Trends in Medical School Application and Matriculation Rates Across the United States From 2001 to 2015: Implications for Health Disparities

Donglan Zhang, PhD, assistant professor , Gang Li, MD, PhD student , Lan Mu, PhD, professor , Janani Thapa, PhD, assistant professor , Yan Li, PhD, associate professor , Zhuo Chen, PhD, associate professor , Lu Shi, PhD, associate professor , Dejun Su, PhD, associate professor , Heejung Son, MS, PhD student , and Jose A. Pagan, PhD, professor

Donglan Zhang

Department of Health Policy and Management, College of Public Health, University of Georgia, Athens, Georgia

Gang Li

School of Medicine and Health Management, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, China, and Department of Health Policy and Management, College of Public Health, University of Georgia, Athens, Georgia

Lan Mu

Department of Geography, University of Georgia, Athens, Georgia

Janani Thapa

Department of Health Policy and Management, College of Public Health, University of Georgia, Athens, Georgia

Yan Li

Department of Population Health Science and Policy and Department of Obstetrics, Gynecology, and Reproductive Science, Icahn School of Medicine at Mount Sinai, New York, New York

Zhuo Chen

Department of Health Policy and Management, College of Public Health, University of Georgia, Athens, Georgia

Lu Shi

Department of Public Health Sciences, College of Behavioral, Social and Health Sciences, Clemson University, Clemson, South Carolina

Dejun Su

Center for Reducing Health Disparities, Department of Health Promotion, Social & Behavioral Health, College of Public Health, University of Nebraska Medical Center, Omaha, Nebraska

Heejung Son

Department of Health Policy and Management and Department of Epidemiology & Biostatistics, College of Public Health, University of Georgia, Athens, Georgia

Jose A. Pagan

Department of Public Health Policy and Management, School of Global Public Health, New York University, New York, New York

Donglan Zhang, Department of Health Policy and Management, College of Public Health, University of Georgia, Athens, Georgia;

The authors have informed the journal that they agree that both Donglan Zhang and Gang Li completed the intellectual and other work typical of the first author.

Correspondence should be addressed to Donglan Zhang, Department of Health Policy and Management, College of Public Health, University of Georgia, 100 Foster Rd., Wright Hall 205D, Athens, GA 30602; telephone: (706) 713-2755; ude.agu@gnahzd; Twitter: @SH_LA_AT.

Abstract

Purpose

Socioeconomic and geographic determinants of medical school application and matriculation may help explain the unequal distribution of physicians in the United States. This study describes trends in MD-granting medical school application and matriculation rates and explores the relationship between county median family income, proximity to a medical school, and medical school application and matriculation rates.

Method

Data were obtained from the Association of American Medical Colleges, including the age, gender, and Federal Information Processing Standards code for county of legal residence for each applicant and matriculant to U.S. MD-granting medical schools from 2001 through 2015. The application and matriculation rates in each county were calculated using the number of applicants and matriculants per 100,000 residents. Counties were classified into 4 groups according to the county median family income (high-income, middle-income, middle-low-income, low-income). The authors performed chi-square tests to assess trends across the study period and the association of county median family income with application and matriculation rates.

Results

There were 581,833 applicants and 262,730 (45.2%) matriculants to MD-granting medical schools between 2001 and 2015. The application rates per 100,000 residents during 2001–2005, 2006–2010, and 2011–2015 were 57.2, 62.7, and 69.0, respectively, and the corresponding matriculation rates were 27.5, 28.1, and 29.8. The ratios of the application rate in high-income counties to that in low-income counties during the 3 time periods were 1.9, 2.4, and 2.8, respectively.

Conclusions

The application and matriculation rates to MD-granting medical schools increased steadily from 2001 to 2015. Yet, applicants and matriculants disproportionately came from high-income counties. The differences in the application and matriculation rates between low-income and high-income counties grew during this period. Exploring these differences can lead to better understanding of the factors that drive geographic differences in physician access and the associated health disparities across the United States.

Recent estimates suggest that the United States will have a projected total physician shortage of between 54,100 and 139,000 physicians by 2033, in large part due to the accelerating population growth of older adults. 1 Physician shortages are already particularly pronounced in rural and remote communities across the United States. 2,3 Addressing this challenge will require not only investments in health care systems given limited resources but also new strategies to attract physicians to practice in communities characterized by relatively low levels of income and pronounced health disparities. 4,5

Previous studies have revealed that investing in residency programs in underserved communities can attract physicians to practice in those communities. 6 However, studies have also shown that more than 60% of medical students end up practicing in their birth state. 7 The number of medical school applicants and matriculants from states or counties where physicians are most needed has been a periodic source of concern. 8,9 This concern has reemerged in recent years, with high tuition being identified as a significant barrier to application for low-income students. 10–12 Understanding the geographic distribution of medical school applicants and matriculants over time may provide some insights into disparities in access to physicians and how to reduce these differences. 13–17

Medical school tuition has been growing rapidly over the past 20 years and, as such, there is likely a stronger correlation between family income and application rates now than 2 decades ago 18 as well as increased differences in the socioeconomic characteristics of college students who apply to medical school by their geographic location over this same time period.

We studied trends in the rates of application and matriculation to MD-granting medical schools across U.S. counties from 2001 through 2015. We also analyzed the association between county median family income, proximity to a medical school, and application and matriculation rates per 100,000 residents in each county. 19–22 We hypothesized that county median family income was positively associated with higher rates of medical school application and matriculation, as high-income communities have more resources, better access to high-quality public school systems, and are more likely to be in proximity to medical schools, all of which aid students pursuing a medical education.

Method

Data sources

We obtained the data for this analysis from the Association of American Medical Colleges’ (AAMC’s) Student Records System, which includes characteristics (i.e., age, gender, and Federal Information Processing Standards code for county of legal residence) of all applicants and matriculants from the entering classes at the MD-granting medical schools in the United States, including those in Washington, DC and Puerto Rico. The AAMC provided data aggregated at the county level and over 5-year intervals covering the 2001–2015 period to protect the privacy of applicants. We also obtained county-level population and median family income data from the U.S. Census Bureau and the American Community Survey 23 and matched these data using the Federal Information Processing Standards codes for all 3,220 U.S. counties in the U.S. Census Bureau’s Geodatabase files. 24 AAMC members were the 145 accredited MD-granting medical schools in the United States and Puerto Rico at the end of 2015. We geocoded the addresses of these medical schools, which we obtained from the AAMC, into the format of latitude and longitude. Our study was deemed exempt from review by the Institutional Review Board of the University of Georgia.

Measures

In our analysis, applicants refer to students who applied to at least one U.S. MD-granting medical school through either the American Medical College Application Service or the Texas Health Education Service. 25 Matriculants refer to students who applied to begin at a U.S. MD-granting medical school in a specific academic year and enrolled in that academic year. 25 Applicant acceptance ratio refers to the ratio of the number of applicants to the number of matriculants. 26 The application and matriculation rates were defined as the number of applicants and matriculants per 100,000 residents at the county level, respectively.

We estimated the average annual county population for each 5-year interval and for the total 15-year time period, using the county population each year from 2001 to 2015. Counties were grouped into 4 categories according to the interquartile range of the county median family income 27 (i.e., low-income [$0–$44,454], middle-low-income [$44,455–$51,113], middle-income [$51,114–$58,970], and high-income [≥ $58,971]). All of the aforementioned county-level data sources were merged into one database.

Statistical analysis

We described the trends in application and matriculation rates among students who applied to MD-granting medical schools in the United States from 2001 to 2015 and compared those rates across age, gender, and county income categories. We performed chi-square tests to assess the trends and the association of county median family income with application and matriculation rates. All statistical analyses were performed using R version 3.6. 28

To explore the geographic density and clustering of applicants and matriculants at the county level, we coded each county as a point feature. We imported the latitude and longitude data and the application and matriculation rates of all counties (n = 3,220) into ArcGIS 10.7 (ESRI, Redlands, California). 29,30 In addition, we used kernel density estimation to explore the geographic patterns in the data and to identify hot spot areas with high application and/or matriculation rates. Hot spots are marked with a darker color on the maps in Figures 1 and ​ and2. 2 . 31,32

Geographic distribution of application rates to MD-granting medical schools per 100,000 residents by time period across counties in the United States, 2001–2015. Hot spots with a higher application rate are marked with a darker color on the map. Data are displayed by applicant’s county of legal residence.

Geographic distribution of matriculation rates to MD-granting medical schools per 100,000 residents by time period across counties in the United States, 2001–2015. Hot spots with a higher matriculation rate are marked with a darker color on the map. Data are displayed by matriculant’s county of legal residence.

The relationship between proximity to a medical school and application rates in a given county could be due to the spatial clustering effect of high-income counties where medical schools tend to be located; as such, we used the same method to explore the geographic patterns of counties with different family income levels. By comparing the distribution of the counties with different application rates and different family income levels, we explored if geographic clustering of high-income counties exists, which would lead to the aforementioned effects. We estimated the Pearson correlation coefficient to measure the strength of the association between the Euclidean distance (length of a line segment between 2 points) to a medical school and county-level application and matriculation rates.

Results

A total of 581,833 students applied to MD-granting medical schools between 2001 and 2015; 262,730 (45.2%) of these applicants matriculated to medical school over the 15-year time period (see Table 1 ). There were 220,837 applicants in 2011–2015, a 31.4% increase from 2001 to 2005. There were 95,404 matriculants in 2011–2015, a 17.9% increase from 2001 to 2005. Overall, the numbers of applicants and matriculants per 100,000 persons were 63.7 and 28.8, respectively.

Table 1

Number and Percentage of Applicants and Matriculants to MD-Granting Medical Schools in the United States, 2001–2015

The numbers of applicants per 100,000 persons during 2001–2005, 2006–2010, and 2011–2015 were 57.2, 62.7, and 69.0, respectively. The numbers of matriculants per 100,000 persons during these same time periods were 27.5, 28.1, and 29.8, respectively. The applicant acceptance ratios during these intervals were 2.1, 2.2, and 2.3, respectively. The percentage of applicants who applied to medical school twice was 23.6%, and the percentage who applied more than twice was 3.8%.

From 2001 to 2015, the percentages of female applicants and matriculants were 48.0% and 47.9%, respectively, and the percentages of applicants and matriculants under the age of 24 were 70.1% and 76.8%, respectively (see Table 2 ). The percentages of female applicants and matriculants and applicants and matriculants under 24 years old declined slightly during the 2001–2015 time period.

Table 2

Characteristics of Applicants and Matriculants to MD-Granting Medical Schools in the United States, 2001–2015

a A total of 38 applicants and 4 matriculants had missing information for gender and were not included.

From 2001 to 2005, the median numbers of applicants per 100,000 persons in the low-income, middle-low-income, middle-income, and high-income counties were 21.2, 23.0, 30.2, and 39.7, respectively (see Table 3 ). The number of applicants per 100,000 persons in the high-income counties was 1.9 times (39.7/21.2) that of the low-income counties. Also, between 2006 and 2010, the median numbers of applicants per 100,000 persons in the same categories of counties were 18.8, 24.4, 31.0, and 45.0, respectively.

Table 3

MD-Granting Medical School Application and Matriculation Rates Per 100,000 Persons Across Counties in the United States by Income Category, 2001–2015

Abbreviations: SD, standard deviation; IQR, interquartile range.

b Ratio of application rate stands for the ratio of the median application rates between high-income counties and low-income counties.

c Ratio of matriculation rate stands for the ratio of the median matriculation rates between high-income counties and low-income counties.

The number of applicants per 100,000 persons in the high-income counties was 2.4 times (45.0/18.8) that of the low-income counties. Similarly, between 2011 and 2015, the median numbers of applicants per 100,000 persons in the 4 income category counties were 17.7, 23.4, 31.7, and 49.6, respectively. The number of applicants per 100,000 persons in the high-income counties was 2.8 times (49.6/17.7) that of the low-income counties.

In 2001–2005, the number of matriculants per 100,000 persons in the high-income counties was 2.2 times (18.6/8.5) that of the low-income counties. The corresponding ratio increased to 2.5 times (19.9/8.1) in 2006–2010. It then further increased to 3.2 times (21.3/6.7) in 2011–2015, revealing a growing gap in the number of matriculants per 100,000 persons between high-income and low-income counties (P < .001). The number and percentage of counties without any matriculants during 2001–2005, 2006–2010, and 2011–2015 were 614 (19.1%), 637 (19.8%), and 667 (20.7%), respectively, and most were low-income or middle-low-income counties.

Figures 1 and ​ and2 2 show the geographic distribution of application and matriculation rates by time period. The hot spots included New York City, New York, and the surrounding regions in the Northeastern United States; Washington, DC; Atlanta, Georgia, and the surrounding regions in the Southern United States; Frankfort, Kentucky; Sioux Falls, South Dakota, and the surrounding regions in the Midwestern United States; and Los Angeles, California, and the surrounding regions in the Western United States.

Although the distribution of application rates varied by time, the rates exhibited common features in all 3 time periods. While the application rates in the hot spot regions increased, the rates in all other regions did not change significantly during the 15-year time period, except for an increase in the application rate in the West North Central states during 2006–2010. It appears that certain areas, such as the rural regions of the West and Southwest, saw no or few applicants or matriculants over time, while there was an increasing concentration of applicants in higher-income regions over time. Furthermore, the geographic distribution of the hot spots for application rates and for matriculation rates was similar, with a greater number of matriculants in the regions with a greater number of applicants.

We also compared the geographic distribution of medical schools and that of the counties with different application and/or matriculation rates. We found that geographic proximity to a medical school was associated with higher county application and matriculation rates (Pearson correlation coefficient: r = −0.088, P < .001).

Discussion

Our study documented trends in application and matriculation rates to MD-granting medical schools in the United States by county from 2001 to 2015. Both rates increased steadily during this time period. However, differences between high-income and low-income counties also widened.

The number of medical school applicants has been fluctuating since 1974. Between 1974 and 2000, that number reached its lowest level in 1988 33 and then rose to its highest level in 1996. 34 From 1996 to 2001, the number of medical school applicants declined before starting to increase again. 35 Between 2001 and 2015, the average annual growth rate of applicants was 3.1%, higher than that of 2.7% between 2003 and 2004, the years closest to those before our study period with available data. 36 This increase is partially attributed to the addition of new medical schools. 37,38 Between 2001 and 2015, 21 new accredited medical schools opened largely as a response to concerns about a future physician shortage. 39,40

During the same period, the applicant acceptance ratio gradually increased to 2.3 between 2011 and 2015; this ratio was much higher than that of 1.7 in 1989. 41 The greater increase in the number of applicants than in the number of matriculants might explain why this ratio increased only slowly and why it was more difficult to be admitted to a medical school later than earlier in our study period.

The percentage of female applicants and matriculants decreased slightly during the study period. However, that percentage reached a historic high from 2001 to 2015 compared with what it was in the 1980s (48.0% vs 30.0%). 33,42 The distribution of applicants and matriculants by age did not change dramatically during our study period. Despite a slight decrease in the percentage of applicants ages 21–24 from 2001 to 2015 compared with the percentage in the 1990s (69.0% vs 59.0%), 41 our data showed that the percentage of matriculants in this age group was greater than that of applicants in this group in the same period.

The hot spot areas we identified showed substantial differences in application and matriculation rates across counties. 43 We found that geographic proximity to a medical school was associated with a higher rate of medical school application and matriculation at the county level. Two reasons may explain this finding. First, college students who live close to a medical school may be more likely to appreciate the value and promise of pursuing a medical degree due to their affiliation or contacts with the medical school, such as family connections, summer internship programs, or personal visits. 44–46 These channels also allow students to get information about scholarships, loan repayment programs, or other training fellowships that may lower the cost of attending medical school. 44,47 Second, college and high school students who study science or related subjects and live close to a medical school may learn from local physicians who can be role models for their careers. 48,49

Another interesting finding concerns the growing gap in medical school applicants and matriculants from low-income and high-income counties from 2001 to 2015. The rising cost of attending medical school might have deterred many students from low-income counties from even applying. At the same time, this growing difference may be related to the increasing concentration of medical schools in high-income areas 39 or to more competitive application processes, both of which might disproportionately affect the enrollment of students from low-income areas. As the literature indicates, students from lower-resourced families and regions have fewer educational opportunities (e.g., limited access to qualified public schools or local colleges), which likely hinders their pursuing higher education. 50,51 Even when they show the potential to be successful in medical education, financial constraints significantly affect their decisions to attend medical school. 52 Black and Hispanic communities are disproportionately affected by these socioeconomic factors. 10

Recently, several leading medical schools have begun offering need-based free tuition to newly enrolled students, which might alter the trends we identified in our study. 8 However, it is unlikely that a free tuition model will be widely adopted in the United States, given the high cost of medical education. 53 Without addressing existing disparities in socioeconomic status and access to public school systems by geography, physician shortages and the lack of access to physicians in rural and remote regions will persist. Simply increasing the number of medical school applicants and matriculants may not eliminate geographic differences and gaps in the physician supply in the short term.

Limitations

Our study has several limitations. First, to protect confidentiality, annual individual-level data were not available. All data were aggregated and organized into 3 time intervals. Thus, we were not able to conduct formal trend tests. Also, the latest aggregate data covering 2016 to 2020 had not been released at the time of our analysis, so our study explored trends before 2016. Second, data on other characteristics such as race/ethnicity were not available, which meant we could not conduct further, highly relevant subanalyses of application and matriculation rates. Third, individual socioeconomic status was not available, so we used county median family income as a proxy measure; as such, the socioeconomic disparities in application rates we calculated might be underestimated. 10,54 Finally, physician shortages might be partially addressed by international medical graduates, Doctor of Osteopathic Medicine program graduates, nurse practitioners, and physician assistants, all programs that could be more accessible to a broader set of applicants. Future studies should examine trends in these programs that may affect geographic differences in the supply of health care professionals.

Conclusions

Our study showed that the number of applicants and matriculants to MD-granting medical schools in the United States increased steadily between 2001 and 2015. Meanwhile, the gap in the number of applicants and matriculants from low-income and high-income counties grew during this time period. Our findings provide an important perspective on the geographic distribution of applicants and matriculants, including high application and matriculation rates in counties that were close to medical schools. By quantifying the geographic variation in medical school applicants and matriculants, we can better understand the mechanisms and factors that drive geographic differences in physician access and the associated health disparities across the United States.

Acknowledgments:

The authors wish to thank the Association of American Medical Colleges for preparing and sharing the data used in this study.

Funding/Support:

This study was funded by the National Institute on Minority Health and Health Disparities (1R01MD013886-01).

Footnotes

Other disclosures: None reported.

Ethical approval: This study was deemed exempt from review by the Institutional Review Board of the University of Georgia.

Data: Data from the Association of American Medical Colleges (AAMC) were used with permission. The AAMC reviewed a draft of this report without any input on the analysis before submission for publication. Data from the American Community Survey are publicly available with no individual identifiers.

Contributor Information

Donglan Zhang, Department of Health Policy and Management, College of Public Health, University of Georgia, Athens, Georgia.

Gang Li, School of Medicine and Health Management, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, China, and Department of Health Policy and Management, College of Public Health, University of Georgia, Athens, Georgia.

Lan Mu, Department of Geography, University of Georgia, Athens, Georgia.

Janani Thapa, Department of Health Policy and Management, College of Public Health, University of Georgia, Athens, Georgia.

Yan Li, Department of Population Health Science and Policy and Department of Obstetrics, Gynecology, and Reproductive Science, Icahn School of Medicine at Mount Sinai, New York, New York.

Zhuo Chen, Department of Health Policy and Management, College of Public Health, University of Georgia, Athens, Georgia.

Lu Shi, Department of Public Health Sciences, College of Behavioral, Social and Health Sciences, Clemson University, Clemson, South Carolina.

Dejun Su, Center for Reducing Health Disparities, Department of Health Promotion, Social & Behavioral Health, College of Public Health, University of Nebraska Medical Center, Omaha, Nebraska.

Heejung Son, Department of Health Policy and Management and Department of Epidemiology & Biostatistics, College of Public Health, University of Georgia, Athens, Georgia.

Jose A. Pagan, Department of Public Health Policy and Management, School of Global Public Health, New York University, New York, New York.

References

1. Association of American Medical Colleges. The Complexities of Physician Supply and Demand: Projections from 2018 to 2033 . https://www.aamc.org/system/files/2020-06/stratcomm-aamc-physician-workforce-projections-june-2020.pdf. Published June 2020. Accessed February 22, 2021.

2. Health Resources and Services Administration. Designated Health Professional Shortage Areas Statistics, Fourth Quarter of Fiscal Year 2021, Designated HPSA Quarterly Summary . https://data.hrsa.gov/Default/GenerateHPSAQuarterlyReport. Published December 2020. Accessed February 22, 2021.

3. Shipman SA, Wendling A, Jones KC, Kovar-Gough I, Orlowski JM, Phillips J. The decline in rural medical students: A growing gap in geographic diversity threatens the rural physician workforce . Health Aff (Millwood) . 2019; 38 :2011–2018. [PubMed] [Google Scholar]

4. Goodfellow A, Ulloa JG, Dowling PT, et al. Predictors of primary care physician practice Location in underserved urban or rural areas in the United States: A systematic literature review . Acad Med . 2016; 91 :1313–1321. [PMC free article] [PubMed] [Google Scholar]

5. Kirch DG, Petelle K. Addressing the physician shortage: The peril of ignoring demography . JAMA . 2017; 317 :1947–1948. [PubMed] [Google Scholar]

6. Rabinowitz HK, Diamond JJ, Markham FW, Wortman JR. Medical school programs to increase the rural physician supply: A systematic review and projected impact of widespread replication . Acad Med . 2008; 83 :235–243. [PubMed] [Google Scholar]

7. Owen JA, Hayden GF, Bowman RC. Influence of places of birth, medical education, and residency training on the eventual practice locations of family physicians: Recent experience in Virginia . South Med J . 2005; 98 :674–675. [PubMed] [Google Scholar]

8. Thomas B Free medical school tuition: Will it accomplish its goals? JAMA . 2019; 321 :143–144. [PubMed] [Google Scholar]

9. Lachs MS, Choi AMK. Eliminating medical school debt: A dean and geriatrician’s view from opposite ends of the training pipeline . Ann Intern Med . 2020; 172 :279–280. [PubMed] [Google Scholar]

10. Cooper RA. Impact of trends in primary, secondary, and postsecondary education on applications to medical school. II: Considerations of race, ethnicity, and income . Acad Med . 2003; 78 :864–876. [PubMed] [Google Scholar]

11. Kinslow CJ. Tuition-free versus need-based, loan-free medical school: Which will accomplish its goals? Acad Med . 2020; 95 :492. [PubMed] [Google Scholar]

12. Millett CM. How undergraduate loan debt affects application and enrollment in graduate or first professional school . J High Educ . 2003; 74 :386–427. [Google Scholar]

13. Eden J, Berwick DM, Wilensky GR; Institute of Medicine (US). Committee on the Governance and Financing of Graduate Medical Education. Graduate Medical Education that Meets the Nation’s Health Needs . Washington, DC: National Academies Press; 2014. [PubMed] [Google Scholar]

14. Cooper RA, Laud P, Dietrich CL. Current and projected workforce of nonphysician clinicians . JAMA . 1998; 280 :788–794. [PubMed] [Google Scholar]

15. Cooper R, Stoflet S. Diversity and consistency: The challenge of maintaining quality in a multidisciplinary workforce . J Health Serv Res Policy . 2004; 1 ( 9 suppl 1 ):39–47. [PubMed] [Google Scholar]

16. Kane GC, Grever MR, Kennedy JI, et al. The anticipated physician shortage: Meeting the nation’s need for physician services . Am J Med . 2009; 122 :1156–1162. [PubMed] [Google Scholar]

17. Health Resources and Services Administration. The physician workforce: Projections and research into current issues affecting supply and demand . https://bhw.hrsa.gov/sites/default/files/bureau-health-workforce/data-research/physiciansupplyissues.pdf. Published December 2008. Accessed February 22, 2021.

18. Wear D, Castellani B. The development of professionalism: Curriculum matters . Acad Med . 2000; 75 :602–611. [PubMed] [Google Scholar]

19. Luo W Using a GIS-based floating catchment method to assess areas with shortage of physicians . Health Place . 2004; 10 :1–11. [PubMed] [Google Scholar]

20. McGrail MR, Humphreys JS. Measuring spatial accessibility to primary care in rural areas: Improving the effectiveness of the two-step floating catchment area method . Appl Geogr . 2009; 29 :533–541. [Google Scholar]

21. Taggart MP, Wartman SA, Wessen AF. An analysis of medical students’ residency and specialty choices . Soc Sci Med . 1987; 25 :1063–1068. [PubMed] [Google Scholar]

22. Xierali IM, Maeshiro R, Johnson S, Arceneaux T, Fair MA. Public health and community medicine instruction and physician practice location . Am J Prev Med . 2014; 47 ( 5 suppl 3 ):S297–S300. [PubMed] [Google Scholar]

23. United States Census Bureau. American Community Survey (ACS) . https://www.census.gov/programs-surveys/acs.html. Published 2019. Accessed February 22, 2021.

24. United States Zip Codes . https://www.unitedstateszipcodes.org. Published 2019. Accessed February 22, 2021.

25. Association of American Medical Colleges. FACTS glossary . https://www.aamc.org/data-reports/students-residents/interactive-data/facts-glossary. Published 2019. Accessed February 22, 2021.

26. Barzansky B, Etzel SI. Educational programs in US medical schools, 2003-2004 . JAMA . 2004; 292 :1025–1031. [PubMed] [Google Scholar]

27. Chetty R, Stepner M, Abraham S, et al. The association between income and life expectancy in the United States, 2001-2014 . JAMA . 2016; 315 :1750–1766. [PMC free article] [PubMed] [Google Scholar]

28. R Core Team. The R Project for Statistical Computing . https://www.r-project.org. Published 2019. Accessed February 22, 2021.

29. Fischer MM, Getis A. Handbook of Applied Spatial Analysis: Software Tools, Methods and Applications . Berlin, Germany: Springer-Verlag; 2009. [Google Scholar]

30. Fischer MM, Wang J. Spatial Data Analysis: Models, Methods and Techniques . Berlin, Germany: Springer-Verlag; 2011. [Google Scholar]

31. Chainey S, Ratcliffe J. GIS and Crime Mapping . Hoboken, NJ: John Wiley & Sons; 2013. [Google Scholar]

32. Silverman BW. Density Estimation for Statistics and Data Analysis . New York, NY: Chapman and Hall; 1986. [Google Scholar]

33. Clark DS, Killian CD, Mitchell KJ. The declining applicant pool and MCAT scores . Acad Med . 1990; 65 :219–220. [PubMed] [Google Scholar]

34. Hall FR, Mikesell C, Cranston P, Julian E, Elam C. Longitudinal trends in the applicant pool for U.S. medical schools, 1974–1999 . Acad Med . 2001; 76 :829–834. [PubMed] [Google Scholar]

35. Barzansky B, Etzel SI. Educational programs in US medical schools, 2000-2001 . JAMA . 2001; 286 :1049–1055. [PubMed] [Google Scholar]

36. Barzansky B, Etzel SI. Educational programs in US medical schools, 2004-2005 . JAMA . 2005; 294 :1068–1074. [PubMed] [Google Scholar]

37. Whitcomb ME. The Development of New MD-Granting Medical Schools in the United States in the 21st Century . Acad Med . 2020; 95 :340–343. [PubMed] [Google Scholar]

38. Association of American Medical Colleges. Results of the 2017 Medical School Enrollment Survey . https://www.aamc.org/system/files/reports/1/medicalschoolenrollmentreport.pdf. Published May 2018. Accessed February 22, 2021.

39. Barzansky B, Etzel SI. Medical schools in the United States, 2018-2019 . JAMA . 2019; 322 :986–995. [PubMed] [Google Scholar]

40. Liaison Committee on Medical Education. Medical school directory . www.lcme.org/directory. Published 2020. Accessed February 22, 2021.

41. Petersdorf RG. The declining applicant pool . Trans Am Clin Climatol Assoc . 1989; 100 :39–46. [PMC free article] [PubMed] [Google Scholar]

42. Monroe A, Quinn E, Samuelson W, Dunleavy DM, Dowd KW. An overview of the medical school admission process and use of applicant data in decision making: What has changed since the 1980s? Acad Med . 2013; 88 :672–681. [PubMed] [Google Scholar]

43. Nesoff ED, Branas CC, Martins SS. The geographic distribution of fentanyl-involved overdose deaths in Cook County, Illinois . Am J Public Health . 2020; 110 :98–105. [PMC free article] [PubMed] [Google Scholar]

44. Parker PD, Jerrim J, Anders J, Astell-Burt T. Does living closer to a university increase educational attainment? A longitudinal study of aspirations, university entry, and elite university enrolment of Australian youth . J Youth Adolesc . 2016; 45 :1156–1175. [PubMed] [Google Scholar]

45. Gill T, Vidal Rodeiro C, Zanini N. Higher education choices of secondary school graduates with a science, technology, engineering or mathematics (STEM) background . J Furth High Educ . 2018; 42 :998–1014. [Google Scholar]

46. Suhonen T Field-of-study choice in higher education: Does distance matter? Spat Econ Anal . 2014; 9 :355–375. [Google Scholar]

47. Hirschl N, Smith CM. Well-placed: The geography of opportunity and high school effects on college attendance . Res High Educ . 2020; 61 :567–587. [Google Scholar]

48. Wouters A, Croiset G, Isik U, Kusurkar RA. Motivation of Dutch high school students from various backgrounds for applying to study medicine: A qualitative study . BMJ Open . 2017; 7 :e014779. [PMC free article] [PubMed] [Google Scholar]

49. Simmenroth-Nayda A, Görlich Y. Medical school admission test: Advantages for students whose parents are medical doctors? BMC Med Educ . 2015; 15 :81. [PMC free article] [PubMed] [Google Scholar]

50. Economic Policy Institute. Education inequalities at the school starting gate: Gaps, trends, and strategies to address them . https://files.eric.ed.gov/fulltext/ED588751.pdf. Published September 2017. Accessed February 22, 2021.

51. Rosenberg CE. Condition of Access: Higher Education for Lower Income Students . Lanham, MD: Rowman & Littlefield Publishers; 2005. [Google Scholar]

52. Kane TJ. The Price of Admission: Rethinking How Americans Pay for College . Washington, DC: Brookings Institute Press; 1999. [Google Scholar]

53. Asch DA, Grischkan J, Nicholson S. The cost, price, and debt of medical education . N Engl J Med . 2020; 383 :6–9. [PubMed] [Google Scholar]

54. Dugger RA, El-Sayed AM, Dogra A, Messina C, Bronson R, Galea S. The color of debt: Racial disparities in anticipated medical student debt in the United States . PLoS One . 2013; 8 :e74693. [PMC free article] [PubMed] [Google Scholar]