Medical students’ perceptions of the patient-centredness of the learning environment

Patient-centred care is an important aspect of quality health care. The learning environment may impact medical students’ adoption of patient-centred behaviours. All medical students at a single institution received an anonymous, modified version of the Communication, Curriculum, and Culture instrument that measures patient-centredness in the training environment along three domains: role modelling, students’ experience, and support for patient-centred behaviours. We compared domain scores and individual items by class year and gender, and qualitatively analyzed responses to two additional items that asked students to describe experiences that demonstrated varying degrees of patient-centredness. Year 1 and 2 students reported greater patient-centredness than year 3 and 4 students in each domain: role modelling (p = 0.03), students’ experience (p = <0.001), and support for patient-centred behaviours (p < 0.001). Female students reported less support for patient-centred behaviours compared with male students (p = 0.03). Qualitative analysis revealed that explicit patient-centred curricula and positive role modelling fostered patient-centredness. Themes relating to low degrees of patient-centredness included negative role modelling and students being discouraged from being patient-centred. Students’ perceptions of the patient-centredness of the learning environment decreased as students progressed through medical school, despite increasing exposure to patients. Qualitative analysis found that explicit patient-centred curricula cultivated patient-centred attitudes. Role modelling impacted student perceptions of patient-centredness within the learning environment.


Introduction
Patient-centred care is a model for collaborative medical interactions. Components of patient-centred care include: 1) exploring both the disease and the patient's illness expe-rience, 2) understanding the whole patient, 3) finding common ground between patient and physician about priorities and management, 4) incorporating prevention and health promotion, 5) enhancing the patient-physician relationship, and 6) being realistic about time and resources [1]. The patient-centred model is associated with improved patient outcomes [2], a finding that prompted the Institute of Medicine to highlight patient-centred care as one of six core aspects of high quality health care [3]. Medical schools have developed curricula to teach patient-centred practices. However, informal aspects of the learning environment may impact medical students' adoption of patient-centred beliefs and behaviours. These include the interpersonal, unscripted interactions that constitute the informal curriculum, and the cultural structures that define the hidden curriculum [4].
The Communication, Curriculum, and Culture instrument is a validated tool that quantitatively measures patient centredness in medical school learning environments. This 29-item questionnaire produces three subscale scores that quantify the extent to which attending physicians and house staff model patient-centred behaviour, the relative patientcentredness of students' experiences, and perceived support for students' patient-centred behaviours [5]. The instrument was developed using a process of item writing and refinement, item selection, and determination of reliability and validity as described by the authors [5]. Although the instrument was found to be reliable and valid, a notable limitation is that there is no gold standard for the definition of patient centredness, and for that reason expert opinion was used for item writing and refinement. Our objective was to use a modified version of the Communication, Curriculum, and Culture instrument that pertained to the experiences of both preclinical and clinical students to evaluate the patient centredness of aspects of medical school training. We sought to measure the perception of the patient-centred nature of the learning environment, determine whether that perception varied by year of training or gender, and elucidate the types of experiences students perceived to have high or low degrees of patient-centred.

Study methods and measure
An online anonymous survey was sent to all 498 medical students at the Johns Hopkins University School of Medicine in Baltimore, Maryland. The survey, sent in the late spring and therefore near the end of the class year, included the Communication, Curriculum, and Culture instrument [5], a validated instrument that quantitatively measures the patient-centredness of the medical study learning environment. This instrument measures three domains on 5-or 7-point Likert scales. Role modelling was scored by averaging of ten items measured on a 7-point scale indicating frequency of patient-centred observations (i. e. 'Please indicate how often you observed chief residents communicate concerns, and interested in patients as unique persons.') The students' experience domain was measured by the mean of 11 items, with 4 items composing the 'learning relationships' dimension being reversed scored, on a 5-point scale indicating how often he/she has experienced a similar situation (i. e. 'You hear students telling stories about patients. These stories tend to portray patients as diagnoses rather than unique human beings.') Support for students' own patient-centred behaviours was measured using a 5-point Likert scale from 'completely encouraged' to 'discouraged.' 'Respondents' are asked to rate the response that was received from instructors by filling in the blank (i. e. 'In general, when I made an effort to develop rapport with patients, my instructors ... me.') A higher score represented greater patient-centredness in the role modelling and the support for students' behaviours domains; a lower score represented greater patient-centredness in the students' experiences domain.
Because the wording of the original Communication, Curriculum, and Culture instrument pertains only to medical students during their clinical years (i. e., years 3 and 4), we created an additional version of the instrument for preclinical students (years 1 and 2) in which titles referenced in particular items (e. g. chief residents, senior residents, or interns) were made relevant (e. g. course faculty or clinical mentors). We also added two qualitative survey items to characterize students' personal patient-centred experiences through the following questions: 1) Describe an encounter or experience that reflected a high degree of patient-centredness, and 2) Describe an encounter or experience that reflected a low degree of patient-centredness. We gathered information on respondents' gender (female, male) and class year [1][2][3][4].
This work was approved by the Johns Hopkins University Institutional Review Board (Study number CR00009001) and was carried out in accordance with the Declaration of Helsinki.

Study setting
The curriculum of the Johns Hopkins University School of Medicine is called the Genes to Society curriculum (available online at http://www.hopkinsmedicine.org/som/ curriculum/genes_to_society/curriculum/interactive_map. html) and encompasses students' full four years of medical school. This curriculum emphasizes an integrative approach to patient care. For this reason, clinical experiences begin early, and include a longitudinal clerkship in years 1 and 2. However, the frequency of clinical experiences increases significantly in years 3 and 4, and the bulk of clinical rotations occur during the second two, traditionally named 'clinical' medical school years. Clinical rotations occur in the 'real world' clinical environment as opposed to classroom and simulation experiences that are planned and can be specifically created to emphasize patient-centred skills.

Study analysis
The modified Communication, Curriculum, and Culture instrument quantitatively measures three patient-centred domains on 5-or 7-point Likert scales: role modelling, students' experience, and support for students' own patientcentred behaviours. The distributions of scores in each domain were within acceptable limits of skewness. Mean differences in domain scores were compared by gender and class year (years 1 and 2 and years 3 and 4) using independent groups t-tests (Microsoft Excel version 14.5.3). In addition, we were interested in determining how often students responded in a desired patient-centred fashion by class year. For the role modelling domain with a 7-point Likert scale, a response of 6 or 7 was considered desired, and for the students' experiences and support for student behaviour domains with a 5-point Likert scale, a response of 4 or 5 was considered desired. For each item we computed chi square tests to examine whether there were differences in the proportion of patient-centred responses between year 1 and 2 vs. 3 and 4 students. The confidence level of all analyses was 0.05.
Qualitative responses were evaluated using an editing style analysis [6]. Three study team members (MW, MO, and LAH) independently reviewed students' survey responses to identify thematic categories and subcategories prior to discussing this with the other two members. Year 1 and 2 and year 3 and 4 responses were reviewed together. Individual coding strategies were developed by each team member to identify themes. Themes were compared and those common to all team members were accepted. Two investigators were medical students at the time of the study; one investigator was a faculty member. All three investigators participated in each step of the analysis. All decisions were made by team consensus.

Results
Thirty-one percent of students (156/498) completed the survey; 49% of respondents were female and 51% were male. The overall student body composition is approximately 50% female. The response rates for year 1 and 2 (preclinical) and year 3 and 4 (clinical) students were similar, with 77 preclinical and 79 clinical student responses.
Preclinical students reported significantly greater patient centredness than clinical students in all three domains. For role modelling (preclinical students' mean = 5.27 vs. clinical students' mean = 4.96, the 0.32) difference in means between groups indicated significantly higher ratings among preclinical students (95% CI [0.03, 0.61], p = 0.03). On students' experience ratings, preclinical students ratings (mean 2.48) were significantly lower than clinical student ratings (mean = 2.79, 95% CI [-0.45, -0.15], p < 0.001). In the domain measuring support for students' own patient-centred behaviours, preclinical student scores (mean 4.24) were significantly higher than clinical student scores (mean = 3.76, 95% CI [0.20, 0.77], p = 0.001). Aggregated across all four years, female students (mean 4.15) reported lower support for their own patient-centred behaviours compared with male students (mean = 3.83, 95% CI [0.04, 0.62], p = 0.03). There were no significant differences in mean scores by gender in role modelling or student experiences domains year 1/2 and 3/4 student responses were also compared by item. Multiple significant differences were observed (Table 1). Some student experiences differed based on year in school. Year 3 and 4 students reported patients were perceived or treated as objects more often than year 1 and 2 students. Clinical students also felt they received less feedback about their bedside manner and listening skills than did year 1 and 2 students. This correlates with the domain of support for patient-centred behaviours as year 3 and 4 students felt less supported than year 1 and 2 students when acting in a patient-centred manner. Notably, year 1 and 2 students reported greater patient-centredness for every item in which a significant difference was found.
Despite their disparate characterization of the patientcentredness of their learning environment, students from all class years qualitatively described experiences which pin our analysis had similar themes emerge regarding experiences with high (Table 2) and low (Table 3) degrees of patient centredness. When asked about highly patient-centred experiences, the following themes emerged from students' responses: (1) explicit patient-centred teaching in the curriculum, (2) positive role modelling, and (3) independent time to interact with patients all foster patient-centredness. Themes that were identified from students' descriptions of experiences with low degrees of patient-centredness are as follows: (1) negative role modelling (with sub-themes of ignoring patient concerns, poor communication, and lapses in professionalism), (2) students discouraged from performing patient-centred behaviours, and (3) objectification/lack of humanism.  Table 2 Themes with representative quotes from medical student experiences that were highly patient-centred

Students are influenced by practitioners who model PC approaches and interactions
'I worked with my attending who was just incredibly understanding and compassionate with parents of a newborn who were unwilling to vaccinate their child ... My attending was truly professional, non-judgmental, and compassionate throughout the encounter.' 'ED resident who would squat next to the bedside of every patient so he would be on eye level.'

Other factors
Students need opportunities to independently practice implementing patient-centred skills 'During LC when taking 30 min interviews independent from supervision. Those were the times when there was the least pressure to get clinical information and then stop, but rather to get to know the patient as a person.

Discussion
In this mixed methods approach to understanding the learning environment at one institution, students' reports of patient-centredness decreased as they progressed through medical school and, curiously, as exposure to patients increased. As has been noted previously in the literature [7], there has been relatively little research on factors that promote or inhibit patient-centredness among medical students. This underscores the importance of soliciting students' opinions and ideas based on their experiences in medical school. Our qualitative analysis of students' responses indicated that formal learning methods in courses and on clinical teams contributed to patient-centred learning experiences. This corresponds with previous students who suggest that patient-centred care can be taught in medical curricula [8][9][10], but does not exclude the importance of other factors.
Our study affirms that both positive and negative role modelling strongly affect students' perception of the learning environment. Some of the most detailed qualitative survey responses involved students' recollections of particular exchanges between clinical mentors and patients. This supports a number of prior evaluations of the undergraduate medical education learning environment and has been described as a major component of the 'hidden curriculum' [4,[11][12][13]. Although the learning environment significantly differs among US medical schools [14], it has been shown repeatedly over the last 50 years that medical students lose empathy and humanism and become more cynical and less patient-centred the further they go in their training [15][16][17][18][19]. Our results suggest that role modelling by clinical mentors plays a major role in this trend.
Female students in our study reported less support than male students for their own patient-centred behaviours during medical school. Previous research demonstrates that female students consistently display greater empathy and patient-centredness than male students during patient encounters [18,20]. It is possible that female students do not feel as supported in patient-centred behaviours as male students do because instructors do not see as much room for improvement in their patient-centred skills. Alternatively, this result may reflect an implicit recognition of the above noted gap between female and male students as female students feel social or professional pressures to decrease their patient-centred behaviours. Additionally, we do not know the gender of the faculty and resident trainees with whom students worked. It is possible that the gender of those supervising physicians may have affected these findings.
The major limitation to our study is the small sample size. Approximately one-third of medical students responded to the survey and it is possible that these students had uniquely strong views on the topic of patient-centredness so that their experiences may not accurately represent the larger student body. Additionally, our study population is limited to one institution during one academic year and our results may not correspond with the entirety of undergraduate medical education. Furthermore, it is difficult to know what results are clinically or educationally significant using the Communication, Curriculum, and Culture instrument. It is possible that although we have found statistically significant differences in learners' experiences, this does not have a clinically or educationally significant impact on them. Fewer students responded to qualitative survey items than quantitative questions and it is possible that our results give more of a voice to students with outlying experiences.

Conclusion
In conclusion, we found that student assessments of the patient-centredness of the learning environment decreases as students' progress from years 1/2 to years 3/4 of medical school and, curiously, as exposure to patients increases.
Students describe positive influences on the learning environment, which include explicit curricula to teach patient-centred care, and role modelling of patient-centred behaviours. They noted that negative role modelling and discouragement from supervisors were negative influences on the patient-centredness of the learning environment. Since patient-centred care is an important attribute for physicians and is a core aspect of high quality health care [3], these findings suggest a need to explicitly teach patient-centredness throughout the medical school curriculum and to focus specifically on the attitudes and behaviours modelled by more senior clinicians in the clinical years. Additionally, at our institution, we found that female medical students reported decreased support for patient-centred behaviours as compared with male medical students. This observation should be further explored to better understand the underlying reasons for this report. Standardized evaluations of institutional learning environments are necessary to effectively plan and execute interventions aimed at increasing the patient-centred behaviours of current and future clinicians.