INTRODUCTION
In recent years, medical schools have engaged in curricular reforms aimed at fostering patient- and learner-centeredness1,2,3. Such curricular transformations are thought to counterbalance the erosion of humanistic and professional values4 and the loss of idealism5 of recently-graduated physicians. Innovations usually involve faculty, students and institutions acting as the modeling process of relationship-centered professional values6,1.
Despite attempts for curricular change, desirable outcomes towards patient-centered care have not been fully analyzed7,8,3. Most studies have assessed punctual interventions in the medical curriculum, addressing longitudinal integrated clerkship2,9 and integrated communication skills activities10,11. However, previous results of the real effects of curricular designs on medical students’ attitudes are controversial12,5.
In line with this trend, Brazilian medical schools are facing the challenge of redesigning medical curricula towards more learner-centered and patient-centered approaches, stimulated by recent national medical education guidelines13. Since the publication of these educational guidelines, Brazilian schools have been involved in different stages of curricular changes. This multicenter cross-sectional study aims at assessing medical students’ attitudes and predictors of patient-centered relationship among different curricular designs (traditional, innovative and advanced).
METHODS
Setting and study population
All medical students enrolled at the medical schools participating in the project for evaluating change and trends proposed by the Brazilian Association of Medical Education were invited to this cross-sectional study.
This national project consisted of the completion of a self-evaluation instrument designed to detect different stages of curricular changes by each academic community (students, faculty and preceptors)13. Based on this self-evaluation, medical schools may be classified into three categories of curricular design: traditional (predominance of teacher- and disease-centered approaches), innovative (with characteristics of both traditional and advanced education) and advanced (predominance of learner-centered and system-based approaches)14.
The institutional review board at each of the 21 participating medical schools approved the study prior to student recruitment. An online website-platform was used for data collection from August 2015 to March 2016. After signing informed consent forms, 1,463 medical students from the 1st to 6th year answered two online instruments: the Brazilian version of the Patient–Practitioner Orientation Scale (B-PPOS) and a questionnaire with sociodemographic and curricular variables.
Study instruments
We used the B-PPOS to assess medical students’ attitudes towards the doctor-patient relationship. The scale contains two 9-item domains: Caring and Sharing. Answers were given in a six-point Likert scale (1 = strongly agree to 6 = strongly disagree). Based on previous studies with the original PPOS, students’ attitudes could be classified as patient-centered (score ≥ 5.00), moderately patient-centered (4.57 < score < 5.00) and doctor-centered (score ≤ 4.57)15. The domains were divided into four subscales: caring – the patient’s perspective and the psychosocial context; and sharing – sharing understanding and sharing power and responsibility16,17.
We used a questionnaire with sociodemographic and curricular variables to assess students’ age, gender, father’s and mother’s educational level, religion and religious involvement, future medical service (public/private), participation in extracurricular activities, future area of specialization (clinical/surgical) and year of medical training. We also assessed school administration (public/private), school self-evaluation of curricular design, presence and duration of communication skills activities in the curricular syllabus.
Statistical analysis
We used descriptive statistics and chi-square tests to compare sociodemographic and curricular characteristics. We predicted the influence of sociodemographic and curricular variables in B-PPOS scores with Ordinary Least Squares Regression Analysis. We included only variables with significant Spearman’s correlation coefficients with the B-PPOS scores in regression models. P-values lower than .05 were significant.
A minimum of 374 medical students was needed to detect a medium effect size18. We used Cliff’s delta (d) values to calculate effect sizes and they were interpreted as follows: negligible (d < 0.147), small (d< 0.330), medium (d < 0.474) and large effects (d ≥ 0.474)19. We used IBM SPSS Statistics (version 21), Gretl (version, 1.9) and Cliff’s Delta Calculation software for our analyses.
RESULTS
Medical students’ mean age was 22.3 years (SD = 3.3), most of them being female (n = 867; 59.3%). Table 1 shows students’ sociodemographic and curricular characteristics.
Sociodemographic characteristics | Total (n = 1,463; 100.0%) | School administration | X 2 | p | |
---|---|---|---|---|---|
| |||||
Private (n = 177; 12.1%) | Public (n = 1,286; 87.9%) | ||||
Average age, years (SD) | 22.3 (3.3) | 23.3 (4.5) | 22.2 (3.0) | – | 0.06 a |
Female / Male gender, n (%) | 867 (59.3)/596 (40.7) | 117 (66.1)†/ 60 (33.9)† | 750(58.3)*/ 536(41.7)* | 3.90 | 0.05 |
Father’s and mother’s educational level, n (%) | |||||
- Higher education | 773 (52.8) / 932 (63.7) | 75 (42.4)†/ 115 (65.0)† | 698(54.3)*/ 817(63.5)† | 11.7/1.18 | 0.01 /0.76 |
- High school | 469 (32.1) / 387 (26.4) | 66 (37.3)†/ 47 (26.6)† | 403(31.3)†/ 340(26.4)† | ||
- Elementary school | 207 (14.1) / 137 (9.4) | 32 (18.1)†/ 15 (8.5)† | 175 (13.6)†/ 122 (9.5)† | ||
- Missing data | 14 (1.0) / 7 (0.5) | 4 (2.3)†/ 0 (0.0)† | 10 (0.8)† / 7 (0.5)† | ||
Curricular characteristics | |||||
Self-evaluation of curricular design, n (%) | |||||
- Advanced | 448 (30.6) | 89 (50.3)† | 359 (27.9)* | 67.4 | <0.01 |
- Innovative | 710 (48.5) | 88 (49.7)† | 622 (48.4)† | ||
- Traditional | 305 (20.9) | 0 (0.0)† | 305 (23.7)* | ||
Communication-skills activities in curriculum, n (%) | 355 (24.3) | 88 (49.7)† | 267 (20.8)* | 71.0 | <0.01 |
Future medical service, n (%) | |||||
- Public service | 599 (40.9) | 65 (36.7)† | 534 (41.5)† | 2.7 | 0.26 |
- Private service | 844 (57.7) | 111 (62.7)† | 733 (57.0)† | ||
- Others | 20.0 (1.4) | 1 (0.6)† | 19 (1.5)† | ||
Extracurricular activities, n (%) | 715 (48.9) | 107 (60.5)† | 608 (47.3)* | 10.8 | <0.01 |
Future area of specialization, n (%) | |||||
- General practice | 519 (35.5) | 74 (41.8)† | 445 (34.6)† | 11.6 | 0.02 |
- Surgical area | 314 (21.5) | 37 (20.9)† | 277 (21.5)† | ||
- Surgical clinic area | 68 (4.6) | 12 (6.8)† | 56 (4.4)† | ||
- Doesn’t know | 521 (35.6) | 46 (26.0)† | 475 (36.9)* | ||
- Other | 41 (2.8) | 8 (4.5)† | 33 (2.6)† |
Abbreviations: n (%) followed by same symbol does not differ statistically according to Chi-Square test (X 2 ). Mann-Whitney Test. Significance level: p ≤ 0.05.
Brazilian medical student scored higher for caring than for sharing (p < 0.001; d = 0.599). Caring scores related to the psychosocial context were significantly higher than scores related to the patient’s perspective (p < 0.001; d = 0.797). Sharing power and responsibility scores were significantly higher than sharing understanding (p < 0.001, d = 0.455) (Table 2).
B-PPOS | Score (SD) (n = 1,463) | pa | δb | ||
---|---|---|---|---|---|
Total score | 4.64 (0.47) | ||||
Caring domain score | 5.02 (0.45) | 0.000 | 0.599 | ||
Sharing domain score | 4.26 (0.66) | ||||
Caring subscale score | |||||
The patient’s perspective subscale (2, 3, 7, 11, 14, 17 Items) | 4.70 (0.58) | 0.000 | -0.797 | ||
The psychosocial context subscale (6, 13, 16 Items) | 5.68 (0.47) | ||||
Sharing subscale score | |||||
Sharing understanding subscale (5, 8, 10, 18 Items) | 3.88 (0.87) | 0.000 | -0.455 | ||
Sharing power and responsibility subscale (1, 4, 9, 12, 15 Items) | 4.57 (0.75) | ||||
B-PPOS | Curricular design | pc | δb | ||
| |||||
Traditional (n = 305) Score (SD) | Innovative (n = 710) Score (SD) | Advanced (n = 448) Score (SD) | |||
Total score | 4.63(0.46) | 4.64(0.48) | 4.65 (0.45) | > 0.05 | – |
Caring domain score | 5.01 (0.48) | 5.02 (0.44) | 5.04 (0.45) | > 0.05 | – |
Sharing domain score | 4.25 (0.65) | 4.27(0.68) | 4.27 (0.64) | > 0.05 | – |
pa / δb | 0.000/ 0.651 | 0.000/ 0.644 | 0.000/ 0.686 |
Abbreviations: B-PPOS: Brazilian Patient-Practitioner Orientation Scale. aWilcoxon Rank-Sum Test. Significance level: p< 0.05. bδ (Cliff’s Delta): effect size: negligible (δ < 0.147), small (δ < 0.330), medium (δ < 0.474) and large (δ ≥ 0.474). cKruskal Wallis Test: scores followed by same symbol do not differ statistically according to Student-Newman-Keuls post hoc test.
There was no statistical difference among total, sharing and caring scores according to curricular designs (p > 0.05) (Table 2).
In the regression analysis (Table 3), we observed a positive association between total B-PPOS scores with public schools, female gender, year of medical training, preference for future medical practice in public services, extracurricular activities, as well as a negative association between total B-PPOS scores with the father’s educational level and choice to study surgical specialties (p < 0.05).
OLS Analysis Regression a | Total B-PPOS score | Caring B-PPOS score b | Sharing B-PPOS score b | |||
---|---|---|---|---|---|---|
| ||||||
R 2 = 0.089, Adj.R2 = 0.082 F (12, 1450) = 11.83, * p = < 0.001 | R 2 = 0.043, Adj.R 2 = 0.035 F= (12, 1450)= 5.38, * p = < 0.001 | R 2 = 0.083, Adj.R 2 = 0.075 F (12, 1450)= 10.93, * p = < 0.001 | ||||
Predictors | B (95% CI) | * p | B (95% CI) | * p | B (95% CI) | * p |
Constant | 4.267 (4.08, 4.45) | < 0.001 | 4.702 (4.51, 4.89) | < 0.001 | 3.833 (3.57,4.10) | < 0.001 |
School administration | 0.132 (0.06, 0.21) | < 0.001* | 0.100 (0.03, 0.17) | 0.008* | 0.165 (0.06, 0.27) | 0.002* |
Self-evaluation of curricular design | 0.014 (-0.02, 0.05) | 0.401 | 0.019 (-0.01, 0.05) | 0.268 | 0.010 (-0.04, 0.06) | 0.683 |
Gender | 0.180 (0.13, 0.23) | < 0.001* | 0.110 (0.06, 0.16) | < 0.001* | 0.250 (0.18, 0.32) | < 0.001* |
Father’s education level | -0.029(-0.05, -0.01) | 0.003* | -0.021 (-0.04, 0.002) | 0.029* | -0.036 (-0.06, -0.008) | 0.010* |
Having a religion | -0.023 (-0.09, 0.05) | 0.542 | 0.013 (-0.06, 0.08) | 0.728 | -0.058 (-0.16, 0.04) | 0.270 |
Religious involvement | -0.015 (-0.05, 0.02) | 0.353 | -0.003 (-0.03, 0.03) | 0.848 | -0.027 (-0.07, 0.02) | 0.240 |
Year of medical training | 0.021 (0.00, 0.04) | 0.019* | 0.018 (-0.00, 0.03) | 0.040* | 0.023 (-0.00, 0.05) | 0.063 |
Communication-skills activities | 0.079 (-0.01, 0.17) | 0.100 | 0.056 (-0.04, 0.15) | 0.236 | 0.102 (-0.03, 0.23) | 0.132 |
Years of communication-skills activities | 0.024 (-0,02, 0.07) | 0.314 | 0.028 (-0.02, 0.07) | 0.235 | 0.020 (-0.05, 0.09) | 0.557 |
Future medical practice | 0.053 (0.03, 0.08) | <0.001* | 0.028 (0.00, 0.05) | 0.019* | 0.078 (0.04, 0.11) | < 0.001* |
Extracurricular activities | 0.068 (0.02, 0.12) | 0.008* | 0.046 (-0.00, 0.10) | 0.070 | 0.089 (-0.00, 0.15) | 0.014* |
Choice for surgical specialties | -0.028 (-0.05, -0.01) | 0.008* | -0.017 (-0.04, 0.003) | 0.095 | -0.038 (-0.07, 0.01) | 0.011* |
Abbreviations: B-PPOS: Brazilian Patient-Practitioner Orientation Scale. OLS: Ordinary Least Squares Regression Analysis. B: unstandardized coefficient. CI: Confidence Interval.Multicollinearity test: Variance Inflation Factor (1.03 ≤ VIF ≤ 3.05). Regression analysis of subscale B-PPOS scores are describe in Results sections. Significance level: *p < 0.05.
For caring scores, we found a positive association with public schools, female gender, year of medical training, preference for future medical practice in public services and negative association with the father’s educational level (p < 0.05). Sharing B-PPOS scores showed a positive association with public schools, female gender, preference for future medical practice in public services, extracurricular activities, and a negative association with the father’s educational level and choice to study surgical specialties (p < 0.05) (Table 3).
DISCUSSION
This cross-sectional study demonstrates that Brazilian medical students are moderately patient-centered. We know that holistic care and sharing decision-making varies among cultures and social norms7. Brazilian cultural and societal values reinforce medical students’ attitude in caring more than sharing in the doctor-patient relationship20. Furthermore, our students consider the patients’ psychosocial context far more than the patients’ perspective. For a more person-centered care, beyond the psychosocial context, individual patient’s goals and preferences must be at the center of the doctor-patient relationship, which requires a great change in the culture of health care and medical education21,22. Understanding care from the patient’s perspective can be fostered by educational strategies that value contact with the patients’ narratives23.
Our results show that students share more power and responsibility than understanding. Doctors and students tend to underestimate patients’ preference to be fully informed24, giving partial information about a diagnosis and limiting the patients’ opportunity for reflection25. Furthermore, physicians may consider shared decision making a simple philosophical principle or value, and may not necessarily adopt or use it to communicate with patients26. Since patients’ attitudes and behaviors influence physicians’ behavior27, patients should be encouraged by physicians to assume an active role as agents for change22, acquiring the knowledge, skills and confidence to manage their healthcare28,29. A truly shared decision approach requires both doctors and patients to be involved in information exchange and treatment decisions30,31.
Gender is one of the factors that influence the doctor-patient relationship32, and the strongest predictor of patient-centered attitudes in our study. Pre-medical school entry factors such as family background and personality have a remarkable influence on the identification of the professional role of women and patient-centeredness33.
We know that both male and female medical students have the capacity to improve their patient-centeredness, but evidence shows a gender-dependent communication style influenced by stereotypes. Female physicians and students report more interest in the doctor-patient communication and the patients’ psychosocial context, sharing more information with their patients34,35. Women also report to be more empathetic, receptive, helpful and sociable than men, whereas men see themselves as robust, overbearing and inhibited. Men tend to establish a relationship of power over the patient, concerned more with medical history and less interested in voluntarily improving their communication skills and attitudes35. In view of these gender-differences, effective strategies to implement gender-sensitive communication skills teaching is crucial to enhance the quality of medical training and care, especially among male students32,35.
Surprisingly, students from advanced curricular designs did not show more patient-centered attitudes in all domains of the B-PPOS. This could be attributed to the low offer of communication skills activities during medical training. Studies show most medical schools concentrate communication skills activities during preclinical years36. It is common knowledge that communication training improves rapport building, shared decision making37 and understanding patients’ perspectives38. Nevertheless, communication-skills learned at the beginning of medical training can be easily forgotten if not reinforced longitudinally throughout medical education39,40,11. This might explain the lack of association between patient-centered attitudes among students and the presence of communication-skills activities in their curricula.
Despite expectations of medical schools to develop humanistic and patient-centered relationships, medical training does not guarantee these desirable learning outcomes7,3. Except for the caring domain, we did not observe an influence of the number of years of training among students’ attitudes. We believe that informal aspects of medical training such as role models and the hidden curriculum contribute to these results.
Medical schools have traditionally depended on role modeling as part of the informal curriculum, in which physicians’ values, attitudes and behaviors influence students’ attitudes and career choice41,42. Depending on the clinical attributes, teaching skills and personal qualities of the faculty43, there may be a negative role model, that resists changes in medical education44 and threatens the acquisition and nourishing of humanistic skills and attitudes in future physicians12. During medical clerkship, much of the doctor-student socialization occurs outside formal places, where informal experiences become more influential, creating a cascade of events that precludes real changes in medical students’ attitude, despite curricular reforms45,3.
Additionally, the hidden curriculum also shapes values, roles and identity of medical trainees, at the level of the organizational culture and structure46,47,48, creating important barriers to effective and positive role modeling. Emotional neutralization, loss of idealism, acceptance of the hierarchical structure and the adoption of a ritualized professional identity may influence enculturation in medical training46,47,49. At the same time, lack of institutional support for faculty development, career development and formal training in teaching methods and reflective practice, associated with work overload50,41 lead to a less effective role modeling, especially in clerkship years.
Extracurricular activities are associated with more patient-centered attitudes. Participation in these activities develops empathy, allows reflection and self-development51, improves interaction with peers and people of different realities, makes students more engaged in the teaching-learning process52 and in the diversity of care53. These experiences may contribute to students’ interest in the patient’s psychosocial, cultural and environmental factors51.
The father’s higher educational level is a predictor of less patient-centered attitudes. A higher educational level is linked to higher socioeconomic conditions and higher engagement in parental education54. Previous studies have also associated physicians who have a high socioeconomic status with more difficulty in sharing information and decision-making with patients55.
Our results indicate that the choice to study surgical specialties may predict doctor-centered attitudes. Personal characteristics and personality traits may be related to the choice of medical specialty56 and can predict cognitive and clinical performance of medical students57. Students who intend to choose a surgical specialty demonstrate less social orientations, preference for acute patients, technology-oriented activities and prestige, and are more hospital-oriented than clinical-oriented students58,59. These findings suggest that students prone to surgical specialties should receive special support to foster their patient-centered attitudes towards the doctor-patient relationship.
Our study has some limitations. Since we chose an online-based questionnaire to assess students from 21 medical schools of different regions in Brazil, there were difficulties concerning response rates. Despite these issues, we achieved a representative sample of respondents. Although the use of self-reported validated scales facilitates assessing a larger sample of participants, social desirability bias inherent to such instruments may have influenced responses. Another limitation may be related to the fact that PPOS measures attitudes rather than real behavior towards patient-centeredness. Nevertheless, we know that attitudes can predict future physicians’ skills and competences60. Inferences of causality and temporality are limited by our study design. Despite such limitations, to the best of our knowledge, our study was the first to assess attitudes of students enrolled in different curricular designs, in a multicenter approach.
CONCLUSIONS
Curricular designs did not predict medical students’ attitudes. Being female and attending a public medical school were important predictors of patient-centered attitudes among medical students. Further research should investigate the direct influence of faculty professionalism development programs on patient centered-attitudes among students. Such programs should foster relationship-centered values, in order to establish a proper and safe learning environment.