EHRs have been implemented in hospitals across the country to help warn physicians of medication errors that could result in adverse drug interactions, allergic reactions, or death—but according to a study published in JAMA last month, the systems fail to detect more than 30% of these errors.

Cheat sheets: How hospital quality ratings work

The study

Researchers from the University of Utah Health, Harvard University, and the Brigham and Women's Hospital used a computerized provider order entry (COPE) tool to analyze the EHR performance of more than 2,000 U.S. hospitals from 2009 to 2018.

For the study, the researchers used the Leapfrog Group's CPOE EHR evaluation tool, which uses simulated scenarios to test EHRs ability to reduce adverse drug events.

As such, the researchers developed more than 8,600 hypothetical harmful scenarios based on real-world medication orders that resulted in a patient injury or death to test whether the EHR detected the errors either "in the form of alerts, messages, guidance," or "soft or hard stops," or if the order was allowed to be filled.

EHRs fail to alert researchers of up to one-third of medical errors

The researchers found EHRs' ability to detect and prevent adverse events did improve over the 10-year study period, Healthcare IT News reports. For instance, warnings about potential medication issues increased from 54% to 66% between 2009 and 2018. However, the researchers noted that the data means EHRs in 2018 failed to detect up to 33% of medical errors in the simulations.

"Despite broad adoption and optimization of EHR systems in hospitals," the researchers wrote, they still found "wide variation in the safety performance of operational EHR systems remains across a large sample of hospitals and EHR vendors." The researchers noted that the type of EHR did not have a big impact on scores. For instance, the researchers found that all nine vendors included in the study were used by at least one hospital that received a perfect safety score.

The bigger picture

While EHRs have been implemented in hospitals across the country largely to reduce human error that could lead to patient injury or death, the study results show that EHRs still fail to detect more than 30% of errors, according to David C. Classen, author of the study and a professor of internal medicine at University of Utah Health.

"In any other industry, this degree of software failure wouldn't be tolerated," Classen said. "You would never get on an airplane, for instance, if an airline could only promise it could get you to your destination safely two-thirds of the time."

But the results indicate that hospitals play a big role in the EHRs' ability to detect medical errors, especially since performance varied so widely between hospitals, USA Today reports. "Hospitals decide what drug-related decision supports to turn on within their systems," David Bates, study co-author and chief of the Division of General Internal Medicine and Primary Care at Brigham and Women's Hospital, said. "They have a great latitude around this."

However, some experts said the study's findings shouldn't be used to judge real-world EHR performance.  Allison Weathers, associate chief medical information officer at the Cleveland Clinic, said, "It's a helpful guidepost to enhance EHR systems but not an exact one-to-one correlation of safety performance,” she said (Jerich, Healthcare IT News, 6/2; Rodriguez, USA Today, 6/2; Commins, HealthLeaders Media, 6/1; Terry, Medscape, 5/29).