Communication skills training and the conceptual structure of empathy among medical students

A total of 295 fourth-year Japanese medical students from three universities completed the Jefferson Scale of Empathy and a newly developed willingness to show empathic behaviour scale twice after communication skills training, pertaining to post-training and retrospectively for pre-training. In other words, students completed the questionnaire once for post-training and once retrospectively on their empathy before training. While the retrospective pre/post self-report approach comes with the limitations of possible memory effects and participants’ tendency to think that some change—preferably learning—must have taken place (i.e., good subject effect) [20], it has the advantage over an actual pre/post design that it does not suffer from a response shift bias due to a relative inability to assess one’s own ability when prior knowledge or experience is lacking [21]. One of the consequences of such a response shift bias could be that students give lower post-OSCE than pre-OSCE ratings on perspective taking, compassionate care or willingness to show empathic behaviour because the OSCE has made them more aware of how much they can still grow rather than that the OSCE results in an actual decrease in any of these factors.

The OSCE is a common achievement test conducted every year in all medical schools in Japan, usually in the late fourth year in a six-year curriculum, to assess students’ clinical competencies. Most medical schools, including the three medical schools that participated in the current study, teach communication skills to fourth-year students using simulated patients in several medical interviewing training sessions before the OSCE takes place. Questionnaire administration took place right after the OSCE.

We used the Japanese student version (i.e., S‑Version) of the Jefferson Scale of Empathy [18]. This questionnaire has 20 items, each to be rated on a seven-point Likert scale. In previous studies, factor analysis revealed three factors: perspective taking (10 items), compassionate care (8 items) and standing in the patient’s shoes (2 items) [22]. As we were interested in how perspective taking and compassionate care relate to willingness to show empathic behaviour, we developed a new three-item scale that aimed to measure students’ willingness to show empathic behaviour and fit in the context of the training. The training has been a regular part of the medical curriculum in the universities that participated in this study. Face validity of the willingness to show empathic behaviour items was acquired from the experiences and comments from tutors engaged in the training as well as from students. Table 1 presents the entire questionnaire of perspective taking, compassionate care and willingness to show empathic behaviour items.

Additionally, we included one open-ended question: ‘How did medical interviewing training sessions and the OSCE influence your communication skills in terms of empathy?’ Finally, we also asked students to indicate the number of medical interviewing training sessions they had attended.

Although 295 students participated in the study, seven students (2.4% of 295 students) completed fewer than 53% of the items. Furthermore, of the 288 students who completed the questionnaire (217 men, 66 women, and 5 did not specify; average age 23.1 years, eight students did not specify), one student omitted 13% of the items, one student omitted 11% of the items, 20 other students omitted less than 7% of the items, and the remaining 268 students responded to all items.

To examine the relation between perspective taking (10 items), compassionate care (8 items) and willingness to show empathic behaviour (3 items), we performed confirmatory factor analysis in Jamovi 0.8.1.18 [23] , conform the questionnaire structure displayed in Table 1, with full information maximum likelihood estimation for missing response [24], for both post-training and pre-training, respectively. Moreover, to gain an additional understanding of the relation between the number of medical interviewing training sessions and each of perspective taking, compassionate care and willingness to show empathic behaviour, we performed linear mixed-effects analysis in Jamovi 0.8.1.18 [23] , treating retrospective pre-training and post-training as repeated measurements, the number of medical interviewing training sessions as between-subjects factor and student-level intercept as random effect. The latter was based on only 234 students (79% of 295 students) due to non-response of the other students to the number of medical interviewing training sessions and only one student indicating 0 medical interviewing training sessions. Thus, the number of medical interviewing training sessions consisted of three categories: 1 (n = 31), 2 (n = 135) and 3 (n = 68).

Finally, responses to the open-ended question on the self-perceived influence of the number of medical interviewing training sessions and OSCE on their communication skills in terms of empathy were subjected to qualitative descriptive analysis [25] to check for recurring themes that could help triangulate the findings from confirmatory factor analysis and/or linear mixed-effects analysis. First author DS conducted the initial coding and three other authors (IS, HI and MA) reviewed that coding. Disagreements were discussed until consensus was reached.

Students received an informed consent form, and those who understood the purpose of the research and agreed to participate received the questionnaire. Questionnaires were administered anonymously. Participants were informed that the survey was not mandatory and had no relation to grading or study progress. The study was conducted under the approval of the Institutional Review Board of the University of Tokyo, Shinshu University, and Nagoya University, where the study was implemented.

The high loadings for willingness to show empathic behaviour indicate that the attempt to create a unidimensional willingness to show empathic behaviour scale has succeeded. The correlation between perspective taking and compassionate care was 0.494 for retrospective pre-training and 0.469 for post-training. The correlation between perspective taking and willingness to show empathic behaviour was 0.631 for retrospective pre-training and 0.699 for post-training. The correlation between compassionate care and willingness to show empathic behaviour was 0.244 for retrospective pre-training and 0.334 for post-training.

Table 4 shows the mean scores and standard deviations of perspective taking, compassionate care and willingness to show empathic behaviour for retrospective pre-training and post-training, and for each number of medical interviewing training sessions (i.e., 1, 2 or 3).

Although there was a modest increase in perspective taking (95% confidence interval: 0.057–0.150 points), compassionate care (95% confidence interval: 0.150–0.240) and willingness to show empathic behaviour (95% confidence interval: 0.179–0.333) from retrospective pre-training to post-training, both the main effect of the number of medical interviewing training sessions and its interaction with occasion were small and not statistically significant for any of these factors. In other words, we failed to find evidence for any relation between the number of medical interviewing training sessions and perspective taking, compassionate care or willingness to show empathic behaviour.

Table 5 highlights the 16 themes that emerged, with 12 themes on positive aspects of learning and 4 themes on difficulties of learning.

On the positive aspects, students reported among others ‘learning how to show empathic attitudes to patients’ and ‘recognizing that communications skills are learnable.’ Recurring difficulties were the ‘difficulty of empathizing with patients’ feelings’ and the ‘difficulty of feeling compassion despite showing empathic behaviour’. Some students elaborated further. For instance, one student proceeded that ‘even if I try to show compassion, sometimes I cannot care about it due to my lack of skills,’ and another student wrote ‘I learned some skills of interviewing patients, but I could not empathize with patients very well; I felt like I had lied when I said emphasizing words to them.’ Taken together, these findings from the open-ended responses appear to indicate that students experience difficulties in empathizing with patients emotionally yet do manage to acquire the cognitive skills that are needed to communicate with patients. This interpretation is in line with the confirmatory factor analysis outcomes in that the correlation between compassionate care and willingness to show empathic behaviour was substantially lower than the correlation between perspective taking and willingness to show empathic behaviour.

Future studies could replicate the current study in a different stage of the medical curriculum and subsequent career as well as in different cultural settings. The associations of medical students’ empathy scores with outcomes such as professionalism, empathic behaviour in later years of the curriculum, or patient satisfaction and compliance may have been studied quite well in Western contexts [16, 28‐30] but remain relatively underexplored in non-Western cultures. In the Japanese context, empathy is also considered a critical component of professionalism, especially linked with the Bushido concept of jin or benevolence [31]. However, future studies are needed with regard to how Japanese physicians or medical students show empathic behaviour to patients and how this is linked to patient outcomes.

Apart from considering cultural context, some new studies should focus on the longitudinal development of perspective taking, compassionate care and willingness to show empathic behaviour. The current study provided some indication that fourth-year medical students may not be able to care a lot about showing compassion because they lack skills that may be needed to facilitate perspective taking. With their rather preliminary cognitive schemas of the type of cases under consideration, the average fourth-year medical student may need to allocate their working memory resources to cognitive processes that are needed for perspective taking to such an extent that fewer resources are available for dealing with emotion or compassionate care [32, 33]. Although this may sound odd at first, research has demonstrated that emotion in patient encounter may—certainly in the early stages of learning—consume working memory resources that are needed for cognitive processes related to learning [34, 35]. Longitudinal studies could provide more insight into how cognitive and emotive processes affect students’ willingness to show empathic behaviour as they learn and enter subsequent stages of training.

Another factor that may influence students’ focus or lack of focus on compassionate care, even when their cognitive schemas are sufficiently developed to engage in compassionate care, is that of the assessment context. There is some evidence that adding a hypothesis-driven component to OSCEs can foster the engagement of clinical reasoning when preparing for an OSCE if students expect that they are going to be assessed on clinical reasoning as well [36]. Future experimental studies could test to what extent such (pre-)assessment effects also occur when adding a compassionate care component to the OSCE assessment. Given the apparent importance of developed cognitive schemas, one might expect an interaction between cognitive schemas and assessment context: the assessment context might not matter that much among students whose cognitive schemas are still limited (e.g., fourth-year medical students), but might well make a difference at a later stage. If so, the interaction might explain why in a previous study fifth-year medical students weighed more on the emotive component of empathy, whereas residents weighed more on the cognitive component of empathy [15]. If the students in that study expected to be assessed predominantly on the cognitive component, any focus on the emotive component might be extraneous activity in the context of their preparation for that assessment.

The training in the current study has been a regular part of the medical curriculum in the universities that participated in this study and, as in the vast majority of cases, did not undergo additional validation. Future studies could consider alternative approaches to training such as the Process Communication Model ® [37, 38]. Based on motivational and personality theory, this model provides an approach to teaching communication and training communication skills.

Finally, our willingness to show empathic behaviour items did not undergo validation additional to the establishment of face validity. Although the willingness to show empathic behaviour items developed for this study fit in the context of the training of focus in this study and are easy to administer, this set of items could be developed further and adapted for different contexts. Additionally, future studies should consider using direct observation and other alternatives to self-reported willingness to show empathic behaviour.