Empathy and its correlates in Iranian physicians: A preliminary psychometric study of the Jefferson Scale of Physician Empathy

Abstract

Background: Empathy is one of the fundamental factors in patient care that is beneficial to both patient and physician.

Aims: To assess the psychometric properties of the Persian version of the Jefferson Scale of Physician Empathy (JSPE) in a sample of Iranian physicians and examine its correlates.

Method: Two hundred and seven general physicians completed the JSPE. The associations of empathy scores with demographic characteristics and practice-related variables were examined.

Results: The scale showed an acceptable internal consistency (α = 0.78). Three of six extracted factors were considered as prominent based on the scree test, which were similar to those obtained in the US samples. Women scored higher than men. Empathy improved with increasing practice experiences. Other practice-related variables did not show a significant association with empathy.

Conclusions: Results support the construct and criterion-related validities and reliability of the Persian version of the JSPE. Score difference between Iranian and American samples may not reflect a genuine difference in empathy trait and can be explained by cultural factors.

Introduction

Physician empathy is one of the fundamental factors involved in patient care and could enhance the therapeutic effects of patient–clinician relationships (Spiro 1992). Empathy has been linked to several desirable attributes in patient care that is prosperous for both sides of the physician–patient relationship. In an empathic relationship, patients would experience more trust, better compliance with clinical decisions, and more satisfaction with medical services (Becker & Maiman 1975; Barnett et al. 1981; Newton et al. 2000). Furthermore, empathic physicians would benefit from better therapeutic relationships (Bertakis et al. 1991; Livinson & Roter 1995), job satisfaction (Suchman et al. 1993), and less malpractice litigation (Levinson 1994). Therefore, training empathic physicians should be one of the major goals in designing programs for education and evaluation of medical students.

A psychometrically sound assessment tool is necessary for evaluation of empathy in medical education and for any possible intervention. The Jefferson Scale of Physician Empathy (JSPE) is the first and the most widely studied assessment tool specifically designed for measuring physician empathy. This scale was developed by Hojat et al. (2001a) in the US and has been shown to have satisfactory validity and reliability (Hojat et al. 2001b, 2002b). It has also been translated to 36 languages including Persian. However, no study has yet examined the psychometric characteristics of the Persian translation of JSPE. Additionally, to our knowledge, no study has been done on the empathy of Iranian physicians. This study was designed to assess the reliability and factor structure of JSPE in a sample of Iranian physicians and to evaluate the relationship of empathy scores and demographic characteristics including age, sex, marital status, place of practice (urban vs. rural), practice type (private vs. general), practice setting (office, clinic, or hospital), and practice experiences.

Table

Practice points

The Persian version of the JSPE has an acceptable validity and reliability among Iranian physicians. Female physicians show a higher empathy score than males. Empathy improves in GPs with increasing practice experience. Lower empathy scores in a country could be due to cultural differences and not necessarily an indicator of a genuine difference in the underlying empathy trait.

Method

Participants

Study participants included 207 general practitioners (GPs; 109 women, 97 men, and one unspecified gender) participating in the Annual Meeting of the Iranian National Medical Council in December 2007. Educational programs are provided during this meeting, and the participating physicians will gain credits for the maintenance of their certification.

Measures

JSPE – Health Professional Version is an instrument developed to measure empathy in the context of patient care among practitioners (Hojat et al. 2002b). JSPE is a self-report instrument and includes 20 Likert-type items answered on a seven-point scale (from strongly disagree to strongly agree). The scale was translated into Persian by Prof. Reza Shapurian, an Iranian psychologist and the first author, and edited by Prof. Mohammadreza Hojat, the developer of the instrument as well as five bilingual Board certified academic psychiatrists. A pilot study on 19 physicians was performed to assess the internal consistency of the scale. The pilot study showed that although the test had an acceptable internal consistency (Cronbach's alpha = 0.88), one of the items (item 4: I consider understanding my patients’ body language as important as verbal communication in caregiver–patient relationships) had a poor item-total correlation. Therefore, the translation of the item was edited and some other items were slightly modified by the first author. The revised version of the translation was then approved by the original translators.

Additionally, questions were included to elicit information on age, sex, marital status, place of practice (urban vs. rural), practice type (private vs. general), or practice setting (office, clinic, or hospital), and practice experiences. Participants were also asked to rate themselves from 0 to 100 on how much they generally empathize with their patients (self-rating of empathy).

Procedure

GPs participating in the meeting were approached (during the final 2 days of the meeting) and were asked to fill out the questionnaires. First, the questionnaires were given to the assistants of the meeting at the entrance of halls. This method was abandoned very early because of its very low response rate; only 20 out of 180 questionnaires were returned (10% response rate). Then, we used two research assistants to distribute the remaining questionnaires. This method had an acceptable response rate of 75%. Totally, 208 completed questionnaires were received; 90% of them were gathered by the latter method. One questionnaire was excluded because of incorrect scoring and 207 questionnaires were included in the study.

Statistical analysis

To test the psychometric properties of the Persian translation of the JSPE, Cronbach's alpha and item-total score correlations were calculated. Principal component factor analysis with varimax rotation was performed to search for the underlying factor structure of the scale. Subsequently, the data were analyzed using descriptive statistics including tests for assessing the normality of the data, as well as independent t-test to assess the difference of empathy between the two sexes, analysis of variance to evaluate the relationship of empathy score with marital status, place of practice, practice type, or practice setting, bivariate correlation to explore the associations between empathy scores and age and practice history, and stepwise linear regression to find out the unique contribution of age or practice experiences in predicting empathy scores. All the analyses were performed using SPSS software (version 11.5) and a p-value of less than 0.05 was considered as significant.

Results

The internal consistency of the scale was acceptable (Cronbach's alpha = 0.78). The mean score of empathy in the sample was 110.1 (SD = 13.6). The range of the scores was from 64 to 139 and the median score was 111; additionally, the histogram showed a bell-shaped distribution and the Kolmogorov–Smirnov test was not significant (p = 0.06). Therefore, empathy score was considered as normally distributed. Descriptive statistics of the items and their corrected item-total correlation are shown in Table 1. Only one item did not show a significant correlation with the total score (item 18: I do not allow myself to be influenced by strong personal bonds between my patients and their family members). Test–retest reliability was not assessed in this study, but in another study on residents of different specialties that was performed with the same version of the translated JSPE, the test–retest reliability of the test was 0.92 after 2 weeks (Shariat & Keikhaveni in press).

Table 1.  Descriptive characteristics and corrected item-total correlation of the items of Persian translation of JSPE on 207 Iranian physicians

Item	Mean	SD	Corrected item-total correlation
My understanding of how my patients and their families feel does not influence medical or surgical treatment (1)	5.0676	2.1622	0.3832
My patients feel better when I understand their feelings (2)	6.5700	0.8942	0.3866
It is difficult for me to view things from my patients’ perspectives (3)	4.9469	1.7824	0.2675
I consider understanding my patients’ body language as important as verbal communication in caregiver–patient relationships (4)	6.0338	1.4694	0.3948
I try to imagine myself in my patients’ shoes when providing care to them (5)	5.3188	1.5656	0.3419
Because people are different, it is difficult for me to see things from my patients’ perspectives (6)	4.8019	1.8629	0.3410
I try not to pay attention to my patients’ emotions in history taking or in asking about their physical health (7)	5.5411	1.8241	0.4485
Attentiveness to my patients’ personal experiences does not influence treatment outcomes (8)	5.1546	1.7081	0.3666
I have a good sense of humor that I think contributes to a better clinical outcome (9)	5.0531	1.7521	0.1851
My patients value my understanding of their feelings which is therapeutic in its own right (10)	5.9130	1.3446	0.4924
Patients’ illnesses can be cured only by medical or surgical treatment; therefore, emotional ties to my patients do not have a significant influence on medical or surgical outcomes (11)	6.2271	1.3940	0.4676
Asking patients about what is happening in their personal lives is not helpful in understanding their physical complaints (12)	6.1546	1.1680	0.4948
I try to understand what is going on in my patients’ minds by paying attention to their non-verbal cues and body language (13)	5.6667	1.4142	0.4045
I believe that emotion has no place in the treatment of medical illness (14)	6.2222	1.3931	0.3871
Empathy is a therapeutic skill without which success in treatment is limited (15)	6.0773	1.3595	0.4044
An important component of the relationship with my patients is my understanding of their emotional status, as well as those of their families (16)	5.7488	1.4329	0.4529
I try to think like my patients in order to render better care (17)	4.0821	1.8612	0.2187
I do not allow myself to be influenced by strong personal bonds between my patients and their family members (18)	3.0338	1.7443	−0.1091
I do not enjoy reading non-medical literature or the arts (19)	6.3333	1.3438	0.2634
I believe that empathy is an important therapeutic factor in medical or surgical treatment (20)	6.2029	1.1936	0.5306

Results of factor analysis are reported in Table 1. As shown in the table, seven factors were extracted each with an eigenvalue greater than one, However, the scree test showed that the magnitude of eigenvalues leveled off after factor three indicating that the first three factors were more prominent than others accounting for 45% of the variance. The first three factors could be entitled as: perspective taking, compassionate care, standing in the patient's shoes which are similar to those obtained in the US physicians. Factors 4 (paying attention to patients’ feelings) and 5 (thinking like the patient) are not grand factors due to their leveled-off eigenvalues. Factors 6 and 7 can be discarded because of only one or two items had considerable factor coefficients under these factors (Table 2).

Table 2.  Rotated factor loadings for the JSPE, based on the responses of 207 physicians

	Factor
Item	1^a	2^b	3^c	4^d	5^e	6	7
15	0.739	0.128	−0.034	−0.067	0.062	0.165	0.008
20	0.717	0.304	0.001	0.054	0.163	−0.023	0.101
2	0.696	−0.014	0.194	0.164	−0.111	−0.165	−0.180
10	0.600	0.157	0.013	0.199	0.158	0.151	−0.079
16	0.593	0.150	−0.064	0.131	0.255	0.095	−0.141
14	0.215	0.717	0.018	0.079	0.016	−0.081	0.038
12	0.222	0.673	0.093	0.165	0.025	0.212	−0.049
11	0.062	0.661	0.326	0.158	0.051	0.143	−0.209
6	0.069	0.226	0.828	−0.007	−0.019	0.091	0.144
3	−0.035	0.073	0.820	0.159	−0.010	0.005	−0.071
1	0.043	0.201	0.049	0.717	0.079	0.121	0.099
8	0.198	0.091	0.096	0.638	−0.199	0.235	0.064
7	0.119	0.454	0.182	0.512	0.084	−0.148	−0.059
17	0.085	0.064	0.041	−0.066	0.837	0.028	0.023
5	0.213	−0.212	−0.050	0.348	0.525	−0.226	−0.213
9	0.323	0.294	−0.099	−0.114	0.517	0.273	0.129
19	0.052	0.090	0.021	0.150	0.048	0.807	−0.194
4	0.404	−0.004	0.241	0.167	−0.004	0.466	0.131
18	−0.088	−0.122	0.067	0.144	0.021	−0.110	0.872
13	0.280	−0.011	0.302	0.308	0.284	0.078	−0.333
Eigenvalue	4.6	2	1.21	1.16	1.12	1.5	1.02
% Variance	23.12	10.01	6.03	5.8	5.6	5.27	5.09

Notes: ^aPerspective taking.

^bCompassionate care.

^cStanding in the patient's shoes.

^dPaying attention to patients’ feelings.

^eThinking like the patient.

Women had a higher mean empathy score than men (t = 2.38, p = 0.018), but marital status, place of practice, practice type, or practice setting did not show any significant association with empathy (Table 3). There was a significant positive correlation between empathy and both age and practice experiences (r = 0.15, p = 0.026; r = 0.16, p = 0.02).

Table 3.  The association of JSPE with gender, marital status, place of practice, practice type, and practice setting

Variable		Number	Empathy score	SD	Statistical significance
Sex	Female	109	112.34	13.74	t = 2.38	p = 0.018
	Male	97	107.87	13.1
Marital status	Single	62	107.58	13.7	F = 3.53	p = 0.063
	Married	140	111.43	13.29
Place of practice	Big city	118	109.67	13.01	F = 0.65	p = 0.52
	Small city or village	62	110.9	13.6
	Other	11	114.18	16.56
Practice type	Private	83	110.06	13.14	F = 0.02	p = 0.97
	General	47	110.32	13.54
	Both	24	109.62	12.86
Practice setting	Hospital	24	107.75	9.95	F = 1.19	p = 0.32
	Office	24	114.79	13.74
	Clinic	36	110.22	14.16
	More than one setting	28	108.82	16.80

As age and practice experiences are significantly correlated (0.91), a stepwise regression was performed to determine the unique contribution of each predictor. Regression analysis showed that only practice experiences could reliably predict empathy beyond the effect of physicians’ age (F = 5.53, df = 201, p = 0.02; B = 0.331, SE = 0.141, t = 2.346, p = 0.02).

The association of self-rating of empathy was also sought with the score of JSPE. As expected, there was a significant correlation between self-report empathy ratings and JSPE score (r = 0.34, p < 0.001). This could be a rough indicator of concurrent validity of JSPE.

Discussion

This study showed that the Persian version of the JSPE has construct and concurrent validities and reliability in a sample of Iranian GPs. Empathy mean score was higher in women and practice experiences were significantly and positively associated with empathy scores. Some of the findings of this study are in line with previous studies, while other findings are not. Regarding the higher empathy score in females, the findings of this study is consistent with most other studies (Hojat et al. 2001b, 2002a, 2002b, 2009; Hojat 2007); however, one study has not indicated a significant difference between genders (Di Lillo et al. 2009).

We did not find a significant association between age and empathy after controlling for practice experience in this study. Previous studies have had different results in this regard; Hojat et al. found no association (Hojat et al. 2002b), whereas DiLalla showed a negative correlation (DiLalla et al. 2004). To our knowledge, this is the first time that an association is reported between empathy and practice experiences. It is probable that the communication skills of the physicians increase as a function of experience and they become more aware of the importance of empathy in relating to their clients.

The mean score of empathy in this study (mean = 110, SEM = 0.94) was much lower than the western studies (Hojat et al. 2002b; Di Lillo et al. 2009) (US: mean = 120, SEM = 0.45; Italy: mean = 115.1, SEM = 0.91; and Poland: mean = 113, SEM = 1.33), but higher than the score reported from Japan, another eastern country (mean = 104.3, SEM = 0.65; Kataoka et al. 2009). However, another study in Iran with residents showed an even lower mean score than that in a Japanese study (mean = 100.8, SEM = 0.84; Shariat and Keikhaveni, in press). It seems that the scores of JSPE in eastern countries tend to be lower than those in the western countries. The difference in the mean scores of empathy does not necessarily mean that the latent trait of empathy in western counties is superior to their Iranian or Japanese counterparts. One can explain the reasons for such differences in terms of cultural peculiarities and the importance assign to notion of each item in different cultures (e.g., I try to think like my patient in order to render a better care). It should be noted that JSPE was originally designed for American physicians. Although empathy is a component of the patient–physician relationship regardless of the culture, it can be argued that its assessment may require slight cultural adaptation in the assessment tools.

For further examination of the observed cultural differences, we compared the item mean scores obtained in our study with those in Japanese (Kataoka et al. 2009) and American samples (Hojat, personal communication). Figure 1 interestingly shows that despite the variation, the mean scores of the items in the three cultures have a relatively similar pattern. However, the correlation of the item mean scores (items as the unit of observation) between Iran and Japan (r = 0.69) was much lower than that between Iran and the US (r = 0.9); and Japan and the US (r = 0.88). At least two points could be inferred from these comparisons: first, the observed difference between the countries could not be ascribed to a possible problem in translation of an item or several items. Instead, it seems that most of the items gain a lower mean in Iran or Japan than in the US. Second, despite having more similar total scores, Iran and Japan item scores are less similar to each other than to the corresponding items in the US. In other words, the items with highest differences with US were not the same in Iran and Japan. In the Iranian sample, the three most different items with the US cluster together in a factor named “paying attention to the patients’ feelings” (fourth factor); but, some other items (items 3, 6, and 10) had this role for the Japanese study. This difference between Iran and Japan shows that although both countries have lower mean scores than the US, the reasons for their lower scores could at least be to some extent distinct in the two countries.

Figure 1. Comparing the item means of JSPE across studies performed in Iran, Japan, and the US.

Kataoka et al. (2009) have suggested that the “heavily science-oriented selection system” in Japan could play a role. We suppose that this is true for Iran too; although other factors including the cultural differences might be implicated in the score differences. For example, Iran is a collectivistic society (Assadi et al. 2007) and a patient is generally taken to doctor by a number of her/his worried family members. Therefore, the Iranian physician learns to filter out the exaggerated emotions of family members, in order to have a rational decision making. This is a possible explanation for the lower scores of Iranian physicians in the items related to “paying attention to the patients’ feelings.”

Regarding item 18, we suppose that the translation of this item has not been completely clear or understandable to the subjects and hence it has had a low correlation with the other items and the total score of the scale. It seemed that the word-by-word translation of the phrase “I do not allow myself to be influenced …” has not been intelligible to the Iranian reader. Therefore, the translation of item 18 needed revision and we changed its wording for future studies.

The three prominent factors in this study are consistent with those reported with American samples (Hojat et al. 2002b), but in different order. Similar pattern of findings was observed among Italian physicians (Di Lillo et al. 2009).

This study has some limitations. Although the physicians who participated in the meeting came from different regions of Iran, we could not claim that our sample was representative of Iranian physicians. Another limitation was the point that item 18 did not show a significant correlation with the total score of the scale; therefore, the item remained “orphan” in factor analysis. Although the translation of this item was revised for the later studies, probably many of our participants have not correctly understood the meaning of item 18. This could have affected, though minimally, both the factor structure and the total score of the scale.

Validation of the Persian translation of JSPE could be the first step for further studies that could be designed to monitor or improve empathy in Persian speaking countries.

Acknowledgments

We thank Prof. Mohammadreza Hojat for his invaluable comments in different stages of the study. We also like to thank Dr. Baradaran and the Iranian Council of Medicine for making this study possible.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the article.