Educational Outcomes for Students Enrolled in MDPhD Programs at Medical School Matriculation, 19952000: A National Cohort Study

Jeffe, Donna B. PhD; Andriole, Dorothy A. MD; Wathington, Heather D. PhD; Tai, Robert H. EdD

doi: 10.1097/ACM.0000000000000071
Research Reports

Purpose: To describe educational outcomes for a national cohort of students who enrolled in MD–PhD programs at medical school matriculation (MD–PhD matriculants).

Method: The authors used multivariate logistic regression to identify factors independently associated with overall MD–PhD program attrition (MD-only graduation or medical school withdrawal/dismissal) compared with MD–PhD program graduation among the 1995–2000 national cohort of MD–PhD matriculants at medical schools with and without Medical Scientist Training Program (MSTP) support.

Results: Of 2,582 MD–PhD matriculants, 1,885 (73.0%) were MD–PhD graduates, 597 (23.1%) were MD-only graduates, and 100 (3.9%) withdrew/were dismissed from medical school by July 2011. MD–PhD matriculants at non-MSTP-funded schools (adjusted odds ratio [AOR], 1.96; 95% confidence interval [CI], 1.60–2.41) and who had lower Medical College Admission Test scores (< 31 versus ≥ 36: AOR, 1.60; 95% CI, 1.20–2.14; 31–33 versus ≥ 36: AOR, 1.31; 95% CI, 1.01–1.70) were more likely to leave the MD–PhD program; matriculants who reported greater planned career involvement in research (AOR, 0.65; 95% CI, 0.51–0.84) and matriculated more recently (AOR, 0.90; 95% CI, 0.85–0.96) were less likely to leave the MD–PhD program. Gender, race/ethnicity, and premedical debt were not independently associated with overall MD–PhD program attrition.

Conclusions: Most MD–PhD matriculants completed the MD–PhD program; most of those who left were MD-only graduates. Findings regarding variables associated with attrition can inform efforts to recruit and support students through successful completion of MD–PhD program requirements.

Dr. Jeffe is research professor of medicine, Washington University School of Medicine, and director, Health Behavior, Communication, and Outreach Core, Alvin J. Siteman Cancer Center at Barnes-Jewish Hospital and Washington University School of Medicine, St. Louis, Missouri.

Dr. Andriole is assistant dean for medical education and associate professor of surgery, Washington University School of Medicine, St. Louis, Missouri.

Dr. Wathington is assistant professor, Curry School of Education, University of Virginia, Charlottesville, Virginia.

Dr. Tai is associate professor, Curry School of Education, University of Virginia, Charlottesville, Virginia.

Editor's Note: A commentary on this article by A.C. Bonham appears on pages 21–23.

Funding/Support: This study was funded in part by grants from the National Institute of General Medical Sciences (NIGMS; 2R01 GM085350-04 and R01 GM094535-02) and the National Cancer Institute (NCI) Cancer Center Support grant to the Siteman Cancer Center (P30 CA091842-07) for use of the Health Behavior, Communication, and Outreach Core data management services. The NIGMS and NCI were not involved in the design or conduct of the study; collection, management, analysis, and interpretation of the data; or preparation, review or approval of the manuscript.

Other disclosures: None reported.

Ethical approval: The study was approved by the institutional review board at Washington University School of Medicine.

Disclaimer: The conclusions of the authors are not necessarily those of the Association of American Medical Colleges, National Institutes of Health, or their respective staff members.

Previous presentations: Preliminary results of this study were presented in part at the Association of American Medical Colleges Graduate Research, Education, and Training (GREAT) Group, MD–PhD section, 2011 Annual Meeting, Minneapolis, Minnesota, July 28–30, 2011; at the 5th Annual Conference on Understanding Interventions That Broaden Participation in Research Careers, Baltimore, Maryland, May 10–12, 2012; and at the 51st Annual Conference on Research in Medical Education, 123rd Annual Meeting of the Association of American Medical Colleges, San Francisco, California, November 2–7, 2012.

Correspondence should be addressed to Dr. Jeffe, Division of Health Behavior Research, Department of Medicine, Washington University School of Medicine, 4444 Forest Park Ave., Suite 6700, St. Louis, MO 63108; telephone: (314) 286-1914; e-mail: djeffe@dom.wustl.edu.

Article Outline

The National Institute of General Medical Sciences (NIGMS) introduced the Medical Scientist Training Program (MSTP) in 1964 to support the training of students in both basic sciences and clinical research and to foster their development as physician–scientists engaged in biomedical research.1 The number of institutional MSTP awards increased from 3 in 19641 to 44 in 2012–2013.2 MD–PhD dual-degree programs were offered at 111 of 131 medical schools in the United States as of 2010–20113; most of these programs were not funded by the MSTP.

The number of students enrolled in MD–PhD programs at U.S. medical schools has grown considerably since 2000.4 The NIGMS (through its MSTP funding) and medical schools have substantially increased the level of support they provide for MD–PhD programs since the mid-1990s. Figure 1 shows annual amounts of MSTP-funded and medical-school-funded, non-need-based (NNB) grants and scholarships awarded, without service commitments, in support of MD–PhD program enrollees in U.S. Liaison Committee on Medical Education (LCME)-accredited medical schools from 1994–1995 through 2010–2011.5–7 As shown, the amounts of NNB MSTP and medical school grant/scholarship awards were about equal in 1994–1995; by 2010–2011, however, MSTP funding had more than doubled to $32.9 million, whereas medical school funding had increased by more than six times the 1994–1995 amount to $93.9 million. Given this level of financial support, educational outcomes for MD–PhD program enrollees are of interest to medical schools throughout the United States.1,8–10

Nationally, MD completion rates among all medical school enrollees (reportedly over 95%11) are much higher than biomedical field PhD completion rates among all PhD program enrollees (reportedly about 50%12). However, educational outcomes for national cohorts of MD–PhD program enrollees have not been reported. Thus, we conducted a retrospective study to describe the educational outcomes of a national cohort of MD–PhD program enrollees who entered their dual-degree programs at matriculation (MD–PhD matriculants) and to examine those outcomes in the context of MSTP funding and other relevant factors. We sought to answer the following question: What are the rates of and factors associated with MD–PhD program noncompletion?

We hypothesized that most of these MD–PhD program matriculants would complete the MD–PhD program and that most of those who did not complete the PhD program would complete the MD requirements. We also hypothesized that dual-degree completion rates would vary in association with institutional MSTP-funding status.13

Back to Top | Article Outline

Method

We conducted this study in 2012 using a database that included individualized, deidentified records for all medical school matriculants who enrolled in U.S. LCME-accredited medical schools in 1995–2000 as well as follow-up data through July 26, 2011. These deidentified records were provided to us by the Association of American Medical Colleges (AAMC) and were linked using unique identification numbers assigned by the AAMC. We had no access to any student or medical school identifiers. The institutional review board at Washington University School of Medicine approved this study as non-human-subjects research.

To analyze variables associated with MD–PhD program noncompletion versus completion, we included data for the medical school matriculants who reported on the AAMC Matriculating Student Questionnaire (MSQ) that they enrolled in an MD–PhD dual-degree program at the time of matriculation (MD–PhD matriculants). The MSQ is a survey administered annually to incoming medical students and is completed voluntarily on an identifiable but confidential basis.14 We included those MD–PhD matriculants who entered medical school during 1995–2000 to allow sufficient time for most to have completed their dual-degree requirements through the follow-up period in 2011. We also analyzed data for the following MSQ items: age at matriculation (categorized by the AAMC as < 20, 20–22, 23–25, 26–28, and > 28 years; we combined the two younger age categories into < 23 years); planned extent of career involvement in research (rated on a five-point scale: 1 = not involved, 2 = involved in a limited way, 3 = somewhat involved, 4 = significantly involved, 5 = exclusively involved); career intention at matriculation (categorized as no choice [for respondents who did not answer this item], undecided, other, full-time clinical practice, and full-time faculty/research scientist); and total premedical debt (categorized as no debt, $100–$19,999, ≥$20,000, and missing data).

In addition, we analyzed the following data from the AAMC’s Student Records System (SRS): year of medical school matriculation, gender, self-identified race/ethnicity, last status description and date, graduation date, and degree program in which student was enrolled at follow-up. The SRS is a secure, Web-based application that contains individualized records for all students enrolled in U.S. LCME-accredited MD-granting medical schools. It is used to track each medical student from matriculation through graduation, with medical schools’ registrars and student records representatives regularly updating and verifying the accuracy of the SRS data.15

We computed medical school duration according to last status date or graduation date and matriculation date recorded in the SRS. We categorized matriculants’ last status description as medical school withdrawal/dismissal (including withdrawal for academic reasons, dismissal for academic reasons, withdrawal for health reasons, withdrawal for other nonacademic reasons, and dismissal for nonacademic reasons), graduated, or still in medical school. We categorized most recent degree program of enrollment as either MD–PhD or MD only; the MD-only category included all other degree programs at enrollment (i.e., MD only, BA–MD, BS–MD, and all other MD–advanced-degree programs—e.g., MD–MA, MD–MPH, MD–JD—excluding MD–PhD).

Race/ethnicity data were self-reported by MD–PhD matriculants in response to a list of options on the American Medical College Application Service questionnaire. We categorized race/ethnicity as follows: Asian/Pacific Islander; underrepresented minority in medicine (URM), including matriculants who self-identified as black, Hispanic, or American Indian/Alaska Native; white; and other/unknown, including matriculants who self-identified as “other” or multiple races or who did not respond.

We obtained Medical College Admission Test (MCAT) scores from the AAMC Data Warehouse and computed composite MCAT scores for MD–PhD matriculants using their most-recent-attempt verbal reasoning, physical science, and biological science subscores. To include those matriculants who did not have MCAT scores available, we created a five-category variable for this analysis. We used the quartile split of MCAT scores for all MD–PhD graduates with scores in our database (< 31, 31–33, 34–35, ≥ 36) and added a fifth category for those without MCAT scores.

To determine institutional MSTP funding during the study period, we used NIGMS rosters of MSTP-funded institutions,16 which are updated annually.2 We categorized all U.S. LCME-accredited medical schools either as having received MSTP funding for some or all years of the study period (yes) or as never having received MSTP funding during the study period (no). We obtained archived institutional MSTP-funding data using NIH RePORTER.17 Of the 129 U.S. LCME-accredited medical schools to which students had matriculated during the study period, 39 schools (30%) had received MSTP funding for some or all of the years 1995–2000 (MSTP-funded schools), and 90 schools (70%) had never received any MSTP funding during this period (non-MSTP-funded schools). The AAMC then linked those institutional MSTP-funding data with each matriculant’s individualized record and provided the deidentified data, without school names, to us.

Finally, we determined the proportion of MD–PhD graduates who were included in our final study sample of MD–PhD matriculants (based on MSQ data, as described above). We used SRS data to identify all MD–PhD graduates in the entire cohort of medical school matriculants who enrolled in U.S. LCME-accredited medical schools in 1995–2000. We then created a three-category variable for these MD–PhD graduates based on their responses to the MSQ item for degree program of enrollment at the time of medical school matriculation (MD–PhD matriculants, MD only [including BA–MD, BS–MD, or MD–other-advanced-degree program matriculants], and unknown [for those MD–PhD graduates who did not complete this MSQ item]).

Back to Top | Article Outline

Outcome measure

On the basis of the last status description (SRS data), we created a three-category educational outcome variable for all MD–PhD matriculants in our study sample: MD-only graduation (for those who graduated from medical school without the PhD), medical school withdrawal/dismissal, and MD–PhD graduation. Students still enrolled in medical school at follow-up were excluded from this analysis.

Back to Top | Article Outline

Statistical analysis

We used chi-square tests to identify associations among categorical variables and analysis of variance to describe differences in continuous variables between groups. We report adjusted odds ratios and 95% confidence intervals from three separate multivariate logistic regression models. These three models identified variables independently associated with each of the three outcomes—MD-only graduation (PhD program attrition), medical school withdrawal/dismissal (medical school attrition), and overall attrition (including both MD-only graduation and medical school withdrawal/dismissal); each outcome was compared with MD–PhD graduation (reference group). All tests were performed using IBM SPSS Statistics version 20.0.0.1 (IBM Corporation, Armonk, New York, 2011). Two-sided P values < .05 were considered significant.

Back to Top | Article Outline

Results

Of the 97,096 matriculants at U.S. LCME-accredited medical schools in the 1995–2000 cohort, 89,948 (92.6%) answered the MSQ item pertaining to degree program of enrollment; of these, 2,645 (2.9%) reported MD–PhD dual-degree program enrollment at matriculation. At the time of follow-up (July 2011), 2,526 (95.5%) of the 2,645 MD–PhD matriculants had graduated from medical school, and 101 (3.8%) had left medical school (withdrawal/dismissal). We excluded from our study the 11 (0.4%) who were still in medical school at follow-up and the 7 (0.3%) who were deceased or had their degree revoked. Therefore, of the 2,627 MD–PhD matriculants eligible for inclusion in our analysis of educational outcomes, our final sample included the 2,582 (98.3%) who had graduated or left medical school and had data for all variables of interest.

The characteristics of these 2,582 MD–PhD matriculants, grouped by educational outcome, are shown in Table l. At the time of follow-up, 1,885 (73.0%) were MD–PhD graduates, including 79.5% (1,384/1,741) of MD–PhD matriculants at MSTP-funded and 59.6% (501/841) of those at non-MSTP-funded medical schools. Of the remaining 697 MD–PhD matriculants, 597 (23.1% of 2,582) were MD-only graduates—including 17.3% (302/1,741) of the MD–PhD matriculants at MSTP-funded and 35.1% (295/841) of those at non-MSTP-funded medical schools; 100 (3.9%) withdrew or were dismissed from medical school—including 3.2% (55/1,741) of those at MSTP-funded and 5.4% (45/841) of those at non-MSTP-funded medical schools (chi-square, P < .001).

As Table 1 shows, the following variables were associated with MD-only graduation but were not associated with withdrawal/dismissal, when each was compared with MD–PhD graduation: gender, race/ethnicity, matriculation year, total premedical debt, MCAT score, career intention at matriculation, and planned career involvement in research at matriculation. The other two variables, age at matriculation and institutional MSTP funding, were associated with both MD-only graduation and medical school withdrawal/dismissal, when each outcome was compared with MD–PhD graduation. Mean (standard deviation) medical school duration in years was 8.0 (1.3) for MD–PhD graduates, 5.6 (1.8) for MD-only graduates, and 4.6 (3.0) for students who withdrew or were dismissed from medical school (P < .001). Of the 100 students who withdrew or were dismissed, 21 were dismissed for academic reasons, 12 withdrew for academic reasons, 6 were dismissed for other reasons, 4 withdrew for health reasons, and the remaining 57 withdrew for other reasons.

Table 2 shows the results of the three multivariable logistic regression models used to identify variables associated with MD-only graduation, medical school withdrawal/dismissal, and overall attrition (including both MD-only graduation and medical school withdrawal/dismissal), when each outcome was compared with MD–PhD graduation. Among the 2,582 MD–PhD program matriculants in our sample, those who had enrolled at non-MSTP-funded medical schools, had MCAT scores < 34, and were ≥ 23 years of age at matriculation were significantly more likely to be in the overall attrition group, whereas those who matriculated in more recent years and reported greater planned career involvement in research at matriculation were significantly less likely to be in the overall attrition group. Among the 2,482 medical school graduates, those who had enrolled at non-MSTP-funded medical schools, had MCAT scores < 34, were ≥ 23 years old at matriculation, and intended full-time clinical practice careers at matriculation were significantly more likely to be MD-only graduates, whereas those who matriculated in more recent years and reported greater planned career involvement in research at matriculation were significantly less likely to be MD-only graduates. Among the 1,985 MD–PhD matriculants who either graduated with MD–PhD degrees or withdrew/were dismissed from medical school, individuals who matriculated at non-MSTP-funded medical schools, were of URM race/ethnicity, and were > 28 years of age at matriculation were significantly more likely to withdraw/be dismissed from medical school. Neither gender nor total premedical debt was independently associated with overall attrition, with MD-only graduation, or with medical school withdrawal/dismissal.

To determine the proportion of MD–PhD graduates among the 1995–2000 MD–PhD matriculants included in our final sample, we also examined program enrollment at time of matriculation for all MD–PhD graduates in the entire 1995–2000 national cohort of 97,096 medical school matriculants. There were 2,629 MD–PhD graduates (2.7%) in this national cohort according to SRS data. On the basis of their responses to the MSQ item about degree program of enrollment at medical school matriculation, we determined that 1,920 (73.0%) of these MD–PhD graduates had enrolled in MD–PhD programs at matriculation (of whom 1,885 [98.2%] were included in our final sample of MD–PhD matriculants), and 399 (15.2%) had enrolled in MD-only or MD–other-advanced-degree programs at matriculation. The remaining 310 (11.8%) had not responded to this MSQ item. Thus, of the 2,319 MD–PhD graduates in our database for whom degree program at matriculation was available from MSQ data, 82.8% (1,920/2,319) had entered MD–PhD programs at the time of medical school matriculation.

Back to Top | Article Outline

Discussion

Most of the MD–PhD matriculants in our study sample were MD–PhD graduates by the time of follow-up, and the majority of those who discontinued MD–PhD program participation graduated from medical school. The overall MD degree completion rate of 95.5% (including MD–PhD and MD-only graduates) that we observed among MD–PhD matriculants aligns with the 10-year medical school completion rate of 96% reported for all U.S. LCME-accredited medical school enrollees.11 The PhD degree completion rate of 73.0% that we observed among MD–PhD matriculants in our sample compares favorably with recently reported PhD degree completion rates among doctoral degree program enrollees in biomedical science fields, which ranged from 41.6% (genetics and genomics) to 56.2% (immunology and infectious diseases).12

The mean medical school duration for MD-only graduates in our sample was 5.6 years, suggesting that these graduates likely completed, on average, two years of preclinical study and one or more years of predoctoral research training prior to PhD program discontinuation and completion of MD degree requirements. This line of reasoning is consistent with results of an earlier NIGMS study indicating that MSTP-funded medical school graduates who enrolled in MD–PhD programs but completed only their MD degree requirements had received a mean of 36 months of predoctoral training support.1 Similarly, a single-institution analysis of MD–PhD program attrition reported that more than half of MD–PhD program enrollees who ultimately discontinued participation had remained in the MD–PhD program for at least three years.18 Thus, our results and others’ observations suggest that more than three years of follow-up after MD–PhD program enrollment is needed to estimate the extent of attrition among enrollees. Indeed, with the minimum 10-year follow-up period in our national cohort study, there were 11 MD–PhD matriculants who were excluded from our final sample because they were still in medical school.

Our findings regarding MD–PhD program attrition should be considered in the context of two previous multi-institutional studies that analyzed MD–PhD program enrollees’ career paths and included attrition data.1,10 First, an NIGMS study examined outcomes for all MD–PhD program enrollees nationally who had received at least 12 months of MSTP support during 1969–1990 and had graduated by 1990.1 Of the 1,430 MSTP-funded trainees included in the study, 1,161 (81.2%) were MD–PhD graduates whereas 269 (18.8%) were MD-only graduates by 1990, indicating 18.8% attrition from the PhD degree component of the program among these MSTP-supported medical school graduates.1 The rate of attrition from the PhD degree component of MD–PhD programs at MSTP-funded schools in our sample of 1995–2000 MD–PhD matriculants who graduated from medical school (17.9%; 302/1,686) was closely aligned with the rate observed in the earlier NIGMS study. Thus, despite the increase in mean duration of training required for MD–PhD program completion (from 6.6 years in the 1980 cohort to 7.3 years in the 1990 cohort of MSTP graduates in the NIGMS study,1 to about 8 years in more recent cohorts of MD–PhD graduates from MSTP-funded institutions19 and in our study), PhD program attrition among MD–PhD program enrollees at MSTP-funded medical schools was relatively constant over these two study periods.

Second, in a survey study of 24 MD–PhD program directors, Brass and colleagues10 found that an average of 10% of MD–PhD program enrollees who entered MD–PhD programs from academic years (AYs) 1998 through 2007 and were followed through AY 2008 had withdrawn from their MD–PhD programs. (Across programs, attrition varied from 3% to 34%.) This average overall attrition rate of 10% is substantially lower than either the 27.0% (697/2,582) overall attrition that we observed among all MD–PhD matriculants in our study sample or the 20.5% (357/1,741) overall attrition that we observed among those enrolled in MD–PhD programs at MSTP-funded medical schools (Table 1). Differences in study design and sample selection can explain the disparities in the attrition rates observed by us and by Brass et al.10 Our study included a national sample of individuals who enrolled in MD–PhD programs at matriculation and were followed for a minimum of 10 years after program enrollment, and our sample comprised only those MD–PhD matriculants who were no longer in medical school at follow-up. By comparison, Brass et al10 included MD–PhD program enrollees (who may or may not have entered MD–PhD programs at the time of medical school matriculation) at 24 selected institutions (20 MSTP-funded and 4 non-MSTP-funded). They followed these students for 1 to 10 years after program enrollment and included in their sample many students who were still enrolled in their MD–PhD programs.

Back to Top | Article Outline

Implications

Our observations have implications for the selection and support of MD–PhD program enrollees. MCAT scores of MD–PhD matriculants in our sample indicate that MD–PhD program enrollees are well prepared academically. Our finding that MD–PhD matriculants with MCAT scores < 34 were significantly more likely than those with scores ≥ 36 to be MD-only graduates provides support for MD–PhD program directors’ use of MCAT scores as one of the criteria for selecting applicants who will most likely complete the dual-degree program. MD–PhD programs currently accept and enroll students with a wide range of MCAT scores; in 2011, for example, MD–PhD program matriculants’ MCAT scores ranged from 22 to 44.20 Thus, MD–PhD program directors may wish to consider using MCAT scores to identify enrollees who may be at risk of attrition from the PhD program component. Some of these students may benefit from additional academic support.

The relationships we observed between MD-only graduation (PhD program attrition) and students’ planned career involvement in research at matriculation and career intention at matriculation provide support for MD–PhD program directors’ selection of applicants with career aspirations that are well aligned with programmatic missions and goals.18 In addition, our observations related to age at matriculation and risk of MD-only graduation (greater for those ≥ 23 years old) or risk of withdrawal/dismissal from medical school (greater for those > 28 years old) indicate that further research is warranted to examine whether older enrollees face particular challenges in completing the MD–PhD program that are amenable to intervention.

That more recent matriculation year was associated with a lower likelihood of PhD program attrition suggests that, during the time frame of our study, MD–PhD program directors may have become better able to recruit and support enrollees who will successfully complete the dual-degree program requirements. This observation seems particularly important in light of the 40% increase in MD–PhD program enrollment nationally from 3,632 in 200221 to 5,097 in 2012.22

The likelihood of attrition also differed significantly by medical school MSTP-funding status, even after controlling for other variables in the models. Attending a non-MSTP-funded school was associated with a greater likelihood of both PhD program attrition and withdrawal/dismissal from medical school. These observations suggest that MD–PhD program directors and administrators at MSTP-funded institutions, in particular, may have expertise in selecting applicants most likely to complete their MD–PhD programs. MSTP-funded institutions may also have particular resources in place (e.g., academic assistance, mentoring and advising programs) to support their students through successful MD–PhD program completion. Differences in financial support also may have contributed to the differences in attrition rates we observed among enrollees at schools with and without MSTP funding. A 2012 AAMC survey on MD–PhD program policies found that MD–PhD program positions (which, at MSTP-funded medical schools, can include both MSTP and non-MSTP-funded positions) were fully funded at 41 (93.2%) of the 44 responding MSTP-funded medical schools but at only 29 (72.5%) of the 40 responding non-MSTP-funded medical schools.13

Although Watt and colleagues’23 earlier single-institution survey study reported that high levels of educational debt (> $50,000) were associated with greater consideration of leaving the MD–PhD program, in our cohort higher total premedical debt (> $20,000) was not independently associated with MD–PhD program attrition. We were unable to examine total accumulated debt during medical school in our national cohort, because those data were available only for medical school graduates who completed the AAMC Medical School Graduation Questionnaire (GQ)24; not all MD–PhD matriculants included in our sample completed or were eligible to complete the GQ, so we did not use data from the GQ in this study. Whether attrition rates would decrease with lower levels of total accumulated debt after matriculation is an empirical question requiring further study, but it would be a plausible hypothesis on the basis of Watt and colleagues’23 findings.

Finally, our results have implications for physician–scientist workforce diversity. Over 30% of the MD–PhD matriculants in our sample were women, and we did not observe any independent associations between attrition and gender. This finding extends earlier observations that women were no more likely than men to consider leaving MD–PhD programs23 and that MD–PhD program attrition rates were not associated with the percentage of female trainees enrolled.10 Thus, as the number of women enrolled in MD–PhD programs increases,4,11 the number of women MD–PhD graduates can be expected to increase similarly. However, we did find self-identified URM race/ethnicity to be independently associated with a greater likelihood of attrition due to withdrawal/dismissal from medical school. Efforts to identify additional factors contributing to medical school attrition among URM MD–PhD program enrollees are warranted so that appropriate interventions can be designed to minimize URM student attrition.

Back to Top | Article Outline

Strengths and limitations

Among the strengths of our study was the inclusion of a national sample of MD–PhD program enrollees who were followed for a minimum of 10 years from the time of program entry at medical school matriculation. Unlike previously published studies that included educational outcomes data for MD–PhD program enrollees (some of whom could have enrolled after matriculation),1,10 our study described the extent of PhD program attrition and medical school attrition for a national cohort of MD–PhD matriculants. We also included data for many variables of interest that have not previously been examined as predictors of MD–PhD program attrition (e.g., MCAT scores, extent of planned career involvement in research at matriculation, institutional receipt of MSTP funding).

Our study also has some limitations. Our sample was limited to individuals who completed the MSQ and indicated their enrollment in MD–PhD programs at the time of medical school matriculation. As with all self-reported survey data, the accuracy of MSQ data is dependent on the accuracy with which the respondents completed the questionnaire. Our educational outcomes of interest were based on SRS data for last status and for degree program at graduation; thus, the accuracy of our educational outcomes data is subject to the accuracy with which medical school registrars and student records representatives updated their students’ records. Our findings thus represent only the best inferences that can be made about educational outcomes for MD–PhD matriculants from these available data and must be interpreted with these limitations in mind. MD–PhD program enrollees who entered their dual-degree programs after medical school matriculation and those who chose not to respond to the MSQ item for degree program of enrollment at matriculation were not included in our analysis. We might reasonably expect that MD–PhD program attrition would be lower for students who are motivated to enroll in these programs after medical school matriculation, when they have already successfully completed some portion of the medical school curriculum, compared with students who enroll in these programs at the time of medical school matriculation, when they have not yet successfully completed any portion of the curriculum. If we were to assume, as a “best case” scenario, that there was 0% attrition among students who entered MD–PhD programs after medical school matriculation and 0% attrition among MD–PhD program enrollees who did not respond to the MSQ item for degree program of enrollment (and who may or may not have enrolled in the MD–PhD program at matriculation), then overall attrition among all MD–PhD program enrollees in our national cohort of 1995–2000 medical school matriculants would be 20.5%, calculated as follows: 679 MD–PhD program noncompleters/3,308 MD–PhD program enrollees (i.e., 2,629 MD–PhD program graduates + 679 MD–PhD program noncompleters).

Given that the overall attrition rate was 27.0% in our sample of MD–PhD matriculants, it appears to be sound policy for MD–PhD program directors to accept applicants from among internal MD students. Indeed, of the 84 MD–PhD program directors responding to a 2012 AAMC survey, 80 (95.2%) reported accepting internal MD student applicants; in addition, 46 (54.8%) reported accepting internal PhD student applicants, and 32 (38.1%) reported accepting transfer students from other MD–PhD programs.13

We also note that program-specific attrition may vary from our observations of this national cohort, and this may be so among both MSTP-funded and non-MSTP-funded institutions. Among MSTP-funded institutions in our study, the duration of MSTP funding ranged considerably, from receipt of initial funding during the study period to continuous funding since MSTP inception in 1964.1,17 MD–PhD program enrollees at MSTP-funded institutions in our study also may include enrollees who were not in positions fully funded by the MSTP.13 In addition, individual MD–PhD programs vary by size,10,25 demographic composition,10 curricular requirements,13 and the extent and nature of advising systems for enrollees.26 Program-specific attrition may be associated with many of these program characteristics, but we lacked such program-level data. We also note that students enrolled in MD–PhD programs may pursue their PhD studies in a variety of fields10,25 with which program attrition might be associated, but we did not have information about students’ fields of study. Moreover, our results may not be generalizable to MD–PhD program enrollees at medical schools in Canada or at DO-granting medical schools. Nonetheless, our findings can inform MD–PhD programs’ efforts to recruit and enroll a diverse pool of students whose educational and professional goals are well aligned with MD–PhD programmatic goals.

Back to Top | Article Outline

Conclusion

In conclusion, although some level of noncompletion is likely inevitable, most MD–PhD program matriculants complete the dual-degree program. Our findings may be of particular interest to MD–PhD program directors and faculty seeking to maximize program completion rates. MSTP funding, in particular, plays an important beneficial role in promoting MD–PhD program completion. Given the steadily increasing divergence in costs borne by medical schools and by MSTP funding for MD–PhD programs and the relatively stagnant MSTP-funding levels from 2007 to 2011 (see Figure 1), the impact that the current federal fiscal crisis27 will have on MSTP-funding levels (and on the medical schools and students who benefit from this funding) remains to be seen.

Acknowledgments: The authors thank their colleagues Paul Jolly, PhD, Gwen Garrison, PhD, David Matthew, PhD, Franc J. Slapar, MA, Susan Gaillard, BS, and Hershel Alexander, PhD, at the Association of American Medical Colleges for their support of these research efforts through provision of data and assistance with coding. They also thank James Struthers and Yan Yan, MD, PhD, at Washington University School of Medicine for data management services and statistical consults, respectively.

Back to Top | Article Outline

References

1. National Institute of General Medical Sciences. . MSTP Study: The Careers and Professional Activities of Graduates of the NIGMS Medical Scientist Training Program. September 1998 http://publications.nigms.nih.gov/reports/mstpstudy/. Accessed September 18, 2013
2. National Institute of General Medical Sciences. . Medical Scientist Training Program (MSTP) institutions: For award period 7/1/2012–6/30/2013. http://www.nigms.nih.gov/Training/InstPredoc/PredocInst-MSTP.htm. Accessed September 20, 2013
3. Barzansky B, Etzel SI. Medical schools in the United States, 2010–2011. JAMA. 2011;306:1007–1014
4. Association of American Medical Colleges. . Table 32. MD–PhD applicants, acceptees, matriculants and graduates of U.S. medical schools by sex, 1999–2011. https://www.aamc.org/download/161868/data/table32-mdphd00-11.pdf. Accessed September 22, 2013
5. Association of American Medical Colleges. AAMC Data Book: Statistical Information Related to Medical Education. 1998 Washington, DC Association of American Medical Colleges
6. Association of American Medical Colleges. AAMC Data Book: Statistical Information Related to Medical Schools and Teaching Hospitals. 1999–2000 (annual editions) Washington, DC Association of American Medical Colleges
7. Association of American Medical Colleges. AAMC Data Book: Medical Schools and Teaching Hospitals by the Numbers. 2001–2012 (annual editions) Washington, DC Association of American Medical Colleges
8. Fang D, Meyer RE. Effect of two Howard Hughes Medical Institute research training programs for medical students on the likelihood of pursuing research careers. Acad Med. 2003;78:1271–1280
9. Andriole DA, Whelan AJ, Jeffe DB. Characteristics and career intentions of the emerging MD/PhD workforce. JAMA. 2008;300:1165–1173
10. Brass LF, Akabas MH, Burnley LD, Engman DM, Wiley CA, Andersen OS. Are MD–PhD programs meeting their goals? An analysis of career choices made by graduates of 24 MD–PhD programs. Acad Med. 2010;85:692–701
11. Garrison G, Mikesell C, Matthew D. Medical school graduation and attrition rates. AAMC Analysis in Brief. 2007;7(2) https://www.aamc.org/download/102346/data/aibvol7no2.pdf. Accessed September 18, 2013
12. National Research Council. Research-Doctorate Programs in the Biomedical Sciences: Selected Findings From the NRC Assessment. 2011 Washington, DC National Academies Press http://grants.nih.gov/training/biomedical_sciences.pdf. Accessed September 20, 2013
13. Association of American Medical Colleges. . Summary of MD–PhD Programs and Policies. Updated October 10, 2012. https://www.aamc.org/students/download/62760/data/faqtable.pdf. Accessed September 18, 2013
14. Association of American Medical Colleges. . Matriculating Student Questionnaire (MSQ). Sample Matriculating Student Questionnaire. https://www.aamc.org/data/msq/. Accessed September 18, 2013
15. Association of American Medical Colleges. . Student Records System (SRS). https://www.aamc.org/services/srs/. Accessed September 18, 2013
16. National Institute of General Medical Sciences. . Medical Scientist Training Program. http://www.nigms.nih.gov/Training/InstPredoc/PredocOverview-MSTP.htm. Accessed March 6, 2013
17. NIH RePORTER. . Research Portfolio Online Reporting Tools (RePORT). http://www.projectreporter.nih.gov/. Accessed March 6, 2013
18. Milewicz DM. Training the physician scientist: Alternative pathways. Presented at: Association of American Medical Colleges GREAT MD–PhD Section Annual Meeting; October 2–4, 2008 Seattle, Wash https://www.aamc.org/linkableblob/62074-3/data/milewicz08-data.pdf. Accessed September 20, 2013
19. Jeffe DB, Andriole DA. A national cohort study of MD–PhD graduates of medical schools with and without funding from the National Institute of General Medical Sciences’ Medical Scientist Training Program. Acad Med. 2011;86:953–961
20. Association of American Medical Colleges. . Table 35. MCAT scores and GPAs for MD/PhD applicants and matriculants to U.S. medical schools, 2008–2011. https://www.aamc.org/download/161878/data/table45-mdphd-mcatgpa-2011.pdf. Accessed September 18, 2013
21. Association of American Medical Colleges. . Table 36. Total active MD/PhD enrollment by U.S. medical school and sex, 2002–2006. https://www.aamc.org/download/161882/data/table36a-mdphd-enroll-sch-sex-0206.pdf. Accessed September 18, 2013
22. Association of American Medical Colleges. . Table 36. Total active MD/PhD enrollment by U.S. medical school and sex, 2008–2012. https://www.aamc.org/download/321554/data/2012factstable36-2.pdf. Accessed September 18, 2013
23. Watt CD, Greeley SA, Shea JA, Ahn J. Educational views and attitudes, and career goals of MD–PhD students at the University of Pennsylvania School of Medicine. Acad Med. 2005;80:193–198
24. Association of American Medical Colleges. . Medical School Graduation Questionnaire: Previous questionnaires. https://www.aamc.org/data/gq/questionnaires/. Accessed September 18, 2013
25. Ahn J, Watt CD, Man LX, Greeley SA, Shea JA. Educating future leaders of medical research: Analysis of student opinions and goals from the MD–PhD SAGE (Students’ Attitudes, Goals, and Education) survey. Acad Med. 2007;82:633–645
26. Lee S, Wiley C. Advising and tracking of trainees and graduates: Results of the June 2008 survey of MD–PhD program directors and administrators. Presented at: Association of American Medical Colleges GREAT MD–PhD Section Annual Meeting; October 2–4, 2008 Seattle, Wash https://www.aamc.org/linkableblob/62062–3/data/sandraee08-data.pdf. Accessed September 18, 2013
27. Executive Office of the President of the United States. . White House Office of Management and Budget (OMB). OMB Report Pursuant to the Sequestration Transparency Act of 2012 (P. L. 112–155). http://www.whitehouse.gov/sites/default/files/omb/assets/legislative_reports/stareport.pdf. Accessed September 18, 2013
© 2014 by the Association of American Medical Colleges