Abstract:
|
While advances in electronic health record infrastructure and statistical computing have made it possible to provide dynamic individualized decision support in a clinical setting, technical, institutional, and cultural barriers to fully realizing learning health systems remain. As a result, many statisticians have been frustrated to find that innovative modeling and prediction methods designed to support more intelligent health care are difficult to implement in the real world. This talk will focus on how statisticians can effectively advocate for data-driven health care and provide the necessary leadership to integrate statistically valid tools into the clinical workflow. We'll discuss an example from Johns Hopkins Medicine, where statisticians have developed and deployed a tool to assist in personalized management of prostate cancer.
|