Monroe, Alicia MD; Quinn, Erin PhD, MEd; Samuelson, Wayne MD; Dunleavy, Dana M. PhD; Dowd, Keith W. MA
In recent years, academic medicine has been shifting from a hierarchical, autonomous, and expert-centered model to one that is increasingly team-based, collaborative, and patient-centered. Rapid scientific and technological advances are also leading to changes in academic medicine, as are changes in the demographic composition of the United States and Canada.1,2 Additionally, health care reform efforts are under way in the United States, and there have been calls for increases in the number and diversity of health care practitioners in the workforce.
These changes signal that it is necessary to review the medical school admission process. We suggest that they have a cyclical relationship with the admission process: Changes in academic medicine affect the admission process through their influence on the applicant pool, the legal and social contexts in which admission decisions are made, and the medical school curriculum. At the same time, the future physician workforce is shaped by the admission process and medical training.
Although researchers have spent considerable time investigating the emergent culture in academic medicine, relatively little is known about the medical school admission process of the early 21st century. In fact, a description of the typical admission process has not been published in more than 20 years.3,4 To better understand the current process and how it differs from that described over two decades ago, in 2008 we surveyed admission deans about their medical schools’ admission policies and practices.In this article, we begin by contrasting the applicant pool, legal and social contexts, and medical school curriculum at the time of our survey with those in 1986, when the Association of American Medical Colleges (AAMC) conducted a previous survey of admission deans that also focused on medical school admission practices. We then describe current admission processes and compare our findings, when possible, with those of the 1986 survey as reported by Mitchell.4
Changes in the Applicant Pool
Since the mid-1980s, the number and composition of medical school applicants have changed dramatically. At that time, there were concerns about a declining applicant pool and a potential decline in the academic quality of applicants.5 Additionally, the percentages of minority and female applicants were relatively low.6 By 2008, the applicant pool had grown and become more diverse with respect to Asian and female applicants.*
In 1986, among the 32,886 individuals who applied to at least one U.S. medical school, there were 2,303 (7%) black or African American, 1,972 (6%) Hispanic or Latino, 2,581 (8%) Asian or Pacific Islander, 125 (<1%) Native American, and 25,250 (77%) white applicants.7 In 2008, the number of applicants increased to 42,231 and included 3,342 (8%) black or African American, 3,086 (7%) Hispanic or Latino, 9,317 (22%) Asian or Pacific Islander, 400 (<1%) American Indian or Alaska Native, and 26,601 (63%) white individuals.8
The number and percentage of women in the applicant pool increased from 11,558 (35%) in 1986 to 20,360 (48%) in 2008.9,10 The majority of the growth among female applicants was from an increase in the percentage of female Asian or Pacific Islander applicants, which nearly tripled from 8% (n = 981) to 22% (n = 4,469); conversely, the percentage of white female applicants decreased, and the percentages of black and Hispanic or Latino female applicants remained stable.7–10
We suggest that these changes in the size and composition of the applicant pool may have affected the admission process in several ways when compared with that of the mid-1980s. Given the increase in applicants, admission committees—especially those with large applicant pools—may add stages to the process to reduce the number of applicants remaining at each stage. Second, with more applicants in the pool, admission officers may rely more heavily on data that are quantifiable and easily incorporated into pre-interview screening tools. Third, in light of the changes in composition of the applicant pool, admission committees may use a combination of quantitative and qualitative (academic and nonacademic) data in order to achieve broad diversity in the student body.
Changes in the Legal and Social Contexts
Increasing the diversity of the workforce and the numbers of minorities and women in medicine have been long-standing concerns in medical education.11 Even so, the legal and social contexts of the mid-1980s limited the consideration of age, race/ethnicity, and gender in admission decisions12 and thus limited admission committees’ options.
The legal context in which admission committees operate has changed substantially, however. For example, the Supreme Court’s 2003 decision in Grutter v. Bollinger13 affirmed the importance of mission-driven, evidence-based admission decisions and introduced the concept of educational benefits of diversity.† It also established that all applicants must be considered through the same admission process, which allowed schools to change their diversity and admission policies. In 1986, only 28% of U.S. medical schools included diversity as a primary goal of their admission process,‡ whereas 57% did in 2008.14,15
In addition, changes in the social context enabled the AAMC to introduce several projects aimed, in part, at improving diversity. For example, Project 3000 by 2000 and the Aspiring Docs program—introduced in 1991 and 2006, respectively—were designed to increase the number of historically underrepresented minority (URM) students enrolled in medical school. The Holistic Review Project was introduced in 2002 with the aim of improving diversity by encouraging admission committees to evaluate applicants’ nonacademic characteristics in addition to their academic achievements. In recent years, the admission community has expanded its definition of diversity to include “personal attributes, experiential factors, demographics, or other considerations”16 such as socioeconomic status (SES) and rural background.
We suggest that, together, these legal and social context changes may have affected the admission process in at least two ways. First, admission committees may now consider more and varied information about applicants in making admission decisions than they did in the mid-1980s. Second, with a slightly more diverse applicant pool and a more permissive legal environment, admission committees may now be more likely to consider information about applicants’ race/ethnicity, gender, and/or SES background in the admission process.
Changes in the Medical School Curriculum
Medical educators have long expressed concern that the medical school curriculum places too much emphasis on the natural sciences at the expense of the psychosocial, humanistic, and professional aspects of medicine.17 In the mid-1980s, most medical schools’ curricula were organized into distinct basic and clinical science years. In general, the first two years were lecture-based and focused on the natural sciences, whereas the third and fourth years focused on the clinical sciences and patient interactions.
In the 1990s and 2000s, a series of structural modifications to the medical school and residency accreditation processes, as well as new curricular resources, paved the way for fundamental changes in medical education. For example, the Liaison Committee on Medical Education18 (LCME) and the Accreditation Council for Graduate Medical Education19 revised their accreditation standards to require medical schools and residency programs to teach and assess professional attributes. LCME standard MS-31-A states, “A medical education program must ensure that its learning environment promotes the development of explicit and appropriate professional attributes in its medical students (attitudes, behaviors, and identity).”18 In addition, the AAMC Medical School Objectives Project series20 and the Institute of Medicine (IOM) report on behavioral and social sciences in medical school curricula21 identified—and, importantly, provided curriculum materials to help medical schools modify their curricula to teach—the broad knowledge, skills, and attitudes that graduating medical students should possess.
These formal accreditation requirements and the availability of the aforementioned AAMC and IOM resources may have enabled medical schools to revise their curricula in fundamental ways. For example, many schools now include both natural and clinical sciences instruction in the first two years and/or offer problem-based rather than lecture-based courses. A growing number also offer behavioral and social sciences, humanities, and professionalism courses.
We suggest that, given these changes in the structure and focus of the medical school curriculum, today’s applicants may be required to demonstrate some new knowledge, skills, and attitudes compared with applicants in the mid-1980s. As such, admission committees may now place more emphasis on different types of information about applicants—such as personal attributes like teamwork, communication skills, compassion, empathy, and integrity—than they did in the past.
In 2008, the committee charged with the fifth comprehensive review of the Medical College Admission Test (MR5 Committee) conducted a two-part study to explore current medical school admission policies and practices. The first part of the study consisted of site visits to medical schools in the United States and Canada; the second part was a survey of admission officials in all MD-granting U.S. medical schools and the Canadian schools that use the MCAT exam.
Medical school site visits
To gather information about current admission policies and practices, in spring 2008 MCAT staff visited eight MD-granting medical schools in the United States and Canada that were selected to be representative of AAMC’s member schools with respect to geographic location and mission/educational philosophy. The site visits were made by seven MCAT staff, who conducted eight interviews with individuals and/or small groups at each site. The interview participants were staff and faculty involved in the admission process (including the dean or associate dean of admissions, chair or co-chair of the admission committee, admission committee members, admission staff, student affairs and academic affairs officials, diversity and cultural affairs officials, the vice dean for education, and office of academic enhancement/counseling/academic coordinator officials) and medical students.
MCAT staff transcribed and reviewed the 30- to 90-minute interviews. The resulting qualitative data (not reported here) were used to inform the development of the survey (i.e., wording of questions and response options).
Survey of medical school admission officials
In spring 2008, admission deans from all MD-granting U.S. medical schools and the Canadian medical schools that use the MCAT exam (n = 142) were invited via e-mail to participate in an online survey regarding admission processes. Three reminder e-mails were sent to nonrespondents at one-week intervals after the survey opened. The survey took approximately 60 minutes to complete and included 69 questions, which were divided into three sections: description of the admission process; use of undergraduate grade point average (UGPA), MCAT score, and other applicant data at each stage of the process; and the importance of such data to admission decisions. Ratings of importance were made using a five-point scale, ranging from 1 = not important to 5 = extremely important. The survey items relevant to this study, with response data, are provided in Supplemental Digital Appendix 1, available at http://links.lww.com/ACADMED/A124.
Demographic and institutional data were not collected on the survey. We drew data on participating schools’ institutional characteristics (e.g., public or private, location) from the AAMC’s Data Warehouse in March 2012 and linked these data to survey responses. Responses were confidential; all identifying information was removed after survey and institutional data were matched, prior to data analysis.
In addition, we drew MCAT score and UGPA data from the AAMC’s Data Warehouse for 2008–2010 applicants who were offered acceptance by one or more MD-granting U.S. medical schools. We computed the percentages of accepted applicants in various MCAT total score and UGPA categories (e.g., MCAT total score = 27–29 and UGPA = 3.20–3.39). We drew data for 2008–2010 to correspond with the period in which the survey was conducted.
Data were analyzed with IBM SPSS Statistics.22 We computed means, standard deviations, and frequencies.
This study was reviewed by the AAMC Human Subjects Research Protection Program and determined to be exempt. It also underwent AAMC Data Collection and Instrument Clearance review to ensure that the survey instrument and methodology complied with AAMC policies and procedures.
Of the 142 admission deans invited to participate in the survey, 129 deans or their designees responded (response rate = 90%). For these analyses, we excluded 9 responses because of incomplete data (i.e., missing data for more than 80% of the questions). The final sample for these analyses was 120 medical school admission officers representing 77 (64%) public and 43 (36%) private institutions, a distribution mirroring that of AAMC member schools. With regard to location, 6 (5%) of the responses were from Canadian schools that use the MCAT exam, whereas the rest were from schools in the United States: 42 (35%) from the southern region, 30 (25%) from the northern region, 27 (23%) from the central region, and 15 (13%) from the western region. There were no differences between responding and nonresponding schools with respect to public/private status or region of the United States or Canada.
The admission process in 2008
Slightly more than half (57%, 68/120) of the respondents reported that their medical school’s admission decisions are made using a two-stage process that includes an initial application and an interview. Admission staff and/or a subset of admission committee members first review application data to select the interview pool. This review is often formulaic, but some schools conduct a holistic review of every application to determine who will be interviewed. After the interviews are completed, the admission committee meets to decide which applicants will be offered acceptance. In general, these meetings consist of a formal presentation of each applicant’s background and qualifications, a discussion of ratings and/or content from the interview, and a formal vote to accept, reject, hold, or refer to a special program.
Slightly less than half (43%, 52/120) of the respondents reported that their schools use a three-stage process to make admission decisions, the same as reported by Mitchell4 in 1986 (43%, 49/113). The three-stage approach includes an initial review of application data by admission staff to select applicants to invite to submit secondary applications. Then, admission staff and/or committee members review data from both the initial and secondary applications to select interviewees. Finally, the admission committee meets to review and discuss all applicant information (i.e., data from both of the applications and the interview) and decides which applicants will be offered acceptance. Of the 52 schools structuring their admission process in this manner, 42 (81%) were public and 10 (19%) were private.
Overall, admission officers rated a wide range of data as important to admission committees’ decisions about which applicants to invite to submit secondary applications, interview, and accept into medical school (Table 1). However, the uses and importance of these data differed by the stage of the process. For example, respondents rated MCAT scores and UGPAs as the most important types of data for deciding which applicants to invite to submit secondary applications and to interview, but these data were less important in deciding whom to admit. Admission officers rated two types of nonacademic data—interview recommendations and letters of recommendation—as the most important in deciding which applicants to accept. There was substantial variation across schools, though, suggesting that admission committees tailor their processes to match their schools’ educational missions and goals.
Table 2 compares the relative importance of 15 types of data in making acceptance decisions in 1986 and 2008. Among the most important types of data, 64% (7/11) were nonacademic in 2008 compared with 50% (5/10) in 1986,4 suggesting that nonacademic data are more important to admission decisions today than in the past. In 2008, admission officers rated interview recommendations and letters of recommendation as the most important, whereas they rated cumulative GPA and science and math GPA as the most important in 1986. Furthermore, personal statements and community service in medical/clinical settings gained in importance in 2008 compared with 1986. The rated importance of MCAT scores was about the same in 1986 and 2008. Demographic characteristics were among the top 15 types of data as rated in 1986, whereas race and gender were among the least important variables as rated in 2008.
The admission interview in 2008
All responding admission officers reported that their medical schools conduct admission interviews. Almost two-thirds (64%, 77/120) reported that admission committee members screen application materials to decide whom to interview; more than half (56%, 67/120) indicated that admission staff are also involved in the screening process. Only 12% (14/120) reported that their schools use computer-based algorithms to make this decision. The majority (69%, 78/113) indicated that two or more people screen each applicant’s information, and 53% (60/114) reported that this review takes 15 minutes or longer.
Interviews were described as one-on-one by 83% (99/119) of the responding admission officers. Most respondents (87%, 104/120) reported that interviews are conducted by admission committee members, whereas 17% (20/120) reported that they are conducted by staff and 68% (81/120) indicated that, in some cases, they are conducted by medical students. (Percentages may exceed 100% because respondents could select multiple types of interviewers.) Many respondents (59%, 71/120) indicated that their schools conduct two interviews with each interviewee. Over 50% of respondents (more than 65 of 119) reported that interviewers are allowed to review personal statements, letters of evaluation/recommendation, MCAT scores, or UGPAs before or during interviews; however, 13% (16/119) indicated that interviewers are not allowed to do so before the interview. Admission interviews were described as typically lasting 30 to 44 minutes each.
Results showed that the admission interview is somewhat structured. The majority of respondents (65%, 77/119) indicated that interviewers are given general guidance about the content of the questions they should ask. Many medical schools use a standard rating process to evaluate applicants during the interview: For example, over 50% of respondents (more than 60 of 119) reported that interviewers use a numeric rating scale to assess applicants on multiple dimensions or on overall interview performance. The interview process and format are nearly identical for faculty-, staff-, and student-led interviews, but fewer data about applicants are made available to student interviewers than to faculty and staff interviewers.
Admission officials indicated that interviews are most often used to assess nonacademic characteristics and skills: Over 85% (more than 100 of 119) reported that interviews include questions about applicants’ motivation for pursuing a medical career, compassion and empathy, personal maturity, oral communication skills, service orientation, and professionalism. Less than 20% (fewer than 22 of 113) reported including questions about applicants’ academic content knowledge (e.g., biology, chemistry, psychology) in interviews.
The role of UGPAs and MCAT scores in 2008
Medical school admission processes typically use MCAT scores and UGPAs to identify the most and least capable applicants and to provide interpretive context for one another. As Figure 1 shows, less than half of respondents indicated that their schools use these data to identify applicants who may need additional academic support or nonnative English speakers who have adequate reading comprehension skills.
As Figure 2 shows, most respondents indicated that their schools use MCAT scores to predict performance in the basic sciences (68%, 82/120) or on the United States Medical Licensing Examination (USMLE) Step 1 exam (77%, 92/120). The majority of respondents also reported use of UGPAs to predict performance in the basic sciences. Less than 33% (fewer than 35 of 120) indicated that their schools use MCAT scores and UGPAs to predict USMLE Step 2 and Step 3 exam performance, clinical clerkship performance, academic distinction, and time to graduation.
Chart 1 shows that although UGPAs and MCAT scores are important factors in admission processes, they are not the sole determinants of acceptance decisions. For example, 105 (9%) of the 1,233 applicants with UGPAs of 3.80 to 4.00 and MCAT total scores of 39 to 45 were not accepted by any of the medical schools to which they applied in 2008–2010. In contrast, 597 (18%) of the 3,324 applicants with UGPAs of 3.20 to 3.39 and MCAT scores of 24 to 26 were accepted by at least one medical school. These findings buttress the importance ratings data presented earlier, suggesting that a wide variety of data are important to admission decisions.
Our findings suggest that several aspects of the medical school admission process remain unchanged since the mid-1980s, whereas others have changed in fundamental ways. Some of the changes may be explained by differences in the applicant pool, legal and social contexts, and medical school curricula. Many admission committees now seem to use a holistic admission process to identify applicants who best fit their schools’ educational missions and goals. Below, we contrast our findings with the 1986 findings reported by Mitchell.4
Aspects of the admission process that have not changed
Certain aspects of the admission process are largely unchanged since the mid-1980s. First, as in 1986,4 admission officers today use a variety of data in making decisions, which suggests that they remain committed to evaluating both academic and nonacademic information. Second, our data suggest that, as in 1986, schools’ admission processes are structured into two or three stages. Third, in both 1986 and 2008, admission officers rated MCAT scores as important to each stage in the process and indicated that they use MCAT scores and UGPAs to provide an interpretive context for each another. Fourth, the number, length, and format of admission interviews are the same as those described by Johnson and Edwards3 in 1991. Similarly, the admission interview continues to be the primary source of information about applicants’ personal characteristics.
Finally, as in the 1986 survey,4 admission officers in our survey rated the importance of demographic characteristics in the admission process as relatively low. We found this to be somewhat surprising given changes in the legal and social contexts that now allow admission committees to consider diversity in the context of their schools’ educational missions and goals. We suggest, on the basis of the data presented in this article and comments made by focus group participants during site visits, that some medical school admission committees may feel that conducting holistic reviews allows URM and rural applicants to show their full potential and precludes the need to consider demographic variables explicitly.
Aspects of the admission process that have changed
As was the case in the mid-1980s, most admission committees use a multistage process to make decisions. However, our data suggest that admission committees now place different emphasis on applicant data at each stage of the process. For example, in summarizing the results of the 1986 survey, Mitchell4 noted that although test scores decreased in importance as decision making proceeded, importance ratings did not differ appreciably across the stages of the admission process. In contrast, our data suggest that admission committees now consider slightly different data when deciding whom to invite to submit secondary applications, interview, and accept. Academic data seem to be slightly more important in deciding which applicants to invite to submit secondary applications and to interview than in deciding whom to accept.
This difference is likely due to the increasing size of applicant pools and the ease of incorporating academic data into automated screening processes. As applicant pools become larger, schools may elect to reduce the number of applications for review at each stage in the process by screening applicants on a subset of admission data. Incorporating computers into the screening process may exacerbate the emphasis placed on quantifiable data because applicants’ data can be entered, sorted, and compared easily. In addition, the emphasis that many medical schools place on their U.S. News and World Report rankings23,24 may also explain why admission officials rated MCAT scores and UGPAs as highly important. However, we interpret the 2008 survey data reported in this article—and the qualitative data from the 2008 medical school site visits—as indicating that the inclusion of multiple stages does not preclude the use of a holistic admission process.
Arguably, the most notable change in the admission process is the increased importance placed on nonacademic data in making acceptance decisions. In 2008, admissions officers rated more nonacademic data as “of high importance” than did admission officers in 1986. Further, all types of academic data dropped in ratings of relative importance in the 2008 survey compared with the 1986 survey (except MCAT scores), whereas nonacademic data such as interview recommendations, letters of recommendation, and personal statements gained in importance.
Acceptance rate data provide additional support for this change, showing that applicants with different levels of academic preparedness (i.e., the combination of cumulative UGPA and MCAT total score) are accepted into medical school. Together, these data and the high importance ratings given to both academic and nonacademic data in the 2008 survey suggest that many medical schools are conducting a more holistic admissions process than they have in the past. This may be due to changes in the legal and social contexts that allow schools to consider demographic information as part of their admissions process; formal training programs such as the AAMC’s Holistic Review Training, which helps schools tailor their admission process to their educational missions and goals16; and changes in the medical school curriculum, health care systems, and the practice of medicine that may require matriculants to possess different knowledge, skills, and attitudes to be successful in medical school today.25 Additionally, grade inflation occurs at many undergraduate institutions.26 It is unclear whether knowledge of such inflation affects admission officers’ use of UGPAs in their decision making; however, we suggest that it may contribute to the emphasis placed on nonacademic factors, where there is more variance among applicants.
Limitations and future directions
Our findings underscore the complexity of the medical school admission process as well as the admission community’s desire to consider the whole applicant when making admission decisions.
This study is not without limitations, however. Despite the high survey response rate, data were not collected from all MD-granting U.S. medical schools and all Canadian schools that use the MCAT exam. Similarly, findings are limited to the questions included in the survey, and admission deans (or their designees) were selected as the only type of respondents. Therefore, the study may not have captured all aspects of the medical school admission process. Gathering data from a larger number of medical schools and expanding the response base to include other admission staff would increase the generalizability of these findings. In addition, the results reported here were informed only by quantitative data, so it is difficult to infer meaning absent from the context from which the data were derived. Finally, this study was conducted in 2008 and may not precisely reflect the medical school admission process in 2013.
Findings from this study will be used to inform AAMC initiatives focused on transforming medical school admissions. Future research should investigate how admission committees learn about applicants’ nonacademic characteristics and whether the emphasis that committees place on such information varies by institutional characteristics (e.g., public versus private status, educational mission, size of the applicant pool). Future research should also explore whether positive and negative data about applicants are “weighted” differently. For example, do unfavorable letters of recommendation carry more “weight” in admission decisions than do favorable ones?
Future research should also explore the relationship between admission variables and the changing demographics of the applicant pool and matriculating classes. During a time when diversity has become an important social value, why have the percentages of Asian and female applicants increased, whereas the percentages of black and Latino applicants have remained relatively stable since the mid-1980s? Future research should explore the factors affecting who applies to medical school, how those factors affect what admission officers consider when making admission decisions, and what effects those factors have on the demographic composition of matriculating classes.
It would be interesting to replicate this study on a regular basis (perhaps every 5 or 10 years) to identify trends and changes in the admission process. For example, if more schools adopt a holistic admission process, will nonacademic factors become more important than academic factors in pre-interview screening decisions? In addition, replication would provide comparison data for schools interested in self-study of their admission processes. Another direction for future research would be to explore the context in which medical school admission decisions are conducted, with a focus on identifying links between admission policies, practices, and specific institutional educational missions and goals. Collecting qualitative in addition to quantitative data would provide additional context for interpreting results.
Acknowledgments: The authors thank the following AAMC personnel for reviewing earlier versions of this article: Henry Sondheimer, Amy Addams, Elizabeth White, Geoffrey Young, and Cynthia Searcy. They also thank the members of the MR5 Committee for their tireless efforts: Steven Gabbe, Ronald Franks, Lisa Alty, Dwight Davis, Kevin Dorsey, Michael Friedlander, Robert Hilborn, Barry Hong, Richard Lewis, Maria Lima, Catherine Lucey, Saundra Oyewole, Richard Riegelman, Gary Rosenfeld, Richard Schwartzstein, Maureen Shandling, Catherine Spina, and Ricci Sylla, as well as consultant Paul Sackett. In addition, they thank Trey Pigg for his contributions to this article. Finally, they would like to thank three anonymous reviewers for their suggestions, which greatly improved the manuscript.
Other disclosures: Medical College Admission Test (MCAT) is a program of the Association of American Medical Colleges (AAMC). Related trademarks owned by the AAMC include Medical College Admission Test, MCAT, and MCAT2015.
Ethical approval: This study was reviewed by AAMC’s Human Subjects Research Protection Program and was determined to be exempt because it was conducted for operational purposes. However, it underwent AAMC’s Data Collection and Instrument Clearance review to ensure that the survey instrument and methodology complied with AAMC’s policies and procedures.
Previous presentations: Some of the data presented in this manuscript were published in Dunleavy DM, Whittaker KM. The evolving medical school admissions interview. AAMC Analysis in Brief. 2011;11(7); and Dunleavy DM, Sondheimer H, Castillo-Page L, Bletzinger RB. Medical school admissions: More than grades and test scores. AAMC Analysis in Brief. 2011;11(6).