As basic research findings are translated into clinical practice, the promise of delivering personalizedmedicine based on genomic information is increasingly being fulfilled. Two themes arise in the area ofcancer: 1) Risk-stratification before disease manifests to optimize prevention and early detection strategies,and after disease manifests to inform treatment decisions; and 2) The impact of testing algorithms (e.g.,testing accuracy, sequencing, and timing) on subsequent management.Our objectives are to develop tools to address the clinical and economic information gaps regarding thevalue of genomic diagnostics and therapies for personalized medicine.
Specific aims are to:
Aim 1 : Develop a generalized framework using flexible, modular cost-effectiveness analysis models toaddress three key issues specific to personalized medicine in cancer:t Key issue #1: Risk stratification to optimize treatment and early detection, as illustrated by (a) UsingGEP in patients with BC to predict the risk of recurrence and guide treatment; and (b) Using genomictesting to risk-stratify and guide screening and surveillance in patients with CRC who may be affectedby Lynch syndrome and in their relatives. Key issue #2: Testing algorithms designed to guide targeted therapy, as illustrated by the impact ofalternative tests for HER2, test sequencing, and timing on guiding targeted therapy with trastuzumab(Herceptin , Genentech) in early-stage BC. Key issue #3: Impact of risk-stratification on individuals and their family members, as illustrated bythe example of testing for possible Lynch syndrome in relatives without cancer.
Aim 2 : Estimate the effectiveness and cost-effectiveness of using genomic information in the applications ofpersonalized medicine stated in Aim 1 by populating the decision models with unique data from ProgramProjects.
Aim 3 : Inform health policy and clinical decisions by characterizing the uncertainty in the model inputs,estimating the decision uncertainty, and calculating the potential value of undertaking additional research toreduce this uncertainty by using state-of-the-art value-of-information methods.
Showing the most recent 10 out of 43 publications
