Dorsey, E Ray MD, MBA; Jarjoura, David PhD; Rutecki, Gregory W. MD
Research has shown that the recent specialty choices of graduating U.S. medical students are strongly associated with a specialty’s lifestyle.1 Specifically, after controlling for income, work hours, and duration of graduate medical education, “controllable lifestyle,” characterized by the physician’s control of time spent on professional responsibilities,2–5 accounted for 55% of the variability in specialty preference from 1996 to 2002.1 The factors underlying this strong association have yet to be determined.
Some authors have speculated that the growing number of women in medical school is driving the demand for specialties with a controllable lifestyle.1,6,7 Over the same time period in which the specialty preferences of U.S. medical students changed, the number of women graduating from medical school increased. From 1996 to 2002, the percentage of U.S. medical school graduates who were women increased from 41% to 44%, and from1983 to 2003, the percentage increased from 27% to approximately 45%.8
Investigators have cited lifestyle considerations as an explanation for specialty choice among women, especially for the relatively small number of women choosing the surgical fields.9,10 In addition, lack of “workplace control” and long work hours have been associated with increased “burn out” among female physicians.11 On the other hand, women account for nearly three-quarters of the residents in obstetrics and gynecology,12 a field with an uncontrollable lifestyle.3
In this study, we sought to determine to what extent sex contributes to the increase in choice of controllable lifestyle specialties.
Specialties for evaluation
We examined only specialties to which graduating U.S. medical students could apply. Based on data available from the 2000–2002 edition of Physician Socioeconomic Statistics published by the American Medical Association (AMA),13 we identified the following 16 specialties for study: anesthesiology, dermatology, emergency medicine, family practice, internal medicine, neurology, obstetrics and gynecology, ophthalmology, orthopedic surgery, otolaryngology, pathology, pediatrics, psychiatry, radiology (diagnostic), surgery (general), and urology.
Specialty choices of graduating U.S. medical students
We determined the specialty choices of graduating female and male medical students using responses to the Association of American Medical Colleges’ 1996–2003 Medical School Graduation Questionnaires (GQs). (We obtained unpublished data from R. Sabharwal, graduation questionnaire project director, April 2004.) The Graduation Questionnaire is administered annually to graduating U.S. medical students. In 2003, nearly 14,000 students completed the GQ, a response rate of approximately 90%.14 In 1996, the response rate was 83% and over 12,000 students completed the GQ.
Table 1 summarizes the lifestyle, income, work hours, and years of graduate medical education required for each of the selected specialties. Based on the work of Schwartz and colleagues,3 we classified the 16 specialties as having either a controllable or uncontrollable lifestyle. Orthopedic surgery and urology were considered surgical subspecialties and thus were classified as having uncontrollable lifestyles.1
We determined the average income for each selected specialty by averaging the mean net income after expenses, but before taxes, from 1993 to 1998 as reported in the AMA’s Physician Socioeconomic Statistics, 2000–2002 edition.13 We used a mean value because we found no statistically significant evidence for changes over these years in the relative rankings of incomes across specialties. The work hours per week for each of the selected specialties reflected the average number of hours in professional activities per week for 1998 and 1999.13,15 Finally, the minimum number of years of graduate medical education required for each of the selected specialties was determined from the AMA’s Graduate Medical Education Directory.16
To control for the influence of the trend away from primary care specialties,17 we conducted a separate analysis by eliminating family practice, internal medicine, obstetrics and gynecology, and pediatrics from the data.
A table of the numbers of graduating medical students who chose each of the 16 specialties within each sex for each of the years 1996–2003 formed the basis for the analysis. Our main focus was on the trends over years. The major portion of the variability in the cell percentages of this three dimensional table (specialty, time, and sex) was potentially explainable by a steady increase over time in preferences for specialties with controllable lifestyles. We used standard chi-square tests to determine the importance of two-way interactions (sex-by-specialty, specialty-by-time, and sex-by-time). The sex-by-specialty interactions are sensitive to differences between sexes in their specialty preferences. Specialty-by-time interactions are sensitive to a time trend in specialty preferences. Sex-by-time is sensitive to a difference in the proportions of female students over time.
We used log-linear models to determine how much of the variability in preferences over the years could be explained by the following four specialty characteristics: controllable lifestyle, income, work hours, and years of graduate medical education required. We focused in particular on the degree to which the trend toward controllable lifestyle was mainly due to preferences of women. In our log-linear model this analysis was represented by an interaction effect of sex, controllable lifestyle, and time. We used deviances from a saturated model (a three-dimensional model of specialty, time, and sex) to determine the explanatory power of sex and choice of controllable lifestyle specialties.18 For clarity, we dichotomized the income variable into higher or lower-than-average income.
We conducted a sensitivity analysis by eliminating the four primary care specialties (family practice, internal medicine, obstetrics and gynecology, and pediatrics) from the data. This analysis tested whether the effect of controllable lifestyle persisted apart from a trend away from primary care specialties. We performed analyses using the SAS Proc GENMOD statistical software program (SAS Institute Inc., Cary, North Carolina).
As Table 2 shows, the specialty choices of both women and men changed from 1996 to 2003. Table 2 also shows that the 16 selected specialties accounted for well over 90% of assigned specialty choices for all years examined.
While the change in specialty choice was large for both sexes, the differences between the specialty preferences of women and men over the time period examined were relatively small. Most of the variability in preferences was attributable to the interaction between time and specialty choice (χ2 = 3,424, 104 df, p < .0001) as opposed to sex and specialty choice. In addition, the three-way interaction among sex, specialty choice, and time was significant but explained much less of the change (χ2 = 202, 104 df, p < .0001). In other words, differences in preferences of women and men do not explain nearly as much of the variability in time in specialty choice as do changes over time in preferences common to both sexes.
As Figure 1 shows graphically, from 1996 to 2003 both women and men increasingly chose specialties with controllable lifestyles, and men consistently chose specialties with controllable lifestyles at a higher proportion than did women. The proportion of women choosing controllable lifestyle specialties rose from 18% in 1996 to 36% in 2003. By comparison, over the same time period, the proportion of men choosing controllable lifestyle specialties rose from 28% to 45%. The differences between women and men with respect to controllable lifestyle as an explanation for the change in preferences over time were much smaller (χ2 for differences in changes = 12, 1 df, p = .0005) than were the changes common to both sexes (χ2 for common changes = 920, 1 df, p < .0001).
The log-linear model demonstrated that controllable lifestyle alone explained 58% of the changes in specialty choice from 1996 to 2003 for women and 42% of the change for men. After controlling for income, work hours, and years of residency, controllable lifestyle explained 41% of the changes in specialty choice for women and 45% for men.
Income explained part of the changes in specialty choice for women and men but to a smaller extent than did controllable lifestyle. By itself, income explained 22% of the changes in specialty choice for women and 21% of the changes for men. After controlling for controllable lifestyle, work hours, and years of residency, income explained 1% of the changes in specialty choice for women and 4% of the changes for men.
When the four primary care specialties were eliminated from the data, controllable lifestyle remained a significant explanatory variable for the changes in specialty choice. After controlling for income, controllable lifestyle explained 33% of the changes in preference among the remaining specialties for men and women combined. This result is in comparison to the 41% and 45% values reported above for all 16 specialties (including primary care).
By all measures examined, controllable lifestyle is an increasingly important factor in the specialty choices of both women and men. The similarities in specialty choice trends between women and men were striking, and the results of the statistical analyses demonstrate that the differences between the specialty choices of women and men were very small relative to their similarities. Thus, the specialty preferences of female medical students alone were insufficient to explain the broader trend toward controllable lifestyle specialties, and alternative explanations must be sought.
One possible explanation of this finding is that the reward structure in medicine has changed. A model of physician satisfaction holds that a physician’s happiness (or utility, in economic terms) is a function of income, practice autonomy, and leisure.19 The recent growth of managed care has retarded the growth of physicians’ income19 and decreased physicians’ perceptions of professional autonomy.19,20 After decades of significant growth in physicians’ income,21 physicians’ real income (adjusted for inflation) from 1990 to 2000 grew by only 2%.22 With constraints on income and physicians’ autonomy, future physicians may be looking to lifestyle to maximize their professional satisfaction.
While changes in the medical profession may be an important factor in the trend toward specialties with controllable lifestyles, broader societal factors may have more influence. The last half century has seen a “massive movement” of women into the paid labor force23 that may be affecting the labor preferences of not only women but also men.
The large movement of women into the paid labor force has dramatically changed households and the family structure.23 A growing body of social science research suggests that one element of that change is the role of fathers.23,24 While the involvement of fathers in childrearing and housework remains limited relative to the involvement of mothers, it has increased in recent years23 as fathers are increasingly expected to be more involved in child-rearing.24
Another result of the higher participation of women in the paid labor force is an increase in time spent at work for married couples. In 1970, the average number of hours worked per week by nonfarm married couples was 52.5 hours. By 2000, the average soared 20% to 63.1 hour.25 The effect of more time spent at work is less time available for leisure or avocational pursuits, including parenting.
As a consequence of these changes in the labor force, achieving a “balanced” lifestyle between work and leisure is increasingly in demand among women and men. A survey of college students found that “achieving a balanced lifestyle (and) having a rewarding life outside of work”26, p.6 was a top career goal, and a study found individuals aged 26–40 years wanted more time to devote to life outside of work.27 A study published in 2002 found that 74% of men would rather have had a “daddy-track” job than a “fast-track” job.24 The desire for a balanced lifestyle is widespread in business,28 engineering,29 and the military.30 This same desire for a balanced or controllable lifestyle on the part of women and men may explain in part the trend toward controllable lifestyle specialties among female and male medical students.
Our results also demonstrate that while the proportion of women and men choosing specialties with controllable lifestyles grew steadily from 1996 to 2003, women chose specialties with uncontrollable lifestyles in greater proportion than did men. Certain specialties with uncontrollable lifestyles attracted women in large proportions. For example, in 2003, 70% of graduating medical students who chose pediatrics were women, as were 80% who chose obstetrics and gynecology. The factors that make pediatrics and obstetrics–gynecology more attractive to women despite their perceived uncontrollable lifestyles are uncertain and likely extend well beyond lifestyle considerations. Researchers have suggested that women have not entered other uncontrollable lifestyle specialties, such as the surgical specialties, in part due to the lack of available role models10 and perceptions of male bias.31
The overall trend toward lifestyle considerations in medicine is likely to evolve in the future. The broader forces working in medicine and society may also influence later decisions, such as which fellowships are selected, whether academics and research are pursued, and how physicians structure their practices. Additional research will be needed to evaluate these issues to help develop educational and workforce policies that address the lifestyle needs of physicians.
Our study had limitations. Chiefly, we dichotomously assigned selected specialties to controllable or uncontrollable lifestyles, which at its roots is subjective.1,3,32 The classification is an oversimplification as lifestyle among and within the specialties is highly variable (e.g., an orthopedist specializing in shoulder surgery may have a more controllable lifestyle than an anesthesiologist covering a critical care service) and does not necessarily reflect recent changes to the structure of practice that influence lifestyle (e.g., the growth of group practices and the emergence of hospitalists). Moreover, medical students’ choices are based on their perceptions of specialties and not necessarily reality. Despite these and other limitations, research does support the differentiation of specialties based on lifestyle. Schwartz and colleagues3 have shown that students choosing a specialty with a controllable lifestyle placed more emphasis on perceived lifestyle factors, such as number of work hours, perceived number of call nights, and adequate time for pursuit of avocational activities, than those who did not choose a specialty with a controllable lifestyle.
Although the classification of specialties (orthopedics is surgical and therefore uncontrollable) has a subjective component, multiple studies have validated the lifestyle designation used by Schwartz and colleagues.3 For example, recent research has noted that an medical students’ increased emphasis on lifestyle factors has led to declining interest in the surgical fields.5,33 In addition, a study of students’ career choices at the University of Florida found that the percentage of graduating students entering a surgical residency was declining and concluded that “lifestyle issues are increasingly important in student career decisions.”33,p.53 The Liaison Committee of Obstetrics and Gynecology, in response to the decreasing number of U.S. applicants, found that medical students were looking for “areas of medicine that provided a better lifestyle following graduation.”34,p.490 Lifestyle considerations have become so central to specialty choice that U.S. medical students have nicknamed “lifestyle” specialties the “E-ROAD,” an acronym for emergency medicine, radiology, ophthalmology, anesthesiology, and dermatology.34,35
Another study limitation was our reliance on GQ data to determine the specialty choices of women and men. While recent response rates have been high, the response rates in earlier years were lower. These data are thus subject to response bias and additionally may not reflect the actual specialty choices of medical students in all cases. Moreover, the data reflect the likely final choice of graduating medical students and not necessarily their first preference of specialty.
Notwithstanding these limitations, in our study we found that the recent trend toward specialties with controllable lifestyles cannot be explained by the unique preferences of women. Rather, both women and men from 1996 to 2003 increasingly chose specialties with controllable lifestyles, and the differences between their trends were very small. Changing dynamics in medicine and the increased participation of women in the paid labor force may be driving the trend toward controllable lifestyle specialties. Further research is required to investigate these relationships and their potentially broad implications for the future of medicine.
1Dorsey ER, Jarjoura D, Rutecki GW. Influence of controllable lifestyle on recent trends in specialty choice by US medical students. JAMA. 2003;290:1173–78.
2Schwartz RW, Jarecky RK, Strodel WE, Haley JV, Young B, Griffen WO Jr. Controllable lifestyle: a new factor in career choice by medical students. Acad Med. 1989;64:606–09.
3Schwartz RW, Haley JV, Williams C, et al. The controllable lifestyle factor and students’ attitudes about specialty selection. Acad Med. 1990;65:207–10.
4Jarecky RK, Schwartz RW, Haley JV, Donnelly MB. Stability of medical specialty selection at the University of Kentucky. Acad Med. 1991;66:756–61.
5Gelfand DV, Podnos YD, Wilson SE, Cooke J, Williams RA. Choosing general surgery: insights into career choices of current medical students. Arch Surg. 2002;137:941–47.
6Richtel M. Young doctors and wish lists: no weekend calls, no beepers. The New York Times 2004 January 7; A1, A19.
7Adams D. Younger doctors less dedicated, hardworking. American Medical News 2004 February 2;1, 4.
8Barzansky B, Etzel SI. Educational programs in US medical schools, 2002-2003. JAMA. 2003;290:1190–96.
9Mayer KL, Perez RV, Ho HS. Factors affecting choice of surgical residency training program. J Surg Res. 2001;98:71–75.
10Baxter N, Cohen R, McLeod R. The impact of gender on the choice of surgery as a career. Am J Surg. 1996;172:373–76.
11McMurray JE, Linzer M, Konrad TR, Douglas J, Shugerman R, Nelson K. The work lives of women physicians results from the physician work life study. The SGIM Career Satisfaction Study Group. J Gen Intern Med. 2000;15:372–80.
12Graduate medical education. JAMA. 2003;290:1234–38.
13Wassenaar J, Thran S (eds). Physician Socioeconomic Statistics. 2000–2002 edition. Chicago: American Medical Association, 2001.
15Zhang P, Thran S (eds). Physician Socioeconomic Statistics. 1999-2000 edition. Chicago: American Medical Association, 1999.
16Graduate Medical Education Directory: 2004-2005. Chicago: American Medical Association, 2004.
17Pugno PA, McPherson DS, Schmittling GT, Kahn NB, Jr. Results of the 2001 National Resident Matching Program: family practice. Fam Med. 2001;33:594–601.
18Agresti A. Categorical Data Analysis. New York: John Wiley & Sons, 1990.
19Hadley J, Mitchell JM. The growth of managed care and changes in physicians’ incomes, autonomy, and satisfaction, 1991-1997. Int J Health Care Finance Econ. 2002;2:37–50.
20Stoddard JJ, Hargraves JL, Reed M, Vratil A. Managed care, professional autonomy, and income: effects on physician career satisfaction. J Gen Intern Med. 2001;16:675–84.
21Starr P. The Social Transformation of American Medicine. New York: BasicBooks, 1982.
22Wassenaar J, Thran S (eds). Physician Socioeconomic Statistics. 2003 edition. Chicago: American Medical Association, 2003.
23Milkie MA, Bianchi SM, Mattingly MJ, Robinson JP. Gendered division of childrearing: ideals, realities, and the relationship to parental well-being. Sex Roles. 2002:21–38.
24Morman MT, Floyd K. A “changing culture of fatherhood”: effects on affectionate communication, closeness, and satisfaction in men’s relationships with their fathers and sons. West J Speech Commun. 2002;66:395–411.
25Jacobs JA, Gerson K. The Time Divide. Work, Family, and Gender Inequality. Cambridge, MA: Harvard University Press, 2004.
26Kelly A. In balance. Woman Engineer 1997 September 30; 6.
27Bond J, Galinsky E, Swanberg J. National Study of the Changing Workforce. New York: Families and Work Institute, 1998.
28Gutner T. A balancing act for Gen X women. Business Week 2002 January 21; 82.
29At issue: work-life balance. Electronic Engineering Times 1999 August 30; C3.
30Roffey AE, Wood FR, Price BR, Kallet MR. Work and family issues affecting early career decisions of the first women graduates of the U.S. Air Force Academy. Minerva: Quarterly Report on Women and the Military. 1989;7(3):25.
31Richardson HC, Redfern N. Why do women reject surgical careers? Ann R Coll Surg Engl. 2000;82(9 suppl):290–93.
32Schwartz RW, Simpson WG, Strodel WE, Jarecky RK, Griffen WO Jr, Young AB. Career change: in quest of a controllable lifestyle. J Surg Res. 1989;47:189–92.
33Lind DS, Cendan JC. Two decades of student career choice at the University of Florida: increasingly a lifestyle decision. Am Surg. 2003;69:53–55.
34Lawrence HC III. The impact of residents’ work-hour restrictions. Curr Womens Health Rep. 2003;3:487–91.