Related Experiment Video
Updated: May 18, 2026

A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
Local development of MCQ tests for evidence-based medicine and clinical decision making can be successful
Gerald E Crites1, Ronald J Markert, D Stephen Goggans
1Department of Medicine, Georgia Health Sciences University/University of Georgia Medical Partnership, Athens, GA, USA. gcrites@uga.edu
Background:
Guidelines for the design of multiple-choice item (MCQ) tests of evidence-based medicine (EBM) and clinical decision making (CDM) have not been published.
Purpose And Methods:
We describe a strategy to develop an EBM/CDM MCQ test database guided by educational theory and used psychometric analyses, including reliability, validity, and item analyses, to judge the strategy's success.
Results:
The internal consistency reliability of tests derived from the database was in the good-to-excellent range (0.74-0.95) and test-retest reliability was fair (0.51). One test discriminated across three levels of EBM/CDM learners (discriminant validity). Tests also predictively correlated with other medical school assessments according to theory (convergent and discriminant validity). The items were infrequently misclassified, had statistics close to historical standards, and were acceptable after no more than one round of revisions.
Conclusions:
Our strategy for developing an EBM/CDM MCQ database was successful and tests derived from it can be flexibly sampled to assess different EBM/CDM knowledge domains and three levels of EBM/CDM learners. Assuming the availability of similar resources to support its application, this strategy should be replicable at other settings.
Related Concept Videos
Clinical Trials
There are four phases in a clinical trial. A phase one...
Critical Thinking II
Preclinical Development: Overview
Patient-centered Care
Methods of Documentation VI: Case Management Model
For example, a patient with a chronic illness...
Clinical Trials: Overview