USMLE performances in a predominantly Asian and Pacific Islander population of medical students in a problem-based learning curriculum.

PURPOSE

To compare the USMLE performances of students of various ethnicities, predominantly Pacific Islander and Asian, at one medical school and to examine the predictive validity of MCAT scores for USMLE performance.

METHOD

A total of 258 students in the graduating classes of 1996-2000 at the University of Hawai'i School of Medicine were classified by ethnicity. Demographic and performance characteristics of the groups were examined, and MCAT scores with and without undergraduate science GPA were used to predict USMLE performance. Under- and over-prediction rates were computed for each ethnic group.

RESULTS

Ethnic groups did not differ significantly by gender or undergraduate GPA. Chinese, Caucasian, and Other Asian students tended to have higher MCAT scores than Hawaiian/other Pacific Islander, and Filipino students. Ethnic groups did not differ significantly in prediction of USMLE Step 1 performance. For Step 2, MCAT scores significantly over-predicted performance of Filipino students and tended to under-predict performance of Caucasian students.

CONCLUSION

Although MCAT scores and science GPA were good predictors of USMLE performance, ethnic differences were found in the degrees of their predictive validity. These findings both replicate and extend results of earlier studies, and again point to the importance of exploring additional predictor variables. The authors encourage future research on the effects of the following factors on success in medical school: reading and test-taking skills, socio-cultural and environmental influences on learning, communication styles, primary language use, family support, and family responsibilities.