With our partners across Mount Sinai, the Mount Sinai ConduITS Informatics Core has enabled translational science activities, with a current focus in the areas of AI, data sharing and training. The Informatics Core enabled nearly 200 publications in 2023.
AI
To help guide researchers to navigate the rapidly evolving AI landscape, we have offered a new area of specialty through the weekly Digital Concierge sessions. We held several courses in the spring on how to use AI tools and GPUs for AI inference, training and data analysis. New modern GPUs have been integrated into the computational and data ecosystem.
To understand the best practices in the national community, Mount Sinai has joined the Coalition for Health AI (CHAI) and the Trustworthy and Responsible AI Network (TRAIN) with other leading academic medical centers. Within Mount Sinai, new committees have been formed that focus on the impact of AI on research, education, risk/ethics/policy and the clinic.
Dr. Girish Nadkarni was invited to present at the National Center for Advancing Translational Sciences (NCATS) Large Language Model Colloquium on November 20, 2023. Planning has begun on a regional symposium to be held in March 2025 on Enabling AI in Clinical Care (Girish Nadkarni, MD, Ankit Sakhuja, MD, Patricia Kovatch).
Data sharing
To help researchers determine feasibility for clinical trials, research projects and proposals, new cancer, pathology and image cohorts have been made available in the self-service Leaf EHR cohort query tool. In 2023, >3,000 cohorts were developed by >300 users with the self-service Leaf, Atlas or TriNetX query tools.
New multi-modal data products linked to the Mount Sinai Data Warehouse include one million de-identified pathology slides available on the computational and data ecosystem for multi-modal data analysis. We have also geocoded of all the current and historic patient addresses in the Mount Sinai Epic EHR and linked of the addresses to the Census Bureau’s American Community Survey. We have engaged with MSHS genomic testing vendors to setup a pathway to receive raw data files that will ultimately reside in the computational and data ecosystem.
Nationally, we have shared de-identified data with the National COVID Cohort Collaborative (N3C). We also share de-identified data with TriNetX for the identification of patients for clinical trials by pharmaceutical companies.
We have implemented a new process for researchers to access to clinical data systems, called the MSHS Researcher Access to Clinical Data and Databases Policy (https://mshs.policytech.com/dotNet/documents/?docid=51638).
ConduITS is supported by NCATS of the NIH’s CTSA Program. Any use of CTSA-supported resources requires citation of grant number UL1TR004419 awarded to ISMMS in the acknowledgment section of every publication resulting from this support. Adherence to the NIH Public Access Policy is also required.
