The objective of the Biostatistics and Research Informatics Shared Resource (BRISR) is to provide collaborative and consultative services to Stanford Cancer Institute investigators and thereby add value in all phases of cancer research. Capabilities include high-quality biostatistics consultation on the use of standard methods, as well as innovation in statistical methods specifically developed to enhance the basic and translational research efforts of a discovery-oriented SCI. The BRISR operates under the direction of Drs. Phil Lavori, Tze Lai and Ying Lu. As a dedicated shared resource, BRISR benefits SCI investigators in need of statistical and analytic collaboration and consultation in their studies. Since 2009 the shared resource has served more than 130 SCI members representing all eight research programs. Significant methodological contributions by faculty and staff in this funding period include the introduction of new methods in high dimensional data analysis and genetic association studies, software tools for distributed computation (to avoid the need to aggregate data), as well as innovative designs for clinical trials, including Point-of-Care randomization methods, seamless Phase I/II designs, novel Phase I Bayesian designs, designs for biomarker-guided therapies, adaptive choice of subgroup, seamless Phase II/III designs and sequential randomization. The Shared Resource has the additional responsibility for expanding and managing the Stanford Cancer Institute Research Database (SCIRDB). This database allows SCI researchers to collect, integrate, and analyze de-identified patient data from the electronic health record (EPIC) and from a wide range of other data sources and provides consultative services to ensure its optimal use. Shared Resource staff also advise on strategic issues with statistical and informatics content, such as criteria for scientific review and development of databases and registries, as well as the re-engineering of clinical and research informatics across Stanford Medicine and its affiliated networks.
Showing the most recent 10 out of 322 publications
