Characteristics of medical students interested in emergency medicine with intention to practice in underserved areas

Abstract

Objectives

Emergency departments serve a wide variety of racial, ethnic, socioeconomic, and gender backgrounds. It is currently unknown what characteristics of students who express interest in emergency medicine (EM) are associated with a simultaneous desire to work in medically underserved areas. We hypothesize that those who are underrepresented in medicine, are female, learn another language, and have more student debt will be more likely to practice in a medically underserved area.

Methods

Data from the National Board of Medical Examiners, Association of American Medical Colleges (AAMC) Student Record System, and the AAMC Graduation Questionnaire were collected on a national cohort of 92,013 U.S. medical students who matriculated from 2007 through 2012. Extracted variables included planned practice area, intention to practice in underserved areas, race/ethnicity, sex, medical school experiences, age at matriculation, debt at graduation, and first‐attempt USMLE Step 1 score.

Results

EM‐intending students who identified as female, non‐Hispanic Black/African American, or Latinx/Hispanic; had a larger debt at graduation; had experiences with health education in the community; had global health experience; and had learned more than one language were more likely to report an intention to practice in underserved areas.

Conclusion

With the increasing importance of physician diversity to match those of the community being served, this study identifies factors associated with a desire of EM students to work in underserved areas. Medical schools and EM residencies may wish to consider these factors in their admissions process.

INTRODUCTION

The current emergency medicine (EM) workforce does not reflect the increasing diversity of the U.S. population.¹, ² In 2018, only 4.5% of active EM physicians identified as non‐Hispanic Black or African American and only 5.3% identified as Latinx/Hispanic.¹ In fact, the number of non‐Hispanic Black/African American residents in EM fell from 5.2% in 2013 to 4.6% in 2018.³

The lack of diversity has significant implications on patient care. Physicians who are underrepresented in medicine (URiM) are more likely to practice in underserved areas, such as areas with shortages of primary care doctors or high poverty rates.⁴, ⁵, ⁶ Regardless of income, communities comprised primarily of non‐Hispanic Black/African American and Latinx/Hispanic residents are four times as likely to have physician shortages, and non‐Hispanic Black/African American and Latinx/Hispanic physicians are more likely to practice in these communities.⁵ Consequently, URiM physicians play an integral role in providing care and access in these areas.

The finding that those who are URiM are more likely to practice in underserved areas has been reported mostly among primary care physicians, and it is not known if this relationship between race and ethnicity and likelihood to practice in underserved areas holds true for physicians working in EM or for medical students intending to pursue EM. Although medical students who expressed interest to work with underserved populations are 70% more likely to plan a career in EM,⁷ little is known about practice location intentions of medical students intending to specialize in EM.

Our study investigated whether graduating students who are URiM and matching into EM are more likely to report intention to practice in underserved areas. We also examined additional factors based on prior literature that may be associated with practicing in underserved areas, such as sex and debt upon medical school graduation.⁸, ⁹, ¹⁰ This is the first study that we are aware of that examines the association between race/ethnicity and practice intentions of medical students entering EM using a national survey of graduating medical students.

METHODS

The study sample of 4324 students was derived from a national cohort of 92,013 U.S. medical students who matriculated in the academic years of 2007–2008 through 2011–2012 and graduated in 2011 through 2017 (Figure 1). Individual deidentified data were obtained from the National Board of Medical Examiners (NBME), the Association of American Medical Colleges (AAMC) Student Record System (SRS), and the AAMC Graduation Questionnaire (GQ). These records were merged across survey years prior to analysis. The following data were obtained: planned practice area, intention to practice in underserved areas, race/ethnicity, sex, participation in electives during medical school (e.g., learned another language to improve communication with patients), age at matriculation, parental level of education, total debt at graduation, scholarship awarded during medical school, degree program, and first‐attempt USMLE Step 1 score. The study was approved by the Albany Medical College Institutional Review Board.

FIGURE 1.

Final study sample size flow chart. *Missing counts for variables were not mutually exclusive

Intention to practice in underserved areas

Students who reported intention to practice in emergency medicine were categorized as EM intending students. Intention to practice in underserved areas was categorized as a dichotomous variable. Students who reported “undecided” were assigned to having no intention to practice in underserved areas.

Race/ethnicity

Race/ethnicity was self‐reported by students and categorized into the following eight ethnoracial groups: Hispanic, non‐Hispanic (NH) White, NH Black/African American, NH Asian, NH Native American/Alaska Native, NH Hawaiian Native/Other Pacific Islander, NH multiracial, and NH unknown/other. Non‐Hispanic students who reported more than one race were categorized as NH multiracial.

Medical school experiences

Graduating medical students completing the AAMC GQ were asked to select the activities they participated in during medical school on an elective or volunteer (not required) basis. Some of the experiences included participated in a community‐based research project, learned another language to improve communication with patients, volunteered in a free clinic for the underserved population, and participated in a global health experience.

Additional covariates

Total debt at medical school graduation, obtained from the AAMC GQ, was categorized into five levels: no debt, less than $100,000, $100,000 to $199,999, $200,000 to $299,999, and greater than $300,000. Degree program was categorized into four levels: MD, BA/BS‐MD, MD‐PhD, and other dual degrees that included MD‐MPH and MD‐MBA. Students who obtained a BA/BS degree prior to matriculation into medical school were categorized as MD students, while those who enrolled in a combined undergraduate and medical school program were categorized as BA/BS‐MD students. Parental level of education was used to create the binary variable of generation status. First‐generation status was assigned to those who reported “some college” or less for both parents. Continuing‐generation was assigned to those who reported “college degree” or higher for either parent. Age at matriculation was used to create a binary variable to identify students who were 23 years of age or older during matriculation. This age was chosen to align with previous research as well as to serve as a proxy to the identification of educational gap years since those who did not have educational gap years are expected to be younger than 23 years old at matriculation.¹¹

All statistical analyses were conducted using STATA 16.1 (StataCorp). Descriptive statistics were calculated for all variables, using means with standard deviations for normally distributed continuous data, medians with interquartile ranges for nonnormally distributed continuous data, and proportions for dichotomous and other categorical data. Chi‐square tests, t‐tests, and Wilcoxon rank‐sum tests were used as appropriate to compare proportions, means, and medians. Multivariable logistic regression was used to estimate the association between student characteristics and medical school experiences with students’ intention to practice in underserved areas. Because of significant (n = 541) missing data for student debt, we compared one multivariable model with student debt and another multivariate model without student debt. Statistics were reported as adjusted odds ratio (aOR) with 95% confidence intervals (CIs).

RESULTS

Of the 48,102 medical students who matriculated between academic years 2007–2008 and 2011–2012, graduated by year 2017, and answered the GQ with complete demographic and academic data (except data on student debt), 4343 (9.0%) reported an intention to pursue EM. Among these EM‐intending students, 4324 (99.6%) had complete data on intention to practice in underserved areas and were included in the final study sample (Figure 1). Of these 4324 EM‐intending students, 2859 (66.1%) reported no intention to practice in underserved areas (including students who were undecided), while 1465 (33.9%) reported intention to practice in underserved areas. Table 1 describes and compares the characteristics of EM‐intending students who did and did not intend to practice in underserved areas. Overall, 148 (3.4%) of the EM‐intending students were Black/African American, 306 (7.1%) were Hispanic, and 1547 (35.8%) were female. However, the distribution of demographic and experiential characteristics differed significantly between students intending and not intending to practice in underserved areas (Table 1). In the bivariate analyses, those intending to practice in underserved areas compared to those not intending were less likely to be non‐Hispanic White (65.0% vs. 75.7%) and more likely to be non‐Hispanic Black/African American (7.3% vs. 1.4%), Hispanic (9.6% vs. 5.8%), and non‐Hispanic multiracial (4.0% vs. 2.7%; p < 0.001). Similarly, EM‐intending students intending to practice in underserved areas were significantly more likely to be female (42.4% vs. 32.4%, p < 0.001). Compared to those not intending to practice in underserved areas, students who intended to practice in underserved areas were more likely to have reported experience in providing health education in the community (48.3% vs. 35.9%, p < 0.001), participated in a community‐based research project (32.9% vs. 21.4%, p < 0.001), participated in a global health experience (47.8% vs. 31.4%, p < 0.001), and learned another language to improve communication with patients (35.8% vs. 23.1%, p < 0.001).

TABLE 1.

Characteristics and experiences by intention to practice in underserved areas among medical students graduating from U.S. medical schools in matriculation years 2007–2008 through 2011–2012 and interested in pursuing EM (N = 4324)

Characteristic	Intention to practice in underserved areas		p‐value
	Yes	No/undecided
Number of students	1465	2859
Race/ethnicity
NHa White	953 (65.0)	2164 (75.7)
NH Black/African American	107 (7.3)	41 (1.4)
NH Asian	162 (11.1)	333 (11.7)
NH Native American/ Alaska native	4 (0.3)	3 (0.1)	<0.001
NH Hawaiian native/ other Pacific Islander	2 (0.1)	4 (0.1)
NH multiracial	58 (4.0)	76 (2.7)
NH unknown/other	38 (2.6)	73 (2.6)
Hispanic	141 (9.6)	165 (5.8)
Female	621 (42.4)	926 (32.4)	<0.001
Global health experience	701 (47.8)	899 (31.4)	<0.001
Learned another language to improve communication with patients	524 (35.8)	659 (23.1)	<0.001
Participated in a community‐based research project	482 (32.9)	613 (21.4)	<0.001
Experience in providing health education in the community	708 (48.3)	1025 (35.9)	<0.001
Experience with a free clinic for the underserved population	1181 (80.6)	2021 (70.7)	<0.001
Experience related to cultural awareness and cultural competency	1105 (75.4)	1821 (63.7)	<0.001
Experience related to health disparities	1140 (77.8)	1827 (63.9)	<0.001
Learned the proper use of the interpreter when needed	1138 (77.7)	2077 (72.6)	<0.001
Participated in educating students about careers in health professions or biological sciences	743 (50.7)	1212 (42.4)	<0.001
Worked on a research project	919 (62.7)	1821 (63.7)	0.53
Age ≥ 23 years at matriculation	996 (68.0)	1765 (61.7)	<0.001
First‐generation college graduate	192 (13.1)	335 (11.7)	0.19
Total debtb
No debt	135 (10.5)	362 (14.4)
<$100,000	165 (12.8)	378 (15.0)
$100,000–$199,999	400 (31.0)	768 (30.5)	<0.001
$200,000–$299,999	427 (33.1)	785 (31.2)
>$300,000	164 (12.7)	226 (9.0)
Acquired a scholarshipa	927 (63.4)	1634 (57.2)	<0.001
Degree program
MD	1370 (93.5)	2728 (95.4)
BA/BS‐MD	8 (0.6)	41 (1.4)	<0.001
MD‐PhD	14 (1.0)	28 (1.0)
Other dual degrees (e.g., MD‐MPH, MD‐MBA)	73 (5.0)	62 (2.2)
Pass on first‐attempt USMLE Step 1	1421 (97.0)	2811 (98.3)	<0.001

Data are reported as n (%).

Abbreviation: NH, non‐Hispanic.

^a n = 1291 for yes and 2519 for no/undecided.

^b n = 1463 for yes and 2858 for no/undecided.

Table 2 describes the results of two multivariable analyses, one excluding total student debt at graduation (model A, n = 4321) and the other including total student debt at graduation (model B, n = 3810 due to 514 observations with missing data on debt). The results of the two models are comparable and indicate that Hispanic (aOR = 1.89, 95% CI = 1.44 to 2.47), non‐Hispanic Black/African American (aOR = 5.4, 95% CI = 3.60 to 8.09), and non‐Hispanic multiracial (aOR = 1.7, 95% CI = 1.14 to 2.53) students were more likely than non‐Hispanic White students to report intention to practice in underserved areas. Males were significantly less likely than females to report intentions to practice in underserved areas (aOR = 0.75, 95% CI = 0.65 to 0.87). Additionally, we found that EM‐intending students were more likely to report an intention to practice in underserved areas if they reported the following medical school experiences: participated in a global health experience (aOR = 1.53, 95% CI = 1.32 to 1.78), learned another language to improve communication with patients (aOR: 1.66, 95% CI: 1.41 to 1.95), participated in a community‐based research project (aOR: 1.36, 95% CI: 1.16 to 1.61), provided health education in the community (aOR = 1.28, 95% CI = 1.1 to 1.49), had experience with a free clinic for underserved population (aOR: 1.23, 95% CI: 1.02 to 1.48), had experience related to health disparities (aOR = 1.41, 95% CI = 1.14 to 1.74), and participated in educating students about careers in health professions or biologic sciences (aOR = 1.23, 95% CI = 1.02 to 1.48).

TABLE 2.

Multivariable adjusted models for the effect of student characteristics and experiences on intention to practice in underserved areas among those interested in pursuing emergency medicine

Characteristics	Model Aa (N = 4321)	Model Bb (N = 3810)
Ethnoracial groups
NH White	(Reference)	(Reference)
NH Black/African American	5.26 (3.58–7.73)	5.40 (3.60–8.09)
NH Asian	1.10 (0.89–1.37)	1.14 (0.91–1.44)
NH Native American/Alaska Native	2.88 (0.60–13.77)	2.20 (0.29–16.84)
NH Hawaiian Native/other Pacific Islander	0.87 (0.15–5.13)	0.90 (0.15–5.42)
NH multiracial	1.68 (1.16–2.43)	1.70 (1.14–2.53)
NH unknown/other	1.12 (0.74–1.70)	1.25 (0.82–1.92)
Hispanic	1.70 (1.32–2.19)	1.89 (1.44–2.47)
Sex
Female	(Reference)	(Reference)
Males	0.72 (0.62–0.82)	0.75 (0.65–0.87)
Medical school experiences
Global health experience	1.57 (1.37–1.81)	1.53 (1.32–1.78)
Learned another language to improve communication with patients	1.52 (1.31–1.77)	1.66 (1.41–1.95)
Participated in a community‐based research project	1.38 (1.18–1.61)	1.36 (1.16–1.61)
Experience in providing health education in the community	1.30 (1.13–1.50)	1.28 (1.10–1.49)
Experience with a free clinic for the underserved population	1.26 (1.06–1.49)	1.23 (1.02–1.48)
Experience related to cultural awareness and cultural competency	1.03 (0.85–1.25)	1.02 (0.83–1.26)
Experience related to health disparities	1.39 (1.14–1.69)	1.41 (1.14–1.74)
Learned the proper use of the interpreter when needed	0.90 (0.76–1.08)	0.90 (0.74–1.08)
Participated in educating students about careers in health professions or biological sciences	1.10 (0.96–1.27)	1.23 (1.02–1.48)
Worked on a research project	0.83 (0.72–0.96)	0.83 (0.72–0.97)
Age at matriculation (years)
<23	(Reference)	(Reference)
≥23	0.77 (0.67–0.89)	0.78 (0.66–0.91)
Generation status
Continuing generation	(Reference)	(Reference)
First‐generation college graduate	0.98 (0.79–1.20)	0.99 (0.79–1.23)
Total debt at graduation
No debt	(Reference)
<$100,000		1.00 (0.75–1.33)
$100,000–$199,999		1.20 (0.93–1.54)
$200,000–$299,999		1.28 (1.00–1.64)
>$300,000		1.68 (1.24–2.27)
Acquired a scholarship
No	(Reference)	(Reference)
Yes	1.06 (0.92–1.22)	1.03 (0.89–1.21)
Degree program
MD	(Reference)	(Reference)
BA/BS‐MD	0.38 (0.17–0.84)	0.32 (0.13–0.78)
MD‐PhD	0.95 (0.48–1.90)	1.08 (0.54–2.18)
Other dual degrees (e.g., MD‐MPH, MD‐MBA)	1.97 (1.36–2.85)	1.97 (1.35–2.87)
First‐attempt USMLE Step 1 result
Fail	(Reference)	(Reference)
Pass	0.63 (0.40–0.98)	0.76 (0.47–1.24)

Data are reported as adjusted OR (95% CI).

Abbreviation: NH, non‐Hispanic.

^a Model A: does not include total debt at graduation; three observations with missing data for scholarship.

^b Model B: includes total debt at graduation; 514 observations with missing data on debt.

EM‐intending students who were 23 years of age or older upon matriculation were 0.78 (95% CI = 0.66 to 0.91) times less likely than younger students to report intention to practice in underserved areas. In addition, compared to EM‐intending students with no debt, those with total debt at graduation of $300,000 or more were up to 1.68 (95% CI = 1.24 to 2.27) times more likely to report intention to practice in underserved areas. Compared to traditional MD students, BA/BS‐MD students were less likely to report intentions to practice in underserved areas (aOR = 0.32, 95% CI = 0.13 to 0.78) and other dual degree EM students (e.g. MD‐MPH, MD‐MBA) were more likely to report intention to practice in underserved areas (aOR = 1.97, 95% CI = 1.35 to 2.87).

DISCUSSION

This study is the first that we are aware of that examines factors associated with EM‐intending students who intend to practice in underserved areas. The major finding of this study is that EM‐intending students who were non‐Hispanic Black/African American, Latinx/Hispanic, or female or who had more than $300,000 in debt upon medical school graduation were more likely to report intention to practice in underserved areas. Key medical school experiences such as learning another language to improve patient communication, participating in global health experiences, and participating in community health education experiences were also associated with intention to practice in underserved areas.

Our findings are consistent with the current literature, which shows that URiM physicians are more likely to practice in underserved areas.⁵, ⁶ However, previous studies have not specified whether this applied to all fields of medicine and most studies have focused on primary care.⁵ In the studies that have investigated different specialties, EM was grouped with other facility‐based specialties, such as anesthesia and radiology,⁶ and it is unclear if the findings would hold true if the specialties had been disaggregated. In addition, studies that have investigated EM and practice locations have only sampled practicing physicians. Our study is unique in that it examines EM intending students with intention to practice in underserved areas.

Consistent with previous literature,⁸, ⁹ we found that female EM‐intending students were more likely to report intention to practice in underserved areas. A previous study found that female medical students regardless of class had more favorable attitudes toward practicing in underserved areas compared to their male peers and that these attitudes remained consistent when comparing these students in their first and fourth year of medical school.⁹ However, these findings were limited to sex and there was no report as to which specialty the students were intending to pursue. The reason why female students are more likely to report intention to practice in underserved areas compared to their male counterparts is still a topic of investigation.

Our study found that students who had certain medical school experiences, such as learning an additional language to improve patient communication, global health experience, or community health education experience, were more likely to report intention to practice in underserved areas. A 2014 study found that learning another language and having experience in the community were associated with positive intention to practice in underserved areas, with students’ intention being significantly associated at matriculation and graduation.⁸ Another study suggested that the association between these educational experiences and intention to practice in underserved areas may be due to increasing cultural competency.⁸ However, like the previous characteristics discussed, these previous studies apply to a broader medical graduate cohort and did not specifically investigate EM‐intending students.

EM‐intending students with over $300,000 in debt were more likely to report intention to serve in underserved areas. This is consistent with previous literature, which shows that medical students with high debt loads have a stronger association with intention to practice in underserved areas.¹⁰ In this study, students with higher debt loads, which was divided in quartiles, also reported intention to use loan repayment programs. This suggests that the option to use loan repayment programs, many of which exist when practicing in underserved areas, may be why students who have more debt are more likely to practice in underserved areas.

Implications

While emergency physicians often work with a diverse population of patients, such diversity has not been reflected among practicing emergency physicians.¹, ² EM is unique in that its clinicians care for patients of all walks of life regardless of ability to pay. Nearly half of all medical care across the United States occurs in emergency departments (ED),¹² and patients from underserved areas report higher level of trust if their physicians identify as URiM.³ In underserved areas, non‐White racial or ethnic minorities are more likely to use the ED for their health care needs due to lack of access to care.¹² URiM physicians are more likely to practice in underserved areas⁵, ⁶ and are therefore critical for providing care for these populations. The use of the ED for acute care needs in underserved areas and the increased trust that patients have in providers that have underrepresented identities further highlights the need to increase racial/ethnic diversity in EM and the imperative to increase the social competence of EM physicians in underserved areas. By examining characteristics of those medical students who intend to pursue EM in an underserved area, such characteristics may be given additional consideration from medical schools and residencies to increase diversity in the EM workforce and more closely align the demographic makeup of emergency physicians with those of the patient population they serve.

ACGME recently published new common program requirements¹³ that mandate a focus on recruitment and retention of a diverse and inclusive workforce. Racial and ethnic minorities in the United States experience various health disparities,¹⁴ and our EDs are places where racial and ethnic minorities experience disparities in care.¹⁵, ¹⁶, ¹⁷ Increasing diversity in the physician workforce has been proposed as a means to help eliminate health care disparities.², ¹⁸ Therefore, in addition to complying with the ACGME accreditation standards, results from our study suggest that recruiting racially/ethnically diverse medical students into EM could have significant implications for enhancing acute care, especially in underserved areas in which racial and ethnic minorities use the ED more frequently.¹²

Further recruitment efforts can be targeted to increase the number of underrepresented minorities in EM and ultimately in underserved areas. The Council of Emergency Residency Directors (CORD) suggested seven recruitment strategies.¹⁹ EM programs that have implemented at least two of the CORD practices have more diverse faculty and are more likely to have a higher racial/ethnic diversity in their residency.²⁰ One of CORD’s recommendations includes broadening selection criteria to include leadership, community service, and other life experiences.¹⁹ In addition to selecting the best students based on medical school grades, standardized test scores, and soft skills demonstrated during interviews, EM residencies may wish to consider a more holistic approach, incorporating characteristics reported in our study such as the key medical experiences that were associated with increased likelihood of intention to practice in underserved areas.

LIMITATIONS

Our study has several limitations. First, our data is from students graduating from 2011 up to 2017. However, studies as recent as 2019 suggest that there is still plenty of work to be done to increase the diversity in EM as evidenced with the declining number of non‐Hispanic Black/African American EM residents.³ Second, our data do not include students from doctor of osteopathic medicine (DO) schools, which make up approximately 16% of all residents who matched in 2020. Data from DO schools are not collected by the AAMC and investigating whether our findings are applicable to DO students is an area of interest to pursue. Finally, the intent to practice in an underserved area upon matching into EM was used as a proxy to assess ultimate practice location. Nevertheless, prior studies have shown that 49% of students who report an intention to practice in underserved ultimately provide care in these areas.²¹

CONCLUSION

Underrepresented racial and ethnic minorities in medicine are more likely to practice in underserved areas. Our key finding that this phenomenon is applicable to students intending to specialize in emergency medicine has significant implications for the care of racial and ethnic minority patients who, due to lack of health care access, disproportionately use the ED for their health care needs. As such, increasing diversity in the emergency medicine workforce is a social justice issue and a way to potentially increase health care access, culturally competent care, and overall health care quality.

CONFLICT OF INTEREST

The authors have no potential conflicts to disclose.

AUTHOR CONTRIBUTIONS

Carina Abreu contributed to the interpretation of the data, drafting of the manuscript, and critical revision of the manuscript for important intellectual content. John Jawiche contributed to the drafting of the manuscript, analysis and interpretation of the data, and statistical expertise. Mytien Nguyen contributed to the study concept and design, analysis and interpretation of the data, statistical expertise, and critical revision of the manuscript for important intellectual content. Andrew K. Chang contributed to the critical revision of the manuscript for important intellectual content. Ashar Ata contributed to statistical expertise and critical revision of the manuscript for important intellectual content. Symone Reid contributed to the interpretation of the data and the drafting of the manuscript. Hyacinth R. C. Mason contributed to the study and design, acquisition of the data, and critical vision of the manuscript for important intellectual content. Daniel Rebagliati contributed to the interpretation of the data and drafting of the manuscript. Joy M. Myers contributed to the interpretation of the data and drafting of the manuscript. Dorcas Pinto contributed to the critical revision of the manuscript for important intellectual content. Donna Jeffe contributed to statistical expertise. Dowin Boatright contributed critical revision of the manuscript for important intellectual content.

Abreu C, Jawiche J, Nguyen M, et al. Characteristics of medical students interested in emergency medicine with intention to practice in underserved areas. AEM Educ Train. 2021;5(Suppl. 1):S65–S72. 10.1002/aet2.10672