In the National Medical Student Career Choice Survey, we surveyed a representative sample of graduating medical students from the class of 1990 [6]. Students responded to a 10-page structured questionnaire using a Likert-scale format regarding perceptions of and influences on a career choice in internal medicine. Results of the study showed that internal medicine was perceived as more stressful and demanding and as less satisfying and rewarding compared with other fields. Key factors turning students away from internal medicine included 1) the nature of the patients seen in internal medicine; and 2) the view that medical residents and practicing internists are overworked and dissatisfied.

Our initial report, however, seemed to tell only part of the story. The data generated several questions. Why was internal medicine felt to be less satisfying and rewarding compared with other professions? Were financial problems more important than we initially found; did a social-desirability bias exist that prevented students from acknowledging this issue in standard Likert-scale format? Did women and men perceive internal medicine differently, and could the experiences or values of women account for their known affinity for primary care careers? And finally, what was the intensity of the perceptions? Our focus groups, which began the project and led to the initial questionnaire, involved many students with powerful feelings about negative experiences during medical school. These experiences seemed not to have been reflected in the Likert-scale responses (such as the experience of a student who told us, "After the way I was treated by my attending during my medical clerkship, I would never have anything to do with the profession again. I'm going into OB!").

We thus became interested in the responses to the open-ended question that concluded our survey instrument. This question asked, "What suggestions do you have for improving the attractiveness of internal medicine?" Eight hundred and twenty-one students wrote lengthy responses (some totaling an entire page of text). The intensity of feeling was clear throughout each essay, and the social-desirability bias appeared to lift, as many students said simply, "Pay us more". Overall, the responses to this question appeared to provide a fuller, more valid, and more personal perspective, with the whole picture coming more into focus.

To analyze these data, we selected 500 responses for a qualitative analysis. This method entails breaking essays into discrete statements or phrases and then tabulating the phrases by frequency in order to assess the prevalence of concerns. Importantly, in a qualitative analysis, direct quotations are presented so that the words of the students are not lost in the frequency tabulations. Qualitative analysis, as a "rich source of descriptions that help explain processes," possesses an "undeniability" [8, 9], a reality that is difficult to ignore. The result of our qualitative research was an in-depth understanding of the student's perspective. This type of research provides access to the meanings that emerge from and guide behavior [9].

Through a synthesis of our previous data [6], new theories on the significance of interpersonal relationships [10, 11], and current literature on the career-choice process [3-5], we developed a model to reveal explanations for career-choice issues that closed-ended questionnaires did not elicit. We present the results of this qualitative analysis and provide new evidence about the importance of relationships among students, patients, and teachers during medical training as a factor in improving the attractiveness of internal medicine.

Methods

Top Methods Results Discussion Author & Article Info References

Students

The study design for the career-choice survey has been described previously [6]. In brief, the original survey targeted graduating students (n = 1650) from the class of 1990 at 16 randomly selected medical schools. A stratified cluster-sampling design was used, with funding status (private or public) and percent of students choosing internal medicine (high or low) as stratification variables. An interval-scale questionnaire was used to elicit students' responses to questions regarding their perceptions of internal medicine and influences on career choice. Students were also asked to tell whether they had ever seriously considered a career in internal medicine but had switched to another career or whether they had decided to stay in internal medicine.

Twelve hundred and forty-four students (76%) responded to the career-choice questionnaire. Thirty-seven percent were female and 63% were male. The results of the closed-ended questions have been published previously [6]. Of the 1244 responders, 821 (66%) wrote responses to the open-ended question at the end of the questionnaire. The gender distribution of these 821 responders was similar to the overall gender mix. The 423 students who replied to the questionnaire but did not complete the open-ended question were similar to the essay responders in terms of sex, demographic characteristics, and other key variables, such as school type (public or private), marital status, and debt. Students were more likely to respond to the essay if they had children (73% compared with 64% without children, P = 0.037). For our analysis, we selected a sample of 500 from among the 821 essay responders. These students were stratified by sex and then randomly chosen from within each sex category. An a priori assumption was made that women and men would differ in their responses, and therefore similar numbers of women and men were chosen for analysis.

Qualitative Analysis

Using a standard qualitative approach [8, 9], we did our analysis as follows: First, we developed a career-choice model. This model incorporated the factor analysis of career-choice influences found in our original report [6], recent literature on the career-choice process [3-5], and evidence for the importance of interpersonal relationships in human growth and development [10, 11]. A 10% sample of the 821 respondents to the open-ended question was independently reviewed by the investigators to further refine the factors that would eventually compose the model. This subset of respondents was subsequently excluded from the study sample. The final model was refined through an iterative process among the investigators.

Model for Determinants of Career Choice

Our career-choice model Table 1 included five main factors: ambulatory care exposure and primary care; attending physician-student interactions and learning climate; workload and stress; income and status; and intellectual stimulation. Based on a preliminary assessment, the five main factors were further subdivided into components to reflect more accurately their content and meaning. The ambulatory exposure-primary care factor included issues relating to the appeal and difficulties of primary care, the value of interpersonal connections with patients, and issues relating to the care of difficult patients. The attending physician-student interaction factor included items regarding role models, feedback, teaching climate, identification, and perceived support from attendings, residents, and peers. The workload and stress factor included issues relating to time demands, the balance between family and career, and worksite-workload problems. Income, status, and prestige composed one category, as they have in previous research [4]. Finally, the intellectual stimulation factor included components that related to thinking and problem-solving and to mastery, competency, and autonomy in the learning environment (Table 1). The remaining comments were grouped under categories labeled "personal preference" and "the trouble with medicine and medical school in general". Such comments were not included in the analysis because either they were not a suggestion for improving the attractiveness of internal medicine (for example, "you either just like medicine or you don't") or were too vague to provide useful information ("medicine is bad because of society these days"). The codes were selected to be exhaustive and mutually exclusive.

Coding

Discrete phrases from the 500 essays were then identified by two of the authors, who were blinded to the sex of the respondent. All phrases were bracketed, omitting only qualifiers and transitional elements. No paraphrasing was used. During a training period, two independent raters practiced coding the 10% sample of responses to the open question that had not been selected for analysis (practice sample). The trained raters then independently assigned one code to each bracketed phrase in the 500 essay responses chosen for the analysis (Figure 1).

View larger version (31K): [in this window] [in a new window]   Figure 1. Sample response to open-ended question that shows bracketing and numbered phrases before coding.

Reliability

Inter-rater reliability was assessed by evaluating the percent agreement between the two raters within each category of coding. The overall agreement on the five codes was 77.9%, with a value of 0.75. The agreement on each factor individually ranged from 90.8 to 96.4%, with values ranging from 0.61 to 0.85. A third rater did partial coding on 25% of the sample, and a fourth rater was used to arbitrate disagreement if all three raters failed to agree. This rater assigned an entirely new code to a small number of phrases (<10) about which there was disagreement.

Statistical Analysis

Phrase frequency was tabulated by gender and by eventual career choice, first using the number of phrases and then the number of students as the unit of analysis. The chi-square statistic was used to calculate the statistical significance of differences.

Results

Top Methods Results Discussion Author & Article Info References

Responses from 249 women and 251 men were analyzed. Students' mean age was 28 years; 38% were married and 17% had children. Thirty-three percent had debt totaling more than $50 000. An ambulatory rotation had been taken by 42% of students. Thirty-eight percent had obtained honors in their medicine clerkship. Women were more likely than men to have chosen a career in primary care medicine, pediatrics, or family practice (37% compared with 23%, respectively; P < 0.001).

Factors Influencing Career Choice in Internal Medicine

We coded 1512 phrases (776 from women and 736 from men) that related to improving the attractiveness of internal medicine. Women and men wrote an average of 3 phrases (range, 1 to 12 phrases per female student and 1 to 10 phrases per male student). Twenty-seven percent of the coded phrases from women concerned relationships with patients and primary care compared with 21% of the phrases from men (P < 0.01) (Table 2). Similar proportions of phrases from women and men pertained to learning climate or relationships with attendings and residents (20% and 21%, respectively) and to stress and workload (13% and 17%, respectively). However, only 12% of women's phrases had to do with income and status compared with 17% of men's phrases (P < 0.01). Overall, when phrases relating to students' interpersonal relationships with both patients and attendings were compared with those about income and status, phrases about relationships were noted approximately four times more frequently than financial concerns by women (47% compared with 12%) and 2.5 times more frequently by men (42% compared with 17%).

The results were similar when the number of students rather than the number of phrases was used as the unit of analysis. Specifically, 65% of female students raised concerns about relationships with patients compared with 50% of male students (P < 0.01). Results for other variables are shown in Table 1 and were not different for women and men. Again, income and status were mentioned by far fewer students than issues relating to professional or interpersonal relationships.

The Students' Words: Support for the Model

Our career-choice model (see Table 1) included five factors that affect career choice in internal medicine. Part of the "undeniability" of a qualitative analysis [8, 9] rests with the primary data themselves. The following section provides examples of students' observations and suggestions regarding internal medicine. The specific quotes are organized according to model category. Responses were predominantly negative because the open-ended question asked for suggestions to improve the attractiveness of medicine. Aside from responses about "intellectual stimulation," students made few positive comments.

Patient Relationships and Primary Care

"The most influential event for me was an elective in outpatient general internal medicine, including offices outside the hospital".

"Take a more humanistic approach to patients similar to the family med docs".

"In the inpatient setting one gets a skewed impression of the problems internists deal with; i.e. prolonging death instead of extending life".

"Dealing with the entire person, not just his kidney, is not rewarded at a place like".

"Outpatient work—more needed".

"As medical students are encouraged to be more humanistic in their approach, more will enter internal medicine".

(The following statements document students' concerns with chronically ill, "difficult" patients. This is a subcategory of the "patient relationship" factor.)

"I was frustrated in internal medicine with the overall lack of being able to cure or fix problems".

"Chronic nature of disease and my inability to make a difference turned me away".

"Some of it is the hopelessness of the patient's situation".

"I think more students would choose internal medicine if they were exposed to patients that were not always chronically ill or dying".

Learning Climate and Relationships with Attendings and Residents

"Unnecessary abuse of the student early on in his clinical experience is all but ‘character building,’ and in this case fostered contempt for the specialty".

"Encourage attendings to abandon humiliation as one of their primary teaching strategies".

"My attending was a journal jockey—inflated ego".

"Delete malignant attendings".

"More motivated teaching!!!!"

"Pimping; demoralizing students".

"Egotistical attendings who teach by humiliation".

"Need role models for women and generalists".

"Students made to feel incompetent and burdensome, non-contributing members of the team".

"My team was great; good humor, considerate, hard working, extremely intelligent. Pimping was kept to a bare minimum".

Workload and Stress

"Residents were all excellent, yet many were negative about their residencies (they didn't seem happy)".

"I didn't go into internal medicine because medicine residents and interns, although often brilliant and caring, were almost always miserable, much more than the house officers from any other specialty I saw".

"Students and residents are slaves".

"Less abusive training".

"Unhappy attending". "Widespread dissatisfaction of faculty".

"Inhumane working hours—decrease scut, schedules, on call".

"Biggest single factor pushing me away from internal medicine was among interns I worked with in internal medicine clerkship. I met few medicine interns and residents who were happy with their experiences and their career choice".

Income and Prestige

"I want 150K a year".

"Financial rewards for time and effort".

"Better pay after residency compared to majority of other fields".

"The old story of better compensation for cognitive services".

"IM clerkship at this institution seems to reinforce the notion that internal medicine residents and internists are overworked, underpaid, and often underappreciated".

Intellectual Stimulation

"I have heard colleagues say that they love the underlying principles and the exciting pathophysiology of science".

"I enjoy the academic aspect of internal medicine".

"Help the patient; don't just be intellectual".

"I wouldn't wish a career of mental masturbation on anyone".

"Have it be more realistic in its intellectual discussions. Talking about 15 more causes of a disease when the actual cause is fairly common".

Students Who Left Internal Medicine

Of the 500 students in the sample, 252, or about 50%, had seriously considered a career in internal medicine but switched to another career. Twenty-six percent of the 127 women who had considered internal medicine but switched to other fields ultimately went into family medicine or pediatrics, whereas only 16% of the 125 men who switched went into these other primary care fields (P = 0.052). When psychiatry and obstetrics-gynecology were considered as well as family medicine and general pediatrics, 34% of women compared with 19% of men switched to these fields (P < 0.01). When ultimate career plans were evaluated for the larger sample of the 1244 respondents to the career-choice survey, these trends confirmed a preponderance of women switching to primary care fields (30% of women compared with 15% of men switched to family medicine and general pediatrics, P < 0.0001). Because of female students' affinity for primary care fields, we looked more closely at the qualitative responses of the female students. Comments made by women who switched to pediatrics or family medicine provide insights into their reasons for leaving internal medicine (Table 3). An absence of female role models, a lack of empathy and sense of humor in internal medicine, and the difficulties of balancing a medicine career and family were important concerns.

The following four subgroups are compared in Table 4: 1) students who never considered a career in internal medicine, 2) students who considered and ultimately selected a career in internal medicine, 3) students who considered a career in internal medicine but switched to a primary care field, and 4) students who considered internal medicine but switched to a surgical or nonmedical specialty such as radiology or anesthesiology

Three points emerged from this comparison: First, despite an anticipated difference in the ways in which men and women experience medical training, there were many similarities in the frequencies of statements made by men and women within the four subgroups. Thus, strategies for improving the attractiveness of medicine are likely to appeal to both men and women. Second, students who switched to other primary care careers rarely mentioned income and prestige as a reason for leaving medicine, whereas students who selected internal medicine had serious reservations about their future income potential. (Concerns about income were noted in 4% of phrases from women who switched to other primary care fields compared with 27% of phrases from women who stayed in internal medicine [P < 0.00001] and in 8% of phrases from men who switched to other primary care fields compared with 27% of phrases from men who stayed in internal medicine [P < 0.008].) Students who expressed these concerns may define a "high-risk" group that may later leave internal medicine for a different career because of inadequate reimbursement. Third, the workload and stress of a career or residency in internal medicine were mentioned far more frequently by men and women who switched to nonmedical specialties (18%) than by men and women who switched to other primary care careers (5%) (P < 0.02) or by men staying in internal medicine (5%, P < 0.00002). Predicted workload and stress of a residency or career in medicine may therefore underlie the "defection" of large numbers of students to fields such as radiology and anesthesiology.

Discussion

Top Methods Results Discussion Author & Article Info References

After completing a time-consuming, 9-page questionnaire, more than 800 students took the time to write a detailed response to a question related to improving the attractiveness of internal medicine. The method of qualitative analysis allowed these responses to become quantifiable data and the voices of the students to emerge. The message was clear: Students want more ambulatory care experiences ("out of hospital," "real world") and more effective and satisfying relationships with patients during medical training. They also want more respect from their teachers, less stress for themselves as medical housestaff, and greater financial reimbursement for their future efforts. Female students cared more often about relationships established with patients and selected primary care careers than did male students. Finally, even students who chose careers in internal medicine were concerned about the poor training environment, lack of esprit de corps and respect among internists, and, most of all, the inadequate financial reimbursement relative to the reimbursements for competing specialties.

What is new in these results? First, many investigators have postulated the findings of our study, but few, if any, have demonstrated them. The use of qualitative analysis preserved the words of the students and thus organized, in a more concrete, vivid way, the flavor of their concerns and the intensity of their comments [8, 9]. Second, negative experiences with medical faculty—not simply the lack of role models but the perceived lack of concern for students' and housestaff's welfare and the hierarchical nature of teaching—may be a crucial factor in losing students to other fields. Third, financial issues finally emerged from the background, perhaps because the open-ended assessment attenuated the social-desirability bias. Financial reimbursement was mentioned less often, however, than other concerns and emerges partly as a compensation for long hours, difficult patients, and a low level of respect. Fourth, our data begin to explain women's affinity for primary care careers (that is, better relationships with patients). Finally, a model of career choice was developed that incorporates five factors that need to be tested in future studies.

Importantly, the first factor in the model combines "relationships with patients" with "more ambulatory exposure". This pairing suggests that simply increasing exposure to ambulatory care experiences may be insufficient to attract more students to medicine. Rather, one must also preserve students' sense of engagement with their patients during these ambulatory experiences in order to provide the personal factor that is missing in current inpatient training environments. This idea is consistent with the findings of Lewis and colleagues [1], who studied career satisfaction in more than 1000 practicing internists. In this large survey, the single most satisfying aspect of practice was "relationships with patients" [1]. Perhaps the deterioration of the hospital training environment (owing to the "intensivization" of current inpatient wards, as detailed several years ago by Schroeder and colleagues [12]) is a key factor in students' turning away from general internal medicine and toward more procedure-oriented specialties.

Negative student-faculty relationships were an unexpected second factor revealed by our analysis. Previous work by our group [6] and others [13-17] has shown the importance of faculty role models (or the lack thereof) in career choice. However, to our knowledge, no other study has shown the vehemence with which the medical students in our survey described their teaching by internal medicine faculty. The Socratic method appears to have gone awry. Whether a cohort effect occurred or teaching effectiveness has declined cannot be determined from our data. We postulate that students' expectations of internal medicine involve images of highly respected, brilliant diagnosticians who are dedicated to patients and students. These images are directly contrary to the technical, non-patient-oriented images associated with the surgical and radiologic specialties. Students' actual training experiences may contradict their images of internal medicine, causing an unsettling sense of disparity. Although our open question was not phrased such that we might have determined whether medicine teachers were worse than instructors in other disciplines, it is clear that students have many angry feelings about the way in which they are taught internal medicine. In a system where unsatisfying relationships are established with hopelessly ill inpatients, the types of teaching behaviors students described may be intolerable.

Indeed, many other fields court students. Students wrote of a sense of being "wanted" by other disciplines, of being treated better right from the start of other rotations and at interviews for residencies. They are lured away from a career in medicine by promises of "controllable lifestyles" [2], less stressful residencies, challenging cases with surgical cures, or, finally, markedly increased income and status. Our data document that financial rewards do matter to students, and, as recent editorials suggest [18, 19], increased income for internal medicine (or decreased salaries for the competing professions) would "level the playing field". However, as society grapples with income disparities, the written words of our students imply that medicine would be a far more attractive field if student-patient and student-faculty relationships were improved and if the medical residency were made a less stressful, "miserable" experience. Lack of attention to these three factors will probably lead to decision making driven by the next factor in the model-that is, income and prestige. It may be a balancing of the multiple factors in the model that ultimately determines a student's career choice.

Internal medicine also loses many students to other primary care fields (family medicine and general pediatrics) with low pay and prestige. As shown in Table 4, these students have serious concerns about the lack of exposure to ambulatory care experiences during medical school and the lack of relationships with patients. However, they have few concerns about income or prestige. Women were predominant in this group and chose primary care careers more often than did men. Although the affinity of women for primary care has been noted before, recent theoretical formulations of women's psychological development [10, 11] suggest that women place a high value on relationships with others and therefore may report a greater interest in patients than their male counterparts [1, 15, 17-27]. Our data imply that the lack of relationships with patients may be a pivotal factor in the eventual career choice of women in medicine.

Our results should be viewed with some caution. Although financial issues finally emerged as a factor, social desirability might have prevented them from assuming even greater importance. Students were not asked to compare the problems of internal medicine with the problems in other fields; thus, although we know that students perceived medical faculty as arrogant and humiliating, we do not know if they are more arrogant or humiliating than other faculty. (We postulate that the students did not expect the internal medicine faculty to be that way and thus were disappointed.) We did not assess the roles of other external factors (such as peer pressure, family, and religion) in determining career choice. These factors have been shown to be important determinants of educational outcomes [28], and the way in which our question was phrased did not elicit responses that would speak to these various issues.

Nevertheless, students have given us ample evidence that there are problems with the medical training environment and have indicated how we might address these concerns. Students want more ambulatory training with continuity experiences. Curricula need to be developed to teach students how to relate to chronically ill, "difficult" patients and to derive satisfaction from caring for patients rather than curing them [29]. Faculty development programs [30] and formal evaluation mechanisms should be instituted so that negative interactions between students and faculty can be identified and corrected. The nature of stressors and the workload of medical residencies should continue to be evaluated and addressed, and career satisfaction among primary care internists should be studied so that strategies to improve the satisfaction of practicing physicians can be constructed. Financial incentives need to be put in place. Money is a sign of respect, and the students recognize a lack of respect in the income disparities among medical specialties. Finally, the concern that women have for fostering better relationships with patients suggests that they may lead the way in the recent push toward generalist careers. The interest of all students in developing better relationships with patients serves both the public and the profession well. We must recognize the importance of these concerns and safeguard the development of these relationships if we are to retain and recruit physicians into internal medicine.