Grant Number:	5R01CA018332-24 Interpret this number
Primary Investigator:	De Mets, David
Organization:	University Of Wisconsin Madison
Project Title:	Statistical Problems in Cancer Research
Fiscal Year:	1998

DESCRIPTION: Therapeutic and prevention clinical trials in cancer and other major diseases with mortality or irreversible morbidity are typically monitored to detect early evidence of benefit or harm for ethical and scientific reasons. However, repeated interim analyses using conventional statistical methods will increase the likelihood of false positive claims of treatment effect. In 1983, Lan and DeMets (Biometrika) extended earlier work of Pocock (1977), O'Brien and Fleming (1979) and others by proposing a flexible group sequential plan using an alpha spending function. In the previous proposal, we extended the application of the alpha spending function to repeated measure designs using linear mixed effects regression models and ordinal regression moods. Trials early may exaggerate treatment benefits or harm even though the hypothesis of no effect has been rejected. In this proposal, we further evaluate the degree of bias in the estimate of treatment effect and propose correction estimators for the linear mixed effects model and the proportional hazards model in survival. However, use of such models in either a sequential or non-sequential setting require that certain model assumptions be met, such as linearity or proportional hazards. We also propose developing sequential versions existing goodness of fit tests for these models such that inadequacy of the models can be identified early in the interim monitoring. Both the bias correction estimates and the sequential goodness of fit tests will be evaluated for common spending functions, various patterns of interim analyses, size of trials and robustness to model assumptions.