This retrospective study was based on real-world data from Nanjing Drum Tower Hospital, Nanjing University Medical School. The participants were first-visit outpatients between June 2014 and May 2019, regardless of age and gender. The patients’ chief complaints included symptoms of insomnia, anxiety, depression, compulsion, panic attack, and social phobia, as well as physical symptoms that were not abnormal during the examination, such as headache, back pain, palpitation, chest tightness, nausea, abdominal distension, fatigue and other symptoms. EPQ and SCL-90 were tested simultaneously. Data such as age, gender, marital status and education were obtained from the medical records. This study was approved by the ethics committee of Nanjing Drum Tower Hospital (approval no. 2021-096-01).

SCL-90, a psychiatric self-report list, is the most widely used outpatient examination scale for mental disorders and other mental diseases in worldwide, with high reliability and validity [23,24]. From the aspects of personal feelings, emotions, thinking, consciousness, behavior, living habits, interpersonal relationship, diet and sleep, the SCL-90 can assess and evaluate whether a person may have some mental disorder related to psychological symptoms and their severity. SCL-90 has a comprehensive description of the psychological symptoms of psychosis as well as the capacity to observe and distinguish the specific mental symptoms. It has been proven to be a vital and effective tool in the identification of primary health care and psychiatric symptoms, measurement and analysis of the overall psychological distress of subjects, detection of changes and reflection of symptom characteristics and comparison of intervention effects in mental and neuropsychiatric diseases [25].

SCL-90 contains 10 self-assessment dimensions and 90 self-report items [26]. The 10 self-assessment factors are somatization, obsessive-compulsive symptoms, interpersonal sensitivity, depression, anxiety, hostility, terror, paranoia, psychoticism and others. Each self-report item is scored on a scale of 1–5. SCL-90 assesses the mental health status of the subjects based on the total symptom score index and factor score. The more severe the frequency and intensity of the symptoms corresponding to the factor, the higher the score of some factors. The national norm of SCL-90 was obtained from the survey of healthy populations in 13 regions in China (n = 1338) [27].

EPQ is a self-assessment scale developed by well-known psychologist Eysenck et al. [28]. The questionnaire consists of four dimensions: extraversion (E), neuroticism (N), psychoticism (P), and lying (L). According to Eysenck’s theory, extraversion is characterized by good communication, a desire for stimulation and adventure, and being easy to be impulsive; neuroticism is characterized by a high degree of negative effects, such as depression and anxiety; and psychoticism is independent of neuroticism and has the characteristics of stubbornness, which are manifested as loneliness, difficulty in adapting to the external environment and strong aggression [29].

The revised EPQ questionnaire contains 88 items, all of which are answered with “yes” and “no”. The national norm of EPQ was obtained from a survey conducted including 13 provinces in 6 regions of China (n = 2517). The original EPQ score were converted to standard score (T-scores) based on the national norm score grouped by age and gender [14].

T=50+10     ×Original score−Mean score of corresponding groupStandard deviation of corresponding group 

If the standard score is between 43.3 and 56.7, the personality characteristics of this dimension are of intermediate type. The standard score of tendency type is between 38.5 and 43.3 or 56.7 and 61.5. The standard score of the typical type is less than 38.5 or greater than 61.5.

In the EPQ questionnaire survey, we use a standard score of 50 as the boundary. A higher score on the E dimension indicates a more significant tendency towards extroversion, while a higher P score indicates greater emotional instability. In Eysenck’s two-dimensional model of personality theory, temperament is classified into four types based on EPQ scores: introverted-unstable, introverted-stable, extroverted-unstable and extroverted-stable, corresponding to, melancholic, phlegmatic, choleric, and sanguineous, respectively. These types reflect the comprehensive effect of psychic temperament and introverted and extroverted personality [30]. Choleric, also known as leadership temperament, is characterized by individuals who are confident, decisive, and aggressive. Sanguineous, also known as lively temperament, refers to individuals who are outgoing, optimistic, and have cheerful and easy-going behavior. Phlegmatic, also known as steady temperament, is characterized by individuals who are thick, steady, and meticulous in their actions. Melancholic is also known as the quiet temperament, which refers to individuals with introverted, quiet, and cautious behavior.

SAS 9.4 was used to analyze data. Continuous variables were described using mean and standard deviation (SD). Categorical variables were expressed using frequency and percentage. Differences in factors of SCL-90 and EPQ tests between the patients and national norms were compared using t-test. One-way variance analysis (ANOVA) was used to compare the differences in SCL-90 scores among four different temperament groups of psychiatric outpatients, Bonferroni post hoc test was used if significant differences were found. The differences between the two groups were expressed using the mean and Bonferroni-adjusted 95% confidence interval (CI). Pearson’s correlation analysis and path analysis were used to investigate the relationship between mental disorders and personality types. We present Person correlation coefficient and 95% CI to account for correlation analysis. A p-value of <0.05 was considered statistically significant.

During the period from June 2014 to May 2019, a total of 8,762 first-visit outpatients received SCL-90 and EPQ tests on the same day. Of these, 353 patients with a high degree of concealment were excluded (if the L score in the EPQ test was greater than 61.5). Finally, 8,409 patients (average age 28.46 ± 9.62 years old) were included in the analysis. There were 3,510 males and 4,899 females. Table 1 summarizes the demographic characteristics of the study population.

Cronbach’s alpha was used to independently analyze internal consistency across all domains. Cronbach’s α of SCL-90 ranged from 0.74 to 0.93 and for EPQ, it ranged from 0.62 to 0.87. Suppl. Tables S1 and S2 show detailed Cronbach’s α coefficients for each dimension, demonstrating that internal consistency was acceptable.

As is shown in Fig. 1 and Suppl. Table S3, the t-test revealed that the total average score and factor scores of SCL-90 in psychiatric outpatients were significantly higher than those in the national norm (p < 0.001) [27].

The composition of the scores of each dimension of the EPQ scale for patients is shown in Fig. 2. t-test results comparing EPQ scores of the patients to the national norms are shown in Fig. 3 and Suppl. Table S4. Compared to the national norms, the P and N scores were significantly higher, whereas the E score was significantly lower.

One-way ANOVA showed the statistically significant differences in temperament among the four groups of patients (p < 0.001). Each factor had a higher score in patients with a melancholic temperament as shown in Table 2. Suppl. Table S5 describes the difference and its Bonferroni-adjusted 95% CI between the two groups.

Pearson’s correlation matrix showed that the total average score and each factor score of SCL-90 were positively correlated with the P score and N score, but negatively correlated with the E score (p < 0.001). The correlation coefficient between the N score and total score, compulsive symptom, interpersonal sensitivity, depression, anxiety, psychotic symptom was 0.69, 0.62, 0.65, 0.66, 0.62 and 0.64, respectively, indicating a strong association. The correlation coefficient between the N score and somatization, hostility, phobia, paranoid and other factors were 0.46, 0.55, 0.49, 0.59, and 0.56, respectively, indicating a moderate association. The correlation between the P score and hostility and paranoia was 0.42 and 0.40, respectively. The correlation coefficients and the 95% CI of EPQ and SCL-90 are shown in Table 3.

The path analysis model was established by considering 10 factors of SCL-90 as endogenous (dependent) variables and 3 factors of EPQ (N, E and P), along with age, gender, marital status and education as exogenous (independent) variables. The model also accounted for the correlation among the 10 factors of SCL-90 and the 3 factors of EPQ. The model fit the data well, with a goodness of fit index (GFI) of 0.9920, an adjusted goodness of fit index (AGFI) of 0.9722, a comparative fit index (CFI) of 0.9938, root mean square error of approximation (RMSEA) of 0.0335, and a standardized root mean square residual (SRMR) of 0.0240.

Supp. Table S6 shows the detailed path coefficients and Fig. 4 shows the path analysis diagram. Due to the large number of paths in the graph, Fig. 4 includes only partial paths with absolute path coefficients greater than or equal to 0.1.

The results revealed that N scores had a significant effect on somatization, obsessive-compulsive symptoms, interpersonal sensitivity, depression, anxiety, hostility, phobia, paranoia, psychotic symptom and others, with standardized path coefficients were all greater than 0.4. The standardized path coefficient of hostility and paranoia on P scores were 0.313 and 0.280, respectively. Interpersonal sensitivity, depression and obsessive-compulsive symptoms were influenced by E scores, with standardized path coefficients of −0.149, −0.138, and −0.105, respectively.

The correlation coefficient matrix of SCL-90 and EPQ indicated that the correlations among SCL-90 factors were strong, whereas the correlations among EPQ dimensions were relatively weak (detailed correlation coefficients are presented in Supp. Tables S7 and S8). Table 4 shows that the path analysis model had a higher explanatory power for interpersonal sensitivity and depression (R² = 0.513, 0.505) but was weaker for somatization and phobia (R² = 0.233, 0.264).

Supp. Table S9 shows the direct, indirect, and overall effects of the parameters influencing the risk of SCL-90. The results indicated that direct effects accounted for more than 80% of the overall effects in all paths, while indirect effects accounted for only a small proportion. The indirect effect of gender on somatization accounted for 18% of the overall effect, while the indirect effect on others accounted for 20%.

Perceptible signs of mental disorders usually include pains, fevers, weaknesses, irritability, depression, sleep disorders, cognitive disorders, and other types of dysfunctions [31,32]. In this study, psychiatric outpatients were examined using SCL-90, examine and subjective symptoms were extensively described. We found that the total average score of SCL-90 and the scores of each factor in psychiatric outpatients were significantly higher than the national norm (p < 0.001). This indicated that psychiatric outpatients had a higher tendency in aspects including somatization, obsessive-compulsive symptoms, interpersonal sensitivity, depression, anxiety, hostility, terror, paranoid ideation, and psychosis.

Somatization is generally considered to be the expression of physical symptoms that cannot be medically explained, reflecting stress and pain, and usually indicating the maladjustment psychological reaction to stress. This connection between somatization and depressive symptoms can explain the association between the two [33]. Defazio et al. [34] reported that the severity of somatization symptoms positively correlated with the score of anxiety and depression scale in patients with functional dyskinesia. Another study observed a high incidence of mental disorders in the Israeli population, with a relatively high rate of depression, generalized anxiety disorder, somatization disorder and neurasthenia [35].

Obsessive-compulsive symptoms are usually characterized by repetitive and distressing thoughts that are difficult to control and can impact social functionality [36,37]. Obsessive-compulsive disorder is often associated with depression, reduced mental function, and even a certain suicidal tendency [38]. Sancak et al. [39] found that more than 40% of schizophrenia patients exhibit apparent obsessive-compulsive symptoms.

Interpersonal sensitivity is internally related to the lack of social functioning. Masillo et al. [40] found that the scores of interpersonal sensitivity measurement and global function: social and role scale of 39 adolescents with ultra-high risk mental state were higher than those of the control group, indicating a worse function in interpersonal consciousness and timidity.

Compulsion and anxiety are often accompanied by hostile thinking [41], with paranoia as the main manifestation.

In this study, psychiatric outpatients exhibited introversion, psychoticism, and neuroticism. The correlation coefficient between the total average score of the SCL-90 test and each factor score of the EPQ test revealed that neuroticism, introversion, and psychoticism had a significant negative impact on the psychological status of psychiatric outpatients. This finding is consistent with the results of numerous existing studies [42–44]. Psychoticism, neuroticism, and introversion are predictors of adverse psychological reactions [45,46].

Neuroticism had a standardized path coefficient of greater than 0.4 for all factors of SCL-90. Pearson’s correlation matrix showed a correlation coefficient of higher than 0.6 between the N score and total score, compulsive symptom, interpersonal sensitivity, depression, anxiety, and psychotic symptom, indicating a strong association. Reportedly, individuals with high neuroticism scores are more likely to experience moodiness and feelings of anxiety, worry, fear, anger, frustration, envy, jealousy, pessimism, guilt, depressed mood, and loneliness [47]. In addition, neuroticism is thought to be related to shifts in attention and biased negative emotional stimuli, lower stress tolerance, and increased negative emotional experience [48]. More psychological symptoms will result from high neuroticism. Neurotic personality traits make people overly detail-oriented and always strive for perfection, blame others for their worries, which can lead to tensions and conflicts in their relationships [49]. Hippocampal dysfunction is considered the pathological basis of high neuroticism, which may be related to trauma and post-traumatic stress disorder (PTSD), as well as the sensitive psychological state of interpersonal relationships [50]. In the SCL-90 test, the factor “others” reflects the diet and sleep status and shows a strong correlation with the N score. Studies have found that patients with functional dyspepsia have higher N scores in the EPQ test [51]. Furthermore, neurotic symptoms are also associated with the nightmare frequency [52]. The psychological status of patients with depressive and choleric temperaments, characterized by emotional instability, is of greater concern compared to patients with stable temperaments [53]. These studies align with our findings.

People with psychotic personalities are known for their cold and strange demeanor, making it difficult for them to adapt to interpersonal communication and social situations. In addition, these individuals are usually stubborn and tend to be hostile towards people around them, making them more prone to encountering challenging life events [54]. As a result, their psychological state is often at the stress level, leading to impaired cortical function and causing emotional changes [55]. Consequently, patients with psychoticism are more likely to exhibit hostility and obsessive-compulsive symptoms, with standardized path coefficients of hostility and paranoia on psychoticism are greater than 0.280.

Extraversion shows weak negative path coefficients for interpersonal sensitivity, depression and obsessive-compulsive symptoms. Noticeable personality traits associated with extrovert tendency types are highly correlated with interpersonal sensitivity and anxiety symptoms. Patients with mental disorders who have introverted personalities struggle with communication and self-confidence [56]. When faced with social stress, such as pressure, they tend to blame themselves, and their interpersonal sensitivity becomes apparent. In times of adversity, introverts find it challenging to take positive measures, making them susceptible to suffering and feeling overwhelmed by negative emotions [57–59].

In Eysenck’s two-dimensional model of personality theory, temperament is classified into four types based on EPQ scores: depressive temperament, mucinous temperament, bilious temperament, and pluripotent temperament. These classfications reflect the comprehensive influence of the N and E dimensions. Patients with higher phlegmatic and sanguine temperaments tend to have stronger emotional stability and are better equipped to cope with stress. On the other hand, the psychological status of patients with depressive and choleric temperaments, characterized by emotional instability, is more concerning compared to stable patients. This observation aligns with the results of the correlation analysis in our study.

The results of the path analysis revealed that gender indirectly affects symptoms of mental disorders. Previous studies focused on the direct effect of gender on SCL-90 scores and consistently found that female patients tend to score higher than male patients in all aspects of the SCL-90 (p < 0.001). This suggested that female patients are more likely to exhibit severe psychological states, especially concerning depression and somatization, with stronger negative tendencies. These differences may be attributed to physiological and social factors, as women are more susceptible to psychological, social, and occupational impairment, leading to increased psychological stress [60].

In clinical practice, the patient’s neuroticism score reflects their emotional stability, while the psychoticism score helps evaluate their interpersonal interaction. Psychotherapy and/or drug treatments for each patient are selected primarily based on the severity of SCL-90 symptoms. During treatment, physicians consider the EPQ personality characteristics to adjust the way of communicating with the patient, providing targeted psychotherapy. Individuals with prominent psychoticism and neuroticism personality characteristics require heightened attention to psychological symptoms to detect early symptoms and provide appropriate interventions.

This study has a few limitations. First is the absence of data on the clinical classification of patients with mental disorders, as well as the treatment modalities received and the effects of treatment. Second, the validation of personalized treatment plans based on different personality characteristics requires further investigation. Further studies should be conducted to understand the impact of personality traits on the mental health of patients with mental disorders.