Gilles Clermont, from the University of Pittsburgh and VA Pittsburgh. I am a critical care physician interested in patient-centered treatments and outcomes in the acutely ill. We have been collecting EHR data for several years now, ~180,000 ICU admissions. These data are very similar to MIMICIII in scope and granularity. We have been conducting analyses and published on this dataset pretty extensively, most recently in JAMA. We are assessing the potential usefulness of mapping to OMOP, as we have already mapped to MIMIC and eICU, the other publically available large repositories of ICU data. It would be outstanding if other participants in this network were aware of similar non-publically available datasets, so we could contemplate analysis of, and analytics on, several datasets, as to assess robustness of analyses and models. I have a use case in mind.
Continuing the discussion from Welcome to OHDSI! - Please introduce yourself:
Hello! I’m a Principal Data Scientist at CDPHP, a payer in the New York Capital region. We are undergoing an exciting transformation in data integration and analytics that have lead a few of us to explore the OMOP CDM as a means to achieve both our goals and serve the broader community. I’m looking forward to learning through participation and offering my thoughts on applying data scientific approaches to understanding and transforming healthcare data. We are actively engaged in using task appropriate tools from Statistics, Machine Learning, and Augmented Intelligence to enable a focused yet flexible technological ecosystem to enable solutions for our business units, our members, and our providers.
Hello! I’m Sergei. I work as a medical knowledge engineer in a healthcare company. Last year I spent working with medical datasets and medical ontologies for building AI model for symptom checker. Now I try to dig into NLP for medical data extraction and coding.
Hello! My name is Pei Lin, Business Insights & Analytics at Astrazeneca, US Oncology. My work interests in patient persistence and outcome research, patient-level (or provider-level) prediction/segmentation and population-level evaluation. In the oncology space, I am particularly interested in how to standardize bio-marker status from various lab testing data, or if we use reverse-engineer patient’s treatment pattern to categorize patient’s bio-marker status when test data is missing or not-done. Very happy to joining this forum and looking forward to closing the information gaps in the Oncology subgroup.