A cross-sectional observational study was conducted amongst undergraduate medical students of University College of Medical Sciences and GTB Hospital with a survey tool.

The JSE-S was used to assess clinical empathy in medical undergraduate students. The English version of the questionnaire was used. Data regarding age, semester, gender, whose decision it was for the student to enroll in undergraduate medical training, choice of specialty, and current place of residence were also collected.

The JSE is a self-administered inventory that contains 20 questions, half of which are negatively phrased, while the other half is positively phrased. The students mark 1 of the 7 options provided on a Likert scale in response to each item (1=strongly disagree, 7=strongly agree). This scale is reversed (that is, 1= strongly agree, 7= strongly disagree) for the negatively-phrased items. It is a 3-factor latent variable scale, with the 3 factors being “perspective taking,” “compassionate care,” and “standing in the patient’s shoes.” Permission to use the questionnaire was obtained [9].

Initially developed in 2001, the scale has since been refined and tailored into 3 versions. Since its development, the scale has been validated elsewhere [7,10].

All medical undergraduate students of University College of Medical Sciences and GTB Hospital in 2017, numbering 600 in total, were invited to participate in the study, and data were collected over a period of 1 month (July 2017). The students were approached in their respective lecture halls at the end of lectures, and printed questionnaires were provided that were to be filled out and submitted in the class itself. Students who could not be contacted in the lecture halls were contacted personally. A total of 3 attempts were made to contact each student.

After obtaining ethical clearance from the Institutional Ethics Board, the investigators distributed and collected the completed questionnaires. Before distributing the forms, the investigators explained the purpose of the study and emphasized that anonymity would be maintained throughout the study period. After completing the questionnaire, the participants were instructed to submit it to the investigators.

The data were entered in a computer-based spreadsheet and analyzed using R version 3.4.1 (https://www.r-project.org/). Missing demographic data were coded as missing and excluded from the analysis. The scoring algorithm allowed for a maximum of 4 blank items (out of the 20), in which case the missing values were replaced by the mean score of the items that the participant responded to. If more than 4 items had no response, the form was considered incomplete and excluded from the analysis. Reverse-scored items were scored accordingly.

After conducting the descriptive analysis, the totaled empathy scores were compared according to gender (male or female), semester (first, third, fifth, or seventh), whose decision it was for the student to enroll in the undergraduate medical curriculum (one’s own or that of parents/relatives), choice of specialty (people-oriented, technologyoriented, or others), and current place of residence (home or other). Comparisons according to gender, who decided for the student to enroll in the undergraduate medical curriculum, and current place of residence were conducted using the Student t-test, while those for semester and choice of specialty were conducted using analysis of variance. The Bonferroni post hoc test was used for semester and choice of specialty. A correlation analysis was also performed between mean empathy scores and the age of the participants. A P-value of < 0.05 was considered to indicate statistical significance.

Prior to conducting the research, ethics clearance was obtained from the Institutional Review Board of the University College of Medical Sciences and GTB Hospital, New Delhi (vide ref no. IEC-HR/2017/ 31/5) after receiving informed consent from the subjects.

In our study, clinical empathy was found to be significantly associated with gender, with females having significantly higher mean empathy scores than males. This difference tended to diminish over the semesters, such that by the seventh semester, no significant difference was seen in the mean empathy scores of female and male participants. Since this is a cross-sectional study, the temporal significance of this finding cannot be definitively explained, but similar findings have not been reported elsewhere. In a study in Pune, it was found that there was no significant change in the mean empathy score of females across the semesters, but males showed a decline [4]. A longitudinal study conducted by Hojat et al. [11] using the same questionnaire found that although mean empathy scores in males and females changed equally over the years, females showed consistently higher scores than males, even when the mean scores dipped in general, and that the difference remained significant. Further studies are warranted to explore the temporal trend of clinical empathy in both male and female medical students.

Many studies have shown that the mean empathy scores of female medical students were higher than that of males, including studies carried out in Pune and Bangladesh [4,7]. The most common explanation for this finding has been said to be the expectations associated with traditional gender roles, but a study conducted by Baez et al. [12] in 2017 found that tools that rely on self-reporting for estimating empathy may induce biases leading the participating individual to assume traditional gender-based stereotypes. In contrast, a review conducted by Christov-Moore et al. [13] asserted that higher empathy in females has not only social, but also phylogenetic and ontogenetic roots. A few studies have also found no differences in the mean empathy scores of female and male medical students [10]. Further studies are therefore needed to explore the associations of sex and gender roles with clinical empathy.

In our study, the mean empathy scores fell from the first to the third semester, then more or less plateaued, and then rose again in the seventh semester. Associations of the number of years of medical education with empathy scores have been explored in many other studies, of which a few have found an increase in clinical empathy with increasing number of years of education [14], some have found a decreasing trend over the years [4,7], and others have found no significant difference in empathy scores across the years of medical education [10]. The findings of our study were unique, in the sense that clinical empathy was seen to increase after the fifth semester in students in the seventh semester. Since this was a cross-sectional study, it is difficult to draw any temporal inferences, but a similar finding in another study from India [4] indicates that further studies should be conducted to explore this phenomenon and its possible causes. It may be the case that a higher empathy score in the seventh semester indicates the positive effects of community medicine (previously called social and preventive medicine) being taught in the sixth and seventh semesters. It has been found elsewhere that doctors of family medicine (loosely an off-shoot of community medicine) are more empathetic than others [15].

In general, the mean empathy score in our study (96.01, standard deviation=14.56) is lower than has been reported in most other studies that have been conducted in Asia [4,7] or in Western countries [11], a finding that is worrisome. It has been asserted that physicians in Asia in general adopt a more paternalistic role in a doctor-patient relationship [16]. This might be partly responsible for explaining our findings, but further investigations are required to identify the factors associated with such low scores, so that steps can be taken to address the situation.

This study has methodological limitations that need to be taken into consideration before interpreting the results. The response rate in our study is a matter of concern; although it is fairly high, non-participants may have been significantly different from those who chose to participate. Our findings must be interpreted in light of this possibility. In addition, self-reporting questionnaires come with their own set of biases, which may have an impact on the results. Setting aside the various reasons the participants might have to underestimate or overestimate their empathy, social desirability may lead them to underreport or overreport empathy. Finally, like all cross-sectional studies, our study cannot be used to comment on causal associations.

In conclusion, the study found significant gender differences in the clinical empathy levels of the participants. The empathy scores were observed to decline initially and then recover as the semester of study increased. The mean empathy level found in this study is lower than has been reported in most similar studies across the world, which is a worrisome finding that requires further analysis.