Related Experiment Video
Updated: Aug 7, 2025
12:18
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
7.6K
Artificial Intelligence Screening of Medical School Applications: Development and Validation of a Machine-Learning
Marc M Triola1, Ilan Reinstein2, Marina Marin3
1M.M. Triola is associate dean of educational informatics and director, Institute for Innovations in Medical Education, NYU Grossman School of Medicine, New York, New York; ORCID: https://orcid.org/0000-0002-6303-3112 .
Summary
A machine-learning algorithm accurately screened medical school applications, matching faculty performance. This AI tool shows promise for consistent and reliable admissions reviews, benefiting diverse applicant groups.
Area of Science:
- Medical education technology
- Artificial intelligence in admissions
- Healthcare application screening
Background:
- Medical school admissions involve rigorous faculty screening.
- Ensuring consistency and reliability in this process is crucial.
- Exploring AI for initial application review can optimize efficiency.
Purpose of the Study:
- To determine if a machine-learning algorithm can accurately screen medical school applications.
- To assess the algorithm's performance against human faculty reviewers.
- To evaluate the algorithm's impact on fairness for diverse applicant groups.
Main Methods:
- Developed a virtual faculty screener algorithm using historical application data (2013-2017).
- Validated the algorithm retrospectively and prospectively on thousands of applications.
- Conducted a randomized trial comparing algorithm-driven vs. faculty-driven application reviews in 2019.
Main Results:
- The algorithm demonstrated strong performance in retrospective and prospective validations (AUROC up to 0.83).
- A randomized trial showed no significant differences in interview recommendation rates between the algorithm and faculty.
- The algorithm performed equitably across female and underrepresented in medicine applicants.
Conclusions:
- A virtual faculty screener algorithm effectively replicates human faculty screening of medical school applications.
- This AI tool has the potential to enhance consistency and reliability in admissions.
- The algorithm offers a promising solution for optimizing the medical school application review process.
