some links,
Correlation of National Board of Medical
Examiners Scores with United States Medical
Licensing Examination Step 1 and
Step 2 Scores
Christopher M. Zahn, MD, Aaron Saguil MD, MPH, Anthony R. Artino Jr, PhD,
Ting Dong, PhD, Gerald Ming, Jessica T. Servey, MD, Erin Balog, MD,
Matthew Goldenberg, MD, and Steven J. Durning, MD, PhD
https://www.aamc.org/download/307774/data/clerkshipeducation-zahn.pdf
"
Academic Medicine, Vol. 87, No. 10 / October 2012
1348
RIME: Clerkship Education
U.S. medical schools granting the
MD degree use scores on Steps 1
and 2 of the United States Medical
Licensing Examination (USMLE) as
objective assessments of medical student
performance.Indeed, passing scores
on all three Steps of the USMLE is a
requirement for medical licensure for
these graduates.
1
While passing these
examinations is not the primary focus
of basic science and clinical education
in medical school, it is an important
curricular outcome.
Clinical clerkships incorporate various
measures to assess student performance,
including objective examinations. Many
clerkships use the National Board of
Medical Examiners (NBME) Clinical
Subject (“shelf ”) examinations as
one component of the final clerkship
examination. Although one may assume
that the scores on subject examinations
in core clerkships correlate with
performance on the USMLE Step 2
Clinical Knowledge (CK) examination,
the data supporting this assumption
are limited.
2-5
The literature addresses
several other variables potentially
associated with Step 1 and/or Step 2
CK scores such as clinical performance,
Medical College Admission Test (MCAT)
scores, medical school grade point
average (GPA), and timing or length
of clerkships. Several studies have also
investigated correlations between subject
exam scores and subtest scores of the
USMLE examination. The only studies
addressing the relationship between
subject exam performance and overall
USMLE scores are in family medicine
and obstetrics and gynecology; these
studies analyzed the correlations of
the USMLE exam with their respective
clerkship subject exam.
2-4,6-8
Only one
study considered the relationship
between student performance on
subject examinations across multiple
clerkships and performance on Step 2
CK examinations; that study addressed
the effect of clerkship timing and
length.
5
Correlating subject exam scores
with performance on the USMLE
examinations could be important to
undergraduate medical educators
interested in predicting outcomes on
national standardized examinations,
and to graduate medical educators for
selecting residents.
The purpose of this study was to
determine whether the NBME Clinical
Subject Examination performances
of medical students from our medical
school’s six core clerkships correlated
with their scores on the USMLE Steps
1 and 2 CK Examinations. We also
sought to correlate these students’ GPAs
for basic science courses (likely more
relevant to Step 1) and clerkships (likely
more relevant to Step 2 CK) with subject
exam and USMLE performance. We
used the theory of context specificity
as a conceptual framework for our
study. Context specificity
9
argues that
performance is uniquely tied to context
and therefore performance on one
subject examination (one context) would
not be expected to be strongly correlated
with performance on other subject
examinations (other contexts) for a given
student.
Purpose
Determine whether the National Board
of Medical Examiners (NBME) Subject
Examination performance from six
clerkships correlated with United States
Medical Licensing Examination (USMLE)
Steps 1 and 2 Clinical Knowledge (CK)
Examination scores. Also, examine
correlations between medical students’
preclinical and clinical year mean
cumulative grade point average (GPA),
subject exam, and USMLE performance.
Method
The sample consisted of 507 students
from the 2008–2010 graduating classes
from the authors’ medical school.
Pearson correlations followed by
stepwise linear regressions were used
to investigate variance in USMLE Steps
1 and 2 CK scores explained by subject
exam scores and GPA.
Results
Data from 484 (95.5%) students were
included. USMLE Steps 1 and 2 CK
scores had moderate-to-large positive
correlations with all subject exam
scores and with both GPA variables.
Correlations between composite
subject exam scores and USMLE Steps
1 and 2 CK exams were 0.69 and
0.77, respectively. Regression analysis
demonstrated that subject exams and
GPA accounted for substantial variance in
Steps 1 and 2 CK exam scores (62% and
61%); when entered into the regression
model first, primary care clerkship subject
examination scores accounted for most
of this variance.
Conclusions
The moderate-to-large correlations
between subject exam performance and
USMLE scores provide reassurance that
subject exam scores are associated with
USMLE performance. Furthermore, the
considerable variance in USMLE scores
accounted for by primary care NBME
scores may be due to primary care topics
being reinforced through all clerkships
and comprising a significant portion of
the USMLE examinations, particularly
Step 2 CK."