There are limited data for how anesthesia providers choose their first job.1 Among anesthesiology residents from Lorraine, France, plans for a first job were influenced by interest in returning to a location previously lived.2 The same applied to Student Registered Nurse Anesthetists training in Iowa.3 Hospitals in Iowa enhanced recruitment of nurse anesthetists by having the students rotate at their facilities because the students lived in their towns.3
Academic anesthesiologists’ careers may be different. For example, when the authors considered jobs after residency, they took national (e.g., FD) and international (e.g., GSD) perspectives. It is not known whether the majority (e.g., >50%) of anesthesiology residents in the United States consider working in areas where they have not lived before.
While previously performing a survey study of perceptions of anesthesiology residents of faculty supervision,4 we included questions to evaluate the predictive effect of prior residence on expectations for locations of future practices to join. Based on the information from European anesthesiology residents and our experience with student nurse anesthetists in Iowa, we hypothesized a significant impact of prior residence on the search locations for future job opportunities for surveyed U.S. anesthesiology residents.
This cross-sectional national study was approved by the Northwestern University Institutional Review Board (STU00085572).
Survey Design and Primary Hypothesis
The first part of the survey involved questions about plans for an academic practice.
“Are you a U.S. citizen or have permanent residency status (‘green card’)?” We considered further the respondents answering “Yes,” and excluded those answering “No.”
“Do you intend to enter academic practice when you graduate from residency and/or fellowship training?” The primary analysis categorized the responses into “surely yes” and “probably” versus “even,” “probably not,” and “surely no.” A secondary analysis compared the respondents answering “surely yes” to those who did not.
“After finishing your residency/fellowship training, are you planning to look seriously (e.g., interview) at jobs located more than a 2-hour drive from a location where you or your family (e.g., spouse or partner/significant other) have lived previously? Before answering, please consider ‘serious’ to imply realistic opportunities (e.g., do not consider an anesthesia group in an idyllic location that offers to pay at the 95th percentile nationwide for a 36-hour per week position with no on-call responsibilities). Consider ‘lived’ to be a period of 4 weeks or longer (e.g., include 1-month training rotations).” Responses were categorized into “very probably” and “somewhat probably” versus “somewhat improbably” and “not probable.” Two-hours was used as a time between nearby metropolitan areas (e.g., Philadelphia and New York City) (see Discussion).
For power analysis, we used the 21% of U.S. anesthesiologists in academic practice as the estimate of the percentage of residents who would be interested in academic practice.1 We used the percentage of the 84 Student Registered Nurse Anesthetists from the Midwest who trained at the University of Iowa and left the Midwest, 10.7%, as a binary independent variable. Expecting a small effect size, we compared the area under the receiver operating characteristic (AUC ROC) curve of 0.556 to random (0.50), since 0.556 corresponds approximately to a Cohen d of 0.20.5 Based on performing multiple comparisons in our secondary analyses, we considered α = 0.01. By Wilcoxon-Mann-Whitney, the total sample size of respondents was approximately 532 for 80% power and 678 for 90% power (see equation A3.3 of Reference 6).6,7 We divided by 40.5%, the percentage of 2773 U.S. anesthesiology residents who completed the final (fifth part) of a similar length survey.4 Thus, inviting 1313 achieved approximately 80% power and 1673 achieved 90% power. The planned sample size was 1500.4,a
Survey Design and Process
The mailing and e-mail lists of 4779 anesthesiology trainees were obtained from the American Society of Anesthesiologists directory available to members. A database was constructed from the list using PostgreSQL ver 9.4.1. A random sample of 1500 residents was selected (see above). The survey was created using SurveyMonkey software (SurveyMonkey Inc., Portland, OR). To assure confidentiality of the respondents, the survey was set up to delink the respondents’ e-mail addresses, but to retain the Internet protocol addresses of the respondents. The software used an internal tracking system to allow only 1 response per survey invitation and generate a list of nonresponders. The residents who did not respond to the electronic questionnaire were mailed a copy of the survey with self-addressed return envelope addressed to the primary investigator. Mailed surveys did not contain specific subject identifiers. Responses (N = 41) were entered by a research assistant (JMB) and verified for accuracy by an investigator (RJM). Residents received $10 for completion.
Survey Design and Secondary Analyses
The first part of the survey involved questions designed to capture characteristics of the respondents and future job preferences including: age, number of hours worked per week, number of residents in their class, gender, and year of training (Table 1 rows 1 to 9).
The first 2 authors (FD, GSD) considered academic jobs nationally after their residencies based on their interests in clinical research (see Introduction). The survey included 15 questions to evaluate residents’ clinical research proficiency, including 12 questions from the Shortened Version of the Clinical Research Appraisal Inventory (Table 1 rows 10 to 19).4,b
Intention to change jobs is positively correlated with burnout.8 The survey evaluated emotional exhaustion and depersonalization dimensions of job burnout: “I feel burned out from my work” and “I have become more callous toward people since I took this job” (Table 1 rows 20 to 21).9
Analysis of the primary hypothesis was confidence intervals of incidences, calculated using the Clopper-Pearson method.10,11
Secondary analyses used respondents’ levels of agreement with statements (Table 1). AUC ROC curves (i.e., WMWodds, the effect size measure for the Wilcoxon-Mann-Whitney test) were used to assess the predictive effects of multiple independent variables on binary dependent variables (e.g., academic practice or not).6,7 The point estimates ± standard errors of the WMWodds was calculated using equations A1.3 and A1.4 of Reference (6) implemented in Excel 2010. Several results were compared with logistic regression, using the binary variable as the dependent variable and the potential predictor as the independent variable, and all were the same to three digits (SYSTAT 13, SYSTAT Software Inc., Chicago, IL). To address the multiple comparisons, we both used α = 0.01 (above power analysis) and, in the tables, show the P values that are significant while maintaining a 5.0% false discovery rate.12,13
Among the 696 U.S. survey respondents, 36.9% (N = 256) would “probably consider” an academic practice and 10.6% (N = 74) would “surely consider” academic practice. For context, 21% of anesthesiologists in the United States are in academic practice1 and 23% of residents surveyed in 2013 applied for at least 1 academic job.14 Only 4.0% of all respondents (N = 28) would “surely consider” academic practice and were “very probably” willing for it to be at a distant location (i.e., greater than a 2-hour drive from a previous residence).
Ranked consideration of a distant location was no different between respondents considering or not considering academic practice (AUC 0.52 ± 0.02, P = 0.27, N = 695).
Among the 256 respondents who would probably consider academic practice, more than half (P < 0.0001) but not the vast majority (<80%, P < 0.0001) would “probably” or “very probably” consider a distant location (62.9%, 99% CI 54.7%–70.6%). Fewer than half (P < 0.0001) would “very probably” consider a distant location (31.6%, 99% CI 24.4%–39.6%).
Tables 2 and 3 consider predictive factors for academic practice at a distant location. The row in bold font shows that respondents with formal research training (e.g., PhD or Master’s) had greater interest in such practices (AUC 0.63 ± 0.03, P = 0.0002).15 There was no effect of year of training including fellowship (AUC 0.51 ± 0.04, P = 0.83).
Based on our observations of responding U.S. anesthesiology residents, when a resident inquires about a job and the resident’s program is not in the region, a question to ask is whether the potential applicant (or the applicant’s family) has previously lived in the area. Except for residents with formal research training, respondents interested in an academic position did not have greater willingness to choose a job at a distant location. Thus, academic anesthesia programs located several hours from other programs (e.g., University of Iowa) may have a substantively smaller pool of future faculty versus programs in large metropolitan areas (e.g., Northwestern University). When addressed by hiring one’s own trainees, this could result in program stagnation and reduced academic productivity.16
Our focus was the incremental effect of prior residence location on job consideration, particularly for academic practice. A previous survey evaluated the academic practice decision.17 Among U.S. anesthesiology residents, each increase in medical school debt from $0 to greater than $150,000 was associated with a significantly lesser “plan to pursue an anesthesiology career” of “academic faculty” (P = 0.002).17 Residents “with a larger amount of medical school debt were more likely to be interested in an anesthesiology group with an education debt repayment program (P < 0.001)”.17 When typical debt is much less than even a few years difference between private and academic compensation, academic practices can benefit by offering debt repayment after a resident remains working there for several years.18
We previously performed a needs assessment for business strategies of anesthesia groups’ practices.19 Porter’s Five Forces were used.20 They describe the competitive forces that shape strategy within industries.20 From prior research by others and us, there is substantial understanding for 3 of the forces: "threat of substitute services," "bargaining power of customers," and "jockeying for position among current competitors."20,21 However, there was negligible prior work investigating the “power of suppliers.” For anesthesia groups, training programs are the principal suppliers.19 A gap in knowledge was whether training programs have substantive power (of the supplier).19 Recognizing this gap, we subsequently studied nurse anesthetists. We found that one of our training programs did not have substantive "power," since the advantage was due to the nurse attending (and living within) the program’s region (i.e., the benefit accrued also to other anesthesia groups in the region) not to the program itself.19 In the current study, we found the same to hold for responding anesthesiology residents.
Response bias is always a concern in cross-sectional questionnaire studies. Therefore, we calculated 99% confidence intervals to be conservative and recommend that the incidences not be considered any more precise than these intervals. Furthermore, we limited our conclusions literally to the respondents. However, our results are extreme in that we did not ask residents whether they were planning a national job search, just whether they would consider one more than a 2-hour drive from a previous residence, and even 2 hours was sufficient to exclude approximately half the interest. Had we used a distance greater than between adjacent metropolitan areas, our results would have been even stronger. Furthermore, the importance of previous residence location, personal and/or family, matched that of the small French survey.2 Future studies could evaluate our observation that respondents with formal research training had different expectations.
Jane M. Bialek, BS, Research Assistant at Northwestern University helped setup the electronic survey and typed in the paper responses. Jennifer Espy, BFA, Administrative Services Coordinator at University of Iowa, helped edit the paper.
Dr. Franklin Dexter is the Statistical Editor and Section Editor for Economics, Education, and Policy for Anesthesia & Analgesia and A & A Case Reports. This manuscript was handled by Dr. Steven L. Shafer, Editor-in-Chief, and Dr. Dexter was not involved in any way with the editorial process or decision.
a The questionnaire included other questions to evaluate the reliability (internal consistency), convergent validity, and discriminant validity of a faculty supervision scale when applied to assessing departments.4
b Available at: http://hsc.unm.edu/research/brep/newforms/KL2/CRAI.pdf. Accessed June 15, 2014.
1. Schubert A, Eckhout GV, Ngo AL, Tremper KK, Peterson MD. Status of the anesthesia workforce in 2011: evolution during the last decade and future outlook. Anesth Analg. 2012;115:407–27
2. Chalot Y, Boileau S, Vedel M, Audibert G, Mertes MP, Bouaziz H. Course of the anaesthesiology and intensive care unit residents in Lorraine, France. Ann Fr Anesth Reanim. 2010;29:209–14
3. Wachtel RE, Dexter F. Training rotations at hospitals as a recruitment tool for Certified Registered Nurse Anesthetists. AANA J. 2012;80:S45–8
4. De Oliveira GS Jr, Dexter F, Bialek JM, McCarthy RJ. Reliability and validity of assessing subspecialty level of faculty anesthesiologists’ supervision of anesthesiology residents. Anesth Analg. 2015;120:209–13
5. Kraemer HC, Kupfer DJ. Size of treatment effects and their importance to clinical research and practice. Biol Psychiatry. 2006;59:990–6
6. Divine G, Norton HJ, Hunt R, Dienemann J. A review of analysis and sample size calculation considerations for Wilcoxon tests. Anesth Analg. 2013;117:699–710
7. Dexter F. Wilcoxon-Mann-Whitney test used for data that are not normally distributed. Anesth Analg. 2013;117:537–8
8. Liljegren M, Ekberg K. The longitudinal relationship between job mobility, perceived organizational justice, and health. BMC Public Health. 2008;8:164
9. West CP, Dyrbye LN, Satele DV, Sloan JA, Shanafelt TD. Concurrent validity of single-item measures of emotional exhaustion and depersonalization in burnout assessment. J Gen Intern Med. 2012;27:1445–52
10. Clopper CJ, Pearson ES. The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika. 1934;26:404–13
11. Hahn GJ, Meeker WQ. Statistical intervals. A guide for practitioners. 1991 New York Wiley:82–4, 100–5
12. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc B. 1995;57:289–300
13. Jones HE, Ohlssen DI, Spiegelhalter DJ. Use of the false discovery rate when comparing multiple health care providers. J Clin Epidemiol. 2008;61:232–40
14. Stein EJ, Mesrobian JR, Abouleish AE. 2013 job market for graduating anesthesiology residents. Am Soc Anesthesio Newsl. 2014;78:44–7
15. Dexter F, Epstein RH, Bayman EO, Ledolter J. Estimating surgical case durations and making comparisons among facilities: identifying facilities with lower anesthesia professional fees. Anesth Analg. 2013;116:1103–15
16. Horta H, Veloso FM, Grediaga R. Navel gazing: academic inbreeding and scientific productivity. Manag Sci. 2010;56:414–29
17. Steiner JW, Pop RB, You J, Hoang SQ, Whitten CW, Barden C, Szmuk P. Anesthesiology residents’ medical school debt influence on moonlighting activities, work environment choice, and debt repayment programs: a nationwide survey. Anesth Analg. 2012;115:170–5
18. Lubarsky DA. Understanding the impact of debt on graduating residents’ employment decisions. Anesth Analg. 2012;115:1–2
19. Scurlock C, Dexter F, Reich DL, Galati M. Needs assessment for business strategies of anesthesiology groups’ practices. Anesth Analg. 2011;113:170–4
20. Porter ME. How competitive forces shape strategy. Harvard Bus Rev. 1979;57:137–45
© 2016 International Anesthesia Research Society
21. Bayman EO, Dexter F, Laur JJ, Wachtel RE. National incidence of use of monitored anesthesia care. Anesth Analg. 2011;113:185–9