Speakers: Monica Gaddis, PhD

Monica Gaddis, PhD

Assistant Teaching Professor
Research Director
Department(s) of Biomedical and Health Informatics, Emergency Medicine
University of Missouri Kansas City

 

Monica Gaddis, PhD is an Associate Professor with the Department of Biomedical and Health Informatics (DBHI) at UMKC School of Medicine. Within the DBHI, Dr. Gaddis teaches Applied Biostatistics (Beginning and Advanced), and is the lead instructor for the Weekly Department Multidisciplinary Seminar. She also oversees and conducts project analysis for the Neuroscience Unit research project, which educates year three medical students in the basics of the research process. Outside of the DBHI, she serves as the Research Director for the Department of Emergency Medicine (DEM), Truman Medical Center – Hospital Hill, Kansas City, MO. Dr. Gaddis is involved as a PI or co-PI in several DEM research projects, leveraging both prospective clinical and retrospective secondary dataset analyses, addressing various questions regarding the diagnosis, treatment and outcomes of patients with sepsis. Dr. Gaddis is also a Co-PI for several Emergency Medical System research projects within the DEM, assessing policy and outcomes regarding delivery of pre-hospital care. Further, Dr. Gaddis oversees the research and scholarly activity of the Emergency Medicine Residents, assists the DEM faculty with their research and produces and co-directs the monthly Department of EM Journal Club. Finally, Dr. Gaddis is a recognized expert in emergency medicine research program development for areas with limited resources, having spoken internationally on three continents.

The Electronic Health Record: What You Ask For May Not Be What You Get

The Electronic Health Record (EHR) can provide critical data for medical research. Clinical trials, outcomes research, longitudinal research and even quality improvement projects (QI) can be informed by EHR data.  Whether from large secondary data sets or directly from hospital systems, an individual hospital, a clinic or an office, data from the EHR can provide a wealth of information for the researcher. However, identifying patients or participants for studies is fraught with difficulty and potential error. These challenges are well documented in published literature examining the use of ICD-9 and ICD-10 codes for identification of study participants. There is a significant disconnect between clinicians, administrators and researchers regarding these codes, in terms of use and understanding.  Clinicians provide clinical care and must state a patient’s diagnosis accurately. However, they are not incentivized or even often able to select diagnoses with the level of detail that EHR researchers may need or want. Administrators have a host of responsibilities including billing and QI and researchers are seeking to operationalize valid and meaningful definitions of diagnoses, procedures, etc., from this data aggregated for the purpose of clinical care and administration. 

As the transition to ICD–11 codes occurs and the number and complexity of available codes increase, this disconnect is likely to only be magnified. This talk will inform non-clinical researchers regarding the selection of diagnoses by clinicians, using real life examples. With this, researchers will better understand why the data that is obtained may lack the precision that is desired.