Do Physician Personality Traits Influence Patient Profiles? A Cross-Sectional Study in Greek Primary Care

Posted on by admin

ACHAIKI IATRIKI | 2026; 45(1):14–24

Original Research Article

Maria Bakola1, Anna Koralia Sakaretsanou1, Georgios Vasileiadis1, Ileana Gefaell Larrondo2, Sara Ares-Blanco2, Grigoris Alokrios1, Eleftheria Kariori1, Athina Lani1,  Hergys Manaj1, Kyriakoula Tsolaki1, Sofia Papachristou1, Maria Kampouraki1,  Vasiliki Karagianni1, Eleni Jelastopulu1

1Department of Public Health, Epidemiology and Quality of Life, University of Patras, Greece
2Federica Montseny Health Care Center, Research Unit, Primary Care, Madrid, Spain

Received: 24 Oct 2025; Accepted: 05 Feb 2026

Corresponding author: Eleni Jelastopulu, Department of Public Health, Epidemiology and Quality of Life, University of Patras, Greece. Tel.: +30 6977 624 636, E-mail: jelasto@upatras.gr

Keywords: Physician personality, Big Five traits, primary care, chronic disease, diabetes

 

Abstract

Background: The personality traits of physicians may subtly influence the type of patients they attract and the disease categories they manage. Although communication styles and therapeutic approaches have been linked to interpersonal traits, the potential correlation between personality and the epidemiological profile of patients in primary care remains underexplored.

Methods: A cross-sectional exploratory study was conducted using a structured online questionnaire distributed to 82 general practitioners (GPs) in Greece. Data collected included physician demographics, estimated patient counts per disease category and personality profiles based on the IPIP-50 Big Five Inventory. Statistical analysis was performed using Python and Excel tools, including descriptive analysis, Spearman’s correlation (selected for robustness to non-normality) and linear regression modeling used in an exploratory manner to assess explanatory power.

Results: Extraversion was significantly associated with a higher number of patients with diabetes (ρ = 0.29, p = 0.005). Other Big Five traits demonstrated no statistically significant correlation with disease categories. Regression models revealed overall low explanatory power, including for diabetes (R² = 0.196). Additionally, gender-based differences were observed, with male physicians scoring higher in Emotional Stability and Openness. No significant associations were found between personality traits and physicians’ age, experience, or chronic illness history.

Conclusions: Physician personality traits, particularly extraversion, may be weakly associated with the type of chronic patients encountered, especially in conditions requiring frequent interaction and adherence strategies such as diabetes. However, organizational, demographic, and health system–level factors are likely to exert stronger influences on patient distribution. This study highlights the need for further exploration using larger samples, objective patient records, and longitudinal or mixed-method designs to better understand the interplay between physician psychology and patient behavior in primary care contexts.

INTRODUCTION

The physician–patient relationship is a cornerstone of effective clinical practice, particularly in primary care settings where longitudinal engagement and trust are critical. While clinical competence and systemic structures are key determinants of care quality, a growing body of evidence suggests that physicians’ personality traits also influence various aspects of medical interaction, including communication patterns, empathy, stress response, and even diagnostic behavior [1,2]. These interpersonal dimensions may be especially salient in primary care, where sustained contact and continuity of care shape both clinical outcomes and patient retention.

The Five-Factor Model (also known as the Big Five) has become a widely accepted taxonomy in personality psychology, encompassing the dimensions of Extraversion, Agreeableness, Conscientiousness, Emotional Stability (the inverse of Neuroticism), and Openness to Experience [3]. These traits have been associated with professional behavior and well-being among healthcare professionals [4], as well as with clinical decision-making under uncertainty, burnout susceptibility, and responsiveness to patient distress [5]. Accordingly, personality traits have been increasingly examined as potential modifiers of clinical practice style rather than direct determinants of clinical competence.

However, despite increasing interest in personality-based studies in medical education and health psychology, limited empirical evidence exists regarding whether a physician’s personality influences the types of patients they attract or retain. Particularly in primary care, where physicians often act as a gatekeeper and longitudinal health managers, it is conceivable that the interplay of personality and patient characteristics might generate non-random patterns in patient distribution by diagnosis or disease type. At the same time, such potential effects are likely to be modest and embedded within broader organizational and system-level determinants of care delivery.

This study aims to address this gap by exploring whether specific personality traits of general practitioners (GPs) correlate with the prevalence of chronic diseases among their regular patients. Conducted within the context of the Greek Primary Health Care (PHC) system, the study uses a structured psychometric instrument to assess personality and investigates associations with patient population characteristics. Adopting an exploratory approach, it seeks to determine whether physician personality can be considered a secondary and non-dominant contributing variable in the epidemiological landscape of outpatient care, complementing traditional determinants such as demographics, training and health system design.

MATERIALS AND METHODS

Study Design and Population

This was a cross-sectional, observational exploratory study conducted between April and May 2025. The sample included 82 GPs working in various Primary Health Care Units across Greece. Participants were recruited through targeted email invitations distributed to professional medical associations, PHC networks, and university-affiliated physician groups. Inclusion criteria were active clinical practice in a PHC setting and willingness to provide data through a structured online survey. Given the sample size and recruitment strategy, the study was designed to generate hypotheses rather than to support causal inference or population-level generalization.

Survey Instrument

The research instrument consisted of a three-part structured questionnaire developed in Google Forms:

  1. Demographic and professional characteristics: age, sex, geographic region, years of clinical experience, chronic illness status, and employment sector (public/private).
  2. Patient disease profiles: physicians were asked to estimate the number of patients they manage regularly in seven chronic disease categories: diabetes (E08-E13), anxiety and stress related disorders (F40-48), depression (F32.9), chronic obstructive pulmonary disease (COPD) (J40-44), osteoarthritis (M15-M19), coronary heart disease (I10-I25), and cerebrovascular disease, stroke (I63).
  3. Personality traits: assessed via the IPIP-50 inventory, a publicly available and psychometrically validated tool that measures the Big Five personality traits using a five-point Likert scale [6].

The questionnaire underwent face validation by a panel of three experts in psychology and primary care medicine before dissemination. Average completion time was approximately 10–12 minutes.

Data Collection and Ethics

All responses were collected anonymously. Participants provided informed consent electronically, and the study protocol was reviewed and approved by the Research Ethics Committee (REC) of the University of Patras, Greece (15926/19-12-2023). No patient-identifiable data were used or collected.

Data Processing

Data were exported to Microsoft Excel for initial cleaning and subsequently processed in Python 3.10 using the Pandas, NumPy, and SciPy libraries. Outlier control was performed by visual inspection, and biologically implausible extreme values (e.g., unrealistically high patient counts) were replaced with the median value of the respective variable to reduce distortion of the results, rather than excluding entire observations. Given the self-reported nature of caseload estimates, no imputation was applied, and descriptive statistics were computed for all variables using complete-case data only.

Statistical Analysis

Spearman’s rank correlation coefficient was used to explore monotonic associations between personality traits and the number of patients in each disease category, as this method is robust to non-normal data distributions and appropriate for exploratory analyses in relatively small samples [7]. Independent samples t-tests were used to examine personality differences by gender, chronic illness status, and employment sector. Where distributional assumptions were not fully met, results were interpreted conservatively. Multiple linear regression models were constructed to evaluate the exploratory explanatory capacity of each Big Five trait on disease-specific patient counts, using one trait per model to avoid multicollinearity and to reduce instability in coefficient estimation. Statistical significance was set at p < 0.05, with emphasis placed on effect size magnitude and overall model performance rather than statistical significance alone. No imputation was applied, and all analyses were based on complete-case data to preserve transparency given the self-reported nature of the dataset.

All visualizations and statistical outputs were generated in Python using the Matplotlib and Seaborn libraries. Figures were designed to facilitate descriptive interpretation rather than inferential generalization. Summary statistics and plots are presented in Figures 1 and 2.

Figure 1. Analysis of Patient Profiles Managed by Physicians.
Figure 1 presents a pairwise scatterplot matrix (with histograms on the diagonal) illustrating the distribution and interrelationships among the numbers of patients managed by participating physicians across multiple chronic disease categories. Each row and column corresponds to a specific patient group, including total patients, diabetes, anxiety-related disorders, depression, COPD, osteoarthritis, coronary heart disease, and stroke. Diagonal panels (histograms) show the univariate distribution of patient counts for each condition. These distributions are generally right-skewed, indicating that most physicians manage relatively small to moderate numbers of patients per condition, while a smaller number report substantially higher caseloads. Off-diagonal panels (scatterplots) display bivariate relationships between pairs of disease categories. Each point represents one physician, plotted according to the number of patients managed in the two corresponding categories. The scatterplots reveal considerable variability across physicians, with no strong linear patterns for most disease combinations. Overall, the figure highlights the heterogeneous nature of GP caseloads and supports the conclusion that patient distributions vary substantially between physicians, without clear clustering or strong interdependence across most chronic disease categories.

Figure 2. Analysis of Physicians’ Personality Trait Scores.
Figure 2 presents a pairwise scatterplot matrix with histograms on the diagonal illustrating the distribution and interrelationships among the five personality traits of participating physicians, as measured by the IPIP-50 Big Five inventory: Extraversion, Agreeableness, Conscientiousness, Emotional Stability, and Intellect/Imagination (Openness to Experience). Diagonal panels (histograms) depict the univariate distribution of each personality trait. All traits show approximately normal to mildly skewed distributions, centered around mid-to-high values on the 1–5 Likert scale. This indicates relatively balanced personality profiles within the sample, with no extreme clustering at the scale boundaries. Off-diagonal panels (scatterplots) display the bivariate relationships between pairs of personality traits. Each point represents an individual physician. The scatterplots reveal weak to moderate positive associations between some traits (e.g., Agreeableness and Conscientiousness), while most trait combinations show substantial dispersion and no strong linear patterns. Overall, the absence of tight clustering or pronounced linear trends suggests that the Big Five dimensions in this physician sample are largely independent, consistent with the theoretical structure of the Five-Factor Model.

RESULTS

Sample Characteristics

A total of 82 GPs completed the survey. The majority were female (n = 54, 67.1%) and worked in the public sector (90.2%). Most respondents were aged 40–60 years, with a mean age of 49.6 years (SD = 8.2) and an average of 13.9 years of professional experience in primary health care. Approximately 34.1% reported a personal history of chronic illness. No statistically significant associations were observed between demographic variables (age, years of experience, chronic illness status) and personality trait scores. Descriptive statistics of participant demographics are presented in Table 1.

Patient Profiles

Respondents estimated the number of patients they regularly manage in seven predefined chronic disease categories (Figure 1). The most frequently encountered condition was diabetes mellitus, followed by osteoarthritis and coronary heart disease. Psychosomatic conditions such as anxiety-related disorders and depression were less prevalent in the reported patient mix. Variation in patient distribution was observed across physicians but showed no systematic association with demographic variables. It should be noted that certain conditions, such as COPD, may be partially managed in specialist settings, potentially contributing to lower representation in GP-reported caseloads. Outlier control procedures were applied through visual inspection, and biologically implausible extreme values were replaced with the median value of the corresponding variable to minimize distortion of descriptive and inferential statistics, without excluding entire observations from the analysis.

Personality Trait Scores

Personality assessment via the IPIP-50 inventory yielded the mean trait scores (on a 1–5 Likert scale) presented in Table 2. Trait scores were approximately normally distributed (Figure 2). Male physicians scored significantly higher than female physicians in Emotional Stability (p = 0.001) and Openness to Experience (p = 0.027) (Figure 3). No statistically significant differences were observed based on chronic illness status or employment sector. The observed gender differences were not associated with systematic variation in patient disease profiles.

Figure 3. Comparative Analysis of Personality Traits by Gender.

Correlation Analysis

Spearman’s rank correlation coefficients were calculated to assess relationships between personality traits and the number of patients managed per disease category. A statistically significant positive correlation was observed between Extraversion and the number of patients with diabetes (ρ = 0.29, p = 0.005) indicating a weak-to-moderate association. No other significant correlations were found across the remaining traits and conditions (Figure 4).

Figure 4: Correlations between Physicians’ Personality Traits and Patient Groups.
Figure 4 presents a correlation heatmap illustrating the strength and direction of associations between physicians’ personality traits (Extraversion, Agreeableness, Conscientiousness, Emotional Stability, and Intellect/Imagination) and the sizes of patient groups across different chronic disease categories, as well as total patient caseload. Correlation coefficients (Spearman’s ρ) are displayed within each cell, with color intensity indicating the magnitude of the association. • The upper-left block of the matrix shows correlations among personality traits. These correlations are generally low to moderate, indicating that the Big Five dimensions are largely independent, consistent with personality theory. • The lower-right block displays correlations among patient groups. Here, correlations are consistently high and positive, reflecting the expected pattern that physicians with larger overall caseloads tend to manage higher numbers of patients across multiple chronic conditions. • The cross-block area—where personality traits intersect with patient groups—reveals mostly weak correlations, indicating limited direct association between physician personality and the size of disease-specific patient groups. • One notable exception is the positive correlation between Extraversion and diabetes (ρ ≈ 0.29), which is visibly stronger than other personality–disease pairings but still falls within the weak-to-moderate range. Overall, the heatmap visually reinforces the statistical conclusion that patient group sizes are strongly interrelated, while personality traits exhibit only modest and selective associations with patient profiles. ILLUSTRATIVE EXAMPLE: For example, the cell corresponding to Extraversion and diabetes shows a correlation coefficient of approximately ρ = 0.29, indicating that more extraverted physicians tend to report managing slightly higher numbers of patients with diabetes. However, this association is substantially weaker than the correlations observed between diabetes and other patient groups (e.g., diabetes and depression or COPD), which exceed ρ = 0.75. This contrast illustrates that disease co-occurrence and overall caseload size explain much more variance in patient numbers than physician personality traits, supporting the interpretation that personality functions as a contextual modifier rather than a primary determinant of patient distribution. Taken together, Figure 4 demonstrates that while physician personality traits may relate selectively to certain patient groups, the dominant structure of the data is driven by caseload size and disease clustering, rather than by individual psychological characteristics.

Regression Models

Seven multiple linear regression models were constructed, each using one Big Five trait as an independent variable and the number of patients per disease category as the dependent variable. Only the model examining Extraversion and diabetes demonstrated a positive coefficient with explanatory value (R² = 0.196). Even in this case, the proportion of explained variance was limited and all other models yielded negligible or negative R² values, underscoring the absence of meaningful predictive capacity of personality traits for patient caseload composition (Figure 4).

DISCUSSION

This study examined whether the personality traits of GPs influence the profiles of patients they manage in primary care, particularly across chronic disease categories. The findings suggest that while Extraversion may be positively associated with the number of patients with diabetes, this association is weak and does not translate into meaningful predictive capacity, and no significant correlations emerged for other personality traits or disease groups. Furthermore, male physicians scored higher than female physicians in Emotional Stability and Openness to Experience, results consistent with certain findings in the medical psychology literature [8] but not associated with systematic differences in patient disease profiles.

The observed association between Extraversion and diabetes may be attributable to the interpersonal demands of diabetes management, which often requires frequent physician–patient interaction, counseling, and behavioral support. Extraverted physicians are more likely to engage actively with patients, employ collaborative communication styles, and foster ongoing therapeutic relationships, factors that can enhance patient retention and trust [9,10]. Consequently, diabetic patients, who frequently require long-term follow-up and lifestyle modification, may gravitate toward or remain longer with physicians exhibiting these traits. Nevertheless, the low proportion of explained variance indicates that such interpersonal mechanisms operate alongside, rather than independently of, broader structural determinants of care.

Contrary to expectations, no significant associations were observed between Conscientiousness, a trait commonly linked with structure, responsibility, and adherence to guidelines, and the number of patients with chronic conditions such as coronary heart disease. This finding may reflect the uniform application of clinical protocols across the PHC system, which likely minimizes the observable influence of individual personality variation in guideline-driven care. In this context, standardized treatment pathways may attenuate any potential effect of physician-level behavioral differences.

Similarly, Emotional Stability and Agreeableness did not correlate with higher numbers of patients presenting with psychosomatic conditions such as depression or anxiety-related disorders. Although previous studies have suggested that emotionally stable or empathetic physicians are more effective in managing patients with mental health needs [11], this was not supported by the present data. A possible explanation lies in the self-reported nature of patient estimation, the underdiagnosis of mental health conditions in primary care and the frequent involvement of specialist services, all of which may obscure subtle physician-related effects.

The higher Openness to Experience scores observed among male physicians may reflect generational or cultural factors rather than clinical preferences, although further qualitative research would be needed to explore these patterns in greater depth. The overall weak correlations shown in Figure 4 indicate that, while physician personality may exert a modest influence on patient composition, organizational characteristics of practices, population catchment profiles, referral mechanisms and health system design are likely to play more substantial roles [12].

The novelty of this study lies in its attempt to bridge two distinct domains, personality psychology and the epidemiology of care, within the real-world context of Greek primary health care. Although the results did not yield strong predictive models, they underscore the complex and multidimensional nature of physician–patient dynamics and highlight new avenues for interdisciplinary investigation. Importantly, the findings support a conceptualization of physician personality as a contextual modifier rather than a primary determinant of patient distribution in primary care.

Comparative Discussion with International Studies

The findings of the present Greek primary care study align with, but also meaningfully diverge from, prior international research examining the role of physician personality, interpersonal characteristics, and contextual factors in clinical practice. Overall, the comparison suggests that physician personality operates primarily as a contextual and relational modifier, rather than as a dominant determinant of patient composition or health outcomes.

McManus et al. (2004), in a large longitudinal cohort of UK physicians, demonstrated that personality traits, particularly Neuroticism and Extraversion, were strongly associated with stress, burnout and approaches to work, but not with objective measures of patient mix or disease epidemiology [13]. This is consistent with the present study, where Big Five traits showed minimal explanatory power for patient disease profiles (R² = 0.196 at best), reinforcing the idea that personality influences how physicians practice rather than which patients they treat.

In contrast, Hojat et al. (2011) reported a robust association between physicians’ empathy and objective clinical outcomes among diabetic patients, including better HbA1c and LDL control [14]. While their study identified empathy as an independent predictor of outcomes, it is important to note a key methodological distinction: Hojat et al. relied on electronic health records and laboratory values, whereas the present study focused on self-reported patient composition and caseload distribution. Thus, the positive association observed in the Greek sample between Extraversion and diabetes caseload appears conceptually compatible with Hojat et al.’s findings, but reflects patient retention and relational continuity, rather than physiological disease control.

System-level explanations are strongly supported by Starfield et al. (2005), who demonstrated that population health outcomes are far more strongly shaped by primary care system characteristics, such as accessibility, continuity, and coordination, than by individual physician attributes [12]. The weak correlations observed in the present study directly reinforce Starfield’s conclusion that structural determinants dominate over individual-level psychological factors in shaping epidemiological patterns in primary care.

Evidence from Wong et al. (2013) further supports this interpretation by showing that even temporary contextual disruptions, such as facemask use, can significantly reduce patients’ perception of physician empathy and relational continuity [15]. This finding underscores that situational and organizational factors may outweigh stable personality traits in influencing patient–doctor relationships, a conclusion consistent with the limited predictive capacity of personality traits observed in the Greek data.

Finally, Barnsley et al. (1999) found that physician sex, specialty, and cohort were associated with communication style and empathic behaviors, but not with disease distribution or clinical caseload composition [16]. The present study similarly identified gender differences in Emotional Stability and Openness, without corresponding differences in patient disease profiles, reinforcing the conclusion that interpersonal style does not translate directly into epidemiological differentiation.

Taken together, these comparisons position the present study firmly within the contemporary literature: physician personality matters, but primarily as a relational lens through which care is delivered, not as a driver of patient allocation or disease prevalence.

Future Research

Future research should build on the exploratory findings of the present study by adopting designs and data sources that allow stronger inference regarding the role of physician personality in primary care dynamics. A first and essential direction concerns the use of objective patient registry or electronic medical record (EMR) data, rather than physician self-estimates of caseload composition. Linking validated personality assessments of physicians with routinely collected administrative or clinical data would enable more precise measurement of patient volumes, diagnostic categories, follow-up frequency, and health outcomes, thereby reducing reporting bias and improving internal validity.

Second, longitudinal study designs are needed to examine whether physician personality traits predict patient-related processes over time. Such designs would allow researchers to move beyond cross-sectional associations and assess whether traits such as Extraversion or Emotional Stability are associated with patient retention, continuity of care, adherence to treatment, consultation frequency, or follow-up intensity across extended periods. Longitudinal data would also make it possible to explore potential bidirectional effects, whereby patient mix and workload may in turn influence physician behavior, well-being, or professional development.

Third, future work would benefit from mixed-method approaches that integrate quantitative analysis with qualitative inquiry. In-depth interviews or focus groups with physicians and patients could illuminate how personality traits interact with communication style, practice organization, team dynamics, and patient preferences in shaping care relationships. Such qualitative insights would help contextualize quantitative findings and clarify the mechanisms through which personality operates as a contextual modifier rather than a direct determinant of patient distribution.

Finally, expanding future studies to larger and more diverse samples, including different primary care systems and organizational models, would enhance generalizability and allow comparative analyses across health systems. Together, these directions would advance understanding of how physician personality interfaces with organizational and system-level factors, contributing to a more nuanced and evidence-based account of relational medicine in primary care.

Limitations

Several limitations should be considered when interpreting the findings of this study. First, the cross-sectional and exploratory design precludes causal inference. The observed associations between physician personality traits and patient caseload composition reflect contemporaneous patterns and cannot determine directionality or temporal dynamics. In addition, the relatively modest sample size, although adequate for exploratory analyses, limits statistical power and may have reduced the ability to detect small or condition-specific associations.

Second, the study relied on self-reported estimates of patient caseloads rather than objective clinical or administrative records. Although participating physicians are expected to have a reasonable overview of their regular patient populations, recall bias and estimation error cannot be excluded. This limitation is particularly relevant for conditions that may not be managed exclusively within primary care.

In this context, COPD warrants specific consideration. In Greece, as in many health systems, COPD diagnosis confirmation, staging, and long-term follow-up are frequently shared with or led by pulmonology specialists, especially for moderate to severe disease. As a result, COPD may be underrepresented in GP-reported caseloads, and the accuracy of self-reported COPD patient numbers at the primary care level may be reduced. Consequently, the absence of statistically significant associations between physician personality traits and COPD caseloads in the present study should not be interpreted as evidence of no relationship, but rather as a reflection of care pathway patterns and shared management arrangements that place part of COPD care outside routine GP caseload accounting.

Third, mental health conditions such as anxiety-related disorders and depression may also be underdiagnosed or variably coded in primary care, further attenuating detectable associations. Finally, unmeasured contextual factors, including practice organization, catchment population characteristics, referral norms, and regional service availability, may have influenced patient distribution independently of physician-level traits.

Taken together, these limitations reinforce the interpretation of physician personality as a contextual modifier rather than a primary determinant of patient distribution and underscore the need for future studies using objective patient registries, longitudinal designs, and mixed-method approaches to more fully capture the complexity of primary care delivery.

CONCLUSIONS

This study examined the potential influence of physicians’ personality traits on the composition of their patient populations in a primary care setting, with particular emphasis on chronic disease categories. The findings provide exploratory and preliminary evidence that Extraversion may be positively associated with the number of patients diagnosed with diabetes, suggesting that personality factors could modestly influence patient retention and communication dynamics in long-term care contexts.

However, the overall predictive power of personality traits was limited, and no consistent associations were observed for other conditions. These results suggest that organizational, health system–level, and population-related factors play a more influential role in shaping patient distributions within primary care.

From a clinical perspective, acknowledging physicians’ interpersonal strengths and behavioral tendencies may facilitate more effective team allocation, medical training, and patient–doctor matching strategies, provided that such considerations are integrated alongside structural and organizational constraints of primary care delivery. From a research standpoint, future investigations using objective patient records, larger and more diverse physician samples, and longitudinal or mixed-method designs could further elucidate how personality influences care delivery and health outcomes.

In summary, while physician personality may not constitute a dominant determinant of clinical caseload, it remains a contextual modifier and an underappreciated dimension of professional identity, one that merits continued exploration within the framework of personalized and relational medicine.

Acknowledgements

The authors would like to thank all the general practitioners who generously took the time to complete the questionnaire. Their participation made this study possible.

Conflict of Interest

None to declare.

Declaration of funding sources

None to declare

Author Contributions

Conception and design MB, IGL, SAB, EJ; collection of data MB, AKS, GA, EK, AL, HM, KT, SP, MK, VK, EJ; analysis of data GV; interpretation of the data GV, EJ; drafting of the article MB, AKS, GV; critical revision of the article IGL, SAB, EJ; final approval of the article all authors.

REFERENCES
  1. McKinstry B, Ashcroft RE, Car J, Freeman GK, Sheikh A. Interventions for improving patients’ trust in doctors and groups of doctors. Cochrane Database Syst Rev. 2006;(3):CD004134.
  2. Roter DL, Hall JA, Aoki Y. Physician gender effects in medical communication: a meta-analytic review. JAMA. 2002;288(6):756–64.
  3. Goldberg LR. The development of markers for the Big-Five factor structure. Psychol Assess. 1992;4(1):26–42.
  4. Boudreau JD, Cassell EJ, Fuks A. Preparing medical students to become skilled at clinical observation. Med Teach. 2009;31(10):e570–6.
  5. Tyssen R, Dolatowski FC, Røvik JO, Thorkildsen RF, Ekeberg Ø, Hem E, et al. Personality traits and types predict medical school stress: a six-year longitudinal and nationwide study. Med Educ. 2007;41(8):781–7.
  6. Goldberg LR, Johnson JA, Eber HW, Hogan R, Ashton MC, Cloninger CR, et al. The International Personality Item Pool and the future of public-domain personality measures. J Res Pers. 2006;40(1):84–96.
  7. Zar JH. Biostatistical Analysis. 5th ed. Upper Saddle River (NJ): Prentice-Hall; 2010.
  8. Costa PT, Terracciano A, McCrae RR. Gender differences in personality traits across cultures: Robust and surprising findings. J Pers Soc Psychol. 2001;81(2):322–31.
  9. DiMatteo MR. The role of the physician in the emerging health care environment. West J Med. 1998;168(5):328–33.
  10. Hall JA, Roter DL. Do patients talk differently to male and female physicians? A meta-analytic review. Patient Educ Couns. 2002;48(3):217–24.
  11. Shanafelt TD, West CP, Sinsky C, Trockel M, Tutty M, Satele DV, et al. Changes in burnout and satisfaction with work-life integration in physicians and the general US working population between 2011 and 2017. Mayo Clin Proc. 2019;94(9):1681-94.
  12. Starfield B, Shi L, Macinko J. Contribution of primary care to health systems and health. Milbank Q. 2005;83(3):457–502.
  13. McManus IC, Keeling A, Paice E. Stress, burnout and doctors’ attitudes to work are determined by personality and learning style: a twelve-year longitudinal study of UK medical graduates. BMC Med. 2004;2:29.
  14. Hojat M, Louis DZ, Markham FW, Wender R, Rabinowitz C, Gonnella JS. Physicians’ empathy and clinical outcomes for diabetic patients. Acad Med. 2011;86(3):359-64.
  15. Wong CK, Yip BH, Mercer S, Griffiths S, Kung K, Wong MC, Chor J, Wong SY. Effect of facemasks on empathy and relational continuity: a randomised controlled trial in primary care. BMC Fam Pract. 2013;14:200.
  16. Barnsley J, Williams AP, Cockerill R, Tanner J. Physician characteristics and the physician-patient relationship. Impact of sex, year of graduation, and specialty. Can Fam Physician. 1999;45:935-42.