ֱ

Predictive Analytics Can Boost Medical School Diversity

— The MCAT isn't the only tool to evaluate future physician success

MedpageToday
Medical school professor lecturing to students in lab.

There's currently a significant lack of diversity among U.S. physicians. According to the Association of American Medical Colleges (AAMC), are African American, 5.8% are Hispanic, and just 5% of medical students' families are in the lowest household income quintile. This lack of diversity, and its accompanying scarcity of cultural competence, has been shown to create vast among minority and low socioeconomic status (SES) patients.

While medical school admissions committees are the gatekeepers that can create a more diverse physician workforce, there's a structural impediment to such positive change: these committees don't employ accurate data tools to help them recognize the potential that racially diverse and low SES medical students possess. And now, the right of medical schools to use a race-conscious approach to the admissions process is being threatened. However, predictive analytics that take into account a broad range of factors -- even if it is eventually ruled that race cannot be considered -- can play an important role in increasing medical school diversity.

At Ponce Health Sciences University (PHSU) -- the medical school where I am president -- we devised a predictive model that seeks to determine whether there's a better indicator of physician capability among minority and low SES students than the MCAT, the test currently used as the principal determinant in the medical school admissions process.

The results thus far have been revelatory: according to our PHSU predictive model, there is a strong positive relationship that shows minority and low SES "pipeline" students will score as well as on the USMLE Step 1 medical board exams as "direct" students from more advantaged backgrounds. The USMLE Step 1 exam is the primary screening tool used by hospitals to select applicants for residency interviews.

Below are two helpful definitions to unpack before I continue my analysis:

  • A pipeline student is one who has entered PHSU after earning a Master of Science in Medical Science (MSMS) degree, which is a 1-year curriculum offered at a wide range of institutions, designed to assist students who would like to enhance their academic credentials for entry into medical, dental, physician assistant, veterinarian, or other professional healthcare programs. We consider MSMS to be a longitudinal performance assessment program because these students must understand and master MD-level curricula. Also, according to the Liaison Committee on Medical Education (an accrediting body for U.S. medical schools), longitudinal performance assessment can be a better measure of a student's knowledge, skills, and abilities compared to a one-time standardized snapshot, such as the MCAT.
  • A direct student is one who has entered PHSU directly from an undergraduate program.

In our analysis, we found that our pipeline students -- especially those from minority and low SES backgrounds who graduated from high schools that didn't have the resources to adequately prepare them for the rigors of undergraduate academics -- have significantly lower average MCAT scores compared to our direct students. However, these students achieved the same score levels on the USMLE Step 1 as did direct students with high MCAT results.

We contend that a pipeline program coupled with predictive analytics (which take into account factors such as MCAT score, undergraduate GPA, and first-generation status, among others) can provide a new metric -- predicted Step 1 pass/fail score -- for medical school admissions committees to consider when evaluating applicants from a pipeline program. The new metric wouldn't replace existing MD admissions processes; rather, it would be supplemental and would only apply to applicants from pipeline programs.

This additional metric could help admissions committees ascertain the academic readiness of pipeline applicants with a reduced emphasis on the MCAT. For example, if a pipeline student has demonstrated mastery of the MSMS curriculum but has a weak MCAT score, the longitudinal performance assessment (pipeline program coupled with other predictive analytics) can be weighted higher than the one-time snapshot of knowledge (MCAT).

It's widely understood that usage of the MCAT as the pre-eminent medical school admissions data point affects national rankings: the higher the accepted students' average MCAT scores, the more prestigious the school is regarded. However, this reliance on a single data point undercuts progress towards a more diverse physician workforce.

According to an AAMC by C.R. Lucey, MD, the executive vice dean at UCLA Medical School, San Francisco, and Aaron Saguil, MD, MPH, the associate dean at F. Edward Hébert School of Medicine and vice chair at AAMC's Medical College Admission Test Validity Committee, "...assigning too much weight to the highest MCAT scores in admissions decision making makes it difficult to build medical school classes that are representative of patient communities."

Two of the recommendations Lucey and Saguil offer to medical schools to boost the diversity of incoming cohorts are:

  • Increase the availability of pipeline programs for aspiring doctors from minority and low SES backgrounds.
  • Implement a more holistic selection approach that reflects a wiser consideration of MCAT scores.

The conclusions of Lucey and Saguil would elevate the diversity of medical school students -- and future doctors -- without compromising the academic aptitude or professional preparedness of student cohorts. The addition of a predicted Step 1 pass/fail score to the medical school admissions criteria would further this effort. It would provide reliable data that pipeline students from diverse and low SES backgrounds can not only adequately handle the rigors of medical school, they will also score as well on the USMLE exam and be as prepared for a hospital residency as direct students.

David Lenihan, PhD, JD, is the president of Ponce Health Sciences University, a medical school with campuses in St. Louis, MO and Ponce, Puerto Rico. He is also the CEO of Tiber Health, a technology company focused on innovating medical school education and helping to solve the global physician shortage.