Predicting performance at medical school: can we identify at-risk students?

Area of Science:

• Medical Education
• Academic Performance Prediction
• Student Risk Identification

Background:

• Assessing medical student academic performance is crucial for identifying those at risk.
• Various indicators, including preadmission scores and grades, are used to evaluate performance.
• Early identification of at-risk students can facilitate timely intervention.

Purpose of the Study:

• To evaluate the predictive power of multiple indicators on medical school academic performance.
• To identify key predictors for early identification of at-risk medical students.
• To inform strategies for supporting student academic success.

Main Methods:

• Analysis of medical student grades over 14 years in a 6-year program.
• Utilized preadmission scores, unit/module grades, and examination scores.
• Employed multivariate linear regression to identify key predictive grades.

Main Results:

• High school scores showed limited predictability (6.8%) for final exams.
• University placement assessment scores offered slightly better predictability (14.9%).
• Course examination scores demonstrated significant predictability (55.8%-64.8%) for subsequent performance.
• Multivariate analysis identified key grades for predicting final integrated examination scores.

Conclusions:

• Continuous data archiving and analysis enable early identification of at-risk medical students.
• Predictive models can be developed using readily available academic data.
• The findings support proactive academic support strategies in medical education.