Abstract
BACKGROUND: MD Anderson (MDA) strives to maximize learning from the > 44,000 patients seen annually. This has historically been challenging due to the diverse, often unstructured, and siloed nature of patient data. In 2018, MDA partnered with Palantir Technologies to integrate disparate clinical and research data sources into a single cloud-based platform enabling both technical and non-technical users to directly access and analyze data. METHODS: IT experts, Palantir engineers and end users collaborated to integrate 150 deidentified datasets from the existing data warehouse, Natural Language Processing pipelines, molecular reports, public data and manually curated data, representing data from > 1.5 million patients. Data integration was tested by answering 12 retrospective queries, of which 5 were specific to neuro-oncology cohorts. RESULTS: All 12 queries were resolved and within the neuro-oncology department we found the cohort identification process to be accelerated with users able to create reports backed by live data, with clear visibility into inputs, analysis and history. In addition, templates were created of common statistical workflows, empowering non-technical users to perform analyses (e.g. KM curve; Volcano plot). Secure access controls enabled increased collaboration between neuro-oncology and other departments. CONCLUSIONS: This work highlights the value of a data foundation and unlocks the potential for new studies, fueling scientific discovery within neuro-oncology and across the institution. To extend the impact we are piloting a combination of (1) Purposeful structured data capture at the time of care, (2) Use of logic to suggest commonly uncaptured but critical patient events such as progression, and (3) Auditing functionality to allow users to manually fill in gaps and correct data in a transparent manner. Upon completion, we anticipate the ability to increase our research throughput and scope.