Global web ALERT button

Electronic Health Records for Research

 Click the arrow at the bottom of the video to navigate between topics or click to view the video in a new window.

Notes

Recorded at the Texas Regional CTSA Consortium Quantitative Seminar Series on February 21, 2018. Presented by Gulshan Sharma, MD, MPH, Vice President, Chief Medical and Clinical Innovation Officer; Director, Division of Pulmonary Critical Care & Sleep Medicine; Sealy and Smith Distinguished Chair in Internal Medicine, University of Texas Medical Branch. This event was sponsored by the UTMB Office of Biostatistics, Clinical & Translational Science Awards, and the Center for Large Data Research and Data Sharing in Rehabilitation.

Learning Module Notes Modules

  1. Overview (00:00 – 00:42)
    • Current Electronic Health Record (EHR) Landscape EHR use nationally, internationally, regionally, locally, UT system Future Directions
  2. Current Landscape (00:43 – 07:58)
    • 50% of the Country’s Electronic Health Records are dominated by: EPIC (Major academic hospitals) & Cerner (community and teaching hospitals)
    • Data aggregation includes claims-based analysis, registry, clinical trials, Electronic Medical Records (EMR)
    • EMR are a record of medical care, maintain by one healthcare organization and an authorized clinician. It is primarily used for billing purposes. An example is EPIC. However, Electronic Health Records (EHR) are available longitudinally in the record, across different organizations. EMR are primarily used for reimbursement, despite additional cost but they were originally designed to indicate “level of visit” (see 32:45)
    • The Medicare Access and CHIP Reauthorization Act of 2015 (MACRA) and Merit-Based Incentive Payment System (MIPS) are physician quality measures.
  3. National (07:59 – 12:45)
    • eMERGE (Electronic Medical Records and Genomics) facilitates genome wide studies.
    • SHARPn (Strategic Health IT Research Program) is working on a better way to identify data and create unified data dictionary.
    • EHR4CR (Electronic Health record for Clinical Research) is predominantly in Europe and it identifies phenotypes for clinical trial criteria.
  4. International (12:46 - 15:31)
    • Health Facts is a longitudinal data warehouse. Observational Health Data Sciences and Informatics (OHDSI) is another international tool.
  5. Local (15:32 – 31:38)
    • The University of Texas Medical Branch (UTMB) is using EMR in research.
    • Issues include that imaging, billing, and EKG results are in a separate system from EHR and they are in PDF format which does not have discrete fields and cannot be studied. One must take caution with how data fields are pulled to be sure a patient isn't counted twice.
    • Take away: Every EHR is configured VERY differently. UTMB Discover UT Health Care Enterprise Healthconnect Service Area
  6. Future Directions (31:37 – 46:34)
    • The ultimate goal is to make EHR data useful for research. Interact with clinicians through feedback reporting. EHR-Based Registries, NLP, Episode of Care.
    • The local goal is to keep all information in cloud locations.
    • A benefit of EMR is that it provides a large sample, lab results, clinical details, near real-time analytics, natural language, and is inexpensive.
    • Bad EMR has incomplete variables, out-of-network utilization, sample attrition over time, incomplete integration of various reports within the health system, is more expensive, and natural language processing not validated.
  7. Questions to ask before going into EHR medical research: What additional info do I need that isn’t available in administrative claims? How is information captured in EHR: discrete fields, free text, or interface? Is documentation across EHR similar? Is there a validated NLP for the free text? Electronic Health Records are here to stay until Comprehensive Health Records become available.
  Learning Modules