Comparison of Educational Methods for Teaching Peripheral Blood Smears

General Approach:

This study will employ a randomized, controlled design and will use as its outcomes two separate post-tests, one after the first teaching session and one 4 weeks after the second intervention.

Design The investigators are attempting to find whether the combination of basic science education using an 1-hour additive microskills- building method along with compare/ contrast model versus a 1-hour traditional case-based format would improve performance on a visual recognition test of hematology questions. A final test at 4 weeks after the cross-over between the two groups will also help investigators analyze whether the order of exposure to the teaching method would have an impact on the final performance.

Sampling, including sample size and statistical power:

Sample size will be between 40-50 resident physicians meeting inclusion criteria for the study. There is not enough data in the literature to effectively determine power, so we will use a convenience sample of the largest feasible group (around 40).

Collection of data:

Data will be collected on an excel spreadsheet and will be permanently de-identified on that spreadsheet; it will be kept on a password-protected site.

Information management and analysis software:

Data will be entered into a password protected excel spreadsheet and each study participant will be assigned a unique identifier code which will be used to link the data. Statistical analysis will be performed using SPSS version 20.

Data entry, editing and management, including handling of data collection forms, different versions of data, and data storage and disposition:

Paper surveys and tests will be collected, recorded and locked; they will be destroyed 2 years after termination of the study. A paired t-test will be used to compare the two samples means for significance.

Quality Control / Assurance:

Pretest and post-tests will be checked against a gold standard, a hospital board-certified hematopathologist, to check the quality of the slides.

Bias in data collection, measurement and analysis:

Given the small numbers of participants expected, a Type 1 error (failure to detect a difference in groups when one exists) is possible, as the study will possibly be underpowered.

Inclusion criteria

• Internal medicine and Family Medicine residents and faculty at The Reading Hospital and Medical Center having completed medical school

Exclusion criteria

• Practicing/ board certified hematologists