EHRs threaten patient safety long after startup
Executive Summary
Research shows that electronic health records (EHRs) can threaten patient safety long after they are first implemented. The cause is a mix of human and technological errors.
Healthcare organizations must carefully monitor EHR-related safety issues long after the first implementation.
The time after an EHR-system upgrade is particularly risky.
Risk managers should seek out harmless-sounding EHR problems that might lead to uncovering a safety threat.
Electronic health record (EHR) patient safety issues persist long after the "go-live" date when everyone expects there to be a learning curve and a period of glitches, according to new research in the Journal of the American Informatics Association. Risk managers should heed the warning and persist in monitoring EHR-related safety issues long after it seems the system has been successfully integrated.
Sophisticated monitoring systems are needed to unearth the complex mix of human and technological causes behind these problems, say the authors from the Department of Veterans Affairs (VA). EHRs can improve the quality of patient care, they say, but recent evidence suggests that they also can prompt new patient safety concerns when computer glitches cause clinical decision support to suddenly stop working, for example, or when network outages occur.
Many of these problems are complex and multifaceted, and they are difficult to detect and prevent, says Hardeep Singh, MD, MPH, patient safety researcher at the Veterans Affairs Center for Innovations in Quality, Effectiveness and Safety in Houston, TX, and one of the study’s authors. To better understand the nature of these patient safety concerns, Singh and his colleagues reviewed 100 closed investigations involving 344 technology-related incidents arising between 2009 and 2013 at the VA. (An abstract of the study is available online at http://tinyurl.com/oeqsgka.) The VA adopted EHRs in 1999 and is a leader in patient safety and the use of health information technology. It runs a voluntary reporting system for health information technology safety reporting and analysis. The authors looked at safety concerns related to technology itself as well as human and operational factors such as user behaviors, clinical workflow demands, and organizational policies and procedures involving technology.
Three quarters of the investigations involved unsafe technology, while the remainder involved unsafe use of technology. Seventy percent of the investigations identified a mix of two or more technical and/or non-technical underlying factors. The most common types of safety concerns were related to the display of information in the EHR; software upgrades or modifications; and transmission of data between different components of the EHR system.
More often than not, Singh says, the concerns arose as a result of the complex interaction between non-technical dimensions, such as workflow, and technical dimensions, such as software/hardware and the user interface.
Long after introduction
Interestingly, the serious patient safety errors still were present long after the go-live date for the EHR, not just in the initial introduction phase.
"EHR-related concerns were still happening many, many years after the EHR had been implemented," Singh explains. "Unless you have a robust infrastructure to detect and analyze these events, you really can’t learn a lot. You have to have that in place so you can learn from the problems you’re experiencing."
Improving the detection of EHR-related threats is of paramount importance, because that’s the first step to correcting them, says another of the study’s authors, Dean Sittig, PhD, faculty member at the University of Texas Health Science Center at Houston, who specializes in clinical information systems and clinical decision support. "With electronic health records, I don’t think people understand that some of the things they do are a lot more dangerous than others," Sittig says.
The patient safety threats come in many varieties. Some types of computer screens can be more difficult to read in certain lighting, and some can be more or less sensitive to touch inputs. On a laboratory results screen, for example, Sittig notes that the digits might be too small or the high and low values appear in the same color, which makes it difficult to tell the difference. (See the story below for more on how errors can creep into an EHR.)
Upgrading an EHR system creates a period of high risk, Sittig says. People in the IT department brace for this period because they know they will receive many calls about errors and breakdowns. "But I’m not sure everyone else in the hospital knows that," Sittig says. "They may think that the upgrade means things are better now, and they don’t realize that they should be on high alert for conflicts and errors. The risk manager may not even know the upgrade is happening."
Risk managers should be aware of upgrades, and they should plan to actively investigate the effects rather than waiting for someone to report patient safety issues, he says. Go to the clinicians on the floor and ask them how the upgrade is going, how the system is different from before, and what problems they are seeing now.
Don’t only ask if there are any patient safety issues as a result of the upgrade, Sittig says. That question will keep you from hearing about how everyone is having trouble printing from the EHR now, for example. If all the printers are now misconfigured for the EHR, you should wonder what else was changed. What else is the EHR not communicating with?
"Risk managers can play a pretty strong role in this because many of these concerns are not easily detected," Singh says. "We think that the reported incidents are only about 5% of the errors, but if we have a more robust detection strategy that involves the risk manager, we might be able to uncover far more EHR-related concerns than we currently know about." (For more information on the role of the healthcare risk manager, see story at the bottom of this page.)
- Hardeep Singh, MD, MPH, Patient Safety Researcher, Veterans Affairs Center for Innovations in Quality, Effectiveness and Safety, Houston, TX. Telephone: (713) 794-8601. Email: [email protected].
- Dean Sittig, PhD, School of Biomedical Informatics, The University of Texas Health Science Center at Houston. Telephone: (715) 500-7977. Email: [email protected].
