The skill and knowledge required of doctors are rapidly increasing, and training in professionalism is regarded an important aspect of becoming a doctor. One of the foundations supporting this professionalism is communication skills (Stern, 2006). An attitude of empathy that endeavors to understand things from the patient’s viewpoint is particularly important in communication in healthcare, and efforts to foster this attitude are being incorporated into medical education (Pedersen, 2010).

The primary context for communication between doctors and their patients is the medical interview, and role-playing with simulated (i.e., standardized) patients (SPs) is an effective technique for training and evaluation (Rees, Sheard and McPherson, 2002; Chan et al., 2003; Yedidia et al., 2003). The development of an empathetic attitude is one aspect of medical education, and empathy training bridges the gap between biomedicine and the humanities (Pedersen, 2010). However, although diverse methods have been used to investigate empathy, no consistently informative measure (‘gold standard’) is available (Eikeland et al., 2014).

Both Japanese and overseas studies have demonstrated gender-associated differences in doctors’ communication abilities and styles. For example, female doctors typically exhibit greater empathy than do male doctors (Roter and Hall, 2004; Swygert et al., 2012; Shin et al., 2015; Vogel, Meyer and Harendza, 2018). However, targeted medical interview training throughout medical school may help decrease this gender-associated gap (Harada et al., 2018).

Objective structured clinical examinations (OSCEs) are widely used to evaluate medical students’ acquisition of proficiencies that are difficult to assess through written or oral tests, including decision-making, clinical skills, bedside manner, and medical interviewing (Harden et al., 2016). In most countries, OSCEs are included in the examinations for graduation or for national medical licensure. In an OSCE, whether the examination candidate displayed an empathetic attitude toward the SP is gauged by a third-party assessor, who applies various defined criteria, chiefly whether the candidate used empathetic phrases when talking with patients. However, a study of evaluations by two different assessors (Kubota et al., 2009) revealed numerous inconsistencies between their evaluations of “listening” and “empathy.” The basis for empathy evaluation in the OSCE should not rest solely on arbitrary judgments by an assessor, and specific parameters for observation during OSCEs should be identified to develop a more objective structure for evaluation. Several scales have been developed for the evaluation of communication, including conveying empathy, in medical interviews. The best known of these, the Kalamazoo consensus statement (Makoul, 2001), was written by a panel of experts. Other scales have similarly been established on the basis of specialist opinions (Mercer et al., 2004; Humphris and Kaney, 2001) or literature reviews (Kane et al., 2007; Kurtz and Silverman, 1996; Huntley et al., 2012; Ishikawa et al., 2006). However, none of these previous studies describe the development of a patient-centered scale. The objective of the current study was to identify factors that influenced whether trained simulated (i.e., standardized) patients (SPs) felt that they had been treated empathetically during medical interviews conducted by medical students.

1. Development of questionnaire

We had developed the questionnaire through semi-structured interviews as follows asking SPs to describe when the student had been empathetic (Harada et al., 2018).

1-1. Subjects

In Japan, undergraduate medical education is six years long. Fifth-year Japanese medical students undergoing clinical training in the Department of Cardiology of Tokyo Medical University took part in medical interview practice with SPs. Standard signs and diseases not limited to cardiology were included in the scenarios used. Role-playing during training interviews was done in groups of 7 or 8 people, with 2 students in each group taking the doctor’s role. The study subjects were 18 individual SPs, who participated in 34 SP interviews with 34 students in 17 groups between 30 March 2015 and 16 February 2016.

1-2. Semi-structured interviews

After the role-playing sessions, each SP was asked (1) whether the student had been empathetic (i.e., had treated the SP empathetically) during the medical interview, and (2) at what points the SP felt that the student had displayed empathy.

1-3. Analysis and questionnaire development

The audio data from the interviews were transcribed, coded, and categorized according to the standard method of content analysis. A questionnaire was developed that asked about each of the items on a 5-point Likert scale, with responses ranging from 0 to 4 (Supplementary File 1).

2. Questionnaire-based survey

2-1. Subjects

The subjects of the survey were 36 SPs who participated in practice interviews with a total of 36 fifth-year medical students. Role-playing and training were done in 18 groups between 12 April 2016 and 14 February 2017.

2-2．Survey process

After the training sessions had been completed, the developed questionnaire (Section 1-3) was distributed to the SPs, with the request to complete it within 2 weeks. After 2 weeks, we collected all completed questionnaires.

2-3. Analysis

The reliability of the questionnaire was investigated by using Cronbach’s alpha coefficient, and Spearman’s correlation coefficient was calculated to assess the association between each item and the statement “I felt that the student was empathetic throughout the interview.” Multiple regression analysis was performed also. Statistical analysis was performed by using IBM SPSS Statistics 24, and a P value less than 0.05 was regarded as significant.

1. Development of questionnaire

During fifth-year training in the 2015 academic year, 34 Japanese students participated in medical interviews with 18 individual SPs in a total of 34 SP interviews. During semi-structured interviews the SPs responded that “I felt that the student was empathetic throughout the interview” for 13 (38%) of those 34 interviews. The SPs’ impressions regarding the statement that “I felt that the student was empathic throughout the interview” were classified into the following categories: “I was given enough time to talk at the beginning of the interview,” “The volume and clarity of the student’s voice were appropriate,” “The student made good eye contact,” “The student’s attitude was appropriate,” “I was able to say enough of what I wanted to talk about,” and “The student used phrases that expressed empathy.”

2. Questionnaire-based survey

Responses were obtained from 36 SPs who had participated in staged medical interviews with 36 medical students. The students received high scores overall, with a few students receiving low scores for some items. The median scores and interquartile ranges for each questionnaire item are shown in Figure 1.

Figure 1: Results of questionnaire survey - Contents of question items and evaluation from simulated patients (SPs).

Multiple regression analysis demonstrated that “The student used phrases to express empathy” was the only independent predictor of “I felt that the student was empathic throughout the interview” (P < 0.001) (Figure 2). 

Figure 2: Results of questionnaire survey - Correlation between each question item and empathy.

Cronbach’s alpha coefficient (0.827) showed good internal consistency.

We investigated factors from the patient (SP) perspective that influenced whether SPs felt that fifth-year medical students in Japan were empathetic during staged medical interviews. Multiple regression analysis demonstrated that “The student used phrases to express empathy” was the only independent predictor of “I felt that the student was empathic throughout the interview.” However, in some interviews, the SPs felt that the students had displayed empathy despite not using empathetic phrases. This finding suggests that empathy cannot be evaluated merely by measuring the extent to which it is communicated verbally. Likewise, during the semi-structured interviews that were used to produce the questionnaire, SPs mentioned various additional nonverbal expressions through which they felt that a student displayed empathy, including “I was given enough time to talk at the beginning of the interview” and “The student made good eye contact.”

Pedersen (2010) reported that as medical students go through their educational program, their frequency to empathize decreases, and in the later years of training, their interest tends to focus more intently on the scientific aspects of medicine. Given that fifth-year medical interview training was conducted during clinical training, students in this study may have been biased toward soliciting medical–scientific information, or they may have experienced the fact that sufficiently empathetic medical interviews are not typically conducted in clinical practice (Swygert et al., 2012). Eikeland et al. (2014) reported that the hidden curriculum of actual clinical practice inhibits empathy.

Diverse methods have been used to investigate empathy, but no consistently informative measure (‘gold standard’) has emerged yet (Eikeland et al., 2014). Scales for objectively evaluating the empathy of medical students (Hojat et al., 2001) and parameters for assessing nonverbal communication in medical interviews (Sugawara et al., 2017) have been proposed. The consultation and relational empathy measure (CARE) has been developed as a scale for the qualitative and quantitative measurement of empathy in actual clinical settings (Mercer et al., 2004). The reliability of CARE has been investigated in Japanese populations as well as in predominantly white populations in the United Kingdom (Matsuhisa et al., 2018). However, CARE includes items regarding explanations and treatments, thereby making it unsuitable for evaluating student medical interviews, and its content is somewhat abstract. The proposed Jefferson Scale of Patient’s Perceptions of Physician Empathy (JSPPPE) was developed on the basis of a literature review. JSPPPE asks patients to identify points at which they perceived empathy during medical interviews (Kane, 2007) and measures five factors that primarily address verbal communication, emphasizing the use of phrases to communicate empathy in its evaluation. On the other hand, we couldn’t find any report of study of staged medical interviews performed by students during their early training.  

Strong nonverbal communication skills are reportedly associated with high scores regarding patient satisfaction (Griffith et al., 2003 Ishikawa et al., 2010). Ishikawa et al. (2006) used previous studies to compile a list of parameters for evaluating nonverbal communication. Those previous studies show it is insufficient to evaluate empathy in medical interviews solely on the basis of the observation by the assessor with phrases expressing empathy. The results of our semi-structured interviews suggest that the evaluation of empathy must incorporate measures of nonverbal as well as verbal communication. In the current study, we developed questionnaire items by analyzing the impressions of SPs who participated in medical interview training. As Huntley et al. (2012) have noted, some aspects of the quality of communication cannot be assessed by an observer (third party), and evaluation by SPs may be useful in terms of improving validity and providing feedback to student trainees. In addition, the content of the items we developed may be useful for future studies of empathy evaluation. Communication skills training programs are effective in raising students’ awareness in the short term but have little long-term effect (Kataoka et al., 2018) and are reportedly insufficient to change actual performance (Ishikawa et al., 2010). Repeated training with the content that we identified in the current study may be effective to achieve long-term effectiveness.

The questionnaire used in this study was judged to be valid because it was developed by using categories identified from semi-structured interview data. In terms of reliability, Cronbach’s alpha was 0.827, indicating acceptable internal consistency. A limitation of our study is its nature as a single-center study of SPs belonging to a single organization. Because the diversity of both the participants and their backgrounds was thus narrow, evidence supporting the results as generalizable may be insufficient. In addition, multiple scenarios were used during the training, and different effects of these scenarios cannot be ruled out. Further multicenter studies of medical students at different stages of their education with the participation of either SPs from several organizations or actual patients are warranted.

The factor that correlated most strongly with whether SPs felt that Japanese fifth-year medical students were empathic during medical interview training was “The student used phrases to express empathy.” However, correlations were observed with open-ended questions and the student’s level of attention at the beginning of the interview as well as the student’s attitude in general. Together, our current findings indicate that various types of both verbal and nonverbal communication determined whether SPs felt that medical students showed empathy during staged medical interviews.