Zoom Improv is accessible and enhances medical student empathy

Abstract

Background

Empathy declines during medical training, despite its importance.

Methodology

In this randomized controlled trial, we assessed the impact of Zoom improv on medical student empathy using a concurrent mixed-methods approach. Quantitative assessment with three survey tools and qualitative assessment by content analysis of Zoom session field notes were conducted.

Results

Zoom improv participants had higher empathy scores in perspective-taking and fantasy and lower scores in personal distress compared with the control group. Medical students who participated in Zoom improv exercised emotional expression, active listening, and giving “gifts,” which apply to healthcare settings in which affirming team members with empathic concern can advance communication, patient rapport, and teamwork.

Discussion

This pilot study highlights promising findings for the incorporation of Zoom improv in medical education, including enhanced empathy, self-reflection, and understanding how these skills impact work in healthcare. Future studies may expand on the optimal timing to teach improv. Future studies conducted on virtual platforms may also further investigate our finding that the fantasy domain of empathy increases after Zoom improv sessions, whereas this increase in fantasy was absent from previous in-person studies. Given the increase in telehealth and virtual medical visits, exercising empathy skills through a screen during training may be an important addition to medical curricula.

Introduction

Medical students must learn an exorbitant amount of information during training, often described as “trying to drink water from a firehose,” [1]. Courses blast students with biological mechanisms, statistics, and frameworks for approaching review of systems, physical exam, and history of present illness. While focusing on retaining new knowledge and checklists along a steep learning curve, medical students’ opportunities to exercise other skillsets decrease [2, 3]. Empathy is one such skillset. Empathy can involve cognitive and/or affective engagement. Though we do not have one agreed upon definition of clinical empathy, two commonly used definitions stem from the development of measurement tools such as the Jefferson Scale of Physician Empathy (JSE) and the Interpersonal Reactivity Index (IRI) [4, 5]. The creators of the Jefferson Scale of Physician Empathy (JSE) define empathy as “a cognitive attribute that involves an ability to understand the patient’s inner experiences and perspective and a capability to communicate this understanding” [6]. Whereas, when creating the IRI, Davis asserted that “empathy can best be considered as a set of constructs, related in that they all concern responsivity to others but are also clearly discriminable from each other” [5]. Therefore, the IRI contains 2 cognitive domains (perspective-taking and fantasy) and 2 affective domains (empathic concern and personal distress) [5]. Davis relied on Mehrahbian et al.’s definition of emotional empathy, “an individual’s characteristic inclination to respond with emotions similar to those of others who are present.” when designing the affective domains [7]. Based on a synthesis of these and other common definitions of empathy, we define the term as: the imaginative reconstruction of another’s perspective and the emotional resonance this causes in the self [8,9,10,11].

Despite its importance in the medical field, empathy declines during medical training [12,13,14]. Physicians are required to support, listen, and advise patients, families, and team members in difficult situations. Therefore, they must be skilled in reading emotions, employing situational awareness, and responding with emotional intelligence [15, 16]. Lower empathy is associated with higher fatigue, which can lead to medical errors, such as incorrect diagnoses, unrecognized medication side effects, inadequate pain control, and unnecessary polypharmacy [17]. Distrust or frustration that fractures the patient-physician relationship [18] could contribute to lack of follow-up, avoiding medical care, and not completing treatment plans. Decreases in empathy among health professionals can cause significant personal distress, decreased work satisfaction, inability to relate or connect with others, feelings of isolation, and premature departure from medicine [19]. Improvements in medical training are needed to prevent the decline in empathy and associated negative consequences.

Medical improv uses improvisational theatre techniques and debrief discussions to teach specific skills and has been shown to benefit medical students. Brief improv interventions can improve medical student communication [20,21,22] and improve satisfaction and physician skills as reported by standardized patients [20]. One study of medical students determined that medical improv led to improvements in well-being, proactivity, and engagement with their studies [21]. A review of seven articles found that medical improv promotes acquisition of professional competencies in health professions education and has utility in interprofessional team development, leadership, wellness, and resiliency [23]. Whether medical improv can provide opportunities for healthcare professionals and trainees to practice empathy in an educational environment that can translate to improved empathy in healthcare contexts is an area of ongoing study [24, 25].

The use of virtual learning in medical education expanded during the COVID-19 pandemic. Even before COVID-19, online learning was shown to be effective in medical education [26]. Given the benefits of in-person medical improv, electives were converted to the Zoom platform [27]. In a health disparities course, medical students participated in a Zoom workshop using improv exercises to discuss identity and biases [28]. First-year medical students participated in a virtual improv workshop and reported improved listening and observation skills as well as opportunities to engage in meaningful health equity discussions [29]. Benefits of virtual medical improv continue at the level of resident education. Residents who participated in a pilot program reported enhanced well-being, reflective capacity, and communication skills [30]. Therefore, we have compelling evidence that medical students who participate in Zoom improv during their training are likely to receive benefits. However, no studies have investigated the effects of Zoom improv on medical student empathy, which is of particular interest given declining empathy during training.

Early interventions may provide medical students with increased awareness and opportunity to exercise empathy skills. Long work hours, resource limitations, and competing obligations make it difficult for medical students to engage in a longitudinal, in-person curriculum. In addition, adding yet another onerous requirement may have a further negative impact on their well-being. Therefore, we assessed the impact of Zoom improv on medical student empathy. Using the empathy/improv framework from our prior work [31], we designed a pilot study to answer: what impact does a medical improv intervention delivered over Zoom have on medical student empathy? We used a concurrent mixed methods design to understand outcome measures and capture student thought processes during the intervention.

Methods

Participants and setting

Medical students from the University of Wisconsin were recruited June-August 2022 by email or flier to participate in an IRB-approved study. Participants completed electronic recruitment and demographic surveys and were screened to meet inclusion criteria (available to attend at least three of five 90-minute Zoom sessions and had not previously taken a medical improv elective). After recruitment, F.H. randomized students to control (n = 25) or intervention (n = 26). Students received a pre-survey which included the Jefferson Scale of Empathy (JSE) Student Version, the Consultative and Relational Empathy (CARE) Measure adapted for self-report, and the Interpersonal Reactivity Index (IRI) [5, 32, 33]. After completion of the pre-survey, intervention group participants attended (n = 10) Zoom improv sessions over a month-long period. During this time, control group participants (n = 22) did not receive any intervention and were enrolled in the standard medical curriculum which consists of didactic lectures and cases in the first 18 months, and clinical rotations in the latter phases of training. Post-surveys were completed after the conclusion of Zoom improv sessions.

Teaching methods

Prior to COVID-19, most medical improv was taught in-person. Since then, while educators have proven its efficacy over platforms like Zoom, no randomized controlled trials have been conducted [27]. A.Z. and M.A. created and taught the sessions adapted from the Watson/Northwestern Medical Improv curriculum [34] and Spolin online improv [35] (Supplemental Table 1).

Mixed methods evaluation

Intervention and control groups completed three empathy measures approximately one week before and after the improv sessions. The IRI (Cronbach’s alphas range from 0.70 to 0.78 in the four domains) measured participants’ dispositional empathy [36], whereas the JSE-S (Cronbach’s alpha = 0.80) measured cognitive empathy [37]. The CARE Measure (Cronbach’s alpha = 0.92) captured self-reported empathy skills [33]. Field notes taken by J.K. during the sessions were analyzed. J.K. also conducted one focus group of intervention participants.

Analysis

Demographic information, summarized as raw counts and percentages, was compared using Fisher’s exact tests for categorical data. All continuous variables were expressed as mean plus standard deviation (SD) and compared using the Wilcoxon sign rank test. All continuous data were analyzed using the non-parametric Wilcoxon to account for small samples and non-normal distribution. Only data from participants who completed both pre and post-surveys for the mean change analyses were used. Mean changes from pre to post for both improv and control groups were compared and reported as Cohen’s d effect sizes (small effect = 0.2, moderate effect = 0.5, and large effect = 0.8). IRI and JSE-S scores were calculated according to their instructions, and the above methods were used for calculating effect sizes. P-values ≤ 0.05 were considered significant. Analyses were conducted using STATA version 17. Content analysis of Zoom session field notes was performed by A.Z., M.A., and J.K. by first conducting open coding via consensus, then meeting to conduct higher level analysis using the empathy/improv conceptual model [31].

Results

Quantitative findings

The intervention and control groups were randomized effectively with no differences in participant characteristics including gender, race, year in school, years of previous healthcare experience, or improv experience (Table 1).

Table 1 Participant characteristics

The IRI has four domains (personal distress, perspective-taking, fantasy, and empathetic concern). Unlike the intervention group, control participants had significantly higher personal distress scores and significantly lower perspective-taking scores at the end of the study compared to the beginning (Table 2). The intervention group had a significantly higher fantasy score (Table 2). Empathic concern increased in both groups, increasing more in the intervention group. These results indicate that intervention participants were better able to manage stressful situations, approach topics from different viewpoints, exercise imaginative and creative thinking, and express humanistic qualities.

Between the beginning and end of the study, there was a significant difference in perspective-taking: the improv group increased while the controls decreased (p-value = 0.04). To explore this, participants were divided into early-stage (first- or second-year medical students) and late-stage (third, fourth- and fifth-year medical students). Early-stage students had a significant difference in perspective-taking in which the improv group increased and the controls decreased (p-value = 0.06); the same held true for late-stage students. The change in empathic concern, however, was significantly different between the improv and control groups of late-stage trainees. Improv participants increased in empathic concern whereas control participants decreased (Table 2). In contrast, both early-stage improv and control participants increased in empathetic concern. Regardless of training stage, there was no increase in personal distress in the intervention group (Table 2). Early and late-stage controls had increases in personal distress throughout the study, with late-stage having a greater increase and large effect size (Cohen’s d = 0.84).

Table 2 IRI scores in improv and control groups at early and late stages of medical training

The JSE-S scores can be summarized in three domains (perspective-taking, compassionate care, and walking in the patient’s shoes). Both the improv and controls had significantly higher perspective-taking scores at the end of the study, which may indicate that the importance of perspective-taking becomes more apparent throughout medical education (Supplemental Table 2). When sub-analysis was conducted for early and late-stage trainees, compassionate care scores increased in late-stage improv trainees but decreased in the controls (large effect size Cohen’s d=-0.81). In early stages, compassionate care scores decreased in the improv group, indicating that trainees may value empathy more in later stages when they are exposed to patient care.

The CARE Measure captures patient ratings of their medical providers’ interpersonal skills and has been adapted for self-report [33]. CARE Measure responses were similar from the beginning to the end of the study in the controls. In the improv group, participants rated themselves better on “really listening,” a skill that one of the Zoom sessions focused on (Supplemental Table 3). Improv participants rated themselves more harshly on showing “care and compassion” at the end of the study compared to the beginning, which may reflect insights gained during the intervention on areas for growth. Intervention participants also rated themselves lower on “helping your patients take control” on post-tests, which may reflect insights gained regarding power dynamics and communication within healthcare teams and between patients and physicians.

Qualitative findings

Upon analysis of field notes taken during intervention debriefs, we noticed that participants related improv activities and personal discoveries to patient care and working in the medical field (Fig. 1). These discoveries were about the difficulty inhabiting a role that was foreign to them and portraying emotions they were not currently feeling. Additionally, participants discussed how the improv activities enabled them to explore the impact of active listening, power dynamics, and varying degrees of autonomy on their interactions with peers and supervisors (Supplemental Table 4).

Fig. 1

Qualitative analysis from Zoom improv sessions fits existing improv/empathy model. White circles demonstrate discoveries from Zoom improv debriefs that pertain to self, dark gray circles are discoveries relating to other (including providers, patients and families, and trainees), and the light gray overlap are the discoveries that focus on their relationship. Venn diagram titles correspond to the improv/empathy model

Discussion

We found that medical improv can provide experiential education on skills such as spontaneity, creativity, compassion, and empathy that a didactic approach cannot offer. Exercising these skills through human interaction can improve students’ abilities to connect with patients and provide more personalized healthcare experiences. Our work indicated that medical improv can enhance empathy and self-reflection among medical students. Notably, participants became familiar with two common improv concepts: “yes, and…” and “giving gifts.” These two concepts emphasize the importance of adding or giving additional information to scene partners so that a scene, relationship, or activity can advance. The “yes, and…” mentality acknowledges constraints and offers potential solutions, thereby leading to flexibility and creativity. Participants related these concepts to patient interactions, expressing the benefit of sharing information: “I like the idea of giving gifts…so we’re on a more level playing field…” Participants found it more difficult to initiate new improv ideas and easier to follow or build off others’ ideas, which is another facet of gift giving. In addition, we demonstrated that Zoom improv is an effective teaching method, which may improve access to this technique.

Benefits of study design

There are many strengths of our study design that add value to the field. Other studies on virtual improv for health professional trainees have evaluated self-reported benefits and enjoyment of an improv workshop [27, 28, 30, 29]. Our study is distinct because we randomized participants into either a control group (no intervention) or an intervention group (Zoom improv). Interestingly, we saw many changes in the control group, which we were not anticipating.

We expected the control group to respond similarly to survey questions at the beginning and end of the study. However, the control group was enrolled in the standard medical school curriculum, which requires long study hours, preparation for exams, and working with patients in hospitals and clinics. IRI survey responses from control participants showed increases in personal distress over a one-month period, which may demonstrate the baseline increase in stress of medical students as the semester continues and more demands are placed on them with regards to exams and patient-care responsibilities. The control group also had a decrease in perspective-taking over a one-month period. There is a negative correlation between empathy and fatigue [17], which could partly explain why as medical students progress in the semester, they engage in less perspective-taking.

Since the medical students were randomized with no differences in year in school or level of experience between groups, we would expect perspective-taking in the intervention group to match those of the controls and decrease over the study period. The intervention group did not match the control, however. Intervention participants had no change in personal distress and increased perspective-taking, indicating the strengths of the Zoom improv intervention, even in the context of an expected increase in personal distress [38] and decrease in perspective-taking [19] during medical training. Therefore, the control group and the two data timepoints for all participants provide a better understanding of the current need for intervention in medical training and the beneficial effects that it can have over a short time.

Self-reflection and power dynamics

In assessing quantitative and qualitative findings together, we found that Zoom improv participants had an improved capacity for self-reflection. The CARE Measure is an opportunity for students to self-report their own interpersonal skills. The control group was largely unchanged, which may imply several things: complacency regarding interpersonal skills, a standard medical school curriculum that does not focus on such skills, or lack of awareness that these skills can be enhanced for improved patient experience, job performance, etc [18]. Zoom improv participants had different responses to the CARE Measure from the beginning to the end of the study, with self-reported improvements in listening and lower self-reported scores on showing compassion and sharing power. From the qualitative findings, participants discussed: difficulty in expressing certain emotions, how emotions can be displayed differently by different people, emotional range, and appropriateness in different contexts. Together these findings demonstrate that Zoom improv participants had a new appreciation for emotional expression in healthcare to affirm or mirror a patient or team member. Expressing the difficulty and nuance of these skills in debriefs may explain why participants rated themselves lower on this skill with a newfound appreciation for the work that remains.

Zoom improv participants also scored themselves lower on “helping your patients take control” at the end of the study. In debriefs, participants discussed how the power differential between patients and clinicians impacts communication. Students see demonstrations of this when clinicians interrupt patients, stick to pre-set agendas, go through checklists of questions, ignore their perceived authority/power, and have to surrender to rigid time constraints. They also discussed ways of sharing power by including all team members, having specific roles, listening to patient’s agendas, mirroring body language, using verbal and nonverbal communication to acknowledge others, and naming emotions. Together, these quantitative and qualitative findings demonstrate that intervention participants were able to wrestle with power dynamics in improv exercises, discuss their relevance in healthcare settings, acknowledge that power dynamics affect them and their patients, and take responsibility to improve their own skills.

Curriculum timing

Combining the quantitative and qualitative data, another interesting finding that would benefit from further examination in a larger sample was timing. Results indicate that late-stage students may garner more benefits than previously recognized in the field. Other studies of Zoom improv have focused exclusively on early-stage students [27,28,29]. Our results showed that all students benefitted from Zoom improv, with large effect sizes for late-stage students. This might be due to the structure of medical training where in the early years of medical school, students have more didactic lectures and less patient contact. Though the medical school we conducted the study in has undergone a curricular transformation as many others have to embed more patient contact in the first couple of years, there is still a higher percentage of classroom learning during this time. In later years, students participate in more clinical rotations, working on healthcare teams with direct patient care. Proximity to severe illness and death, long hours, and unknown work environments may all be distressing. The improv sessions gave students an opportunity to share experiences, leading to common understanding, rather than isolation.

Fantasy

Our findings regarding fantasy would also benefit from further examination with a larger sample. The IRI domain of fantasy is rarely examined in detail. There is some disagreement about how this domain of empathy interacts with others, however we define empathy as both the imaginative reconstruction of another’s perspective and the emotional resonance this causes in the self. In our prior in-person study, we did not see such an increase in fantasy in the improv group [31]. We wonder if engaging in improv activities over Zoom required a higher level of imagination. Not only did students need to create connection with each other in fictitious relationships and environments, they had to do so through screens. The limitations of using Zoom rarely came up in the debriefs. In contrast, a virtual platform provided many benefits including ease of attending sessions, attention to what the camera was able to capture and not capture, trial of exaggerated body language to communicate, and comfort with muting and unmuting. It is possible the students felt freer to be vulnerable in their own spaces and control what they presented to others.

Strengths and limitations

There were many strengths of this study’s methodology, data collection, and analyses. The randomization into control and intervention groups is a strength of this study because improv participants were not self-selected, and the control group provided a comparison for understanding and contextualizing the data. Mixed methods was the ideal study type because the quantitative and qualitative data analyzed together allowed for further interpretation and understanding. Development of a code book and consensus coding among diverse team members with backgrounds in education, empathy, information sciences, and medicine was another strength of this study’s design and execution. Including medical students across all years of training and using several different empathy measures allowed us to investigate differences in early and late trainees in different empathy domains.

We also acknowledge study limitations. This pilot study was limited in size. While it was a strength of this study to include a control group for comparison, a limitation of this study is that it did not directly compare in-person improv with Zoom improv or Zoom improv with another Zoom medical humanities course.

Next steps

Future work could expand this pilot study by recruiting larger samples of trainees. Rigor and reproducibility could be assessed by investigating the impacts of Zoom improv on medical student empathy at other institutions. Recruiting medical students from diverse backgrounds would determine whether these findings persist in a more heterogeneous population. Additionally, this study was limited to medical students as participants, but other health professional trainees would likely benefit, and studies in the future could expand to include nursing, social work, pharmacy, genetic counseling, and physical therapy students. Future work may expand on our initial findings, investigating the best time to interweave empathy training. As healthcare transitions to more virtual visits, Zoom improv could also be used to teach skills specific to communicating and connecting with patients through a computer screen.

Conclusions

This randomized controlled study demonstrated that Zoom improv is an effective way to enhance empathy, communication, and self-reflection in medical students. Medical students across all years of training benefitted from Zoom improv. Zoom improv was protective against increasing personal distress. Medical students who participated in Zoom improv had increased perspective-taking, creativity, and understanding about power dynamics in healthcare. The Zoom format provided many advantages, including an increase in fantasy, an empathy domain that did not increase in previous in-person improv studies.