We conducted a cost effectiveness analysis using Markov modeling to determine which of three treatment options approved for a specific disease (follicular lymphoma) would fare best over short and longer term horizons. In this analysis, we used data from prospective single arm trials and the efficacy and safety of the various therapies and incorporated health care quality of life data and cost data to help patients and decisions make clinical care decisions.
For another project- we recognized that while there is an abundance of electronic health care data that is encoded as discrete fields, it is often necessary to use non discrete field data for clinical research. We developed a natural language processing model that allowed us to decipher sites of metastatic breast cancer in patients treated at the Stanford Hospital. This algorithm can be used to help conduct retrospective studies involving large datasets to accurately ascertain clinical relevant information without manual chart review.
I am excited about the potential to integrate data science and AI into health care by improving a physician's ability to provide high level quality care and also a patient's ability to receive appropriate support throughout their cancer journey.