Nathanael Fillmore, PhD

Biography

Nathanael R. Fillmore, PhD, is Deputy Director of the VA Boston Informatics Group at the VA Boston Healthcare System, Director of the Coordinating Center for the VA's Partnership for Real-world Outcomes, Findings and Observations Using National Data on Veterans (PROFOUND-VET), and Site Director for the VA/NCI Big Data-Scientist Training Enhancement Program (BD-STEP). He also serves as Assistant Professor of Medicine at the Boston University Chobanian & Avedisian School of Medicine and Lecturer, Part-Time, at Harvard Medical School, and is a Faculty Affiliate of the Rafik B. Hariri Institute for Computing and Computational Science and Engineering at Boston University. His work applies computational science to precision oncology and the learning healthcare system, using artificial intelligence and causal inference methods on large-scale clinical, genomic, and imaging data to improve cancer treatment and prevention.

Dr. Fillmore leads a data science research group whose portfolio spans four connected areas: AI-based prediction of cancer risk and outcomes; real-world evidence on treatment effectiveness using causal inference and comparative effectiveness methods; clinical decision support delivered through the electronic health record to enable a learning healthcare system; and tools to extract, curate, and share clinical, genomic, and imaging data for research. His group has developed and validated phenotyping algorithms and outcomes-ascertainment pipelines for VA cancer cohorts across lung, prostate, colorectal, pancreatic, renal, and hematologic disease, combining structured records with information extracted from pathology, radiology, and clinical notes through rule-based, machine learning, and large language model methods.

This work has had clinical and international reach. Real-world evidence generated through PROFOUND-VET was incorporated into the VA Oncology Clinical Pathway for Prostate Cancer. Models and measures his group has contributed to, such as EHR-CAT for cancer-associated thrombosis and the VA-FI-10 frailty index, have been validated nationally and internationally. Applications developed and deployed by his group are in use across the VA to identify patients for clinical trial enrollment and cancer screening. He has published more than 100 peer-reviewed articles, including work in the Journal of Clinical Oncology, JAMA Oncology, and Nature Medicine, and his research has been covered by outlets including The Atlantic and the Washington Post.

Dr. Fillmore trained as a computer scientist, earning his PhD and MS in Computer Sciences from the University of Wisconsin-Madison and completing a postdoctoral fellowship in biostatistics and computational biology at the Dana-Farber Cancer Institute and the Harvard T.H. Chan School of Public Health. He received his BA in Philosophy and Classical Languages from Carleton College. He has mentored more than 30 trainees and contributes to national committees on data science and artificial intelligence infrastructure.