|
Abstract:
|
Biomarkers that identify patients who have an early clinical benefit from a treatment are important not only for patient survival and quality of life, but also for health care cost. Due to reasons such as biological variation or limited machine precision, biomarkers are sometimes measured with large errors. Adjusting for this error in the proportion of treatment effect explained by markers has been a research subject. The proportion of information gain (PIG), a new quantity to measure a biomarker's importance, has not yet been studied for with the measurement error case. We provide methods to account for the measurement error in the PIG calculation for continuous, binary and time-to-event outcomes. Simulation shows that the adjusted estimator has less bias and variability than the one ignoring the measurement error. Data from an osteoporosis clinical study is used for illustration.
|
