Eye Care Providers Must Improve Electronic Health Record Data Usability

Visual acuity (VA) data contained in electronic health records (EHRs) may not always be usable, and the percentage of unusable data varies across eye care subspecialties,  according to a study published in Ophthalmology Science. 

Researchers retrospectively reviewed EHR data from patient eye encounters (n=513,036) of 166,212 patients from 9 clinical centers between August 2013 and December 2015. The team extracted, categorized, recoded, and assessed data from 13 of 21 available VA fields, which accounted for 93% of the VA data.

Data was classified as either usable or unusable, and usable data was further categorized based on conformance to an internal data dictionary. Usable data were classified as an exact match, having conditional conformance, convertible conformance, or plausible but unable to be conformed. Non plausible data were deemed to be unusable. 

According to the report, 91.4% of entries contained usable data. Among the usable data, 1,196,720 (76.1%) entries were an exact match, 185,692 (11.8%) exhibited conditional conformance, 40,270 (2.6%) had convertible conformance, and 15,979 (1.1%) were plausible but not conformed. 

Usability varied according to subspecialty with the highest rates of unusable VA data  found among retina (17.5%), glaucoma (14.0%), neuro-ophthalmology (8.9%), and low vision (8.8%) practitioners. VA entries from comprehensive eye care (86.7%), oculoplastic (81.5%), and pediatric strabismus (78.6%) specialists yielded the highest proportions of exact matches.

“The magnitude of usable EHR VA data, including the VA entries requiring additional coding, vary by eye care subspecialty,” according to the researchers. “Further work will need to set external standards, further define data quality categories and understand the implications of unusable data on both the encounter and patient level study outcome.”

Study limitations include potential retrospective biases from the unmasked assessment, variability in measurement and documentation, evaluation of only 1 EHR platform, and failure to determine reasons for VA data unusability.