and articulate them in tactful and sensitive communications to the SP. This may yield useful data about the trainee’s communication skills, but if this communication is not rooted in the accurate appreciation of the other’s experience, then it is misleading to describe this as empathy. Real patients may notice when the psychiatrist’s response seems like a performance, and feel alienated if they perceive that the clinical encounter is only a simulation of compassion.
A third inherent tension with simulation is that in many psychiatric interviews there is a gap between the clinician’s goal to accurately and deeply understand and the patient’s agenda to get the clinician to do something. Some researchers argue that one advantage of SPs is that they do not press the students about treatment issues, freeing up the interview to focus on clarity of diagnosis (10). Real patients, however, are of course concerned with treatment issues—in the real world every clinical encounter involves a negotiation about the patient’s own agenda (realistic or not) for relief of distress or validating explanation of the illness. The word patient is derived from the Latin pati—to suffer (31), and patients are willing to make themselves extremely vulnerable to their physicians in their hope that they will find relief from suffering (32). Empathy with a real patient includes the clinician appreciating and bearing the emotional pressure for some kind of action, whether this action is in the form of prescribing, diagnosing, redefining, or reassuring.
Fourth, it may also be problematic to accurately simulate certain common kinds of clinical encounters precisely because those situations already contain simulation in their pathology (e.g., “As-If” personalities, conversion disorders, malingering), and the clinician’s task is to understand the mixture of veracity and fiction. Some patients with narcissistic pathology suffer from their inability to accept the limits of their capacity to dictate the world they live in; they maintain neurotic stances that are constructed around grandiose fabrications, and may even lapse into psychotic conditions that provide the secondary comforts of fantasy (33). When one of the clinician’s tasks is to identify a patient’s (sometimes unintentional) simulation, it may be confusing to teach or assess this by adding another layer of fabrication.
The literature on whether trainees and clinicians can differentiate simulated patients from real patients yields mixed results. The failure to discern the simulation is sometimes cited as evidence for the validity of SP-based evaluations (4). However, such difficulty in differentiating the real from the simulated is also a potential argument against the use of SPs in psychiatry examinations and training. The capacity to deceive, and the capacity to discern such deception, is deeply rooted in normal psychology (34). Medical students and residents need to be trained to discern different levels of dissembling. One important learning goal for a developing psychiatrist is to be able to sort through many levels of deception, often in the same interview. The patient may at one point in an interview be consciously lying to the clinician, while at another moment that patient may be telling only part of the truth and withholding the rest. The same patient, at another point, may be lying to themselves as well; they might be taken in by their own self-deception, or they might be simultaneously aware of the truth while disavowing this knowledge. Sorting out such possibilities entails a complex and sophisticated skill set on the part of developing clinicians.
Attempts to incorporate SPs into courses teaching psychotherapy further illuminate the distinction between simulation and standardization. Studies of these exercises may have been found useful precisely because they did not attempt to standardize the patients, as allowing the actors to “be themselves” made the exercise feel closer to the experience of therapy. The SPs in one study were told they could give a completely true history or make it up if they felt uncomfortable or did not want to reveal themselves in this manner. This differs, however, from the experience of a dynamic therapy, where staying with the truth is the goal, discomfort is the norm, and patients are ambivalent about revealing themselves in an initial session with a new therapist. Not surprisingly, in this complex amalgam of truth and make-believe, some of the SPs noted that the “issue of acting was a big topic of discussion in the session” (25, p. 171).
Psychodynamic psychotherapy is itself centrally concerned with the challenge of teasing apart what is real and what is projected fantasy in the interpersonal field, and what elements of both the real and the fantasy originate in each party (35, 36). This is the foundation of the exploration of transference as a means of helping patients learn about themselves. In addition, many patients come to psychotherapy with a chief complaint that they are living a lie, and with the hope that they may develop a more authentic sense of self. The use of a model of learning that involves the insertion of another element of pretense into the field might be potentially helpful as a means to practice certain specific skills, but we should be cautious in asserting that the essentials of a psychotherapeutic encounter can be reliably replicated using SPs.
Simulated patients have generated a great deal of interest and enthusiasm in both undergraduate and postgraduate medical education, and have been increasingly used for teaching as well as for high-stakes assessments. Simulated patients have clear value in psychiatric education when used to provide exposure to the range of psychopathologic disorders within the limited time frame of a medical student clerkship or limited scope of a particular clinical setting. Simulated patients can also be effective in the assessment of discrete, operationalized skills and behaviors, such as whether or not an examinee asked about a specific symptom or obtained various historical data. Examinations of these discrete skills by SPs are a potentially valuable complement to traditional observations by faculty on the wards, and they add a level of consistency and reliability that is difficult to otherwise achieve in the course of a trainee’s clinical work. Psychiatry, however, provides distinct challenges to the use of SPs for high stakes examinations (with the possible exception of assessing very well delineated skills such as reviewing specific symptoms) and teaching complex interpersonal skills such as psychotherapy. Encounters with psychiatric patients require subtle empathic skills and the capacity to discern nuanced levels of veracity and self-deception in the patient’s presentation. The “suspension of disbelief” that is necessary for successful performance with an SP (30) is not clearly analogous to the skills required of a psychiatrist. Simulated patients, in my opinion, are therefore unlikely to be effective in assessing complex skills such as empathy in psychiatric interviews or in teaching complex interventions such as dynamic psychotherapy.
The difficulty validating the use of SPs in the assessment of such skills may be rooted in aspects of the psychiatrist-patient relationship that are antithetical to simulation. Future research should focus on clarifying the nature of the skill set described by “empathy” and defining the extent to which empathic responsiveness is important in psychiatric assessment and formulation. More research is needed to validate specific uses of simulation in psychiatry, with careful attention to the qualities of the psychiatrist-patient encounter that may be resistant to simulation. In light of this, it is prudent to continue to explore ways our trainees might profit from increased learning opportunities with SPs, while retaining a strong degree of caution about substituting actors for actual patients in high-stakes evaluations such as board certification and in the learning of psychotherapy.