The study was approved by the ethical committee of Hadassah Medical School. Informed consent was signed by all participants.
Context- Structure of Medical studies
The Hadassah Hebrew University of Jerusalem Medical School offers a six-year program. The first three years includes basic sciences and preclinical studies. During the first three years of studies, exposure to patients and everyday hospital life is limited and occasional. During the following three years, students attend hospitals or outpatient clinics in small groups on a daily basis. Thus, the first students' significant clinical experience occurs at the fourth-year of studies.
The study included three consecutive cohorts, differing in humanities curriculum and admission system. The participation in the humanities program in each year was mandatory, thus each class was obligated to participate in the program offered in that year (limited/extended). The three consecutive cohorts included (Figure 1):
(1) PI/MH(lim) cohort: Personal Interview (PI), limited Medical Humanities program (MH(lim) ) (n=91).
(2) MMI/MH(lim) cohort: multiple mini interviews (MMI), (MH(lim) ) (n=86).
(3) MMI/MH(ext) cohort: MMI, extended Medical Humanities program (MHext) (n=85).
In the admission process, candidates to the Hadassah Hebrew University of Jerusalem Medical School are invited to an interview based on their Psychometric entrance test scores (Israeli version of SAT's) and undergraduate academic scores. The first cohort (PI/MH(lim)) went through an admission process that included a panel-style personal interview, in which three examiners interviewed each candidate for about 45 minutes. The second and the third cohorts (MMI/MH(lim) and MMI/MH(ext)) went through a multiple mini interviews (MMI)-based admission process. The MMI is an OSCE-style exercise, consisting of multiple, focused interviews that , in which candidates have a limited time to discuss an issue with an interviewer, or to demonstrate the capacity to work through a challenging interpersonal situation presented by an actor. The MMI is designed to dominantly focus on ethical and communication issues, and to evaluate traits such as motivation, responsibility, self-awareness and interpersonal skills. It is intended to assess skills that are inadequately assessed by the personal interview. Another advantage of the MMI over the panel-style interview is that multiple interviews may dilute the effect of chance and various biases .
The PI/MH(lim) and the MMI/MH(lim) cohorts included a pre-clinical humanities curriculum that was limited to the first year of medical school. The third cohort (MMI/MH(ext)) participated in an extensive and comprehensive three-year pre-clinical curriculum.
The extended pre-clinical humanities program was taught during the first three years of medical school, including the following courses:
During the first year of medical studies the program included the 'Human and Medicine' course on patient- doctor communication, cultural competency in medicine, basic principles of medical ethics, and physician professionalism. In addition, the first year course included early clinical exposure and community medicine. During the second year of medical studies the program included courses such as history of medicine, medicine and literature, physicians and holocaust, narrative medicine. During the third year of studies the program included the following courses: medicine and law, introduction to human sexuality and introduction to breaking bad news .
342 medical students consented to participate in the study. 264 (77%) of these provided full answered questionnaires on the beginning of the 4th year. Among these, two students had a repetitive filling pattern which wasn’t consistent with the content of the questionnaire and were excluded. Out of the remaining 262 medical students, 35 (13.4%) students did not fill the end of year questionnaires. Thus, 227 students were included in the longitudinal analyses.
Interpersonal Reactivity Index (IRI)
The IRI is a validated 28-item self-report measure consisting of four 7-items subscales, each tapping some aspect of the global concept of empathy. The Perspective-Taking scale assesses the tendency of spontaneously adopting the psychological point of view of others; the Fantasy scale taps respondents’ tendencies to identify with feelings and actions of fictitious characters in books, movies and plays. The Empathic Concern scale assesses “other oriented” feelings of sympathy and concern for unfortunate others, and the Personal Distress scale measures “self-oriented” feelings of personal anxiety and unease in tense interpersonal setting .
The Hebrew version of the IRI has been widely used in research in Israel [26-28]. In the present study the alpha Cronbach coefficient of internal reliability of the IRI was very good: α = 0.81.
Jefferson Scale of Physician Empathy – Student version (JSPE-S)
The JSPE-S was developed to measure empathy specifically within the context of the physician–patient relationship . It includes 20 Likert scale items which are scored from 1 (strongly disagree) to 7 (strongly agree). The questionnaire was validated in numerous studies and is frequently used in medical education research and has been translated to more than 42 languages [1,3, 29, 30]. The questionnaire encompasses 3 components of empathy; perspective taking (considered the core component of empathy), compassionate care and standing in the patient's shoes [3,5]. The English version of JSPE-S was translated to Hebrew by four physicians who speak English fluently, and was translated back to English by native English speakers who is also fluent in Hebrew ('forward-backward' procedure). Once the preliminary Hebrew version was obtained, the questionnaire was administered to 3 other physicians to achieve a consensus regarding its final version. We examined convergent validity with a similar instrument (the Interpersonal Reactivity Index (IRI), see above). The correlations between the JSPE-S and the IRI total score administered at the beginning of the study was r = 0.31; p < 0.001. The correlations between the JSPE-S and the IRI subscales were r = 0.43; p < 0.01 for Perspective Taking, and r = 0.30; p < 0.05 for Empathic Concern, and no correlation with the Fantasy and Personal Distress subscales. The findings regarding the Hebrew version of the JSPE-S are very similar to the results reported by Hojat  in the original English version, who also found that Perspective Taking and Empathic Concern subscales had the greater correlations with JSPE. In addition, the alpha Cronbach coefficient of internal reliability of the JSPE-S was excellent: α = 0.86 for pre-clinical JSPE-S scores, and α = 0.90 at the end of the fourth-year.
Socio demographic questionnaire
Participants were asked to provide information regarding their gender, age, ethnicity, marital status, religiosity, and preferences regarding future residency.
The design of the study was longitudinal: investigators distributed questionnaires in two different time points to each cohort during the three study years. The JSPE-S, the IRI and the socio-demographic questionnaire were administered at the beginning of the fourth-year, before attending clerkships. The JSPE-S was distributed to students again at the end of the fourth-year. For all three cohorts the interval between the first and the second questionnaires was 8 months. All questionnaires were filled anonymously. Each participant received a random number, which was written on his/her questionnaire and was used to identify the individual pre-post questionnaires. Students were allowed to return the questionnaires during the following seven days.
One way and two-way ANOVAs and t-tests were used for continuous variables, and χ2 tests for demographic categorical variables. For all data analyses, the dependent variable in the present study was level of empathy as assessed by the JSPE-S that was measured twice: at the beginning (preclinical) and at the end of the fourth year. Dependent samples t-tests and ANOVA with repeated measures were used for longitudinal analyses of pre-clinical JSPE-S vs. end of the fourth-year JSPE-S scores. Tukey post-hoc comparison tests were used to examine the differences among the three groups in variables for which one-way ANOVA tests were significant. A two-sided significance level of 0.05 was established for all analyses. Data analysis was performed using Statistical Package for Social Sciences (SPSS) software, Version 26.0 for Windows.