Awarded in 2017
Home Department: Computer Science
Faculty Advisors: Mohsen Bayati (Operations, Information, and Technology) and Michael Baiocchi (Medicine - Stanford Prevention Research Center)
Research Title: Adaptive Experimental Designs for Clinical Trials
Research Description: The medical community relies on randomized controlled trials to evaluate treatment effects. However, these trials may unnecessarily over-experiment because they do not utilize partial outcomes information. This issue is exacerbated by precision medicine, where treatments are targeted at subgroups. By applying computer science and statistics techniques to clinical settings, Carolyn will develop adaptive experimental designs that strategically adjust treatment assignment probabilities during the trial as outcomes information becomes available. These designs will (i) improve patient outcomes while (ii) identifying high-dimensional heterogeneous subgroups with statistical significance. Carolyn will retrospectively evaluate the designs on Women's Health Initiative clinical trials and hopes to implement them in cardiothoracic surgery and pediatrics at Stanford hospitals.