The relationship between second-year medical students' OSCE scores and USMLE Step 2 scores

J Eval Clin Pract. 2007 Dec;13(6):901-5. doi: 10.1111/j.1365-2753.2006.00768.x.


Rationale, aims and objectives: A valid tool to measure clinical competency early in medical school could identify students who may require special educational attention. The overall aim is to assess the relationship between students' scores on an objective structured clinical examination (OSCE) given in the second year of medical school and their subsequent performance on Step 2 of the United States Medical Licensing Examination (USMLE Step 2).

Methods: Participants were 390 second-year medical students participating in a required OSCE; complete data (Medical College Admission Test, OSCE, USMLE Step 1 and Step 2 scores) were available for 340 students (87%). Univariate correlations and linear regression analyses were performed.

Results: Total OSCE score was moderately correlated with USMLE Step 2 score (r = 0.395, P < 0.001), as were two skills subscores of the OSCE, differential diagnosis (r = 0.343, P < 0.001) and identification of abnormality (r = 0.322, P < 0.001). In linear regression analysis, neither OSCE total score nor any of the subscores independently predicted Step 2 scores; only Step 1 score (beta = 0.687, P < 0.001) and female sex (beta = 0.152, P < 0.001) remained independent correlates of Step 2 score.

Conclusion: OSCEs early in medical school can be useful in the early assessment of clinical competence.

MeSH terms

  • Adult
  • Clinical Competence / standards*
  • Clinical Competence / statistics & numerical data
  • Congenital Abnormalities / diagnosis
  • Diagnosis, Differential
  • Education, Medical, Undergraduate / standards*
  • Education, Medical, Undergraduate / statistics & numerical data
  • Educational Measurement / methods*
  • Educational Measurement / statistics & numerical data
  • Female
  • Humans
  • Licensure, Medical
  • Male
  • Medical History Taking
  • Sex Factors
  • Students, Medical*
  • Teaching / methods