Abstract:
|
Patients who have suffered a stroke are at a high risk of developing recurrent stroke and cardiovascular events, even under pharmacologic therapies. To study whether this residual risk is attributable to novel biomarkers, we conducted a case-cohort sub-study within a large double-blind placebo-controlled clinical trial. Biomarkers were measured in all patients who developed recurrent stroke or other clinical events, and in a random sub-cohort of the entire study population. This allowed us adequate sample size to examine all clinical events separately, assess their associations with biomarkers, estimate relative risk directly, and apply findings to the entire study. We used the Cox proportional hazards model to examine the association between clinical events and novel biomarkers, with/without adjustment for risk factors. Methods based on maximizing a pseudo-likelihood function (Prentice, Self & Prentice, Lin & Ying) were used to analyze this case-cohort study, differing in the definition of risk set and weights assigned to individual patients. The discrimination ability of risk factors and novel biomarkers to separate patients with events from those without was also assessed.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.