Study Recommends Enhanced Monitoring and Testing of Computerized Prescriber Order Entry
Thursday, March 19, 2015
With funding from NPSF Research Grant, researchers create taxonomy of CPOE errors
Boston, MA, February 19, 2015—Over the past decade, Computerized Prescriber Order Entry (CPOE) has become a standard tool in most electronic medical record systems. Yet even as the use of those systems has grown, many acknowledge the potential risk of introducing errors via health IT. In a new study published in BMJ Quality & Safety, researchers reviewed more than 63,000 errors related to CPOE, concluding that enhanced review, monitoring, and testing of systems is needed to reduce the risk of error.
Gordon Schiff, MD, a practicing internist and director of the Center for Patient Safety Research and Practice at Brigham and Women's Hospital in Boston, was principal investigator of the study, which was funded in part by a National Patient Safety Foundation (NPSF) Research Grant. Dr. Schiff and his co-investigators used data from the United States Pharmacopeia (USP) MedMARx medication error reporting system (now owned and administered by Quantros), which has collected more than 2 million medication errors from more than 800 hospitals and health systems. Beginning in 2003 MedMARx added “Computerized Prescriber Order Entry as a Cause of the Error” as one of the variables in their error reporting tool.
“One thing that surprised us was all of the effort that people had put into making reports across a 10-year period, but how little had been done to review the reports,” Dr. Schiff said. “There was a lot of very rich data with paucity of effort to learn from it.”
Once they identified errors related to CPOE, the researchers studied a subset to develop a taxonomy of what went wrong and why. They created 101 codes for what specifically went wrong (for example, “missing quantity or wrong number ordered”); 67 codes for why the error occurred (such as “typing error”); and 73 codes for potential strategies to prevent these errors in future (such as “dose range checking”).
To test the vulnerability of current systems to the types of errors reported, the researchers then enlisted typical users to test 13 different CPOE systems at 16 sites. They developed a set of 20 erroneous test orders based on errors reported to MedMARx and attempted to replicate the errors by entering the erroneous orders. Overall 79.5% of the erroneous errors were able to be entered, with 28% of these easily going through and 28% going through with only minor workarounds.
“CPOE systems have gotten better—we’ve all seen that,” Dr. Schiff said, “but we were surprised at the level of vulnerability that still exists.”
He said that a decade ago, clinicians saw CPOE as a big advance in patient safety. “We’re not saying it’s not very useful,” he said. “CPOE is necessary, but not sufficient on its own, to improve safety. We need to do more to monitor and improve these systems.”
“Clinicians know that CPOE systems are not perfect,” said Tejal K. Gandhi, MD, MPH, CPPS, president and CEO, NPSF. “Dr. Schiff and his co-investigators have demonstrated how critical it is for organizations to evaluate and monitor these systems with an eye toward optimizing their value.”
The study was published online January 16, 2015 and is available open access at