Residents' Perceptions of Their Own Professionalism and the Professionalism of Their Learning Environment

Trust in physicians and faith in the medical profession is thought to contribute to good health care outcomes by promoting patient engagement in healing. Important to this trust is a belief that physicians demonstrate professionalism and act in the patient's best interest. To promote this trust, residency programs strive to ensure that residents are able to practice as professionals, but they struggle with how to operationalize and document the learning of professional behaviors. Without being able to accurately and precisely assess professionalism, programs find it difficult to set clear standards. The challenges associated with assessing professionalism have been widely discussed1 and include questions regarding content (what is the definition of professionalism?), source (who should evaluate professionalism?), and context (within which settings and situations is professionalism most accurately assessed?). This paper reports on the development and use of a survey that incorporates a multifaceted assessment of professionalism, elicits the perspective of the resident, and focuses both on residents' abilities to be professional and on the professionalism exhibited in the residents' learning environment.

The Accreditation Council for Graduate Medical Education (ACGME) professionalism competency is defined as follows2:

This conceptualization of professionalism is complex, both in terms of focus (eg, patients, society, profession) and required behaviors, which include actions in the interpersonal (respect), ethical (integrity), practical (accountability), and personal (sensitivity) domains. As a result, many assessments and measures focus on only one or two aspects (frequently ethics) or combine professionalism with communication and interpersonal skills,3–5 perhaps because residents typically manifest their professionalism through these skills. We view professionalism as a complex set of values, skills, and behaviors that although related are not reducible to either adherence to ethical principles or communication and interpersonal skills.

Few studies have addressed professionalism as a comprehensive construct, despite some evidence of its multiple domains.6–8 In addition, assessments of professionalism usually do not focus on patient care,6,8,9 possibly to avoid overlap with other core competencies. However, professionalism in patient care is important, and residents strongly prefer clinically oriented methods for learning and assessing professionalism.10 

Professionalism is often evaluated, at least in part, based on faculty observation and interpretation of residents' clinical practice. However, studies suggest that faculty should not be the sole source of such data: they may not observe enough of particular residents' behaviors,11 faculty may refer to generalized impressions of residents,12,13 and some aspects of resident competence may unduly influence others.14,15 The ACGME recommends expanding perspectives (from nurses, physician assistants, care attendants, patients, and peers)4,16 to provide a 360° assessment of professional competence.17 And yet the resident perspective is often neglected, even though professional development requires strong skills in self-awareness and reflective practice.18 While self-assessment is known to be inaccurate for summative purposes,19 it can hone reflective ability and formatively help residents set goals and evaluate their progress.20 This process is what many see as the archetype for the development of the professional self21,22 within the self-regulating nature23 and inherent complexity24 of the medical profession.

The clinical environment also serves as the learning environment, and as such occupies a central role in the development of professionalism and should be included in assessments,25–28 following ACGME18 as well as Liaison Committee on Medical Education29 requirements that residency programs measure professional standards within learning environments and provide residents with a clinical environment that is conducive to learning. The environment can influence residents' perceptions of what constitutes acceptable clinical practice (sometimes considered the “hidden curriculum” or the “ecology of professionalism,”29 or the context of care30), and it can also determine their exposure, through role modeling,31 to a range of strategies for practicing professionally.

The goals of this paper, therefore, are to report on the development and use of a survey that asks residents to report on their own professional competence and on the professionalism of their learning environment and then to describe the relationship between these two perspectives on professionalism.

Surveys were distributed to senior residents in 5 residency programs at our institution: emergency medicine, internal medicine including categorical and primary care, pediatrics, psychiatry, and surgery. Because our primary care program only has 8 residents per year, we surveyed both the current (postgraduate year 3) and recently graduated residents. The survey was fielded online in the summer of 2006; due to scheduling, surgery residents completed a paper version. Eight waves of e-mail reminders were sent. Response rate, which did not differ by specialty or mode of administration, was 77% (85/111) (table 1).

Use of these data for research purposes was approved by the New York University School of Medicine Institutional Review Board.

The conceptual framework for the professionalism survey items was based on the consolidation of ACGME/American Board of Medical Specialties, American Board of Internal Medicine, and Association of American Medical Colleges definitions of professionalism and a comparison with other efforts to establish a professionalism framework.6,32,33 We identified actual behaviors representing each domain of professionalism by reviewing the literature and written standards of professionalism. Specific behaviors, rather than global dispositions or attitudes, were chosen because of evidence that a behavioral approach offers multiple advantages in assessment: professionalism may be best represented as a set of context-dependent behaviors30 rather than as a stable trait; evaluators are reluctant to describe individuals as unprofessional30; items reflecting professionalism should be measurable32; and assessment is more accurate when individuals are asked to rate specific aspects of performance rather than global, often value-laden abstracts,19,34,35 even in the context of self-assessment.36,37 Therefore, we identified behaviors that, when performed, reflected professionalism (categorized as perceived competence) and those that, when observed, indicated unprofessional behaviors (categorized as professionalism of residents' learning environments).

Sixty-one professional competencies were identified (11 accountability, 12 ethics, 2 altruism, 13 excellence, 14 respect, and 9 sensitivity to patients), and 32 behaviors representing lapses in professionalism in the learning environment were identified (3 accountability, 18 ethics, 5 excellence, and 6 respect). Our review failed to identify any examples of unprofessional behaviors related to 2 ACGME professionalism domains: altruism and sensitivity to patient needs. Stem questions were based on frequency: perceived professionalism was framed as ability to perform the behavior when required by circumstances (using a 4-point Likert scale: rare (1), some of the time (2), most of the time (3), and all of the time (4)), and professionalism of the learning environment was framed as frequency of observing professional lapses among fellow residents in the past 6 months (using a 5-point scale: 1, not at all; 2, once; 3, 2 to 3 times; 4, 4 to 5 times; 5, 6 or more times).

The initial draft was then systematically reviewed by the professionalism curriculum steering committee (n  =  15), representing most of the residency programs, through an elicitation questionnaire asking members to rank, delete, and/or add items. Combined with efforts to maintain representation of original domains, ensure adequate sampling of behavior within domains, and reduce respondent burden, this process led to the final selection of 20 items for assessing perceived professionalism and 11 items for assessing the professionalism of the learning environment. Tables 2 and 3 provide descriptions of and sources for these items.

Distributions of responses for each item, organized by domain, were described (tables 2 and 3). Internal consistency of items was estimated using Cronbach α. Given the nonparametric nature of our ordinal survey data, differences in distribution between items within a domain were assessed using Wilcoxon signed rank test for paired comparisons (2 items) and Friedman χ² test for repeated measures (3 or more items). In order to identify strengths and weaknesses in professional competence across domains, a professionalism competence score for each resident was calculated as the percent of items within that domain that residents reported “always” being able to perform. Similarly, scores for learning environment domains were calculated as the percent of unprofessional behaviors within that domain that residents reported witnessing at least once in the past 6 months. Differences in mean scores for these domains were then compared using repeated measures analysis of variance (with Bonferroni-corrected post hoc comparisons). Correlations (Spearman ρ) were used to explore associations between residents' professionalism domain scores and the professionalism of the learning environment domain scores.

Internal consistency of the 20 items assessing professionalism competence was .93 (Cronbach α). Alphas within the domains with multiple items were as follows: accountability  =  .61; ethics  =  .71; excellence  =  .82; respect  =  .86; sensitivity to patients  =  .90. Overall, most residents reported being able to perform the professionalism behaviors most or all of the time (table 2). However, small percents of residents reported being able to perform these behaviors only some of the time or rarely (range 1.2% [n  =  1] to 22.4% [n  =  19]), including 2 behaviors within the patient-sensitivity domain: 9.4% of residents (n  =  8) reported being able to rarely or sometimes “be sensitive to patients' immediate physical and/or emotional needs,” and 22.4% (n  =  19) reported being able to rarely or sometimes “ask patients and families about their beliefs, practices, and values when relevant to the medical issues.” The distributions for these 2 items were significantly different from those of the other domain items (F  =  50.23, P < .001). Significant differences in the distributions of items composing each domain were not found for any of the other 5 domains.

Mean percent professionalism competence scores (% of domain items the resident reported “always” being able to perform) differed significantly across domains (Friedman χ²  =  50.2, P < .001) (table 4). Accountability scores were higher than all other domain scores (F  =  10.3, P < .001): residents, on average, were “always” able to perform 51.4% of the accountability behaviors compared with a mean range of 20.8% to 39.6% for the other domains. Residents also felt able to perform a mean of 39.6% of the behaviors in the respect domain compared with a mean of only 23.9% of the altruism and 20.8% of the excellence behaviors.

Internal consistency of the 11 items assessing professionalism of the learning environment was .86 (Cronbach α). Alphas for the 3 domains with multiple items were as follows: accountability  =  .64; ethics  =  .74; respect  =  .80. Residents' report of the frequency with which they witnessed unprofessional behaviors in the past 6 months is shown in table 3. Few residents reported witnessing most of the unprofessional behaviors more than 4 times. However, there were a few exceptions: 9.4% n  =  8 of residents reported seeing someone fail to ensure transfer of responsibility, 14.1% n  =  12 of residents reported witnessing a fellow resident referring to himself or herself as more qualified than he or she was, 17.6% n  =  15 reported seeing a resident being abusive and critical during times of stress, and 21.2% n  =  18 reported seeing a resident being disrespectful, all more than 4 times in the past 6 months. Significant differences among items were found in 2 domains. In the domain of commitment to ethical principles, “falsifying medical records or misrepresenting a clinical scenario” was witnessed the least often, and “referring to oneself as more qualified than one is” was witnessed the most often (Friedman χ²  =  75.9, P < .001). And in the domain of respect, residents witnessed verbal abuse less often than disrespect or abuse and criticism during times of stress (Friedman χ²  =  30.2, P < .001). Respect scores were higher than in the other domains (F  =  8.1, P < .001). Residents witnessed a mean of 41.1% of the “disrespectful” behaviors at least once in the past 6 months, compared to witnessing a mean of 25.3% of the possible lapses in accountability, 22.3% of the possible ethical breaches, and 21.9% of the possible lapses in commitment to excellence.

Scores for the 4 domains of professional lapses in the learning environment were each significantly negatively correlated (Spearman ρ) with at least 2 of the domains of residents' perceived professionalism (table 5). Lapses in accountability in the learning environment were negatively correlated with residents' perceived competence in 5 of the 6 domains.

Our assessment of self-reported competence meets minimum requirements for reliability in terms of internal consistency. At our institution, experienced residents across specialties reported feeling fairly capable of consistently performing professionally across the 6 ACGME competency domains of professionalism. However, variations across domains and among items within domains suggest that professionalism is multifaceted, and the distribution of responses highlights some specific domains where our residents' performance could improve. For example, within the area of sensitivity and responsiveness to patient needs, residents were least able to consistently ask patients about their beliefs and be sensitive to their needs—two principles central to providing high-quality care. This may reflect the challenge of providing care within our current health care system, or it may serve to identify residents who are in need of further supervision.

Our residents' assessment of the professionalism of the learning environment also, in the aggregate, suggests that unprofessional behaviors do not occur with great frequency. However, some problematic areas were identified, particularly in terms of demonstrating respect. Additionally, a small but troubling number of residents reported that they witnessed, in a 6-month time frame, multiple failures in ensuring transfer of responsibility and multiple instances of residents' misrepresenting their qualifications. Results from this survey suggest that the climate in some clinical settings may not communicate clear standards of respect and may allow specific lapses in professionalism that require further attention. The next steps are to pinpoint which settings and rotations represent less ideal standards of professionalism than others and to include observations of the professional behavior of not just residents but also faculty and staff.

We found that residents' self-assessment of their professionalism is related to their reports of the professionalism of their learning environment. Accountability in the learning environment may be particularly important. The frequency of witnessing accountability lapses is negatively associated with residents' perceived competence in being accountable, altruistic ethical, respectful, and sensitive to patient needs. We cannot establish the direction of causality; residents who self-report a lack of professionalism may also be likely to make such attributions about their learning environment. Our findings suggest that measuring professionalism in this way provides intriguing and potentially actionable information. Establishing the link between professionalism at the individual level and at the environmental level is essential to understanding how to ensure the professionalism of resident physicians.

The relatively small sample of senior residents drawn from a single institution limits the generalizability of our results. A single source (the resident) for information on both perceived competence in professionalism and the professionalism of the learning environment creates dependencies and bias in our assessment. Further, our newly created assessment tools need additional evidence of their validity and of their reliability in additional samples. Future studies may benefit from including a larger and multi-institutional sample to enhance generalizability of results and for conducting more sophisticated analyses of the underlying structure of professionalism (eg, exploratory and confirmatory factor analysis); they may also benefit from incorporating the resident perspective into a comprehensive, 360° assessment of professionalism of residents and of the learning environment that could include peer, faculty, staff, and especially patient perspectives.46 

While these assessments need further development, they, and others like them, provide a starting point for (1) ensuring that residents are attaining adequate levels of professional competence, (2) identifying program needs and training gaps at both individual and environmental levels, and (3) enhancing residents' ability to assess and reflect upon their professionalism, an imperative for the development of professionalism.

Until we can reliably and validly assess professionalism, both in individuals and as a feature of clinical settings, we risk sending the message that professionalism is either not as important as the other competencies or is simply too complex for setting clear standards. This study provides some direction for moving us closer to effective and constructive assessment of professionalism and toward being able to identify the ways in which the learning environment may shape residents' professional development.

The authors would like to thank the Health Resources and Services Administration (HRSA), Academic Administrative Units in Primary Care (Grant #121-191-1077) for funding this project.