Characterizing the early relationship between covert narcissism and cognitive empathy in medical students: a cross-sectional study

Abstract

Background

In medicine, empathy refers to a predominantly cognitive attribute (rather than an emotional one), which is important as a foundation for positive physician–patient relationships. Physicians with a narcissistic personality trait have an assortment of characteristics that undermine their interpersonal functioning in clinical encounters with their patients. Evidence suggests an inverse relationship between empathy and certain characteristics of a narcissistic personality trait in general population. The aim of this study was to characterize the relationship between cognitive empathy and covert narcissistic personality trait in newly enrolled medical students.

Methods

A cross-sectional study was conducted in a Mexican school of medicine during the COVID-19 lockdown. The study sample included first-year medical students attending on-line classes. The Spanish version for medical students of the Jefferson Scale of Empathy (JSE-S), and the Hypersensitive Narcissism Scale (HSNS), were used as measures of cognitive empathy and covert narcissism, respectively. In addition, gender, age, speciality interest, and semester of enrolment were collected. Comparative, correlation and multiple regression analyses were performed among the variables analysed.

Results

In a sample of 283 students (207 females), differences in cognitive empathy and covert narcissism were observed by gender (p < 0.001). Comparison analyses showed that covert narcissism was greater in students enrolled after one semester attending on-line classes than those who were starting their first semester (p = 0.01). A correlation analysis confirmed an inverse association between empathy and covert narcissism (ρ=–0.23; p < 0.001). Based on this, a multiple regression model was created explaining 12% of the variance of covert narcissism based on a lineal regression with empathy (p < 0.001), gender (p = 0.01), and semester (p = 0.003). This model complied with the necessary conditions for statistic inference and showed an effect size from medium to large.

Conclusions

These findings provide novel information of the relationship between cognitive empathy and covert narcissism in newly enrolled medical students.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12909-024-06626-1.

Introduction

Narcissism and empathy in medical contexts

Narcissism could refer to a developmental phase, a personality trait, or a personality disorder. Traditionally, this term has been associated with personality traits of individuals with a self-centric orientation, followed by thoughts of unlimited power and success, and excessive need of encouragement and special treatment [1]. However, the most recent development of the Narcissism Spectrum Model suggests that a narcissistic personality trait and a narcissistic personality disorder can be treated as two extremes of a narcissism continuum [2]. According to this model, the core components are an excessive focus on the self, a high self-esteem, and a belief that one’s own needs (or goals) have more importance than those that others have. Based on this, narcissism acquires two apparently contradictory variants: an approach-oriented or “overt” variant (expressed as grandiosity–exhibitionism), and an avoidance-oriented or “covert” variant (expressed as vulnerability–hypersensitiveness). While individuals with covertly narcissistic traits tend to express more hypersensitiveness, anxiety, timidity, and insecurity, individuals with overtly narcissistic traits tend to express extraversion, aggressiveness, self-assuredness, and the need to be admired by others. However, both groups share common characteristics related to conceit, self-indulgence, and disregard for the others’ needs [3–6].

In medicine, both forms of narcissism may affect the physician’s interpersonal patterns even if they do not reach the threshold of a psychiatric disorder [7]. Fortunately, the medical community as a whole is usually less narcissistic than the general public [8]. However, such difference becomes less clear in societies where medicine is placed above other disciplines [9–11]. These social environments tend to be supportive or permissive with conducts opposite to the medical professionalism, which can be very harmful in medical students’ perception of medicine [12–14]. Differences on measures of narcissism also emerge when physicians from different specialities are compared [8, 15]. On this regard, some authors have suggested that working in a speciality in which lives can be saved or rapidly changed for the better demands a degree of self-assurance that allows challenging decisions to be made with cool confidence and prompt action [8]. It can explain, for instance, a greater narcissism in physicians from surgical branches in comparison with others, such as family doctors, geriatricians, or cardiologists. It has been proven that a narcissistic personality trait also encompasses lack of empathy [16, 17]. Some of the causes explaining it are an excessive sense of entitlement and the difficulty to understand other individual’s views [18]. These two aspects acquire special relevance in medical education considering that they involve either a misunderstanding of the role that altruism plays in medicine, or a deficit in social abilities that are necessary for establishing an empathic interaction with the patients.

In clinical encounters, empathy refers to a predominantly cognitive (rather than an affective or emotional) attribute that involves abilities of understanding (rather than feeling) of experiences, concerns and perspectives of the patient, combined with the capacity of communicating this understanding, and supported by a permanent intention to help [19]. Given its predominantly cognitive nature, this form of empathy in clinical settings is usually referred to as clinical empathy [20]. In medical students, especially those who are in an earliest stage of their training, is well documented the existence of differences on this clinical empathy according to personal qualities, speciality interest, and gender. However, different studies have demonstrated that such differences can be reduced in time with adequate interventional approaches in tutored face-to-face environments [21–27]. Based on this, it is expected that an improvement in clinical empathy can be particularly beneficial for medical students with a narcissistic personality trait since it can provide a type of knowledge and abilities that are missing or deficient. Unfortunately, as some other authors have recently stated, research in this field is still scarce [28].

Although there is a consensus on the important role that technology has in medical care, an excessive reliance on its use in doctor-patient interactions can be detrimental on clinical empathy [29]. This issue acquired more relevance during the first year of the COVID-19 pandemic, when countries introduced strict and extended lockdowns. In many cases, this situation derived in a drastic change in medical programs that were radically adapted into fully non-face-to-face methodologies. Among other problems, a prolonged exposition in such learning environments could derive in less opportunities to enhance social abilities related to social understanding and communication skills. As previously mentioned, both aspects are not only ingredients of clinical empathy, but also abilities that are missing in medical students with a narcissistic personality trait.

Study purpose

Although the empirical evidence suggests a negative association between cognitive empathy and overt narcissism in medicine [15, 20, 30], this relationship has been scarcely studied in medical learning environments. Based on this, the main purpose of this study was to assess the association between covert narcissism and cognitive empathy in the earliest stage of medical studies. The COVID-19 pandemic offered a privileged opportunity to perform this type of assessment. During this period, classes and social contact between trainees and educators was entirely limited to Internet. In addition, considering the existence of evidence supporting gender differences in empathy and narcissism measures [19, 31], this variable was also included into the analysis with the main purpose of characterizing its role. It was hypothesized that a poor development of empathetic orientation in first-year medical students positively correlates with a covert narcissistic personality trait. It was also expected that this association could be exacerbated in medical learning environments lacking personal contact. Three research goals were pursued: (i) to measure clinical empathy (as a measure of cognitive empathy) and covert narcissism in two groups of first-year medical students (one in their first semester, and another in their second semester) who were attending fully-online classes from home; (ii) to determine differences in measures of covert narcissism and cognitive empathy according to gender, speciality interest, age, and semester of enrolment (as indicator of time of exposition in a learning medical environment lacking of personal-contact); and (iii) to characterize the type of association between covert narcissism and cognitive empathy.

Methods

Participants and procedures

A cross-sectional study, using on-line questionnaires, was carried out in the School of Medicine of the Autonomous University of Nuevo Leon (UANL), in Mexico. The study included two cohorts of first-year medical students matriculated during the 2020–2021 academic year. One cohort was composed of students who were starting their first semester of classes in an on-line mode, and the other one was composed by students who were starting their second semester after completing one complete semester in an on-line mode. Altogether, 350 students (40% of all the entire group of first-year students enrolled at this medical school) agreed to participate in the study.

All participants attended on-line classes from home. Attendees were informed about this study in their virtual classes by an external professor who was part of the research team. They were informed that this study was developed with the aim of assessing personal attitudes related to specific aspects of medical behaviour in clinical encounters. To reduce possible bias in students’ participation derived from social desirability, all students were informed that their participation was voluntary, anonymous and confidential, and that it was not related with their ordinary academic evaluation. Those who accepted to participate signed an electronic informed consent prior to accessing a multiple-choice questionnaire with psychometric measurements and a socio-demographic form, specifically developed for this study.

Undergraduate medical students enrolled in second-year or more advanced courses, were excluded from this study. In addition, students who were not active in the system (web platform), students whose residence was abroad, or those who did not fully complete their questionnaires, were excluded from the analyses.

An independent ethical committee for clinical research (Comité de Ética en Investigación del Hospital Universitario “Dr. José Eleuterio González”) approved the study design (Ref. ON19-00002). The study was carried out in accordance with recommendations and authorization from the participating institution’s administration in Mexico.

Measures

Cognitive empathy

The Spanish version for medical students of the Jefferson Scale of Empathy (JSE-S) was used as a measure of cognitive empathy in medical learning contexts [32]. Authors of the JSE-S describe it as a validated measure of cognitive empathy (also called clinical empathy) used by healthcare professionals in their clinical work with the patients [19]. The JSE-S has been originally developed as a psychometric instrument for measuring the empathetic orientation of medical students towards the patients according to the operational definition of clinical empathy [20, 33]. The JSE-S consists of 20 items that are answered in a Likert scale from 1 (strongly disagree) to 7 (strongly agree). The possible score range for the JSE-S is from 20 to 140. Higher scores indicate a higher empathetic orientation in clinical encounters with patients.

Covert narcissism

Covert narcissism was measured using the Spanish version of the Hypersensitive Narcissism Scale (HSNS), originally developed by Hendin and Cheek [34]. The HSNS is a psychometrically sound instrument developed specifically to measure covert narcissism in the normal range of individual differences [5]. The HSNS correlates items from Murray’s Narcissism Scale [35] with a composite of the two measures of narcissism based on the Minnesota Multiphasic Personality Inventory (MMPI) used by Paul Wink in his research on narcissism [3]. The HSNS includes 10 items answered in a Likert scale from 1 (very uncharacteristic or untrue, strongly disagree) to 5 (very characteristic or true, strongly agree). The HSNS covers different aspects of covert narcissism, such as, hypersensitivity, anxiety, withdrawal, and feelings of being neglected. The possible score range for the HSNS is from 10 to 50. Higher scores indicate a higher tendency towards a covert narcissist personality trait.

Other variables

Information regarding gender, age, speciality of interest (focused on direct patient contact or “people-oriented” specialities, such as primary care, paediatrics, or psychiatry, vs. specialities with limited patient contact, such as surgery, radiology, or genetics), country of birth, and current semester of enrolment, were collected in a demographic form (see Supplementary Material).

Data analysis

Global scores on the JSE-S and HSNS were used as measures of cognitive empathy and covert narcissism, respectively. Only fully answered scales were included into the statistical analysis. The reliability of each scale was measured by calculating Cronbach’s alpha coefficient [36].

Once the normality was studied, using Pearson’s chi-squared and Lilliefors-Kolmogorov-Smirnov tests, non-parametric tests were used in further analyses. Analyses based on U Mann-Whitney test were used to compare differences in cognitive empathy and covert narcissism measures by gender, semester, and speciality interest, separately. In those cases, in which two or more variables showed statistical significance, an analysis of variance (ANOVA) was used to determine possible interactions following the recommendation performed by Schmider et al. [37]. According to them, ANOVA can be used as a robust statistical tool when the normality assumption is not observed but the sample size is still large (variables with, at least, 25 observations by condition in the study sample). These analyses were accompanied with the calculation of the partial eta-squared value to estimate the size effect. Effects sizes between 0.01 and 0.058 were interpreted as small, between 0.059 and 0.137 were considered medium, and equal to or greater than 0.138 were interpreted as large [38]. For cognitive empathy and covert narcissism scores, Spearman’s coefficients were calculated to determine statistical correlations.

All variables that showed statistical significance in the above-mentioned analyses were included in two groups of multiple linear regression analyses. In the first one, empathy was used as an explained variable (dependent), while all the others were treated as possible explanatory variables (independent). In the second one, narcissism was used as a dependent variable, while the others were used as independent ones. This procedure was performed with the purpose of exploring the type of relationship existing between cognitive empathy and covert narcissism. A model was accepted only when it met the necessary conditions for statistical inference: normality, zero mean, constant variance and uncorrelatedness of the residuals, in addition to linearity and absence of multi-collinearity. Models created from those analyses enabled the identification of variables that acted as influencing factors for those used as dependent ones. Finally, to quantify the degree of practical significance of the observed findings, the effect size (Cohen’s f²) was calculated for the obtained model. An effect size equal to 0.02 was interpreted as small, equal to 0.15 was interpreted as medium, and equal to 0.35 was interpreted as large [39].

All analyses were done in R language and programming environment for statistical and graphical analysis, RStudio version 2024 for Windows, and with the statistical analysis packages nortest, ApaTables, multilevel, lsr, and rstatix.

Results

Participants

From the 350 first-year medical students who initially accepted to participate in the study, 283 (76 males and 207 females) provided at least one of the two scales used fully answered and were included into the analysis. All these participants were singles and currently living in Mexico. The two psychometric measures used showed acceptable reliability, given by Cronbach’s alpha coefficients higher than 0.70. A summary of the descriptive analysis and the sample characteristics is shown in Table 1.

Table 1.

Descriptive analysis of measures of empathy and covert narcissism and sample characteristics

	Female group (n = 207)	Male group (n = 76)	Entire sample (n = 283)
Cognitive empathy (JSE-S)
Mean (SD)	120.2 (10.6)	113.8 (11.0)	118.4 (11.0)
Median [Min, Max]	122 [83, 139]	114 [92, 137]	120 [83, 139]
Missing	1	0	1
Reliability (Cronbach’s coefficient)	--	--	0.73
Covert narcissism (HSNS)
Mean (SD)	22.5 (5.6)	25.4 (6.0)	23.3 (5.8)
Median [Min, Max]	22 [10, 38]	25 [13, 40]	23 [10, 40]
Missing	17	8	25
Reliability (Cronbach’s coefficient)	--	--	0.73
Age
Mean (SD)	18.6 (1.3)	18.7 (1.3)	18.6 (1.3)
Median [Min, Max]	18 [17, 25]	18 [17, 22]	18 [17, 25]
Speciality interest
Specialities with limited patient contact	113	57	170
People oriented specialities	94	19	113
Semester of enrolment
First	90	25	115
Second	117	51	168

JSE-S, Jefferson Scale of Empathy S-Version; HSNS, Hypersensitive Narcissism Scale; SD, standard deviation

Preliminary findings

Findings from the comparative analyses showed a higher score on empathy in female students (p < 0.001) and in students interested on specialities mainly focused on direct patient contact (p = 0.01). No differences were observed in the global score of empathy by semester of enrolment (p = 0.99). On the other hand, male students (p = 0.001) and students starting the second semester (p = 0.01) showed a higher score on covert narcissism, while no differences were observed by speciality interest (p = 0.83). Based on these preliminary findings, two two-way ANOVA tests were performed. However, no differences were observed from these analyses either in “speciality interest” or in the interaction of “gender and speciality interest” in the case of empathy, nor in the interaction of “gender and semester” in the case of covert narcissism. The summary of this analysis is reported in Table 2. An inverse association between covert narcissism and cognitive empathy (ρ=–0.23; p < 0.001) was confirmed in a correlation analysis.

Table 2.

Two-way ANOVA of empathy and covert narcissism by gender, speciality interest, and semester

Source of variation	Empathy (JSE-S)
	F(1,278)	η2	η2p	p
Main effects
Gender	16.44	0.05	0.06	< 0.001
Specialty interest	3.50	0.01	0.01	0.06
Two-way interaction
Gender–Specialty interest	2.49	0.01	0.01	0.12
Source of variation	Covert narcissism (HSNS)
	F (1,254)	η 2	η 2 p	p
Main effects
Gender	12.25	0.04	0.05	< 0.001
Semester	5.69	0.02	0.02	0.02
Two-way interaction
Gender–Semester of enrolment	1.33	0.00	0.00	0.24

JSE-S, Jefferson Scale of Empathy S-Version; HSNS, Hypersensitive Narcissism Scale; F_(df), F-value_{(degrees of freedom)}; η², eta-squared; η²_p, partial eta-squared; p, p-Value

Main findings

Based on the above-mentioned findings, two linear regression analyses were performed. In the first one, cognitive empathy was used as a dependent variable, and in the second one, covert narcissism was used as a dependent variable. The regression model obtained in the first analysis did not comply with all the conditions that are necessary for statistical inference (normality, zero mean, constant variance and uncorrelatedness of the residuals, in addition to linearity and absence of multi-collinearity). However, a regression model explaining 12% of variance in the measurement of covert narcissism was obtained in the second analysis (R²-adjusted = 0.11; F_(3,254)=; p < 0.001). According to this model, “gender” (male) and “semester” (second) presented a positive linear relationship with covert narcissism, while “empathy” showed a negative one (Fig. 1). A summary of this second analysis is shown in Table 3. This model did fulfil the necessary conditions for statistical inference and showed a medium effect size (Cohen’s f² = 0.14).

Fig. 1.

Variation in the global score of covert narcissism by empathy (A), gender (B), and semester of enrolment (C). JSE-S, Jefferson Scale of Empathy S-Version; HSNS, Hypersensitive Narcissism Scale

Table 3.

Multiple regression model for covert narcissism of medical students

Predictors	β	SE	t	p
Empathy (JSE-S)	–0.12	0.03	–3.56	< 0.001
Gender (male)	+ 2.02	0.81	+ 2.50	0.01
Semester of enrolment (second)	+ 1.50	0.51	+ 2.94	0.003

JSE-S, Jefferson Scale of Empathy S-Version; β, beta coefficient; SE, standard error; t, t-experimental; p, p-Value

Discussion

The greater proportion of female students reported in this study is similar to the one observed in the entire medical school in which the study was performed, and in other medical schools from Mexico [40, 41]. The calculation of Cronbach’s alpha values confirms an adequate reliability for the two psychometric instruments used in this study. This value, in the case of the HSNS, was slightly higher than those initially reported in the United States [3] and in Spain [41]. In the case of the JSE-S, this value was slightly lower than the one previously reported in the United States [40], and in Mexico [32].

The greater empathetic abilities of female students in comparison with their male peers, measured by the JSE-S, is consistent with findings previously reported in previous studies in different cultural settings [12, 22, 24, 32, 33, 42–44]. These findings are in consonance with the idea that gender is a contributor to empathetic response [20], especially in medical students who are at the beginning of their medical studies. Such differences can be attributed to inborn characteristics, and to evolutionary, social learning and sociocultural factors, which make women tend to be more endowed towards social relationships than their male counterparts [20, 45]. In medical learning environments, targeted training and mentored working experience have been proven to help in reducing such gender differences in time [22, 24]. On the other hand, male students showed higher scores in covert narcissism in comparison with their female peers. This finding is also consistent with previous studies where gender differences in narcissism measures are reported in general population [46, 47].

Differences in narcissism, but not in empathy, also appear when medical students enrolled in the first semester were compared with those enrolled in the second one. An interpretation of this finding is that medical students, forced to limit their training to non-face-to-face classes, tend to perceive a poor self-confidence on their medical abilities as trainees in time. In consequence, this detrimental effect could be greater as more time of their training is limited to on-line learning methodologies. On the other hand, in pre-pandemic circumstances it could be expected that students’ empathetic abilities varied in time due to the training experience and the direct contact with their mentors and trainers. Since all medical education is performed in non-face-to-face environments, learning opportunities for improving empathetic abilities become null. In the Mexican society, where medicine is placed above other healthcare disciplines [12], is plausible that medical students with narcissistic personality traits could increase their feelings of anxiety and depression, vulnerability, hypersensitivity, and decreased zest for work due to the lack of self-confidence in abilities that were expected to be acquired in time. A greater perception of these feelings can also explain that those students become in one hand more egotistical, while in the other more apathetic about others’ needs. The inverse correlation observed between empathy and narcissism can be an indicator of this phenomenon.

Limitations and strengths

The main purpose of this study was to confirm an association between a covert form of narcissism and cognitive empathy in medical students who are at the beginning of their medical studies. Authors are aware of the limitation of this study. Findings reported in this study explain a small part of the variance of covert narcissism. It suggests that there are other factors, not covered in this study, that are affecting this form of narcissism in first-year medical students. Some of those factors are probably related with individual features (such as individual characteristics associated with certain personality traits, life experience, social skills, and emotional regulation, students’ social interest, or principal career motivations), as well as characteristics associated with social and family backgrounds.

However, findings reported in this study help to clarify the type of relationship between empathy and covert narcissism in medical students. As described before, the relationship between narcissism and empathy is not new and it was initially reported years ago [16], but without clarifying the role that each element played. Findings reported in this study confirm that a poorer development in cognitive empathy is predictive of a greater covert form of narcissism, and not vice-versa. Similarly, findings reported in another study suggest that empathy in healthcare professionals plays an important role to build adequate human connections at their workplaces [48], a difficulty that is commonly observed in individuals with covert narcissistic personality traits. Furthermore, these findings are also in consonance with those reported in another study, where the lack of guilt proneness – a characteristic of non-empathetic individuals– is described as an explanatory variable of the covert narcissism trait [18].

Conclusions

Findings reported in this study bring empiric evidence supporting the role that cognitive empathy plays in the early development of covert narcissism in newly enrolled medical students. Findings reported confirm that the lack of empathy is explanatory of a greater covert narcissistic trait. Gender, in the case of male students, and having a prolonged period of dedication on non-face-to-face classes during the pandemic appear as other two predictive elements of a covert form of narcissism trait. On the other hand, being female student and having an orientation toward specialities with more contact with patients were associated with greater development of empathetic abilities.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

LV and AAG oversaw the study’s overall design. OVG, JFGG, FAN and GAPG oversaw coordination with the management department in the participating institution, students’ recruitment, and data collection. MSM and LV performed the statistical processing of data. LV, RCD, AAG and FAN oversaw drafting the manuscript. All authors contributed during the interpretation process of the results and approved the final manuscript.

Funding

This study was supported by the Autonomous University of Nuevo Leon, in Mexico.

Data availability

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

All methods were performed in accordance with the relevant guidelines and regulations of the BMC Medical Education’s policies. In addition, an independent research ethics committee approved the study design prior to be executed (Comité de Ética en Investigación del Hospital Universitario “Dr. José Eleuterio González”; Ref. ON19-00002). The electronic informed consent was presented first and must have been accepted by the participant to proceed further in the survey. All participants were aware that they could leave the study at any time.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.