A total of 626 trainee nurses from a general hospital in Hangzhou, China, were initially invited to participate in the anonymous online survey. A total of 614 valid responses were ultimately obtained (556 female and 58 male; mean age 21.85 ± 1.12, range 19 ~ 26 years). The remaining 12 responses were excluded on the grounds that they exhibited a standard score in excess of 65 on the Lie factor of the Parker Personality Measure, thereby indicating a low degree of accuracy in self-report. All the nurses were self-reported to have no history of psychiatric or neurological abnormalities, and they were not currently suffering from any acute or chronic illnesses. All participants were of Han Chinese ethnicity, had been engaged in hospital practice for a period of less than six months, and were assigned to rotate between different departments. Table 1 presents their demographic characteristics and corresponding comparisons of empathy levels. The study was approved by a local ethical committee (No. 2024029) and all participants gave their informed consent to participate.

Statistical analysis was performed using SPSS IBM Statistics v.26.0. Descriptive statistics were used to describe demographic variables such as age, gender, and education level of participants. Comparisons of the composite empathy scores between participants of different gender and education levels were made using Student’s t-test and one-way ANOVA, respectively, and the correlation between empathy and age was tested using Pearson correlation analysis. The significance of the P value was set at < 0.05. Once the influence of demographic variables on total JSE-HP scores was identified, these were used as covariates in further analyses.

For hypothesis 1, Pearson correlation analysis was used to test the relationships between scores on the PERM, DAP-R, and JSE-HP total score in the trainee nurses. For hypotheses 2 & 3, least squares regression (bootstrap method) was used to analyze the mediating and moderating roles of death attitudes in PROCESS plug-in of SPSS 26.0 (51). Breusch-Pagan test was conducted to detect the heteroscedasticity between personality styles, attitudes toward death, and empathy, and VIF and tolerance were calculated for test of linearity. Then each of the 11 personality disorder functioning styles (i.e., Paranoid, Schizoid, Schizotypal, Antisocial, Borderline, Histrionic, Narcissistic, Avoidant, Dependent, Obsessive-compulsive, and Passive-Aggressive) was used as an independent variable, and the composite empathy was used as the dependent variable, as illustrated in Figure 1 . To eliminate the multicollinearity effects between the variables, the data were centralized prior to analysis. 95% confidence intervals (CIs) were estimated by the bias-corrected bootstrapping procedure, and the number of iterations was set to 5000. A non-zero 95% CI for the interaction term and indirect effect indicated that specific attitudes toward death had a significant moderating and mediating role in the interaction terms and indirect effects.

Table 2 showed Pearson’s correlations with age as a covariate. All PERM personality disorder functioning styles were negatively correlated with JSE-HP total scores of the trainee nurses (all Ps < 0.001). The study found positive correlations between all personality disorder functioning styles and DAP-R Fear of Death, Death Avoidance, Escape Acceptance, and Approach Acceptance (all Ps < 0.001), and negative correlations between all personality styles except the Obsessive-compulsive and Neutral Acceptance (all Ps < 0.001 except for schizoid personality disorder; P < 0.01 for schizoid personality disorder). Furthermore, Fear of Death, Death Avoidance, Approach Acceptance, and Escape Acceptance exhibited negative correlations, while Neutral Acceptance exhibited a positive correlation with the JSE-HP total scores of the trainee nurses (all Ps < 0.001). See Supplementary Table 2 for more detailed descriptive statistics on scores of PERM, DAP-R, and JSE-HP among the participants. No other significant relationships were found between personality disorder functioning styles, death attitudes, and empathy in the trainee nurses.

There was no significant heteroscedasticity between personality styles and attitudes toward death or empathy (Ps > 0.05), while that between personality styles and empathy was significant (Ps < 0.01). Meanwhile, there was acceptable linearity between personality styles and attitudes toward death (VIFs = 1.95 ~ 8.92, tolerance = 0.11 ~ 0.51), between attitudes toward death and empathy (VIFs = 1.10 ~ 2.28, tolerance = 0.44 ~ 0.91), and between personality styles and empathy (VIFs = 1.95 ~ 8.92, tolerance = 0.11 ~ 0.51).

Table 3 displayed the results of the mediating effect with age as a covariate. When Fear of Death, Escape Acceptance, or Approach Acceptance was taken as the mediator, direct associations were found between all the personality disorder functioning styles (except for the Obsessive-compulsive) and JSE-HP total scores in trainee nurses (Ps < 0.05). The paths from Schizoid, Schizotypal, Histrionic, and Narcissistic styles to all five DAP-R death attitudes, as well as from these death attitudes to the JSE-HP total score, were included. Additionally, the paths from Paranoid, Antisocial, Borderline, Avoidant, Dependent, and Passive-Aggressive styles to all death attitudes except for Death Avoidance, as well as from these death attitudes to the JSE-HP total score, were also included. Significant associations were found between Obsessive-compulsive style and all death attitudes except for Neutral Acceptance, as well as between these death attitudes and JSE-HP total score (all Ps < 0.05). The bootstrap method revealed that the 95% confidence intervals of indirect effects in all models did not contain zero. This indicated that Fear of Death, Death Avoidance, Escape Acceptance, Approach Acceptance and Neutral Acceptance (–) partially mediated the negative associations of Schizoid, Schizotypal, Histrionic, and Narcissistic styles with empathy levels in the trainee nurses respectively. Fear of Death, Escape Acceptance, Approach Acceptance and Neutral Acceptance (–) partially mediate the negative associations of Paranoid, Antisocial, Borderline, Avoidant, Dependent, and Passive-Aggressive styles with their empathy levels respectively. Death Avoidance (-) partially mediated the negative association of Obsessive-compulsive style with their empathy levels. Moreover, Fear of Death, Escape Acceptance, and Approach Acceptance fully mediated the negative association of Obsessive-compulsive style with their empathy levels respectively (also see Figure 2 ).

No other significant mediating effect of death attitude was found on the relationships between personality disorder functioning styles and empathy levels in the trainee nurses.

Table 4 displayed the results of the moderating effect with age as a covariate. The JSE-HP total score was negatively predicted by both Narcissistic and Dependent styles (Ps < 0.001). Additionally, the interactions between Narcissistic style and Fear of Death, as well as between Dependent style and Death Avoidance, positively predicted the JSE-HP total score (Ps < 0.05). These results indicated that Fear of Death and Death Avoidance strengthened the negative predictions of Narcissistic and Dependent styles on empathy levels in trainee nurses respectively, as presented in Figure 3 . No other significant moderating effect of death attitude was found upon the predictions of personality disorder functioning styles on their empathy levels.

The current study revealed a negative association between each personality disorder functioning style and the empathy level of trainee nurses. Previous studies have shown that high level of non-adaptive personality traits, such as neuroticism, can be detrimental to the career development of nurses (56). Additionally, patients with personality disorders, such as schizoid, antisocial, narcissistic, or avoidant types, have been reported to have decreased empathetic concerns (30, 57–59). Such evidence supported our finding above. However, it is important to note that not all individuals with personality disorder lack empathy all the time. According to Ritter et al. (31), patients with narcissistic personality disorder underestimated their cognitive empathy and empathic behaviors on tests of empathic ability due to insufficient motivation, while in fact their cognitive empathy was normal, and it was their emotional empathy that was deficient. In addition, according to a review, patients with borderline personality disorder exhibited enhanced empathy in some social interaction situations, although these findings were contradicted by some other studies (60). Such evidence, on the one hand, conflicted with our results, and on the other hand, was suggesting that there were other factors influencing the relationship between personality disorder functioning styles and empathy, as we discussed below.

We found that attitudes toward death also influenced the empathy levels in trainee nurses. To be more specific, their JSE-HP total scores were negatively correlated with DAP-R Fear of Death, Death Avoidance, Approach Acceptance and Escape Acceptance, while positively correlated with Neutral Acceptance, in line with our previous findings on trainee nurses (61). Among the death attitudes, Fear of Death and Death Avoidance reflect negative thoughts and feelings about death, and a defense mechanism to keep death out of consciousness, respectively. On the contrary, Approach Acceptance and Escape Acceptance reflect a perception of death as a pathway to a blissful afterlife, and an escape from a painful existence, respectively. And unlike the attitudes above, Neutral Acceptance reflects a perception of death as a reality that is neither frightening nor desirable (32). Therefore, such findings implied that rational attitude like Neutral Acceptance of death instead of irrational ones, either fearing or desiring, was favorable for maintaining empathy at the right level. Although evidence showed that patients with schizophrenia with suicidal ideation had higher empathy levels compared to those without (62), indicating a positive correlation between an increase in tendency to death and higher empathy in patients with serious mental illness, other evidence supported the negative association between fear or avoidance of death and empathy in medical staff. For example, a study proved that the motivation of oncology nurses to care for terminally ill patients was negatively associated with death avoidance (63). Furthermore, we findings revealed that unlike empathy, the various personality disorder functioning styles were positively associated with the four irrational attitudes toward death mentioned above while negatively Neutral Acceptance, partly supported by the previous reports of increased fear of death in patients with narcissistic personality disorder, and higher suicidal ideation, which suggested higher desire for death, in patients with histrionic and antisocial personality disorders (40).

Our findings that DAP-R Fear of Death, Death Avoidance, Escape Acceptance, Approach Acceptance, and Neutral Acceptance (–) partially mediated the negative associations of Schizoid, Schizotypal, Histrionic, and Narcissistic styles with the empathy levels of trainee nurses; and Fear of Death, Escape Acceptance, Approach Acceptance and Neutral Acceptance (–) partially mediated the negative associations of Paranoid, Antisocial, Borderline, Avoidant, Dependent, and Passive-Aggressive styles with their empathy levels indicated that the decrease in empathy among nurses with specific disordered personality traits could be attributed to their loss of neutrality towards death and the adoption of irrational attitudes towards it. Nurses were indeed a population that needed to be frequently exposed to high life risk and death. For example, studies have reported mortality rates as high as 40% in emergency and ICU units (64, 65). And such frequent exposure has been found to increase compassion fatigue in nurses (66), not to mention susceptible nurses who already have non-adaptive personality traits. Furthermore, classic study has shown that patients with schizophrenia spectrum disorders experienced social and emotional withdrawal when negative symptoms intensified and suicidal ideation emerged (67). It could be plausible that desire for death may reduce empathy in individuals with cluster A personality styles which associated with the schizophrenia continuum. This evidence supported the present findings. The risk of suicide and self-harm was also common in individuals with Cluster B and Cluster C personality disorders (68, 69), which might leave them with no room to empathize with others. For example, in narcissistic patients, when their narcissism was severely threatened, patients would develop increased approach or escape acceptance of death to get rid of shame, which in turn led them to suicidal thoughts and behaviors (70), and they have been reported to lack the capacity for real empathy (71). Similarly, borderline personality disorder with high suicide rates also had difficulty understanding the emotions of others, leading to an inability to meet the needs of others (72). However, a correlation has also been found between their empathic impairment and the severity of mood and psychotic symptoms (73), suggesting the mediation of factors other than attitudes toward death.

Furthermore, Death Avoidance partially mediated the negative association between Obsessive-Compulsive style and empathy, whereas Fear of Death, Approach Acceptance and Escape Acceptance fully mediated the negative association of Obsessive-Compulsive style and empathy, implying that Obsessive-Compulsive style could impair the empathy of trainee nurses partly through psychological defenses against death-related information, or entirely by causing them to be afraid of death, desire death, or use death as an escape from the suffering of reality. Individuals with obsessive-compulsive style often alleviate their internal anxiety by adhering to strict rules and striving for perfection (74), and their attention is narrowed (75). Death, on the other hand, symbolizes ultimate loss of control and uncertainty (76). When they are filled with the fear of death, it is easy for them to neglect the emotional needs of others because of their obsession with maintaining inner order. Conversely, when they can no longer cope with the pressures of reality, they may simply turn to death as the most effective way to solve problems (77), thus eliminating the need to focus on the suffering itself, which can also lead to a drop in empathy for others (78, 79).

The findings further revealed that Fear of death and Death Avoidance also negatively moderated the JSE-HP total scores of nurses with Narcissistic or Dependent personality style, which suggested that such death attitudes further weakened the empathic competence of these nurses. Previous researches suggested that individuals diagnosed with narcissistic personality disorder exhibit lower acceptance of death (40) and a deficiency in empathy (80), they may engage in helping behaviors solely for the purpose of gaining attention rather than out of empathy or morality (71). On the other hand, patients with dependent personality disorder had high levels of insecurity and lack of self-confidence. When they perceived a relationship to be in jeopardy, they would resort to suicidal or abusive behavior to gain the attention of others (81). They were therefore prone to a lack of empathy.

It is important to consider the limitations of the study. Firstly, it was a cross-sectional design, so the relationships between personality disorder functioning styles, attitudes toward death and empathy were inferred and need further validation through longitudinal studies. Secondly, although the sample of trainee nurses was obtained from a large general hospital, subsequent multicenter, cross-cultural validation of the findings is encouraged, particularly in light of recent evidence on the influence of culture on empathy (82–84). Thirdly, as the data were collected from self-report questionnaires, there may be a reporting bias. Fourthly, different aspects of empathy were not modelled separately as the study focused on the impact of different personality disorder styles and attitudes toward death on composite empathy. However, further modelling and validation of emotional, cognitive, and behavioral empathy respectively using other-assessment tools is warranted. Finally, we have primarily examined the impact of different personality disorder styles on empathy in early-career nurses, and further validation is also warranted in more experienced nurses and other healthcare professionals who have greater workloads and potentially greater exposure to death.