Search tips
Search criteria 


Logo of jgimedspringer.comThis journalToc AlertsSubmit OnlineOpen Choice
J Gen Intern Med. 2008 July; 23(7): 1095–1100.
Published online 2008 July 10. doi:  10.1007/s11606-008-0579-z
PMCID: PMC2517937

Changes in U.S. Medical Students’ Specialty Interests over the Course of Medical School

Michael T. Compton, MD, MPH,corresponding author1,2 Erica Frank, MD, MPH,1,3 Lisa Elon, MS, MPH,4 and Jennifer Carrera, MS4



Studies have examined factors affecting medical students’ specialty choice, but little research exists on stability of these specialty interests.


To describe patterns of change in specialty interests during medical school and examine associations between specialty change patterns and gender, desire for a high-prestige career, and interest in prevention.


Medical students (Class of 2003) at 15 representative US schools were invited to complete surveys during freshman orientation, entry to wards, and senior year.


This analysis used data from 942 students who completed all 3 surveys.


In addition to a number of other items, students were asked to choose the 1 specialty they were most interested in pursuing.


The most common specialty choices among freshman students were pediatrics (20%) and surgery (18%); least common choices were psychiatry and preventive medicine (1% each). General internal medicine was the initial specialty choice for 8%. Most students changed their specialty choices, regardless of initial interest. Only 30% of those initially interested in primary care (PC) remained interested at all 3 time points, compared to 68% of those initially interested in non-PC. Female versus male students were more commonly interested in PC at all 3 time points. Senior students interested in non-PC specialties were more likely to desire a high-prestige career (48%) than those interested in PC (31%).


Medical students may benefit from more intensive introduction to some specialties earlier in pre-medical and medical education. In addition, increasing the prestige of PC fields may shape the physician workforce.

KEY WORDS: medical students, changes in specialty choices, prestige


The distribution of physicians across specialty areas has crucial implications for access and availability of the breadth of medical services, overall costs of the health care system, and optimal health of the nation. Considering the importance of the supply side of this equation—especially in light of recent trends toward a relative lack of primary care (PC) and general internal medicine (GIM) physicians—the topic of medical students’ specialty interests is worthy of research.

A number of recent studies have documented determinants of specialty choices of medical students. Putative predictors include: institutional characteristics, including private versus public school1,2; medical school experiences and curricular features, such as influence of faculty members/mentors/role models35; demographic variables68; specialty characteristics, including types of patient problems79; lifestyle factors, such as workload and prestige4,7,8; and financial issues, including income potential and level of indebtedness.7,10,11 However, a dearth of research exists on stability of initial specialty interests during medical education.

The primary purpose of this study was to describe patterns of change in medical students’ specialty interests. A secondary purpose was to examine the association of specialty selections and change patterns with: (1) gender, (2) desire for a high-prestige career, and (3) interest in prevention. Some prior research has suggested that gender may influence stability of specialty choice; for example, fewer female than male students who are interested in surgery at entry to medical school match in that specialty.12 Prestige within the medical profession has been studied as a potential factor influencing medical students’ specialty choice, especially among students selecting surgical specialties.9,10 The third factor was selected because no prior research has examined the potential association between interest in prevention and changes in specialty interest.



All Class of 2003 medical students at 15 US schools were invited to complete 3 questionnaires, at first year orientation (1999), orientation to clinical rotations/wards (typically between the second and third years), and senior year (2002–2003). A nationally representative convenience sample of schools reflected all U.S. medical schools in terms of age, school size, National Institutes of Health research ranking, private/public school balance, underrepresented minorities, gender, and geographic distribution.1317 The following schools were included: Creighton University School of Medicine, Duke University School of Medicine, Georgetown University School of Medicine, Loma Linda University School of Medicine, Medical College of Georgia School of Medicine, Mercer University School of Medicine, Morehouse School of Medicine, Texas Tech University Health Sciences Center School of Medicine, Tulane University School of Medicine, University of California at Los Angeles School of Medicine/Charles R. Drew University of Medicine and Science, University of Colorado School of Medicine, University of Medicine and Dentistry of New Jersey Robert Wood Johnson Medical School, University of Pennsylvania School of Medicine, University of Rochester School of Medicine and Dentistry, and Wayne State University School of Medicine.

At freshman orientation, 2,080 students were eligible to complete the survey and 1,846 responded; 1,982 were eligible at entry to wards and 1,630 responded; and 1,901 were eligible during their senior year and 1,469 responded. Therefore, the overall response rate was 83% (4,945 observations out of 5,963 eligible observations). A total of 971 students responded at all 3 time points.


The survey focused primarily on medical students’ health and wellness, and specific sections included: personal characteristics, health status and behaviors, and health care opinions. Prior publications have examined key research questions related to medical student health.1820

At each of the 3 survey administrations, students were asked, “Choose the one specialty you are now most interested in pursuing.” Family medicine, GIM, obstetrics/gynecology, pediatrics, and preventive medicine/public health were categorized as PC, which is generally consistent with prior definitions of PC in other studies of medical students’ career choice,7 although obstetrics/gynecology was classified as PC. Emergency medicine, medicine subspecialties, psychiatry, surgery, urology, anesthesiology/pathology/radiology, and “other” were deemed non-PC. The “undecided” responses formed a separate category. For the second and third survey administrations, the category “pediatric subspecialty” was added and included in the non-PC category.

The survey given to fourth year students included the following item, “Having a high-prestige career is important to me.” Likert scale response choices were dichotomized as affirmative (“strongly agree” or “agree”) versus the neutral or negative responses. Also during the fourth year, students were asked to respond to the statement “Prevention is less interesting to me than treatment.” Response choices were trichotomized as strongly agree/agree, neutral, and disagree/strongly disagree.


Using an Institutional Review Board (IRB)-approved protocol, surveys were usually administered after semi-mandatory activities (e.g., orientation lunches), or at the end of a class. When occasioned based on low response rates, Dillman’s 5-stage mailing process was used to maximize response rates.21 Approximately one quarter of the senior-year questionnaires were filled out in the first half of the school year, whereas over half were completed toward the end of the school year (April to June). Three schools chose to administer the questionnaire in a mandatory rotation, collecting data throughout the year, whereas the remainder primarily collected data on 1 or 2 specific dates.


Data from the 942 students who returned all 3 questionnaires and responded to the query on specialty interest were used in all analyses. A number of exploratory analyses were conducted to ensure that bias would not be introduced into the main data analytic plan. First, possible bias because of not including those who responded to only 2 questionnaire administrations was examined. Analyses assessed whether or not the estimate of those reporting interest in PC in both the freshman year and at ward orientation would be different if the 366 students who responded only at those 2 times were included. Similar comparisons were made for those responding only at the first and third administrations and those responding only at the second and third administrations; in no case was it found that the estimates differed by more than 2 percentage points from estimates based on the 942 full responders used in the main analyses. Second, first-year respondents who later completed the other 2 questionnaires were somewhat less likely to report a non-White ethnicity, but equally likely to be female and report interest in prevention compared to those who did not. Third, time-specific prevalences of individual specialties based on all responders did not differ from those based on complete responders. Fourth, to ensure that the 942 respondents had equal representation from the 15 schools (i.e., that the majority did not come from just a few of the schools), numbers of respondents from the various schools were examined. The 4 smallest schools (14% of the full cohort) comprised 13% of the subset analyzed here, and the 4 largest schools (39% of the full cohort) comprised 40%.

For the main analyses, associations between specialty choice patterns and key variables of interest (gender, desire for a high-prestige career, and interest in prevention) were examined using chi-square tests of association. These tests took into account the non-independence between observations (because of clustering of students within schools); variance estimates were adjusted using the SUDAAN statistical software for correlated data,22 which uses the Taylor linearization method. Although students were assessed at 3 times, for the purposes of this paper each student was assigned a code that indicated their choice pattern over time; therefore each student appears only once in these analysis.


Data on gender and race/ethnicity for the sample of 942 students are shown in Table 1. The prevalence of specialty intentions at freshman orientation, ward orientation, and during the senior year are displayed in Figure Figure1.1. Pediatrics and surgery were the most popular initial choices, followed by family medicine. By ward orientation, the numbers choosing pediatrics had declined sharply, and it remained a less popular choice during senior year. Among the other PC specialties, the general pattern was one of slight change. In contrast, nearly all non-PC specialty choices were more common with time. One fifth of students were undecided on a specialty choice until senior year, when the number then dropped to 1%, and urology was chosen by 1% or fewer at all times points (not displayed).

 Figure 1
Medical students’ (n = 942) intended specialty choices over the course of medical school, percentages shown (time point 1 = freshman orientation, 2 = ward orientation, 3 = senior year). Pediatric subspecialty was not queried at ...
 Table 1
Description of the Sample of 942 U.S. Medical Students (Senior Year, 2003)

Table 2 displays the patterns of specialty choice over the course of medical school. As mentioned previously, the most common specialty choices among freshman students were pediatrics (185, 20%) and surgery (165, 18%). A similar proportion reported being undecided on a specialty choice (176, 19%). The majority of students changed their specialty choices, either by ward orientation or senior year. For example, only 15% initially interested in pediatrics were interested consistently in that field across all 3 surveys, 11% switched to another choice at ward orientation and reverted to pediatrics during senior year, 10% later chose a pediatric subspecialty, 44% switched to a non-pediatric, non-PC choice, and the remaining 20% switched to another PC specialty. The specialties in which students were most likely to maintain their initial interest included emergency medicine (37%), psychiatry (33%), obstetrics/gynecology (25%), surgery (24%), and family medicine (23%). Regardless of the initial specialty interest, when changing to another specialty by senior year, a non-PC specialty was the most likely new choice. Among the 176 students who were “undecided” when starting medical school, 71% were interested in pursuing a non-PC specialty by senior year. Whereas Figure Figure11 refers to overall numbers with interest in the noted specialties, not the pattern for individuals, Table Table22 indicates that relatively few individuals starting with an interest in GIM remained interested, but other individuals became interested in PC over the course of medical school.

 Table 2
Changes in US Medical Students’ Choice of Specialty, from Freshman Orientation to Senior Year, 1999–2003 (n = 942)*

Table 3 shows patterns of specialty choice when specialties were categorized in the more general terms of PC (including GIM), non-PC, and undecided based on specialty interest in freshman year. The most common patterns over the 3 time points were those consistent in choice (i.e., PC at all 3 time points or non-PC at all 3 time points) and those switching from PC to non-PC. Only 30% of those initially interested in PC remained interested at all 3 time points, and an additional 17% chose a non-PC specialty or were undecided at ward orientation but reverted to a PC specialty by senior year. In contrast, 68% of those initially interested in non-PC remained in a non-PC specialty across all 3 surveys, and an additional 16% settled by senior year on a non-PC specialty. Of those initially undecided, 40% remained undecided at ward orientation, and <1% remained undecided by senior year. At freshman orientation, 418 students (44%) were interested in pursuing PC, compared to 298 (32%) by senior year. About half of those changing from an initial choice of PC had done so by ward orientation (107/220).

 Table 3
Specialty Choice (PC, Non-PC, and Undecided) Patterns Among US Medical Students, 1999–2003 (n = 942)

Because the reputation of some schools for concentrating on either PC or specialties may attract students wanting to go into those respective areas, school-specific patterns were examined. The basic pattern of choice (changing from PC to specialties being much greater than the opposite) was roughly similar for all the schools in the sample: PC being an initial choice for roughly half (school range, 43–72%, median 57%), with large proportions making non-PC choices by senior year (school range, 31–82%, median 53%). Conversely, little movement out of non-PC specialty choices was observed in any school (school range for change to PC, 0–38%, median 14%).

Among females starting in a non-PC specialty, 73% remained in that category as seniors, compared to 90% of males (p = .008). (Note that this is slightly different from the 68% of all students initially interested in non-PC who remained in a non-PC specialty across all 3 surveys mentioned above.) The percentages of females interested in PC at each of the 3 time points were 57%, 41%, and 44%, respectively, compared to 34%, 17%, and 21%, respectively, for males. The differences between genders were significant at all time points (all p < .001).

Basic descriptive data on importance of having a high-prestige career and interest in prevention are shown in Table 1. There was no significant association between consistency of specialty category (i.e., the same category choice at all 3 times versus some change) and prevention interest (p > .05). Male students were more likely than female students (48% versus 36%, p = .005) to report prestige as being important to them. Students indicating interest in a non-PC specialty were more likely to state a desire to have a high-prestige career than those interested in a PC specialty (48% versus 31%, p = .002). Specifically, ≥50% of students intending to specialize in a medical subspecialty, surgery, or anesthesiology/pathology/radiology reported that having a high-prestige career was important to them. Fewer of those pursuing family medicine (18%), general pediatrics (30%), or GIM (40%) stated that prestige was important.


Several interesting findings emerged. First, the most common specialty choices among freshman students were pediatrics (20%) and surgery (18%), and the 2 specialties least likely to be chosen were psychiatry (1%) and preventive medicine/public health (1%). Surgery has been reported previously to be one of the most popular specialty areas among medical students at matriculation.5,8,23 Students may enter medical school relatively familiar with some disciplines, such as pediatrics and surgery (and thus having a high level of interest in these areas), but may be less familiar with other disciplines, like psychiatry and preventive medicine. In addition, stigma, perceived lack of prestige, and inadequate inclusion in the premedical undergraduate curriculum could account for this initial low level of interest. Given both the well-recognized burden of psychiatric disorders and the established importance of prevention, medical students may benefit from more intensive introduction to these specialties earlier in premedical and medical education.

Second, while there was some stability in specialty choice, most students changed their specialty choices, either by the time of ward orientation or by senior year, regardless of the initial specialty interest. When changing to another specialty, a non-PC specialty was the most likely new choice. Furthermore, among students starting medical school undecided on a specialty interest, the majority (71%) were planning to pursue a non-PC specialty by senior year. Less than one third of those initially interested in PC remained interested at all 3 time points, compared to more than two thirds of those initially interested in non-PC. Consistent with earlier reports, only 20–45% of medical students ultimately choose the specialty that they had been initially most interested in.8,2325 Others have noted that most students do not ultimately choose the specialty they originally prefer, and that the direction of change is typically away from PC.26,27 Also noteworthy was the finding that at all 3 time points, female medical students were more likely to choose PC fields.

Third, consistent with prior research,8,9 students wanting a high-prestige career were more likely to be interested in a non-PC specialty than a PC specialty at all 3 time points, and almost half of students reported that prestige was important. Earlier regional studies have shown similar relationships.2830 This association has implications for efforts to shape the physician workforce, such as increasing the prestige of GIM and other PC fields, which may be accomplished if greater respect were afforded to PC disciplines by academic medical centers, the medical profession generally, and society at large. Further survey research involving medical students could elucidate the extent to which such strategies (and others, including increased reimbursement for PC services) could impact medical students’ specialty choice.

A number of interventions have been suggested, and some have been studied, in efforts aimed at effecting change, especially in terms of producing more PC physicians.3 The current findings appear to support some of these prior suggestions, such as enhanced medical school support of students interested in PC. Efforts to increase the number of US medical students going into PC will be more successful when they are less interested in prestige, or when PC specialties are given more prestige, as reflected in both widespread intra- and extraprofessional attitudes, and compensation.

Several methodological limitations of this analysis should be considered. First, the study focused on patterns of change in specialty interest among medical students, and data on the residency match, final career choices, and future career changes were unavailable. Students may respond differently before and after the match, although this could not be examined. Second, only a few key variables were examined in relation to stability of specialty interest, although a large number of potentially important factors can be enumerated.3,27,31 Third, estimates may not reflect population values because the school sample was non-random and a considerable portion of the sample did not contribute information at all survey points, although sensitivity analyses revealed little effect of analyzing only students responding to all 3 surveys. Fourth, despite the fact that a desire to have a high-prestige career was associated with specialty choice, desired prestige is a very complex construct, and its measurement in the present survey was very limited. Fifth, data were not available on students’ potential intentions to pursue careers as hospitalists. Given the growth and increasing recognition of this relatively new field,32,33 future research should include it in surveys of medical students.

To more fully understand specialty interest decision-making, research should assess changes in specialty choice before medical school matriculation, as well as changes that occur after medical school graduation, in addition to further examining patterns of change in medical students’ specialty interests during medical school. Specialty decisions often change during residency,34 and a substantial portion of physicians change specialties after entering practice, caused in part by general dissatisfaction with their choice because of lifestyle incompatibility and negative practice experiences.35 Specialty interest is a complex developmental process, but a greater understanding of the decision-making process may inform efforts to optimize the distribution of physicians across specialty areas; this may, in turn, have implications for access and availability of medical services, overall costs of the health care system, and population health.


This research was supported by grant funding from the American Cancer Society and the Centers for Disease Control and Prevention.

Conflict of Interest None disclosed.


1. Weissman SH, Haynes RA, Killian CD, Robinowitz C. A model to determine the influence of medical school on students’ career choices: psychiatry, a case study. Acad Med. 1994;69:58–9. [PubMed]
2. Sierles FS, Dinwiddie SH, Patroi D, Atre-Vaidya N, Schrift MJ, Woodard JL. Factors affecting medical student career choice of psychiatry from 1991–2001. Acad Psychiatry. 2003;27:260–8. [PubMed]
3. Rosenthal MP, Rabinowitz HK, Diamond JJ, Markham FW. Medical students’ specialty choice and the need for primary care. Prim Care. 1996;23:155–67. [PubMed]
4. Cochran A, Melby S, Neumayer LA. An internet-based survey of factors influencing medical student selection of a general surgery career. Am J Surg. 2005;189:742–6. [PubMed]
5. Sanfey HA, Saalwachter-Schulman A, Nyhof-Young JM, Eidelson B, Mann BD. Influences on medical student career choice: gender or generation? Arch Surg. 2006;141:1086–94. [PubMed]
6. Pamies RJ, Lawrence LE, Helm EG, Strayhorn G. The effects of certain student and institutional characteristics on minority medical student specialty choice. J Nat Med Assoc. 1994;86:136–40. [PMC free article] [PubMed]
7. Newton DA, Grayson MS, Whitley TW. What predicts medical student career choice? J Gen Intern Med. 1998;13:200–3. [PMC free article] [PubMed]
8. Barshes NR, Vavra AK, Miller A, Brunicardi FC, Goss JA, Sweeney JF. General surgery as a career: a contemporary review of factors central to medical student specialty choice. J Am Coll Surg. 2004;1999:792–9. [PubMed]
9. Kassebaum DG, Szenas PL. Factors influencing the specialty choices of 1993 medical school graduates. Acad Med. 1994;69:164–70. [PubMed]
10. Ruhnke GW. Physician supply and the shifting paradigm of medical student choice. JAMA. 1997;277:70–1. [PubMed]
11. Tonkin P. Effect of rising medical student debt on residency specialty selection at the University of Minnesota. Minn Med. 2006;89:46–9. [PubMed]
12. Novielli K, Hojat M, Park PK, et al. Change of interest in surgery during medical school: a comparison of men and women. Acad Med. 2001;76:S58–61. [PubMed]
13. Stoll M. The Best Medical Schools. 2000 Edition. New York, NY: Princeton Review Publishing; 1999.
14. Association of American Medical Colleges. Applicant age at anticipated matriculation, 1992–2001. Accessed 26 February 2008.
15. Association of American Medical Colleges. Total enrollment by gender and race/ethnicity, 1992–2001. Accessed 26 February 2008.
16. Association of American Medical Colleges. Medical schools. Accessed 26 February 2008.
17. National Institutes of Health. Support to U.S. Medical Schools, Fiscal Year 2000. Accessed 26 February 2008.
18. Frank E, Carrera JS, Elon L, Hertzberg VS. Basic demographics, health practices, and health status of U.S. medical students. Am J Prev Med. 2006;31:499–505. [PubMed]
19. Frank E, Carrera JS, Elon L, Hertzberg VS. Predictors of US medical students’ prevention counseling practices. Prev Med. 2007;44:76–81. [PubMed]
20. Henderson SJ, Elon L, Frank E. Self-report of quality of medical student health care. Med Educ. 2007;41:632–7. [PubMed]
21. Dillman DA. Mail and Telephone Surveys: The Total Design Method. New York, NY: Wiley; 1978.
22. Shah BV, Barnwell BG, Bieler GS. SUDAAN User’s Manual, Release 8.0: Software for the Statistical Analysis of Correlated Data. Research Triangle Park, NC: Research Triangle Institute; 2001.
23. Markert RJ. Change in specialty choice during medical school. J Fam Pract. 1983;17:295–300. [PubMed]
24. Zeldow PB, Preston RC, Daugherty SR. The decision to enter a medical specialty: timing and stability. Med Educ. 1992;26:327–32. [PubMed]
25. Kassebaum DG, Szenas PL. Medical students’ career indecision and specialty rejection: roads not taken. Acad Med. 1995;70:938–43. [PubMed]
26. Babbott D, Baldwin DC, Jolly P, Williams DJ. The stability of early specialty preferences among U.S. medical school graduates in 1983. JAMA. 1988;259:1970–5. [PubMed]
27. Bland CJ, Meurer LN, Maldonado G. Determinants of primary care specialty choice: a non-statistical meta-analysis of the literature. Acad Med. 1995;70:620–41. [PubMed]
28. Rogers LQ, Fincher RM, Lewis LA. Factors influencing medical students to choose primary care or non-primary care specialties. Acad Med. 1990;65:S47–8. [PubMed]
29. Kassler WJ, Wartman SA, Silliman RA. Why medical students choose primary care careers. Acad Med. 1991;66:41–3. [PubMed]
30. Fincher RM, Lewis LA, Rogers LO. Classification model that predicts medical students’ choices of primary care or non-primary care specialties. Acad Med. 1992;67:324–7. [PubMed]
31. Weinstein P, Gipple C. Some determinants of career choice in the second year of medical school. J Med Educ. 1975;50:194–8. [PubMed]
32. Freed DH. Hospitalists: evolution, evidence, and eventualities. Health Care Manager. 2004;23:238–56. [PubMed]
33. Wachter RM. Reflections: the hospitalist movement a decade later. J Hosp Med. 2006;1:248–52. [PubMed]
34. West CP, Popkave C, Schultz HJ, Weinberger SE, Kolars JC. Changes in career decisions of internal medicine residents during training. Ann Int Med. 2006;145:774–9. [PubMed]
35. Jennett PA, Kishinevsky M, Bryant H, Hunter KL. Major changes in medical careers following medical school graduation: when, how often, and why. Acad Med. 1990;65:48–9. [PubMed]

Articles from Journal of General Internal Medicine are provided here courtesy of Society of General Internal Medicine