Tom Lasko, MD, PhD
Tom Lasko is an MD with a PhD in Computer Science and an interest in large-scale machine learning from EMR data. In a previous life as a Google Software Engineer, he developed a completely data-driven computational diagnosis engine that used the entire World Wide Web as its source data to compute a list of diseases associated with any arbitrary search string (such as a list of symptoms), and did not rely on the usual simplification of assuming conditional independence of symptoms given a disease. At Vanderbilt, he is interested in building computational tools that improve clinical diagnosis, prognosis, or treatment decisions. One research direction focuses on providing visual abstractions of EMR data that reduce the cognitive load of searching the record to answer clinical questions. Another direction focuses on population-scale data-driven discovery of disease phenotypes that align more closely with actual pathophysiologic mechanisms than our current, expert-driven models of disease do.