Cancer Control Research
5P01CA053996-23
Prentice, Ross L.
STATISTICAL METHODS FOR MEDICINE STUDIES
AbstractThe aim of this application is the development and evaluation of improved
statistical methods for the design, conduct and analysis of a range of
medical and biomedical follow-up studies. There are three projects and
one core unit.
Project 1 is concerned with statistical methods for disease prevention
and risk factor intervention trials. The development and evaluation of
flexible and comprehensive methods for the analysis of multivariate
failure time data, as arises in trials with multiple disease outcomes and
in group randomized trials, is a particular emphasis area. Methods for
assessing overall benefits versus risks in prevention trial monitoring
and reporting is a new emphasis area, while methods for community
intervention trials is a continuing emphasis area. There will also be
emphasis on the methods for extracting additional information from
prevention trials, including methods for identification and use of
surrogate and auxiliary endpoints, for explanatory analyses, and for
extending trial results by using related observational data.
Project 2 is concerned with statistical methods for epidemiologic
studies. There will be an enhanced emphasis on genetic epidemiologic
methods, including the development of a multistage design for family
studies and of related estimating equation- and frailty-based analysis
procedures. There will also be continuing emphases on methods for
accomodating covariate measurement errors in cohort and case-control data
analysis, and/or methods for the design and analysis of aggregate data
studies of disease rates and risk factor survey data, as well as a new
emphasis on nonparametric methods for relative risk estimation.
Project 3 is concerned with statistical methods for the efficient design
and analysis of clinical studies. A particular emphasis will include
strategies for multi-arm trials, including factorial designs and ordered
alternative designs, and on monitoring issues in the design and analysis
of Phase III (comparative) trials. Other emphases will include the
development of exploratory methods in survival analysis, multistate and
multivariate survival analysis methods, and other topics, such as study
of the small sample properties of the logrank test and group sequential
strategies for testing biological specimens, that are pertinent to the
design or analysis of clinical studies. A small administrative and
computing core will serve to coordinate, standardize and facilitate the
methodologic work.
The proposed developments have the potential to increase the efficiency
and reliability, and to decrease the cost of various important types of
medical/biomedical studies, and to enhance the use of routinely collected
biomedical data.
|