Cognitive and behavioural empathy: are 4th and 6th medical students different?

Abstract

Background

Several authors have reported that medical students’ empathy declines over time as they start their clinical work. However, such observations, performed in cross-sectional or longitudinal studies, were mainly based on self-reports measuring cognitive/affective empathy. The aim of the study was to measure both the cognitive and affective dimensions of empathy as well as and behavioural empathy between fourth- and sixth-year medical students.

Methods

Fourth- and sixth-year medical students at the Faculty of Medicine in Geneva were invited to partake to a videotaped formative objective structured clinical examination (OSCE) with standardized patients (SPs). They were asked to fill in the Cognitive and Affective Empathy Scale (QCAE) (31–124 scale) and Jefferson Scale of Empathy (JSE- Student version; 20–140 scale) measuring the cognitive/affective dimensions of empathy. Based on the videotapings, students’ verbal empathy was measured using the Verona Coding Definitions of Emotional Sequences (VR-CoDES-P). Nonverbal behaviour was assessed along five dimensions, using a 1–5 Likert scale. Differences of scores between fourth- and sixth-year medical students were analysed using analyses of variance (ANOVA).

Results

26 fourth-year and 30 sixth-year medical students participated. Their mean scores on two different self-reported cognitive/affective empathy scales were similar (mean ± standard deviation: Jefferson: 114.65 ± 8.17 vs. 116.97 ± 8.56; p = 0.308) (QCAE: 91.46 ± 9.98 and 94.93 ± 9.64; p = 0.192). There is some evidence that fourth year medical students expressed more verbal empathy in response to patients’ cues and concerns than sixth year medical students (0.28 ± 0.19 vs. 0.18 ± 0.15; p = 0.028). There was no difference in nonverbal behaviour between the two groups of students (18.90 ± 2.33 vs. 18.80 ± 2.79; p = 0.92).

Conclusions

There was no difference of scores in self-reported scales of cognitive/affective empathy between fourth and sixth-year medical students. Differences in behavioural empathy observed between fourth- and sixth- year medical students seemed to be due to changes in verbal empathy more than in nonverbal behaviour.

Introduction

Empathy plays a key role in the patient-physician communication and is central to physicians’ interpersonal competences. Expression of empathy during clinical encounters benefits both patients and physicians as it reduces patient pain and satisfaction and decreases physician burnout [1]. Although empathy is influenced by the learner’s personality traits, it has been identified as an essential element of clinical competence and professionalism in medicine that can be trained [2–4]. In medical education, it is commonly defined as a multidimensional construct including cognitive, affective and behavioural dimensions. Cognitive empathy consists of recognizing and understanding a person’s perspectives and feelings/emotions [5]. Affective empathy is defined as being emotionally sensitive and concerned for another person and refers to vicarious experience of other people’s emotions [6–8]. Behavioural empathy involves both verbal and nonverbal ways of expressing understanding, respect, and support toward another person [9]. Some authors also identify a fourth dimension—moral empathy—which closely aligns with compassion. This refers to an inner drive to care for others, motivated by concern for their well-being and a desire to alleviate their suffering through altruistic actions [10].

Several studies have reported that medical students’ empathy changes throughout medical school, especially as they start their clinical work and tends to decrease [1]. Factors influencing the development and expression of empathy have been described and relate to role models, hands-on experience, previous work experience, patients’ profile as well as stress [11]. Reasons for empathy declining have been explored and include the following elements: increased complexity in patients and diseases together, addition of clinical responsibility together with time pressure, prioritisation of biomedical knowledge, focus on acquisition of clinical reasoning and knowledge at the expenses of patient-centredness, stressful clinical environment with (sometimes) poor role models [1, 12]. These elements seem to lead to coping strategies such as cynicism, desensitisation and professional distancing among medical students [1].

However, Ferreira-Valente et al. demonstrated in a scoping review that studies on medical students’ empathy varied in their design, sample size and measures used [13]. Cross sectional studies showed higher or similar scores of empathy in later years while longitudinal studies presented mixed results or decline. Culture seems to play a role since scores of empathy decreased over years in US while they increased throughout medical school in Asian countries [13]. Interestingly, all studies relied on self-reporting questionnaires such as the JSE-S or the IRI which are known to measure the cognitive and affective components of empathy [6, 14]. Although most researchers assume a positive relationship between cognitive/affective and behavioural empathy, results are mixed when measures of behavioural empathy are used [15–19]. Behavioural empathy can be measured through external global ratings, analysis of verbal interactions in dyadic clinical communication, cognitive tasks that assess individual’s sensitivity to suffering of others [17, 19–23]. Assessment of behavioural empathy is important since because it is this dimension that finds expression during clinical encounters and will impact patients’ outcomes such as satisfaction and anxiety [24, 25]. Knowing how to behave does not necessarily mean behaving in an adequate manner in day-to-day practice. Little is known about the extent to which cognitive/affective empathy measures throughout student self-assessment reflect behavioural empathy and whether expression of behavioural empathy differs between medical students according to their level of clinical training.

The first aim of the study was to compare measures of cognitive/affective and behavioural empathy among medical students. The second aim was to evaluate whether these measures differed between fourth- and sixth-year medical students.

Methods

Setting and participants

This study took place at the Faculty of Medicine of the Geneva University in Geneva, Switzerland in 2022. This Faculty offers a six-year curriculum (3 pre-clinical and 3 clinical years) to 160 medical students per year. Basic communication skills training including empathy takes place during the 2nd and 3rd pre-clinical years. During clinical years which are composed of 1 or 2-month clinical rotations, students attend seminars regarding more complex communication issues such as shared decision making, delivering bad news and motivational interviewing. With 60 h dedicated to communication skill training (sometimes included in more global clinical skills training), it represents 1.1% of the total curriculum [26]. Undergraduate medical training concludes with a licensing exam which consists of a written exam and 12 OSCE stations in which communication is assessed.

All fourth-year medical students received an email inviting them to attend an optional formative OSCE as part of a two-months clerkship in primary care. All sixth-year students received an e-mail inviting them to take part into an optional formative objective structured clinical examination (OSCE) a month before the final licensing exam in August 2022. Participating students gave a written and informed consent to the study. The study project was approved by the University of Geneva’s Ethics Committee (CUREG MM 22-01-07).

Development of formative OSCE

Author NJP created four clinical cases, two were designed for fourth year students’ level (subacute abdominal pain and thoracic pain) and two others for sixth year students (subacute dizziness and headache). The clinical cases were adapted to the students’ level of clinical skills training and portrayed clinical situations commonly encountered in everyday practice. In all four cases, patients presented with non-specific symptoms that did not indicate a clear diagnosis or life-threatening disease, and they shared similar socio-professional characteristics, worries and sources of stress. Five female standardized patients (SPs) were trained for two hours prior each session to simulate a 30-year-old patient (a single scenario for each patient) and respond in a standardized way to students’ queries. A special focus was put on the number and type of cues and concerns to express during the encounter. We included only female SPs because female patients have shown to express more emotional intensity than male patients and we wanted to facilitate students’ identification of patients’ emotional distress [27]. SPs were unaware of the study aims and were not informed about the students’ level of training.

Procedure

The students attended an 18-minute formative OSCE that was filmed. After the formative OSCE, the students and SPs completed self-questionnaires on an online platform [28]. The filmed interaction was used to assess behavioural empathy (Fig. 1).

Fig. 1.

Study design, participants and outcome measures. * OSCE=objective structured clinical encounter. ** JSES-S=Jefferson scale for students; QCAE=Questionnaire of Cognitive and Affective Empathy

Outcome measures

Students’ self-reported cognitive/affective empathy

We chose two scales to measure this dimension. The first one was the Jefferson scale for students (JSE-S) − 20 items on a 1–7 Likert scale including: ten items assessing cognitive empathy (perspective taking), eight items measuring affective empathy (compassionate care) and two items referring to “standing in patients’ shoes” [5, 14]. This scale was developed to assess both the cognitive and the affective domains of empathy but mostly emphasizes the cognitive dimension. We used a French validated version [29]. The second one was the cognitive and affective empathy scale (QCAE) [8]. This questionnaire, initially tested among a population of university students and employees has been validated in French [30]. It reliably assesses the different components of cognitive (19 items, Cronbach’s alpha =.83) and affective empathy (12 items, Cronbachs’ alpha =.79) and their sub-components, on a 1–4 Likert scale. Sub-components of cognitive empathy are perspective taking (putting one-self in another person’ shoes) and online simulation (to put oneself in another person’ position by imagining what that person is feeling); sub-components of affective empathy relate to emotion contagion (automatic mirroring of others’ feelings), proximal responsivity (affective response when witnessing the mood of others in a close social context) and peripheral responsivity (affective response when witnessing the mood of others in detached context) [8]. It has been used to assess medical students’ empathy [31]. We chose these two scales because in one study, medical students’ scores evolved differently over time according to the scale used [31]. Some authors suggested that the scales might measure different dimensions of medical students’ cognitive/affective empathy [32].

Student’s behavioural empathy

Global scores of perceived or self-perceived empathy

We asked medical students, SPs and an external observer to evaluate empathy using a 1–10 Likert scale as a global measure of behavioral empathy at the end of the station: “How would you rate the student’s empathy (verbal and nonverbal)? (circle the corresponding number 1 = not at all and 10 = totally)”.

Verbal empathy

Verbal empathy can be defined as acknowledging/reflecting patients’ feelings and legitimizing them [33]. Verbal empathy was assessed using the Verona Coding Definitions of Emotional Sequences, developed by the Verona Network on Sequence Analysis. This system is used to code concerns (a clear and unambiguous expression of an unpleasant current or recent emotion where the emotion is explicitly verbalized) and emotional cues (a verbal or nonverbal hint which suggests an underlying unpleasant emotion and would need a clarification from the health provider) [34, 35]. This tool codes the way in which the students respond to these concerns or cues (unit of analysis = any turn or part of turn said by the student). Cues and concerns are either elicited by health providers (students in this case - they may be given as response to health provider’s questions or statements) or by patients (SPs in this case – introduced by the patients without having been solicited or invited to do so). Health professionals’ answers are neutrally coded as explicit - any response which specifically mentions either the content/topic or the emotion in the cue or concern or both or non-explicit - any response which does not explicitly mention either the content or the emotion of the cue or concern as specified above; providing space - gives space for further disclosure of the cue/concern vs. reducing space - reduces the space for or closes down further disclosure (Fig. 2) [34].

Fig. 2.

Providers’ answers to patients’ cues and concerns adapted from the VR-CoDES-P coding manual (in bold: space providing; in white: verbal empathy) [34]. * Refers to the topic of the cue or concern. ** Refers to the affective or emotional aspect of the cue or concern

Similarly to other dyadic coding systems [17], verbal empathy is not strictly operationalized as a code in the VR-CoDES-P. Students’ verbal empathy can be specifically considered when the student explicitly acknowledges the affect related to SPs’ cues and concerns (“you are worried”), explicitly expresses empathy (“I imagine that this must be really hard for you, especially as you are so afraid of …”), or acknowledges and legitimate SPs’ cues and concerns or emotions in an implicit way (“I see”,” I understand”) [33, 36] (Fig. 2). Students’ responses so SPs’ cues and concerns as well as manifestations of verbal empathy are expressed in proportions (i.e. 0.26 = 26% of the total number of responses to SPs cues and concerns during the clinical encounter). As indicated at the right side of Fig. 2, other responses included exploring, switching, postponing, silence, active invitation etc….

Nonverbal behaviour

Nonverbal behaviour was coded using the Verona Coding Definitions of Emotional Sequences (based on the Verona manual) which included five dimensions: eye contact, facial expression, head movement, bodily posture and tone of voice using a five item Likert scale (1 = poor and 5 = rich) [37].

The filmed student-SP interactions were coded using The Observer Video Analysis Software XT X [38].

Analysis

Coding

NJP and AH first coded two videos together, then coded height videos separately, followed by a comparison of their respective coding to calibrate their coding. Cohen’s Kappa coefficients, used to check inter-coder reliability on a subsequent sample of five videos, were excellent (k = 0.87for cues/concerns and 0.80 for provider responses). For nonverbal behaviour, the intercoder reliability obtained on 10% of the videos was also good (intraclass correlation = 0.872).

Statistical analysis

The potential difference between fourth- and sixth-year medical students regarding the different scores (global empathy, Jefferson and QCAE and the corresponding sub-scores, nonverbal behaviour), and of the emotional cues and concerns was investigated with one-way (group of students i.e. either fourth-year or six-year year) analyses of variances (ANOVA). Statistical analyses were performed by BC on R 4.3.3 (the R Foundation for Statistical Computing, Vienna, Austria). Unless specified, reported results indicate the mean ± standard deviation.

Results

Out of 58 students who volunteered to participate, 56 students attended the formative 18-minutes OSCE with a SP (26 4th year – female = 19 (73%) and male = 7 (27%); 30 6th year medical student – female = 25 (83%) and male = 6 (17%). Case distribution was the following: abdominal pain n = 16, thoracic pain n = 10, headache n = 15 and dizziness n = 15. Sixteen interactions required mask wear because of Covid19 institutional requirements prevailing at the beginning of 2022. The nonverbal dimension “facial expression” was then not coded for these interactions.

There was no difference in global empathy perceived by students (Table 1) or an external observer between the two groups, but there is some evidence that SPs perceived 6th year medical students to be more empathetic than 4th year medical students (6.77 ± 2.30 vs. 7.97 ± 1.83; p = 0.035) (Table 1). Similarly, there was no intergroup difference in the JSE-S and QCAE scores as well as on sub-dimensions of these scales except for perspective taking (QCAE) for which sixth-year medical students scored higher than fourth-year medical students (27.96 ± 4.04 vs. 30.37 ± 4.29; p = 0.036).

Table 1.

Students’ JSE-S scores, patients’ cues and concerns and students’ responses (ratio) to cues and concerns

Scale (min-max) or number of occurrences	4th year medical students N = 26	6th year medical students N = 30	p- value
Global empathy (1–10)
- Student	7.19 (1.06)	7.40 (0.86)	0.421
- SPs	6.77 (2.30)	7.97 (1.83)	0.035
- External observer	6.11 (2.07)	5.70 (1.64)	0.406
Self-reported cognitive/affective empathy	Mean (SD)	Mean (SD)
Jefferson mean score (20–140)	114.65 (8.17)	116.97 (8.56)	0.308
- Perspectives (10–70)	57.65 (5.84)	59.33 (6.23)	0.305
- Compassion (8–56)	48.62 (3.81)	49.07 (3.32)	0.637
- Shoes (2–14)	8.38 (3.37)	8.57 (2.18)	0.809
QCAE mean score (31–124)	91.46 (9.98)	94.93 (9.64)	0.192
- Cognitive empathy (19–76)	55.31 (6.09)	57.37 (6.52)	0.230
- Perspective taking (10–40)	27.96 (4.04)	30.37 (4.29)	0.036
- Online simulation (9–36)	27.35 (3.45)	27.00 (3.32)	0.704
- Affective empathy (12–48)	36.15 (5.75)	37.57 (4.85)	0.323
- Emotion contagion (4–16)	12.12 (1.90)	12.74 (2.02)	0.246
- Proximal responsivity (4–16)	12.57 (1.78)	12.60 (1.88)	0.899
- Peripheral responsivity (4–16)	11.50 (3.00)	12.23 (2.08)	0.287
Patients’ cues and concerns	Mean (SD)	Mean (SD)
Mean cues total	4.58 (2.06)	5.47 (1.74)	0.085
- Student elicited	2.81 (1.50)	2.80 (1.49)	0.985
- Patient elicited	1.77 (1.53)	2.67 (1.30)	0.021
Mean concerns	3.15 (1.19)	2.60 (1.10)	0.076
- Student elicited	1.88 (1.14)	1.53 (0.90)	0.204
- Patient elicited	1.27 (1.00)	1.07 (1.01)	0.457
Students’ responses to SPs’ cues and concerns	Ratio (SD)	Ratio (SD)
- Reducing Space	0.39 (0.19)	0.54 (0.22)	0.00948
- Providing Space	0.61 (0.19)	0.46 (0.22)
Explicit Providing space	0.48 (0.18)	0.37 (0.19)	0.0341
- Explicit Providing Space Content	0.26 (0.16)	0.19 (0.12)	0.0994
- Explicit Providing Space Affect	0.23 (0.15)	0.18 (0.15)	0.26
Behavioural empathy	Ratio (SD)	Ratio (SD)
Verbal Empathy	0.28 (0.19)	0.18 (0.15)	0.0276
- Explicit acknowledging	0.11 (0.10)	0.05 (0.07)	0.0136
- Explicit empathy	0.08 (0.11)	0.04 (0.07)	0.0933
- Non explicit acknowledging	0.03 (0.08)	0.03 (0.05)	0.631
- Non explicit empathy	0.06 (0.08)	0.06 (0.10)	0.876
Nonverbal behaviour	Mean (SD)	Mean (SD)
Nonverbal total with facial expression (5–25)	18.90 (2.33)	18.80 (2.79)	0.918
Nonverbal total without facial expression (4–20)	15.04 (1.68)	15.33 (2.29)	0.591
- Non verbal - eye contact (1–5)	4.50 (0.71)	4.37 (0.93)	0.553
- Non verbal - facial expression (1–5)	3.70 (0.82)	3.47 (0.68)	0.379
- Non verbal - head movement (1–5)	3.62 (0.90)	3.93 (0.87)	0.184
- Non verbal - body position (1–5)	3.23 (0.51)	3.47 (0.68)	0.155
- Non verbal – tone of the voice (1–5)	3.69 (0.79)	3.57 (0.73)	0.538

There was no difference in SPs’ global expression of cues and concerns during the fourth and sixth year formative OSCE (but there is some evidence that cues were more often elicited by SPs with sixth year medical students than with fourth-year medical students (1.77 ± 1.53 vs. 2.67 ± 1.30; p = 0.021).

Fourth year medical students used a higher percentage of responses providing space when reacting to patients’ cues and concerns than 6th medical students (61% vs. 46%; p = 0.00948) (Table 1). However, there was no clear difference between the two groups in the type of responses used to address either the content (0.26 ± 0.16 vs. 0.19 ± 0.12; p = 0.0994) or the affect (0.23 ± 0.15 vs. 0.18 ± 0.15; p = 0.260) related to the cues and concerns.

There is some evidence that fourth-year medical students expressed a higher proportion of verbal empathy compared to sixth-year medical students (0.28 ± 0.19 vs. 0.18 ± 0.15; p = 0.0276) by using more explicit acknowledgement of patients’ emotions. Finally, there was no difference in nonverbal behavioural between the two groups of students (Table 1).

Discussion

This study shows that fourth- and sixth-year medical students did not have different cognitive/empathy scores during a formative OSCE, whatever the type of scale used. Regarding behavioural empathy, fourth- year medical students expressed a higher percentage of verbal empathy than sixth- year medical students, while simulated patient expressed the same number of cues or concerns. However, SPs perceived sixth-year medical students to be more empathetic than fourth- year students. There was no difference in non-behavioural communication between the two groups.

Medical students’ scores of cognitive/affective empathy (JSE) measured in this research were quite similar to scores reported in other studies but higher than scores of other Swiss medical students [22, 39–41]. This difference may be explained by the fact that the percentage of female students included in the study was higher than in the student population of our medical school. Indeed, female students are known to score higher on empathy scores [42, 43]. In contrast to studies which showed differences between JSE and QCAE scores [31], there were no differences of scores between these different scales who have been both used to measure cognitive/affective empathy among medical students. In addition, there were no differences of scores between fourth and sixth year medical students. These findings are aligned with other studies including a Swiss study which showcased a linear increase in JSE and QCAE scores from the beginning to the fourth year followed by a plateau effect between the fourth and sixth year [44, 45]. However, they contrast with some studies showing either an increase or a decline in cognitive empathy [13, 46]. As suggested by Ponnamperuma, these variations in scores across the world suggest that there may be a geo-sociocultural pattern in empathy curves according to the context the measures take place [46].

Although cognitive empathy seems to predict clinical performance, especially related to communication [47], reliance on self-perceptions of empathy is largely insufficient to assess students’ empathy during encounters with standardized or real patients. Very few studies have used objective measures such as analysis of sequences of verbal interactions to assess behavioural empathy [17, 19] We are unaware of studies which measured differences in expressions of behavioural empathy between junior and senior students. A study showed that there was no difference in providing space and verbal empathy amongst students scoring high or low on self-reported cognitive empathy [19]. Our findings suggest that there are some differences between junior and senior medical students with fourth-year medical students more often opening space in response to SPs’ cues and concerns and expressing verbal empathy than sixth-year medical students. Reasons for such differences have been mentioned in the introduction and may relate to coping strategies such professional distancing and desensitisation that students acquire as they are confronted to more complex clinical patient situations, higher responsibility and stressful environments, and poor role models. Senior students may also have learned to work faster during clinical encounters. Differences may also be linked to the type of station, the timing of the formative OSCE or the play of the SPs involved. Although patients’ cues, concerns and psychosocial context were similar, SPs presented different complaints to fourth year and sixth-year medical students (cardiac or abdominal vs. neurological complaints). Neurological complaints are usually more challenging in terms of history taking, physical and clinical reasoning skills than cardiac and abdominal complaints and 6th year medical students may have been forced to focus more on biomedical issues because of time constraints. The timing of the formative OSCE may have had some influence as it took place a month before the final high-stakes exam for sixth year medical students while fourth year students attended the OSCE during their first year of clinical clerkships at distance of any exam. It is possible that sixth year medical students focused more on diagnostic and treatment issues than patient perspectives as the OSCE was close to the final high-stakes exam. Finally, although SPs were trained to provide standardized responses to students’ queries, their reactions may have differed as three interacted exclusively with fourth-year or six-year students while two were involved with both groups.

Regarding nonverbal manifestations of empathy, we showed in this study that nonverbal behaviour did not change between fourth- and sixth-year medical students. Nonverbal behaviour can commonly be divided into speech-unrelated (facial expression, gazing, nodding and posture) and speech-related (tone of voice) behaviours [48, 49]. Several studies have shown the importance of nonverbal cues as signs of empathy in physician-patient communication [50, 51]. Other studies have reported that physician’s nonverbal expressiveness correlates with higher patient satisfaction [48, 49, 52]. Finally, a few studies have shown correlations between nonverbal behaviour such as eye gaze and cognitive empathy (measured by JSPPPE scores) while others did not [53–55]. The fact that scores of nonverbal behaviours were generally average except for eye gazing suggest that more attention could be paid to improving nonverbal empathy during medical training. Indeed, much focus is put on the impact of verbal communication and empathy on physician-patient interactions and educational strategies tend to primarily focus on improving verbal empathy [54, 56].

Finally, the fact that SPs perceived senior students to be more empathetic is of importance since the way patients perceive students’ attitude and empathy will impact on the therapeutic alliance and potential clinical outcomes. Reliance on medical students’ self-perceptions of empathy is hazardous as several studies reported the absence of correlation between self-assessment of empathy by physician and patient assessment [57, 58]. Involving patients in assessment of students’ empathy and allowing them to share their perspectives seems to be complementary to other measures of assessment since it provides the opportunity for students to discover the singularity of patients’ experiences and allows to test the impact of the training interventions regarding empathy and communication [58, 59]. Giving voice to real patients would be even more powerful. However, patients from various backgrounds should be regularly integrated in this process since their perceptions of empathy may vary, influenced by several factors that are not specifically put forth in communication skills training.

Limitations

The fact that study included a small sample of students who partook to formative station on a voluntarily basis and involved a higher percentage of women than the student gender distribution at the Geneva Faculty of Medicine might have had an impact on the results. The fact that all SPs were female and that we used only four clinical cases may influence the generalizability of our findings. However, our aim was not to establish external validity but to focus on assessing empathy within a controlled setting.This was a cross-sectional study, and a longitudinal investigation assessing students’ empathy with a two-year interval would have provided more robust insights.A scoping review published in 2017 revealed that cross-sectional studies showed higher or similar scores of empathy while longitudinal studies revealed mixed results or empathy decline [13]. However, these findings were not confirmed in a subsequent review published in 2020 [42]. Furthermore, in European settings, empathy does not seem to diminish during the medical curriculum, whatever the design used [60]. Finally, the study started at the beginning of 2022, when institutional requirements regarding mask wear were still in enforcement (COVID 19 pandemic): the analysis of the students’ facial expression was not possible for 16 interactions (involving fourth year medical students) and might have been a source of additional biases regarding the measure of nonverbal behavior.

Conclusions

Measurement of empathy is complex since it is a construct with cognitive, affective, behavioral dimensions that can be assessed through different perspectives. This study indicates that perceptions of empathy varied according to the type of measure used but there was no clear trend towards decline of students’ empathy over time. Reliance on self-assessment of empathy to measure the effect of a medical curriculum on empathy is insufficient since it does not seem to constantly reflect behavioral empathy. Objective analysis of behavioral empathy could sometimes be used to revisit the link between oneself, patient and external observers’ ratings. More attention should definitely be given to nonverbal behavior given its impact on patient perceptions of empathy in physician-patient communication. Finally, inclusion of patients’ perspective is crucial to ensure that empathy training meets patients’ needs.

Abbreviations

JSE-S

Jefferson scale for students

OSCE

Objective structured clinical examination

QCAE

Questionnaire of Cognitive and Affective Empathy

RIAS

Roter Interaction Analysis System (RIAS)

SPs

Standardized patients

VR-CoDES-P

Verona Coding Definitions of Emotional Sequences

Authors' contributions

NJP, AH & BC conceived and designed the study; NJP, AH and LH collected the data; NJP, AH and BC analysed and interpreted the data; NJP drafted a first version of the article; AH, BC and RL critically revised the article. All authors have read and approved the manuscript.

Funding

Open access funding provided by University of Geneva. The study was supported by the Geneva Faculty of Medicine (Fonds Mimosa). Clinical trial number: not applicable.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study project was approved by the University of Geneva’s Ethics Committee (CUREG MM 22-01-07). Informed consent was obtained from all the participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.