1265 Welch Rd
Stanford, CA 94305
USA
David Ouyang, M.D.
Stanford Health Care
Stanford Medicine
ABSTRACT:
Many traditional data sets in healthcare, such as billing data and administrative outcome registries, are detached from direct patient care. This limitation can make analysis fraught, as these data sets can be insufficiently granular, skewed by patient selection, or biased by mixed incentives. With the wide-spread adoption of EMRs, a tremendous amount of meta-data is generated in through usual clinical practice. Such passively generated data can answer focused clinical hypotheses and explore previously unanswerable questions. In this talk, I present the use of physician EMR time-stamps and intra-provider text communication as organic data sets to explore physician behaviors at Stanford. Using 1 million text pages, we show variation and bias in provider usage of brand name vs generic name medications. With more than four million EMR actions, we characterize physician workhours and it’s correlation with patient outcomes.