DeZee, Kent J. MD, MPH; Maurer, Douglas DO, MPH; Colt, Ross MD, MBA; Shimeall, William MD, MPH; Mallory, Renee MD, MPH; Powers, John MD, MSS; Durning, Steven J. MD
Robust primary care, defined as accessible, comprehensive, coordinated, and continuous care with a designated usual provider or clinic, is strongly associated with improvement in a wide variety of health outcomes.1 In the United States, however, medical students' interest in pursuing primary care specialties (internal medicine, family medicine, and pediatrics) has declined since 1998.2 Some have suggested that lack of financial rewards is the primary reason.3–12 One study showed a strong correlation (r = 0.82) between income and percentage of residency positions filled by U.S. MD medical graduates in 2008.3 Only one prior study attempted to quantify the annual salary required to influence career decisions, finding that 17% of students who did not select primary care would have done so if the annual salary were 18% higher than their expected specialty income.11
Additional research has suggested that lifestyle is a major determinant of residency choice.4,6,8,11,13–16 In one study, lifestyle explained 55% of the variance of changing specialty choice.13 In particular, students have been forgoing noncontrollable lifestyle specialties (obstetrics–gynecology, general surgery, orthopedics, neurosurgery, urology, and primary care) in favor of controllable lifestyle specialties (all others).13,17 However, no study examined whether a specific financial reward, either as a bonus or an increase in annual salary, could influence career choices away from controllable lifestyle specialties toward undermanned primary care specialties.
Thus, the primary goals of this study were to (1) determine what factors were perceived by students as barriers to applying for primary care, (2) determine which type and amount of monetary incentives would impact this decision, and (3) explore characteristics of students who would accept financial inducements to apply.
Design and sampling frame
This study is a cross-sectional quantitative and qualitative survey of all fourth-year medical students who were applying for residency training with the Military Healthcare System. All fourth-year U.S. medical students with a service obligation to the U.S. Army, U.S. Navy, or U.S. Air Force were eligible to participate. Students either attended the Uniformed Services University in Bethesda, Maryland, were participants in the Health Professional Scholarship Program (HPSP, a national military scholarship program for students attending civilian U.S. MD-granting and DO-granting medical schools), or had military service obligations from their undergraduate education (military service academy or Reserve Office Training Corps). The e-mail addresses of this national sample were obtained through program coordinators. On April 23, 2009 (after the military residency match but before medical school graduation), we sent students an e-mail invitation to participate in the survey and the link to the Web site hosting the survey. Students received two reminder e-mails before the survey was closed on June 14, 2009.
The institutional review boards at William Beaumont Army Medical Center in El Paso, Texas, Carl R. Darnall Army Medical Center in Fort Hood, Texas, and the Uniformed Services University in Bethesda, Maryland, reviewed and approved the protocol.
Questionnaire development and content
The authors developed a questionnaire to assess medical student career choice using available literature2,3,5,7–11,13–15,18,19 and their expert opinions as program, course, and clerkship directors in internal medicine and family medicine. The questionnaire was converted to electronic form using the online survey tool http://www.surveymonkey.com. The survey was vetted with practicing primary care physicians and also with physicians who trained in primary care but changed to another specialty. Several of the authors then reviewed the survey, question by question, with fourth-year medical students and current interns to ensure that the questions conveyed the intended meaning and that all necessary items were included. Finally, several senior educators (each of whom had more than 20 years' experience) reviewed the questionnaire for additional input. The final version of the questionnaire is provided in the Supplemental Digital Appendix. http://links.lww.com/ACADMED/A34
The survey assessed medical students' initial specialty choice for residency training, even if they were not ultimately selected for it, followed by up to three other choices considered. Students who did not indicate a primary care residency (defined as internal medicine, family medicine, and pediatrics) as their initial choice were asked about practice characteristics (e.g., appointment length of 20–30 minutes, flexible work hours) and increased training opportunities to determine whether these factors might have an effect on applying for primary care. Two open-ended questions allowed the students to list a “reasonable and realistic” amount of money that would have influenced them to apply for primary care (or “999” if no amount of money would have changed their minds). The first was for a hypothetical bonus paid exclusively to all primary care residents at the start of internship and at the conclusion of residency, and the second was for a military annual salary as an attending primary care physician after training. Additional barriers to choosing primary care were assessed by two more open-ended questions, which gauged (1) the most important reason(s) students didn't choose primary care and (2) the most important factor that could be altered that would have changed their mind to apply for primary care. Students were also asked to rate the likelihood that they would have applied for primary care if that one determinant had changed.
The remainder of the questionnaire included descriptive items for all respondents regardless of specialty choice, including questions about length and personal experiences in primary care clerkships, attitudes about primary care, lifestyle preferences, ratings of lifestyles of the various specialties, and participation in IM interest groups. Students were asked to give their own estimates of the median civilian salary nationwide of attending physicians two to three years after training in internal medicine, family medicine, pediatrics, and their selected future specialty (open-ended). Finally, students were asked for demographic information about their age, gender, marital status, presence of children, type of medical school, and education debt for themselves and their spouse. Except where noted above, all items provided a Likert scale or a list of choices for the answer field.
The primary objective of this study was to determine the proportion of non-primary-care residents whose decision to apply for primary care would have been affected by the residency bonus or the increase in attending pay. The raw responses for both financial-incentive questions were reviewed to exclude students who suggested unreasonably high amounts. By consensus of the authors, we chose a cut point for both of the financial incentives around the 75th percentile, where a clear cut point was seen. We considered the decisions of students who listed any amount of money at or below these cut points to be affected. We classified the decisions of students who listed an amount above these cut points as not affected by any amount of money. In a model to determine characteristics of students whose decisions would be affected by a financial incentive, all variables of a priori interest in primary care (consideration of primary care among secondary residency choices; reasons they did not apply and those that, if changed, would convince them to apply; attitudes about lifestyle and primary care) and each demographic variable were screened for inclusion using chi-square or Student's t test as appropriate. Those with a P value <.25 were included in a logistic regression model, and the model was fitted by backward stepwise regression. The final models for each regression (residency bonus and annual attending salary) were checked for overall model fit by Hosmer–Lemeshow methodology. Individual P values <.05 were considered significant.
Two of the authors (K.D. and R.M.) independently reviewed and developed a coding scheme for all open-ended responses to determine students' reasons for not applying to primary care residencies and what could be changed to influence those decisions. After discussion, a common coding scheme was developed and subsequently reviewed separately by a third author (R.C.) to ensure the scheme was reasonable. The original two authors (K.D. and R.M.) then independently coded the database. For each student, every phrase was coded, so each theme was coded as present or absent for each student. Discrepancies were resolved by consensus.
A sensitivity analysis was performed for lifestyle issues, based on whether or not the student's choice of specialty was uncontrollable (obstetrics–gynecology, general surgery, orthopedics, neurosurgery, urology) or controllable (all others) as defined by Dorsey et al13 and Schwartz et al.17 Students' willingness to accept a financial incentive to apply for primary care was compared against uncontrollable and controllable lifestyle specialty choice while controlling for interest in primary care (defined as listing a primary care residency as one of their other three specialties considered). All calculations were made with STATA 8.0, College Station, Texas.
The response rate was 56% (447 responded of 797 students applying for residency in the Military Healthcare System). One hundred fifty-four (34%) chose a primary care residency (internal medicine, pediatrics, or family medicine) (Table 1). Demographic information is included in Table 2. None of these demographic characteristics were different between primary care and non-primary-care applicants, with the exception of age (primary care applicants' mean age was 28.6 years versus 27.9 for nonprimary care, P = .02, Student's t test).
The question “How much of a monetary bonus [before internship and after residency] would it have taken for you to apply for primary care?” was answered by 272 (93%) of the students who did not apply for primary care. Of these, 82 (30%) would have accepted such a bonus to apply for primary care, ranging from $6,000 to $100,000, with a median of $27,500 (interquartile range [IQR] $15,000–$50,000). An additional 11 students who suggested amounts ranging from $150,000 to $1,000,000 were reclassified as not accepting a bonus to apply for primary care, as the amount suggested was unreasonably high. The question “For what annual salary as an attending physician in internal medicine, pediatrics, and family medicine would you have considered applying for primary care?” was also answered by 272 (93%) of the students who did not apply for primary care. Of these, 112 (41%) listed an annual salary for which they would have considered applying for primary care, ranging from $10,000 to $250,000, with a median of $175,000 (IQR $150,000–$200,000). An additional 15 students who suggested amounts ranging from $280,000 to $750,000 were reclassified as not accepting a higher salary to consider applying for primary care, as the amount suggested was unreasonably high. Students were also asked which type of bonus they preferred. Of those indicating a preference (n = 153), the residency bonus alone was the least popular choice (n = 17; 11%), followed by an increase in attending pay (n = 61; 40%) and finally a mixture of both (n = 75; 49%).
Students were asked at the end of the survey to estimate (1) the annual salary of civilian primary care physicians (general internal medicine, family medicine, and pediatrics) two to three years after residency and (2) the annual salary of their chosen field if it was not a primary care specialty. The median estimated civilian annual salary for general internal medicine was $25,000 less than the participants' median requested military annual salary ($150,000 versus $175,000). For family medicine and pediatrics, the difference was $50,000 ($125,000 versus $175,000). The median civilian expected annual salary for students who did not choose primary care was $230,000 (IQR $200,000–$290,000). Thus, some students were willing to consider primary care for a median income that was $55,000 less or 76% of expected income for their chosen specialty ($175,000 requested salary versus $230,000 anticipated civilian salary).
These financial considerations were also explored in other ways. Equal wages (same pay for every hour worked per week) for all attending physicians regardless of specialty did not seem to entice students to apply for primary care (mean rating 3.1 [SD 2.3] on a nine-point Likert scale, where 1 = still would not have applied and 5 = possibly would have applied). However, students who listed an annual salary that would influence their decision to apply for primary care were more likely to support this concept (mean 4.3 versus 2.2, P < .0001, Student's t test). The typical higher salaries of civilian physicians compared with military physicians also had a minimal effect on students' decisions to apply for primary care (mean 3.1 [SD 2.7] on a nine-point Likert scale, where 1 = no effect and 5 = some effect), but students who listed a salary for which they would have considered applying for primary care endorsed the concept more than those who did not list a salary (mean 4.2 versus 2.2, P < .0001, Student's t test).
The question “What is the most important reason you didn't apply for a primary care residency?” was answered by 261 (89%) of the 293 students who didn't apply for primary care residency (Figure 1). Ten themes (in italics in the following sentences) emerged from the responses (listed in Figure 1). The most common reason was that they preferred another specialty (n = 94; 36%), with students stating factors such as “I love OB-GYN” or “never wanted to do anything except surgery.” This was closely followed by the not interested theme (n = 77; 30%), which was illustrated in students' comments by terms like “not interested,” “monotony,” “bored,” and “hate clinic.” The next group included practice aspects (n = 36; 14%; e.g., “workload,” “appointment length,” “not enough time with each patient,” “conveyor belt feel,” “health care factory”), patient characteristics (n = 33; 13%; e.g., “don't like bread-and-butter cases,” “social issues of primary care,” “not managing disease processes” “noncompliant patients,” “preventative medicine”), and financial reasons (n = 32; 12%; e.g., “pay,” “pay after the military,” “increase pay in the civilian sector,” “reimbursement equivalent to the other specialties”). Ten percent (n = 26) of students cited lifestyle issues, stating “flexibility to put family first,” “longer work hours,” “did not want to be on call,” “time off,” “time with family,” “control of hours,” and “quality of life.” Another 10% (n = 26) of students referred to preferring procedures (e.g., “lack of procedures,” “I want to operate,” “not enough hands-on, procedure-based stuff”). The remaining three themes were cited by less than 7% of students: expert status (n = 15; 6%; e.g., “having to consult too much,” “enjoy specialty work,” “prestige,” “too broad”), military specific (n = 6; 2%; e.g., “deployments,” “lack of combined programs [family medicine–psychiatry] in military”), and poor educational experience (n = 3; 1%).
The question “What is the most important thing(s) that could be changed that would have changed your mind to apply for a primary care residency (if nothing, please state so)?” was answered by 257 (88%) of the 293 students who didn't apply for primary care. The most common answer was nothing, cited by more than half (n = 147; 57%) of the students (Figure 2). The next most common answer was financial reasons (n = 57; 22%), with students answering “pay,” “reimbursement equivalent to the other specialties,” “increased annual pay,” “more money,” “financial compensation,” and “increased pay in the civilian sector.” Lifestyle issues was next (n = 30; 12%), illustrated by terms such as “quality of life,” “decreased work hours,” “half-time work,” “less call,” “change of work schedule,” and “easier scheduling.” Practice aspects, which centered on appointment length and workload, was cited by 9% (n = 23) of students, who gave such reasons as “decrease patient load so able to devote more time to each patient,” “longer appointment times,” “less patients seen per day,” and “more of a teamwork approach.” No other theme was listed by more than 5% of students. These remaining themes included patient characteristics (n = 13; 5%; e.g., “different patient population,” “less having to see unnecessary visits from nonsick to hardly ill patients”), procedures (n = 8; 3%; e.g., the field would have to be much more procedure based,” “more procedures,” “start doing surgery”), less interest (n = 8; 3%; e.g., “if I were more passionate about [pediatrics] than OB-GYN,” “found desire to do so,” “didn't find general surgery so appealing”), military specific (n = 6; 2% “deployments,” “option for combined residencies”), additional training (n = 4; 2%; e.g., “true generalist training,” “surgery, ENT, urology mini-fellowships,” “more fellowship opportunities”), and better educational experiences (n = 3; 1%).
Students were then asked to rate on a nine-point Likert scale their answers to the question, “If the most important thing you just listed were changed, how would this have affected your applying for primary care?” Students who answered something other than “nothing” to the previous question (n = 109) had a mean rating of 5.6 (SD 1.9), which was between “somewhat” (an answer of 5) and a “great deal” (an answer of 7) on the scale.
Multivariate analysis for financial questions
Using logistic regression, we assessed responses to financial questions about the residency bonus and the attending salary to determine underlying factors that were associated with accepting a financial remuneration. For the residency bonus question, four factors were associated with indicating any amount: citing a lifestyle reason for not selecting primary care (OR 3.5; 95% CI 1.4–9.0; P = .01), considering primary care in one of the top four residency choices (OR 3.3; 95% CI 1.8–6.0; P < .001), indicating that a flexible work schedule would have resulted in at least some consideration toward applying for primary care (OR 2.5; 95% CI 1.1–5.8; P = .03), and restriction of deployment to combat zones to a maximum of six months at time for all physicians, not just specialist physicians (OR 2.0; 95% CI 1.0–4.1; P = .05). For the attending salary question, four factors were associated with listing any amount: indicating that a flexible work schedule would have resulted in at least some effect toward applying for primary care (OR 6.4; 95% CI 3.2–13.2; P < .001), citing a lifestyle reason for not selecting primary care (OR 3.2; 95% CI 1.1–9.6; P = .04), desiring a residency with time for outside interests (OR 2.7; 95% CI 0.9–7.7; P = .07; retained in the final model due to confounding with lifestyle), and considering primary care in one of the top four residency choices (OR 2.6; 95% CI 1.4–2.8; P = .003). For both analyses, no demographic factor was significant, nor was the amount of educational debt or type of medical school (MD granting versus DO granting).
Lifestyle sensitivity analysis
As the factors associated with accepting a financial incentive centered around interest in primary care and lifestyle issues, additional logistic regression analyses were performed using only two variables: consideration of primary care and type of specialty selected in the Match (controllable versus noncontrollable). Both consideration of primary care and selecting a controllable lifestyle specialty were associated with listing an annual salary to consider applying for primary care (considered primary care: OR 2.9; 95% CI 1.7–5.0; P < .001; controllable lifestyle: OR 3.4; 95% CI 2.0–5.9; P < .001). However, for listing a residency bonus, only consideration of primary care was significant (considered primary care: OR 3.5; 95% CI 2.0–6.1; P < .001; controllable lifestyle: OR 1.5; 95% CI 0.9–2.7; P = .15). In other words, students who considered primary care but chose a controllable lifestyle specialty were nearly four times more likely to list a salary to consider applying for primary care than were students who did not consider primary care and selected a noncontrollable lifestyle specialty (38/57 or 67% versus 16/93 or 17%).
We found that most students who did not apply for primary care (internal medicine, family medicine, and pediatrics) did so simply because of lack of interest in those specialties. Students indicated financial and lifestyle aspects were the primary factors that could have been changed to convince them to apply for primary care. Many (41%) who did not apply for primary care appeared as though they would have considered applying if the annual salary were higher, at a median military annual salary of 76% of the expected civilian median salary from their chosen specialty. Students who considered primary care but chose a controllable lifestyle specialty were particularly likely to accept a salary to consider applying for primary care, implying a willingness to make trade-offs between lifestyle and financial remuneration. Almost one-third of students appeared as though they would have applied for primary care if a bonus payment were made at the beginning of internship and end of residency. The only modifiable nonfinancial and lifestyle factors that would entice more fourth-year students into primary care were practice aspects, particularly “appointment length.”
Our finding that financial incentives can influence students' career choice is consistent with other findings from the last 15 years.3–6,8,9,11,12 However, only one previous study (in 1994) attempted to quantify the amount of financial compensation necessary to change students' minds.11 In this study, 17% of students not choosing primary care would have chosen primary care for an annual salary of $180,000, which was 18% higher than their expected specialty income of $152,000.11 In contrast, our study found that more non-primary-care students (41%) would have considered switching to primary care for a median salary of $175,000, which was 76% of their expected specialty income of $230,000. The reasons for these differences are not clear, but might be explained by a lower percentage of students entering primary care in the 1994 study (27% versus 34% in our study), differences in survey technique (a pick list of incomes in the 1994 study compared with open-ended questions in our study), our students' expectation of a military salary, and changes in students' motivation over time.
The other major determinant of what could be changed was lifestyle, which again is consistent with the previous literature.4,8,11,13–16 Our finding that 67% of students who considered primary care but chose a controllable lifestyle specialty would have considered applying for primary care for a higher salary suggests that some students can be given financial enticements to offset the perceived lifestyle disadvantages of primary care. Exactly which of these lifestyle concerns would be most attractive to students is not clear, though our respondents seemed to favor a flexible work schedule and perhaps time for outside interests. Further research is needed to better describe the relative impacts of these lifestyle concerns.
In addition to financial rewards and lifestyle changes, we found a variety of other factors that, if changed, might have enticed students to apply for primary care. Most of these, such as changing the patient population or including opportunities to perform surgery, would change the dynamic of the practice to be something other than primary care. Many potentially modifiable factors within primary care practice are contained within the domain of “practice aspects,” as shown in Figure 1. Several quotes from the students are particularly enlightening about this factor:
Primary care physicians, I felt, are overworked, underpaid, and underappreciated. They are supposed to take care of all of there [sic] patients [sic] basic medical problems, give advice, council on preventative med, etc., all within 15 minutes.
Too many patients to see with not enough time to devote to each patient.
The expectations for the amount of patients seen per day in most primary care practices [are] too high. Seeing that many patients every day is exhausting.
It seems that some students perceived the workload of primary care to be too high, both in the number of patients seen and the expectations for what should be accomplished during visits. Perhaps students are seeing a “chaotic work pace,” as was seen in nearly half of primary care practices in a recent study.20 Our finding lends support to the recommendation of others that students need a different training paradigm in outpatient medicine.21–23
Our study had several limitations. The response rate was only 56%, leaving the possibility for nonresponse bias. However, this response rate is typical for physician studies.24 We also determined that the specialties of the Army HPSP nonresponders were nearly identical (within 2%) to those of the responders for all but two specialties. The demographic characteristics of students with a military obligation may be different from nonmilitary students (e.g., gender, debt), but unlike lifestyle issues, none of these factors were significantly associated with accepting a financial reward in the regression models. In addition, 85% of the students attended civilian medical schools, indicating that the results can possibly be generalized to all U.S. medical students. However, some students, possibly those with longer service obligations, may have been less aware of or less concerned with civilian physician pay. As with other studies in this area, we did not follow students who selected primary care to determine whether they ultimately pursued subspecialty training, but this is of lesser importance because our study focused on students who did not initially choose primary care. Finally, because the survey took place after the Match, recall bias is a possibility.
Our data suggest that until financial and lifestyle issues are adequately addressed, some students who may have selected primary care will choose a more lucrative and lifestyle-friendly specialty. If the goal is to correct the shortage of primary care providers by increasing the number of primary care trainees, policy makers should consider altering reimbursement such that primary care physicians earn at least 75% of the salary of most other physicians for equivalent practice-related time and advocate for delivery systems that are sensitive to provider satisfaction.
The authors would like to thank Dr. Patrick G. O'Malley for his assistance in revising the manuscript.
The institutional review boards at William Beaumont Army Medical Center in El Paso, Texas, Carl R. Darnall Army Medical Center in Fort Hood, Texas, and the Uniformed Services University in Bethesda, Maryland, reviewed and approved the protocol.
The opinions in this manuscript are those of the authors and do not represent the official policy of the U.S. Government, the Department of Defense, or any unit within the Department of Defense.
This study was presented at the 2009 Army–Air Force regional meeting of the American College of Physicians, San Antonio, Texas, November 19, 2009.
1Starfield B, Shi L, Macinko J. Contribution of primary care to health systems and health. Milbank Q. 2005;83:457–502.
2Newton DA, Grayson MS. Trends in career choice by US medical school graduates. JAMA. 2003;290:1179–1182.
3Ebell MH. Future salary and US residency fill rate revisited. JAMA. 2008;300:1131–1132.
4Fincher RM. The road less traveled—Attracting students to primary care. N Engl J Med. 2004;351:630–632.
5Fincher RM, Lewis LA, Jackson TW. Why students choose a primary care or nonprimary care career. The Specialty Choice Study Group. Am J Med. 1994;97:410–417.
6Hauer KE, Fagan MJ, Kernan W, Mintz M, Durning SJ. Internal medicine clerkship directors' perceptions about student interest in internal medicine careers. J Gen Intern Med. 2008;23:1101–1104.
7Kahn MJ, Markert RJ, Lopez FA, Specter S, Randall H, Krane NK. Is medical student choice of a primary care residency influenced by debt? Med Gen Med. 2006;8:18.
9Newton DA, Grayson MS, Whitley TW. What predicts medical student career choice? J Gen Intern Med. 1998;13:200–203.
11Rosenthal MP, Diamond JJ, Rabinowitz HK, et al. Influence of income, hours worked, and loan repayment on medical students' decision to pursue a primary care career. JAMA. 1994;271:914–917.
12Morra DJ, Regehr G, Ginsburg S. Medical students, money, and career selection: Students' perception of financial factors and remuneration in family medicine. Fam Med. 2009;41:105–110.
13Dorsey ER, Jarjoura D, Rutecki GW. Influence of controllable lifestyle on recent trends in specialty choice by US medical students. JAMA. 2003;290:1173–1178.
15Hauer KE, Durning SJ, Kernan WN, et al. Factors associated with medical students' career choices regarding internal medicine. JAMA. 2008;300:1154–1164.
16Schwartz MD, Linzer M, Babbott D, Divine GW, Broadhead E. Medical student interest in internal medicine. Initial report of the Society of General Internal Medicine Interest Group Survey on Factors Influencing Career Choice in Internal Medicine. Ann Intern Med. 1991;114:6–15.
18Durning SJ, Dorrance K, Denton D, Poremba J, Roy M. Do residents benefit from participating in internal medicine interest groups? A study of resident perceptions from two institutions. Mil Med. 2007;172:210–213.
20Linzer M, Manwell LB, Williams ES, et al. Working conditions in primary care: Physician reactions and care quality. Ann Intern Med. 2009;151:28–36.
21Fitzgibbons JP, Bordley DR, Berkowitz LR, Miller BW, Henderson MC. Redesigning residency education in internal medicine: A position paper from the Association of Program Directors in Internal Medicine. Ann Intern Med. 2006;144:920–926.
22Hauer KE, Alper EJ, Clayton CP, Hershman WY, Whelan AJ, Woolliscroft JO. Educational responses to declining student interest in internal medicine careers. Am J Med. 2005;118:1164–1170.
23Whitcomb ME, Cohen JJ. The future of primary care medicine. N Engl J Med. 2004;351:710–712.
24Asch DA, Jedrziewski MK, Christakis NA. Response rates to mail surveys published in medical journals. J Clin Epidemiol. 1997;50:1129–1136.