Davis, Beth Ellen MD; Nelson, David B. MD, MSc; Sahler, Olle Jane Z. MD; McCurdy, Fred A. MD, PhD; Goldberg, Rhonda MA; Greenberg, Larrie W. MD
During the 20th century, the emphasis in health care shifted from the hospital-based treatment of acute conditions to the ambulatory management of chronic illnesses. In the 1990s, over 100 million Americans suffered from incurable, prolonged illnesses, which accounted for 75% of the health care expenditures in the United States each year.1 Despite this shift, the medical education model developed in 1910 by Flexner is still the dominant philosophy in U.S. medical schools. Students learn medicine in a disease-oriented, hospital-based system intent on cure, often at the expense of acquiring the necessary knowledge, skills, and attitudes to care for the chronically ill.
Studies have reported medical students' and residents' ambivalence toward the management of chronic illness2–5 even though these negative attitudes were not present prior to the students' entering medical school. A previous study found first-year medical students had positive attitudes toward caring for chronically ill patients and documented that the students had had significant pre-matriculation experiences6 with this population. The students' subsequent disinterest in these patients appears to have arisen during medical school. In fact, a review of medical students' attitudes at completion of their medical school studies concluded that aspects of direct patient care may have contributed to the prevalence of negative, cynical attitudes in residents and clinicians.7 Other studies examining specific groups of chronically ill patients (those with diabetes, cancer, chronic obstructive pulmonary disease) have found that students' attitudes are influenced more specifically by the context and environment in which chronic illness is presented.2,3,5,8–10 For example, students who encountered chronically ill patients exclusively on hospital wards inaccurately believed their patients to have higher disease complication rates and more pessimistic outcomes than did students whose experiences included seeing these patients in ambulatory care settings. Because the attitudes with which students and physicians approach particular patient problems are believed to influence the quality of physician—patient relationships and willingness to change health care behaviors, identifying characteristics of attitudes of physicians-in-training is an important step in restructuring medical education for the next century.
In this study, we asked clerkship directors from five U.S. medical schools to follow students longitudinally over four years to examine their attitudes toward caring for chronically ill patients. We hypothesized that students would enter medical school with positive attitudes that were based on their personal experiences with chronically ill patients and that during clinical (predominantly third-year) rotations, they would develop more negative attitudes that could influence their decisions in specialty selection. Additionally, we wanted to compare responses from multiple institutions to see whether differences among schools could be used to identify a specific curriculum that protects against the development of students' negative attitudes toward the chronically ill.
The cohort for our study was an accessible sample of all students enrolled in the classes of 1997 at five diverse (private, public, military, urban, and rural) U.S. medical schools: F. Edward Hébert School of Medicine, Uniformed Services University of the Health Sciences (USUHS); Georgetown University Medical School (GT); George Washington University School of Medicine (GWU); the University of Nebraska School of Medicine (UN); and the University of Rochester School of Medicine (UR). Pediatrics clerkship directors from these schools agreed to collect survey instruments that had been designed to be distributed three times over the period from autumn 1993 to spring 1997.
The questionnaires, which were similar to the survey used by Fitzpatrick,6 were first distributed on orientation day at each medical school. The survey was re-administered at the completion of second-year basic science classes. Students' participation was voluntary. The questionnaire used for these two surveys requested demographic and career choice information and asked about personal, school, and work experiences with chronically ill patients. Students were provided with a definition of “chronic care” developed by the Department of Health and Human Services, “a condition or disorder with a protracted course that can be progressive and fatal or that is associated with a relatively normal life span despite impairment of physical or mental functioning,”11 and were then asked to identify chronic illnesses from a list of acute and chronic conditions. The remainder of the questions on these two questionnaires were related to caring for chronically ill patients and to whether the students' specialty choices were affected by perceptions of caring for patients with chronic illnesses. The questionnaires were linked to students by social security numbers to permit longitudinal matching of answers. At the completion of all required clinical rotations, the students were given a third, modified questionnaire that specifically sought information about clerkship experiences with chronically ill patients.
We performed statistical analysis using McNemar's chi square. Statistical significance was placed at alpha = .05. The data presented here reflect participants who completed both the first and the third (final) questionnaires.
Of 723 medical students enrolled at the five participating medical schools, 695 (96%) voluntarily completed the first questionnaire and 573 of the same students (79%) completed the second questionnaire. Responses to the first and second surveys demonstrated strong retest reliability and no significant difference in existing attitudes among respondents toward chronically ill patients. Results of the second survey were not used for further analysis. Responses from the 502 students who completed the third questionnaire were individually linked to their responses to the first questionnaire, for an overall participation rate of 69%, and these data were used for the analysis. The demographics for the 193 students (27%) who completed the first questionnaire but who did not complete the final questionnaire (non-participants) were similar to the participants' data. No information was available about the 28 students (4%) who did not participate in the original survey. Response rates by school for the matched data were 85% (USUHS), 35% (GT), 86% (GWU), 67% (UN), and 54% (UR).
Table 1 shows the characteristics of the respondents for the matched data. Of them, 59% were men, approximately 17% defined themselves as members of a minority group, and 72% were less than 25 years old. Thirty-six percent of the respondents answered that a family member or close friend had a chronic illness, 11% stated that a close relationship with a chronically ill individual had contributed to their decisions to become physicians, and 4% identified themselves as chronically ill. Almost 70% of the respondents reported having had academic or work experience related to chronic care prior to medical school. A fifth of the students had at least one parent in the health care profession. Approximately 50% of the participants answered similarly at the beginning and end of their medical education that they planned to pursue a primary care specialty.
A total of 34% of the students responding to the first questionnaire correctly identified all seven chronic illnesses from a list of acute and chronic conditions, and the percentage doubled (68%) for students responding to the third questionnaire (p < .0001).
All participants from all the schools had recently completed their required clinical rotations at the time of the final survey, and 73% recorded favorable perceptions (using a five-point Likert-scale) of interactions with chronically ill patients. Additionally, 90% of all students felt involved in the care of their patients. Significantly fewer students documented positive experiences with residents (57%) or attending staff (59%) in their interactions in the management of chronically ill patients (p = < .01). Of the students who had had positive experiences with attending physicians and residents managing chronically ill patients, 90% indicated they had had positive experiences with direct patient contact (see Figure 1). Of the students who reported negative experiences with their attending physicians and residents, only 10% reported positive experiences with direct patient contact. Of all respondents, 93% described the majority of their patient contacts as inpatient experiences.
Even though an overwhelming majority of respondents to the first and third questionnaires (98% and 99%, respectively) stated they would accept a patient with a chronic illness into their practices, 25% of the students when seniors, versus 9% of the students in their first year, would seek another specialty in medicine if the incidence of chronically ill patients increased in their chosen field (p < .001). Gender, age, school site, and specialty choice were not associated with this finding. In response to the third questionnaire, the senior medical students with the least favorable attitudes toward chronic illness also perceived the quality of care received by chronically ill patients as poorer than that received by other patients (p = .056), based on their clerkship experiences.
The attitudes of medical students and physicians play a large role in the quality of the physician—patient relationship and the types of health care strategies developed for chronically ill patients.12 Our objectives in this study were to (1) describe medical students' attitudes toward caring for chronically ill patients, (2) determine whether medical school experiences influenced these attitudes and, if so, when, and (3) see whether responses differed by school.
In other studies, several demographic factors have been implicated as influences on health providers' attitudes toward specific patient populations. For example, one study found women physicians had more positive attitudes toward the medically indigent than did men, although in that study, the attitudes of both men and women deteriorated during medical school.2 In our study, however, we found that gender, age, ethnicity, specialty interest, and specific medical school curriculum were not associated with changes in attitude about caring for the chronically ill. In the past, a more rigorous admission selection process was suggested as a solution to maximize graduates' interest in caring for chronically ill patients. However, the results from our first questionnaire indicated that students admitted under the current criteria for medical school admission have strong, positive attitudes toward caring for the chronically ill upon matriculation. Thus, our data would not support the pursuit of additional admission tools. Instead, the context and environment in which chronic illness is presented during medical school need to be scrutinized.
From previous reports3,13,14 and our own experiences as students and educators, we believe students' attitudes are influenced during medical school by the attitudes of their role models, particularly teachers and residents. This was the reason we designed the surveys to precede and immediately follow students' experiences with patients during required clinical rotations that are closely supervised by residents and attending physicians. Whereas our data, along with many previous studies, showed negative attitudes toward chronic care patients emerging in students as medical school experiences progressed, our findings also demonstrated that some of the changes in students' attitudes were temporally related to their clinical clerkship experiences and thus, newly acquired. While professional attitudes are considered to be relatively fixed, if the influences causing them are recent, they may be more malleable. Indeed, if educators can identify specific factors likely to result in the development of undesirable attitudes, it may be possible to prevent negative impressions entirely.
Role modeling by faculty and residents in the care of chronically ill patients appears to be one area that requires evaluation. Overall, the students had enjoyed the relationships they had had with chronically ill patients and felt that they had learned a lot, especially when they involved in direct patient care. However, they rated their experiences with their teachers significantly lower. Students who had had negative experiences with attending physicians (or residents) while participating in the care of chronically ill patients were less likely to report positive experiences with their patients. Issues contributing to this negative environment might be the role models' inexperience with the psychosocial issues in chronic care, greater confidence in providing inpatient crisis management than with mobilizing outpatient community support, and use of cynical comments. While cynicism often serves to relieve tension rather than state beliefs, students are often left alone to interpret strong comments in the face of major complicated physical and mental health issues. A study looking at the erosion of ethics among clinical clerks found that 98% of the 665 students interviewed reported hearing unfavorable comments made about patients by supervising physicians.13 In fact, the disparaging label, “Gomer” was created in the late 1950s by house officers for patients presenting with “complex, chronic, medical, psychological and social problems in the setting of the acute care technologically-oriented hospital.”15 Perhaps the informal labeling of patients has an impact on care, management, and student perceptions.
Of the medical students responding to the third questionnaire who had the least favorable attitudes toward chronically ill patients, 25% perceived the quality of care received by the chronically ill to be poorer than that received by other patients. In other words, as students' perceptions of quality of care diminished, the likelihood of selecting an alternative specialty to avoid a high prevalence of chronically ill patients increased. Although not reaching statistical significance, this perception of discrepancy of care may have contributed to negative attitudes toward chronic care providers and their patients.
The students in this study, like most medical students across the United States, continue to receive most (more than 75%) of their exposure to direct patient care in hospital settings. For the chronically ill, a hospital admission often means a breakdown of normal support systems or further irreversible deterioration. When chronically ill patients are seen away from the ambulatory context of community, family, and cultural support, the level of care that must be provided can appear overwhelming, which results in the possibility of a negative message's being conveyed to students. Often, the structural support that provides adequate care to these patients resumes after discharge, but students rarely have the time or connection with chronically ill patients in ambulatory settings to learn about and participate in that environment.
Very few students indicated they would refuse to care for a patient in their practices because of a chronic illness and this remained consistent over the four years. However, a finding that concerned us was that one fourth of the senior medical students stated they would change their specialty choices if the incidence of chronically ill patients increased in their chosen field. The survey that substantiated this attitude was administered only weeks prior to their residency selection process. Their recent negative experiences with chronically ill patients may have played a role in their decision making about specialty choice, but without studying this specific variable longitudinally, it is difficult to predict whether this response affects ultimate career choices.
The demographic makeup of our study's sample and the strong response rate of the first survey (96%) and the matched longitudinal follow-up survey (69%) support the generalizability of our data. A limitation of our study, however, was the variability of the individual schools' response rates. This appears to have been a logistic issue. Many of the schools required a mandatory class activity before starting fourth-year rotations, which provided clerkship directors with the opportunity to schedule time with students to personally request completion of the final survey. Other schools did not have a formal meeting between academic years, which resulted in students who finished their third-year clerkships completing their questionnaires in physical proximity to individuals in charge of survey distribution and retrieval. This logistic issue appears to have contributed to differences in the final response rates, which might have been mitigated by a mail option. Had the statistical analysis demonstrated school-specific attitudes, it would have been difficult to eliminate sample bias as a contributing factor. However, no response pattern by school emerged in the analysis model, despite varying emphases on curricular issues. No program emerged as a specific model to preserve early positive attitudes toward chronic illness.
Another potential limitation of our study was the survey instrument itself. Although it had been piloted previously with good test—retest reliability among individuals, an ordinal (Likert) five-point scale was used to measure attitudes. Intervals between rankings are not necessarily interpreted equally between participants. To minimize this effect, we performed data analysis on extreme responses by groups of participants, or maximal changes in responses among individuals, wherever possible. Specific examples of patients' experiences (e.g., patients' logs) were not obtained, thus a single negative experience may have influenced multiple positive interactions, thus affecting student responses and the conclusions drawn.
By longitudinally following medical students, we were able to analyze changes in responses over time. We conclude that (1) most students enter medical school with positive attitudes toward caring for chronically ill patients, (2) this perspective deteriorates during the clinical experiences, and (3) newly developed attitudes may influence their decision making about specialty choice. It is encouraging that direct student—patient relationships were reported to be strong and usually positive. It is essential to identify and remediate those factors that possibly contribute to the negativity that arises during clinical training. These factors appear to include residents' and attending physicians' ambivalence toward or inexperience with chronic illness, disproportionate hospital-based experiences, and perceptions of inequality of care.
Medical education curricula need to formally incorporate structured, inter-departmental, and longitudinal concepts of caring for chronically ill patients. In response to the Association of American Medical Colleges' Graduation Questionnaire, 43% of the 1997 graduates of North American medical schools reported inadequate instruction time in “long term health care” compared with 2% who felt they had received inadequate instruction time in hospitalized care. Our study cohort, co-incidentally, became a subset of participants in this national graduation questionnaire.
Despite the recent trends in medical education to teach ambulatory-based care, design home visitation programs, and prioritize primary care, educators cannot assume that the unique competencies required in managing chronic illness are being addressed. Two recent reports provide the impetus for incorporating specific objectives into already existing curricula. First, in a report by academic deans from U.S. allopathic and osteopathic medical schools prioritizing current and ideal curricular topics,16 the three highest ranked ideal topics were effective patient—provider relationships, ambulatory care, and health promotion. Care of the elderly, interdisciplinary teamwork, psychosocial care, community problem solving, and patients-as-partners were in the top 12 ideal curricular topics to be included in future programs. Second, Dr. Andrew Cohen, vice chancellor for education, University of Massachusetts Medical Center, recently proposed a list of competencies for managing the chronically ill.14 He described a learning context that precisely matches the medical school deans' visions of ideal curricular topics. Although specific objectives are outlined, they are generic to all training specialties. Dr. Cohen challenged medical educators to implement the chronic illness competencies in an environment that (1) is ambulatory-based, (2) is collaborative with other health care disciplines, and (3) promotes understanding of the patient perspective. These two reports provide a framework for identifying how changes in the current student learning context and environment could occur and specifically address the development of positive professional attitudes toward chronically ill patients.