[Skip to Navigation]
Invited Commentary
July 1, 2021

The Rigorous Work of Evaluating Consistency and Accuracy in Electronic Health Record Data

Author Affiliations
  • 1Department of Learning Health Sciences, University of Michigan Medical School, Ann Arbor
  • 2Department of Internal Medicine, University of Michigan Medical School, Ann Arbor
  • 3Institute for Healthcare Policy and Innovation, University of Michigan, Ann Arbor
  • 4Department of Ophthalmology and Visual Sciences, W. K. Kellogg Eye Center, University of Michigan Medical School, Ann Arbor
JAMA Ophthalmol. 2021;139(8):894-895. doi:10.1001/jamaophthalmol.2021.2042

Ophthalmology researchers often seek out collaborations, especially when a diverse study population will help the researchers create generalizable results. Now, with the increasing availability of electronic health record (EHR) data, 5 research teams in 5 health care networks are able to integrate data around 1 research question without cumbersome manual medical record review. They can access and generate large patient cohorts by combining forces. Large EHR cohort data are especially helpful to answer questions about disease and symptom prevalence, practice variation, and clinical outcomes. This large amount of data is also essential if the research team wants to create clinical prediction models, which generally require larger cohorts than clinical trials.1 However, the relative ease of combining EHR data can create a false sense of confidence about the simplicity of study methods and technical processes. Health services and data science researchers recognize that studies leveraging EHR data make a number of necessary assumptions that, when unmet, may lead to spurious results. Some of the core assumptions exist around the consistency and accuracy of clinical data.

Add or change institution