Format

Send to

Choose Destination
Am Heart J. 2011 May;161(5):848-54. doi: 10.1016/j.ahj.2010.12.026. Epub 2011 Apr 7.

Refining clinical trial composite outcomes: an application to the Assessment of the Safety and Efficacy of a New Thrombolytic-3 (ASSENT-3) trial.

Author information

1
University of Alberta, Edmonton, Canada. paul.armstrong@ualberta.ca

Abstract

BACKGROUND:

Traditional time-to-event analysis assigns equal weight to the first event in the composite end point. This is counterintuitive to many stakeholders.

METHODS:

We constructed weights for components of a composite efficacy end point and a net clinical outcome by including metrics of safety and efficacy and compared the weighted with the traditional approach. Through an externally validated, clinician-investigator Delphi panel, the relative severity of individual components of a composite end point (30-day death, recurrent myocardial infarction, cardiogenic shock, and congestive heart failure) was determined. The net clinical outcome was assessed through the incorporation of risk thresholds for safety events (intracranial hemorrhage and major systemic bleeding). These weights were then applied to a modified analysis of the ASSENT-3 trial.

RESULTS:

The weights for the efficacy composite were as follows: death, 1.0; shock, 0.5; congestive heart failure, 0.3; and recurrent myocardial infarction, 0.2. The traditional time-to-first-event approach demonstrated a comparable advantage for both enoxaparin (enox) and abciximab (abx) over unfractionated heparin (P = .05), whereas the weighted efficacy analysis suggested an advantage for enox and similar outcomes between unfractionated heparin and abx (P = .2). The apparent advantage of enox was attenuated when the net clinical outcome was examined; the apparent efficacy of abx combination therapy was also diminished by an elevated major systemic bleeding rate (P < .001).

CONCLUSION:

This novel approach adds an alternative dimension to treatment evaluation by more efficiently incorporating the differential value of all events in each patient. Further development and application of this approach to future trial design and analysis are warranted.

PMID:
21570513
DOI:
10.1016/j.ahj.2010.12.026
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Elsevier Science
Loading ...
Support Center