Medical Records

About

Summary Researchers in Prof. Nigam Shah’s laboratory have developed ATLAS (Advanced Temporal Language Aided Search), a fast and powerful health informatics search engine to enable high-throughput clinical evidence generation by rapidly building cohorts of patients from millions of electronic health records (EHR) and insurance claims. Currently, treatment guidelines are not available for the vast majority of the patients in a physician’s care. Furthermore, it would take days to weeks to search their own hospital’s records for similar patients’ outcomes to help guide decisions. ATLAS solves this search problem using an innovative patient-centric database architecture and a novel query language that supports text search, Boolean (AND, OR, NOT) as well as temporal commands (INTERSECT, UNION, SEQUENCE), searching of laboratory test results and demographic data.  ATLAS enables users to identify similar patients (e.g., age, race, symptoms, test results, medications) in a matter of seconds and then generate a succinct composite summary of their outcomes (e.g., complications, clinical course). This analysis could help determine effective treatment options for patients in ~6 hours by aggregating outcomes from a cohort and providing a report for clinical interpretation. ATLAS provides a mechanism to tap into large volumes of EHR and claims data to quickly link physicians to the information they need to make real-time decisions to provide the best care for their patients.   Applications Precision medicine - to harness existing EHR and claims data to find similar patients and analyze outcomes to help guide clinical decisions, use cases include: predicting prognosis based on specific patient criteria evaluating alternative interventions identifying quality-related outcomes stratifying risk  Clinical trial recruitment: quickly identify a group patients that meet the criteria for a given clinical trial randomize therapy for certain types of patients when there are no existing treatment guidelines   Advantages Fast searching - temporal query language and patient-centric database architecture accelerates cohort identification and analysis: ATLAS users can identify a patient cohort for analysis in seconds whereas the same query would take days to write with current query-language (e.g., SQL) reduces time needed to perform high impact data analysis to answer clinical questions from months to hours Versatile data model - designed to operate using a variety of data sources, including a commonly used public data model called OMOP Common Data Model version 5.