create new tag
, view all tags, tagging instructions

SuccessStoriesForm edit

Title Strategies for Ruling Out Myocardial Infarction
Year 2010
Contributor/Contact David Schoenfeld
CTSA Harvard
Success Topic Grants, Methods/core grants
Description of Success Many patients with chest pain come to the Emergency Department (ED), fearing an impending heart attack. Identifying individuals who are not in danger quickly can avoid the cost and inconvenience of a long hospital stay and multiple testing procedures.

Toward this end, investigators are conducting the Rule Out Myocardial Infarction using Computer Assisted Tomography (ROMICAT) II, a diagnostic, randomized multicenter clinical trial to determine the effectiveness, measured by the length of hospital stay, of cardiac computed tomography (CT) evaluation, compared to the standard evaluation, by adding cardiac CT to the initial ED evaluation of subjects with acute chest pain.

Before embarking on this large randomized trial, the investigators needed to know whether they could successfully answer the question of whether they could detect an improvement in the time to hospital discharge by the use of a cardiac CT. They decided to simulate the clinical trial that they planned to conduct. Harvard CTSC biostatisticians at the Massachusetts General Hospital, Hang Lee and David Schoenfeld, had the opportunity to create this simulation.

The key information required for the simulation was obtained from the results of ROMICAT I, an observational study to determine usefulness of cardiac CT as a diagnostic test for acute cardiac syndrome (ACS) among patients with acute chest pain.

To simulate the trial, Lee and Schoenfeld drew 500 LOS values from one of three log-normal distributions of which the first subgroup was to represent the patients with normal LV function and without CAD (48.6%, mean SD of 6 1.2 hours); the second was to represent the patients with normal LV function and with non-obstructive CAD (28.8%, mean SD of 10 2.0 hours); and the third group was to represent all other possible conditions (22.6%, mean SD of 57.9 60.4 hours.

They also had to consider the possibility that even with the CT results, some clinicians would not interpret it correctly and fail to appropriately treat patients who did not have CAD (triage inaccuracy) The following table summarizes, for each ED triage accuracy rate assumption, the estimated mean LOS and the power at a Type-1 error rate of 5% for an independent samples t-test using 500 patients in each arm. The proposed sample size of 1000 enrolled patients will attain a sufficient power to allow assessment of the true clinical benefit within a small margin of error in mean LOS between the two groups even in the presence of some degree of ED triage inaccuracy.

(See Table 1 below.)

The trial is currently ongoing, and the statisticians will soon see how well the simulation corresponds to what actually happens when the CT is used in the Emergency Department.
Description Image HarvardSuccessTable1.png
Open to Public Yes
BERD Watch Yes
BERD Watch Suggestions

Published_BERD_Watch Seeking Better Study Results Through Improved Study Designs
Additional Information

Disclaimer The views expressed within CTSpedia are those of the author and must not be taken to represent policy or guidance on the behalf of any organization or institution with which the author is affiliated.
Topic revision: r7 - 21 Jun 2012 - 13:28:31 - MaryBanach

Copyright & by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding CTSPedia? Send feedback