We are being asked to be professional in an unprofessional environment.
This statement emerged from a trainee during an invitational conference on professionalism in medicine organized by the Association of American Medical Colleges and the National Board of Medical Examiners in 2002.1 Indeed, 98% of students surveyed at six medical schools reported witnessing unprofessional behavior among faculty who were involved in their teaching.2 The impact extends far beyond medical education: It is the lack of professional conduct, rather than inadequate knowledge or technical skills, that tends to drive complaints against physicians.3 Along with escalating demands on physician time and a more complex practice environment, there is a perception that lapses in professionalism are becoming more frequent.4,5
Although professionalism has always been a core value in medicine, it has received increasingly explicit attention over the past several years. Since the late 1990s, several influential organizations have endeavored to increase awareness of professionalism as a competency to be achieved not only in medical education but also in practice.6–16 The Physician Charter—a product of collaboration between the American Board of Internal Medicine Foundation, the American College of Physicians Foundation, and the European Federation of Internal Medicine—is a well-known document that highlights fundamental principles and responsibilities central to professionalism in medicine.9,13,14
Unfortunately, the terms that have been used to explain this competency have been rather abstract and, therefore, hard to characterize: Altruism, excellence, duty, honor, integrity, and respect are lofty concepts that may be difficult to translate into practice. Only by defining the behavioral signs of professionalism can we make this competency more tangible for providers and the patients they serve. Moreover, reliable assessment requires definition of observable behaviors rather than attributes.17–19 Accordingly, we conducted a series of qualitative and quantitative studies to systematically identify and prioritize behaviorally based signs of medical professionalism that are relevant to patients, physicians, and nurses. Our approach was to ask each of these groups what professionalism means to them.
We began by convening a national advisory board of individuals with experience and expertise in investigating and promoting professionalism in medicine (see Acknowledgments). We held two meetings to guide the research process: the first in spring 2004 to discuss the plan and procedure of the study, and the second in fall 2004 to provide a progress report and discuss options for organizing our findings.
The qualitative foundation for this project was a series of 22 focus groups designed to explore behavioral signs of professionalism in medicine. Seventeen of these groups were conducted in 2004. The initial group was with patients recruited from the lobby of a major urban tertiary-care teaching hospital in Chicago. We then held a series of 15 focus groups: five separate groups (patients, inpatient nurses, outpatient nurses, resident physicians, attending physicians) from each of three subspecialties (internal medicine, general surgery, physiatry). Patients, nurses, and physicians were recruited in the Chicago area through inpatient and outpatient facilities at the tertiary-care teaching hospital (internal medicine and general surgery) and a rehabilitation hospital (physiatry). To broaden the perspective, we conducted one additional focus group with patients who live in a rural setting in Connecticut and had received care from community physicians as well as the local community hospital. In an effort to incorporate the pediatric context, we conducted a set of five pediatric focus groups at the very beginning of 2005. These sessions with inpatient nurses, outpatient nurses, resident physicians, attending physicians, and the parents of patients, respectively, were recruited from a children's hospital. The goal of these pediatric focus groups was to ensure that the project's scope would not be limited to adult medicine.
The focus group protocol was approved by the Northwestern University institutional review board. Focus groups had two trained facilitators and an average of six to seven participants. Patient focus groups ran for 90 minutes; nurse and physician focus groups lasted 60 minutes. At the beginning of each session, participants completed an information sheet to provide demographic data. Table 1 summarizes the data regarding participant age, gender, race/ethnicity, education, and whether participants reported having seen a doctor act unprofessionally. The item about witnessing unprofessional behavior was included to prompt discussion in the group. As shown in Table 1, the vast majority of participants reported at least one instance of seeing a doctor act unprofessionally.
The 22 group discussions were videotaped to facilitate content analysis and constant comparative analysis; all three authors participated in a debriefing session after each group.20,21 When there was a discrepancy or question, we viewed the appropriate portion of the video together to reconcile differences. Items generated—whether by patients, nurses, or physicians—were included for discussion in all subsequent groups. The first 17 focus groups generated a total of 68 behaviorally based items ranging from the very basic (e.g., good hygiene) to communication-oriented behaviors, accountability, and service to the profession. The five pediatric focus groups reinforced these items; they did not generate new ones. The 68 items are displayed in List 1.
Lexile analysis for readability
The 68 items provide a foundation for surveys and potential assessment tools. Accordingly, we sought to keep the reading skills needed to comprehend the items at an eighth-grade level, a decision consistent with the Institute of Medicine's observation that individuals who have difficulty reading above this level may face problems understanding and acting on health care information.22 Each of the 68 items was subjected to a Lexile analysis for readability. Lexiles are based on sentence length and word frequency in popular literature and range from below 200L for beginning-reader material to above 1700L for advanced text.23,24 The individual items have Lexile values ranging from 180 to 960, with a mean of 607 (standard deviation [SD] = 199). These values correspond to a third- to fifth-grade reading level,25 which increases the likelihood that items can be appropriately understood and used whether self-administered or interviewer-administered.
To broaden the sample and refine the perspective gained in focus groups, we conducted national surveys of patients and physician leaders and a statewide survey of nurses. The introductory text for all surveys indicated that the topic was professionalism in medicine and noted that the list of behaviors was generated through focus groups with patients, nurses, resident physicians, and practicing physicians. All surveys asked respondents to rate the importance of each behavior as a sign of professionalism, using a four-point rating scale (1 = not at all important, 2 = slightly important, 3 = moderately important, 4 = very important). In addition, the patient survey asked respondents to indicate whether they could know if their doctor exhibited the behaviors (yes, no). The nurse survey asked whether a nurse could know this about a physician colleague, and the physician leader survey asked whether a physician could know if a physician colleague exhibited the behaviors. The survey protocol was approved by the Northwestern University institutional review board.
National patient survey.
We designed a national survey to gauge the importance Americans attach to the behaviors as signs of professionalism and to determine whether patients are in a position to see these behaviors in the clinical setting. In an effort to reduce respondent burden, we reasoned that items identified in the focus groups should be included in this survey only if they could reasonably be seen as relevant for most patients. Thus, we applied three criteria for determining whether items should be included: (1) the behaviors should be relevant to the physician-patient relationship, not to “backstage”26 or back-office behaviors that most patients would not be able to see (e.g., communicates clearly and effectively was included; communicates orders clearly and effectively was not), and (2) the behaviors should always be transparent to patients (e.g., keeps personal reactions to self is difficult to judge unless a physician expresses a personal reaction).
We endeavored to be conservative in terms of excluding items and determined, after reviewing the 68 items, that a subset of 41 signs of professionalism met the criteria outlined above for inclusion in the patient survey. We also included a behavior that is not associated with professionalism (i.e., speaks two languages) as a control item to determine whether respondents were paying attention to the items and using the scale appropriately. Thus, the patient survey included a total of 42 items. We conducted a cross-sectional, random-digit-dial, computer-assisted telephone survey of adults in the 48 contiguous United States. Patient surveys were conducted by trained telephone interviewers at Northwestern's Institute for Healthcare Studies between December 2004 and February 2005. There were 1,489 known active residential numbers in the sampling frame. Callers made up to seven attempts at each number to reach a respondent; any English-speaking adult who answered the telephone was considered a potential respondent.
Physician leadership survey.
In 2005, we conducted an online physician survey for directors of the 24 boards associated with the American Board of Medical Specialties (ABMS). Because all items are relevant to physicians, this survey included all 68 items generated during the focus groups, as well as the control item. Employing the same response scales used in the patient survey, this survey asked physician leaders to rate item importance and indicate whether a physician could know if a colleague exhibits the behaviors. The ABMS sent an e-mail message to 411 physician leaders (directors, officers, and subboard chairs of each member board), asking them to complete the survey.
In 2006, we distributed a parallel online survey to inpatient and outpatient nurses across specialties; it included all 68 items and the control item. Using the same response scales as did the patient and physician surveys, this component of the study asked nurses to rate item importance and indicate whether they could know if a physician with whom they work exhibits the behaviors. Potential respondents were recruited via fliers distributed at our academic medical center and a note in the Illinois State Nurses' Association newsletter; we do not know how many nurses saw the invitation to participate in the survey.
Criteria for importance and observability
We set a priori criteria for considering items as signs of professionalism that are meaningful to a majority of patients, physicians, and nurses. More specifically, we were interested in items that were deemed very important by at least 75% of a survey sample. Anticipating development of professionalism assessment tools, we also examined which items were considered observable by at least 75% of each group. We used SPSS to generate basic descriptive statistics to determine which professional behaviors are considered very important and observable by study participants.
Consistent with other recent random-digit-dial surveys about health, the call cooperation rate for the national survey was 28%, yielding a total of 415 patient surveys for analysis.27 Respondent age ranged from 18 to 88 years, with a mean of 47.5 (SD = 16.3). Of the 415 patient respondents, 263 (63.4%) were female, 314 (75.7%) described their race as white or Caucasian (non-Hispanic), and 40 (9.6%) self-identified as black or African American (non-Hispanic). Regarding education, 230 (55.5%) respondents had less than a college degree, and 185 (44.5%) were college graduates or had some postgraduate study.
As shown in Table 2, 30 behaviors on the national patient survey were rated as very important by at least 75% of respondents; Table 3 includes items that did not meet this importance criterion. Sixteen of the 30 items met the criteria (i.e., ≥75%) for both importance and observability (see Table 4). These included behaviors related to hygiene, privacy, communication, and follow-up. The control item (speaks two languages) was rated as very important by 69 (16.7%) patients.
All 24 specialty boards of the ABMS were represented in the sample of 214 physicians responding to the online survey. The overall response rate was 52.1%. Respondents ranged in age from 43 to 72 years (mean = 56.4, SD = 5.5); 179 (83.6%) were male. A total of 195 (91.1%) physician leaders reported direct patient-care activity, and 153 (71.5%) had a full-time academic affiliation. Ninety-eight (45.8%) of the physicians reported association with a multispecialty group practice, and 55 (25.7%) were in a single-specialty group practice.
Table 2 indicates that 28 behaviors were rated as very important by at least 75% of physicians; items that did not meet the importance criterion are listed in Table 3. Nine of the 30 behaviors deemed important by at least 75% of patients did not meet the importance criterion for doctors. Some of these items approached the 75% mark, whereas others were rated as very important by far fewer physicians: communicates with other health professionals to coordinate care (n = 159, 74.5%), pays attention to detail (n = 156, 72.9%), follows up to ensure proper care (n = 154, 71.8%), keeps patient and/or family informed and up to date (n = 135, 63.0%), explores the patient's needs and concerns (n = 131, 61.0%), is approachable (n = 129, 60.4%), is open to patient getting a second opinion (n = 110, 51.6%), pays attention to the cleanliness and comfort of patient areas (n = 96, 45.0%), and prepares before seeing the patient (e.g., reviews chart) (n = 75, 35.1%). As shown in Table 4, the items on hygiene and controlling emotions were the only two that met the criteria for both importance and observability in a physician colleague. The control item about speaking two languages was rated as very important by one doctor.
A total of 237 nurses responded to the online survey. The age of these nurses ranged from 22 to 66 (mean = 41.9, SD = 11.8), and 227 (95.8%) were female. Of the 237 nurses, 124 (52.3%) reported working in an inpatient setting, 60 (25.3%) in ambulatory settings, and 14 (5.9%) in nursing management; 35 (14.8%) noted their work setting as “other.” The five most frequently reported areas of practice were (1) pediatrics (n = 62, 26.2%), (2) intensive care/critical care (n = 51, 21.5%), (3) surgery (n = 33, 13.9%), (4) internal medicine/family medicine (n = 24, 10.1%), and (5) oncology (n = 18, 7.6%). One hundred seventy-seven (74.7%) of nurses in the sample reported working with 10 or more doctors.
As shown in Table 2, 50 of the 68 behavioral signs of professionalism were seen as very important by at least 75% of nurses in the sample. Items that did not meet the importance criterion are listed in Table 3. Two of the behaviors deemed important by at least 75% of patients were not rated similarly by nurses: is open to the patient getting a second opinion (n = 141, 59.6%) and pays attention to the cleanliness and comfort of patient areas (n = 115, 48.5%). Fourteen items met the importance criterion as well as the criterion for observability in a physician with whom the nurse works (see Table 4). The control item was rated as very important by 12 (5.1%) nurses.
Our study demonstrates that it is both possible and instructive to define professionalism in terms of tangible behaviors. Focus groups with patients, resident physicians, attending physicians, inpatient nurses, and outpatient nurses across different specialties generated a list of 68 items and provided a rich qualitative sense of professionalism in medicine. Whereas the focus groups were drawn from three different hospitals affiliated with an academic medical center and one rural site in another state, the surveys broadened the scope in terms of geography, specialty, and experience. In addition, the surveys included a control item (speaks two languages), which is not considered a sign of professionalism. This item received low importance ratings from all three groups of survey respondents, lending credibility to the quantitative results. Accordingly, survey results help refine the definition of professionalism: 53 of the 68 items were deemed very important signs of professionalism by at least 75% of respondents to the patient, nurse, and/or physician surveys. There was considerable consistency in terms of how the survey samples viewed the importance of these behaviors, but the discrepancies may be instructive as well.
Several of the behaviors listed as signs of professionalism focus on effective communication. In a few specific instances, there were clear gaps in the extent to which patients, physicians, and nurses viewed communication behaviors. For instance, whereas keeps patient and/or family informed and up to date and explores the patient's needs and concerns met the importance criterion (i.e., seen as very important by at least 75% of sample) for both patients and nurses, these behaviors were considered very important by only 63.0% and 61.0% of physicians, respectively. Recognizing these, and other, gaps in perception may help physicians and health care organizations more effectively address and improve the patient experience. At a broader level, it is important to note that communication is not a domain featured in the Physician Charter.13,14 This observation suggests that there may be value in integrating the perspectives of patients, physicians, and nurses. Taken together, the different perspectives form a well-rounded and tangible description of professionalism in medicine.
There was a wide gap in perceived importance of prepares before seeing the patient (e.g., reviews chart): 79.4% of patients deemed this very important, as compared with only 35.1% of physicians. Other studies have confirmed that patients expect physicians to prepare before the visit and that office visits without perceived preparation result in decreased patient satisfaction.28,29 Additional discrepancies between what patients and providers viewed as professional behavior included being open to the patient getting a second opinion, rated very important by 78.9% of patients, 59.3% of nurses, and 51.6% of physicians, and pays attention to the cleanliness and comfort of patient areas, which was considered very important by 90.7% of patients but only 49.1% of nurses and 45.0% of physicians. Although it seems that providers may not see attending to patient areas as their responsibility, patients clearly consider the physical environment as part of the professional environment. This finding is relevant because patients' unmet expectations predict decreases in satisfaction with care, symptom improvement, and adherence to prescribed care.28
Two items that did not meet criteria for importance as behavioral signs of professionalism warrant further discussion: dress and relationships with pharmaceutical companies. Students have traditionally been taught that “professional dress” is an important part of becoming a physician and meeting their patients' expectations. However, in our study, only 35.4% of patients, 53.0% of nurses, and 37.1% of physicians felt that dresses appropriately was an important sign of professionalism. Our study adds credibility to Brandt's30 conclusion in his review of the literature that a “neat clean appearance is more important than attire.” Indeed although dresses appropriately was not considered important, has good hygiene (e.g., washes hands, wears clean clothes) was one of the most highly rated behaviors in our study. Finally, the controversy about physician relationships with pharmaceutical companies is highlighted by our finding that fewer than half of physician leaders who responded to the survey felt that avoids nonscientific relationships with pharmaceutical companies was a very important sign of professionalism.
Data in this study are derived primarily from opportunity samples. Focus group and survey participants—even those in the random sample obtained for the national patient survey—decided on their own to accept an invitation to take part in the study. Although this raises the possibility of self-selection bias, the topic of professionalism in medicine has broad appeal.19 In terms of location, most of the focus groups were drawn from one city, albeit three different hospitals. For the surveys, we had access to national samples of patients and physician leaders but not nurses (i.e., respondents to the nurse survey were from one state). It is possible that the way patients, nurses, and physicians view professionalism in medicine differs by region. In future studies, gauging the extent of differences within groups may be a useful complement to comparing the views of patients, physicians, and nurses. Finally, it is important to note that, despite asking focus group participants for behavioral signs of professionalism, some of the items generated are not behaviors per se but are revealed through behavior (e.g., is honest). The strategy of asking survey participants to rate both the importance and observability of each item yielded a “short list” of behaviors that can inform the development of assessment tools (see Table 4).
Assessment of professionalism will become increasingly commonplace for medical students, residents, and practicing physicians.15,19,31 By documenting perceived importance and observability, the results offer a set of 16 items that may be useful in patient assessments of physician professionalism—although the item focusing on the cleanliness and comfort of patient areas might be properly considered a measure of the professional environment rather than a sign of physician professionalism per se—and 14 items that may prove useful for nurse assessments of physician professionalism. Although there is a degree of overlap between these sets, it is clear that some items are better suited for patient assessment and others for nurse assessment. It seems that peer assessment (i.e., physician assessment of colleagues) will be more challenging, as physician leaders considered only two behaviors both very important and observable. Accordingly, self-assessment may be an important adjunct for physicians; the 53 items in Table 2 offer a substrate for this approach. Perceptions regarding observability may differ across the continuum of medical education. Further study is needed to determine the extent to which peer assessment would be more viable for medical students and residents.
The input of patients, nurses, and physicians in defining behavioral signs of professionalism is likely to make assessment instruments tangible, relevant, and valuable for the people who use them. Although the plan and procedure of item generation lends considerable content and construct validity,32 the feasibility, reliability, and validity of assessment tools derived from this research must be tested in a variety of clinical practice settings before being put into widespread use. Focusing on behaviors rather than attributes can facilitate not only assessment but also discussion and modeling of professionalism in both medical education and clinical care. In this way, specific information about professional behavior may improve the educational environment and, eventually, the quality of patient care.
The authors are grateful to members of the advisory board: Mark Albanese, PhD, Louise Arnold, PhD, Linda Blank, Stephen Clyman, MD, Raymond Curry, MD, Richard Frankel, PhD, Geoffrey Gordon, MD, Paul Haidet, MD, MPH, Angela Nuzzarello, MD, Joseph O'Donnell, MD, David Stern, MD, PhD, and Patricia Surdyk, PhD.
This study was funded, in part, by a grant from the American Board of Medical Specialties (ABMS) Research and Education Foundation (PI: Dr. Makoul). The ABMS had no role in the design of the study, analysis and interpretation of data, or preparation of the manuscript. For the online physician survey, a staff member in the ABMS office (Evanston, Illinois) sent an email to directors of each specialty board, asking them to complete the survey.
1 Embedding Professionalism in Medical Education: Assessment as a Tool for Implementation. Baltimore, Md: Report from an Invitational Conference Sponsored by the Association of American Medical Colleges and the National Board of Medical Examiners; May 15–17, 2002.
2 Feudtner C, Christakis DA, Christakis NA. Do clinical clerks suffer ethical erosion? Students' perceptions of their ethical environment and personal development. Acad Med. 1994;69:670–679.
3 Papadakis MA, Osborn EH, Cooke M, Healy K. A strategy for the detection and evaluation of unprofessional behavior in medical students. University of California, San Francisco School of Medicine Clinical Clerkships Operation Committee. Acad Med. 1999;74:980–990.
4 Ludmerer KM. Instilling professionalism in medical education. JAMA. 1999;282:881–882.
5 Relman AS. Education to defend professional values in the new corporate age. Acad Med. 1998;73:1229–1233.
6 Frank J, Jabour M, Tugwell P, et al. Skills for the New Millennium: Report of the Societal Needs Working Group. Ottawa: Royal College of Physicians and Surgeons of Canada; 1996.
7 Association of American Medical Colleges. Learning Objectives for Medical Student Education—Guidelines for Medical Schools: Report 1 of the Medical School Objectives Project. Washington, DC: Association of American Medical Colleges; 1998.
8 Batalden P, Leach D, Swing S, Dreyfus H, Dreyfus S. General competencies and accreditation in graduate medical education. Health Aff (Millwood). 2002;21:103–111.
9 Blank L, Kimball H, McDonald W, Merino J. Medical professionalism in the new millennium: A physician charter 15 months later. Ann Intern Med. 2003;138:839–841.
10 General Medical Council. Good Medical Practice. London, UK: General Medical Council; 1995.
11 Horowitz SD. Evaluation of clinical competencies: Basic certification, subspecialty certification, and recertification. Am J Phys Med Rehabil. 2000;79:478–480.
12 Jolly B, McAvoy P, Southgate L. GMC's proposals for revalidation. Effective revalidation system looks at how doctors practise and quality of patients' experience. BMJ. 2001;322:358–359.
13 ABIM Foundation; American Board of Internal Medicine; ACP-ASIM Foundation; American College of Physicians-American Society of Internal Medicine; European Federation of Internal Medicine. Medical professionalism in the new millennium: A physician charter. Ann Intern Med. 2002;136:243–246.
14 Medical Professionalism Project. Medical professionalism in the new millennium: A physicians' charter. Lancet. 2002;359:520–522.
16 General Medical Council. Good Medical Practice. London, UK: General Medical Council; 2006.
17 Assessment of Professionalism Project. Washington, DC: Association of American Medical Colleges; 2002.
18 Measuring Professionalism in Medicine: Summaries of Methods Reported in 134 Studies Published 1982–2003. Philadelphia, Pa: ABIM Foundation; 2003.
19 Stern DT. Measuring Medical Professionalism. New York, NY: Oxford University Press; 2006.
20 Lincoln YS, Guba EG. Naturalistic Inquiry. Newbury Park, Calif: Sage Publications; 1985.
21 Strauss A, Corbin J. Basics of Qualitative Research: Grounded Theory Procedures and Techniques. Newbury Park, Calif: Sage Publications; 1990.
22 Institute of Medicine. Health Literacy: A Prescription to End Confusion. Washington, DC: National Academies Press; 2004.
23 Stenner AJ. Measuring reading comprehension with the Lexile Framework. Paper presented at: Fourth North American Conference on Adolescent/Adult Literacy; February 1996; Washington, DC.
24 Stenner AJ, Horabin I, Smith DR, Smith M. The Lexile Framework. Durham, NC: MetaMetrics; 1988.
26 Goffman E. The Presentation of Self in Everyday Life. Garden City, NY: Anchor Books; 1959.
27 Hembroff LA, Rusz D, Rafferty A, McGee H, Ehrlich N. The cost-effectiveness of alternative advance mailings in a telephone survey. Public Opin Q. 2005;69:232–245.
28 Bell RA, Kravitz RL, Thom D, Krupat E, Azari R. Unmet expectations for care and the patient-physician relationship. J Gen Intern Med. 2002;17:817–824.
29 Kravitz RL, Callahan EJ, Paterniti D, Antonius D, Dunham M, Lewis CE. Prevalence and sources of patients' unmet expectations for care. Ann Intern Med. 1996;125:730–737.
30 Brandt LJ. On the value of an old dress code in the new millennium. Arch Intern Med. 2003;163:1277–1281.
32 Nunnally JC. Psychometric Theory. New York, NY: McGraw Hill; 1978.