Skip to content Skip to navigation

What Does the Evidence Say about EHRs and Patient Safety Data?

December 19, 2014
| Reprints
The past 12 months were memorable for health IT enthusiasts. There were major mergers of vendors and health systems, forward AND backward shifts in regulatory action, noteworthy healthcare hacks and data breaches, movement at the top of the Office of the National Coordinator for Health IT (ONC), and every week, it seemed the industry was somehow in the mainstream news cycle.
I won't get too far into it since I'm planning on blogging about the year in review next week, but one element I'll touch upon for this blog that gained a lot of attention this year was the role EHRs play in patient safety. If you recall, The Boston Globe featured an article over the summer that essentially pointed the finger at an electronic health record (EHR) system for a significant adverse event that led to a patient's death. 
The article basically called out the government for not requiring mandatory reporting on injuries, deaths, and unsafe conditions from EHRs. While the article was fairly biased, as I had written then, it very much encapsulated many people's feelings on EHRs and patient safety. Just recently, the ECRI Institute, a nonprofit research organization that aims to improve care, named a couple of EHR-related errors among the top health technology patient safety hazards.
Even though the article was misguided on several fronts, I do think a collection of adverse events, as related to EHR miscues, would be a good thing. It would help the industry better understand pain points in systems. As I said then, EHRs aren't perfect. No sane person would try to argue otherwise. 
The good news is that there seems to be progress on this data collection front. This past week, the ONC put on a webinar, "The Role for the EHR in Patient Safety - What does the Evidence Tell Us?" It featured two speakers, Bill Marella of the Pennsylvania Patient Safety Authority and Gerard Castro, Ph.D of the Joint Commission, who have led extensive research efforts in determining the role EHRs have played in adverse and sentinel events. 
So what does the evidence tell us? Well the fact that ONC put this on is a bit of a spoiler, if you ask me. Also, if you've followed the evidence-based research of the Pennsylvania Patient Safety Authority, you know that the nonprofit has proven before that the role of EHRs in being the cause of adverse events is not as dramatic as some would make it out to be.
In fact, on average, Marella revealed that acute-care facilities with an advanced EHR system had seen a net benefit in overall patient safety events from 2005 to 2012 (after HITECH was enacted). The nonprofit used the Healthcare Information Management and Systems Society (HIMSS) database, which he admits is incomplete. He also says there with limitations with observational data. Still, when controlling for different variables, it was clear that the so-called problems caused by EHRs on patient safety were not only overstated, but paint a reverse picture of what's actually happening. Even when the EHR is mentioned, Morella said during the webinar, the event wasn't always about the system. 
Later in the webinar, Dr. Castro unveiled his research on the impact EHRs have on sentinel events. Sentinel events result in a death or severe loss of function. In their analysis, Castro and co. found roughly 3,500 sentinel events nationwide in the Joint Commission over a four-year period. Of those events, 121 were related to health IT (most of which were in the EHR). Let's be honest, approximately 3 percent is not exactly a harrowing figure.  
I can hear what the dissenters are saying. That's 121 too many. They're right. Of course, there is room for improvement, as both researchers said during the webinar. Alert fatigue is still a serious issue, Morella said. Castro's research revealed that most health IT events were caused by issues with system design and user interface. Although, he noted that the vulnerabilities within systems that led to a sentinel event were multiple. 
The biggest problems, Castro surmised, could be solved by matching technology with workflow. Indeed, this is a recurring theme we hear all the time in health IT discussions. 
Yet, listening to the webinar I couldn't help but think of the ridiculous hysteria that we saw in 2014 on this issue. The Boston Globe article was the most prominent example, but as I said, it wasn't the only one. It reflects the feelings of many health IT stakeholders. In fact, there have been Congressmen who have made this one of their issues. 
As we head into a new year, let's make 2015 the year we shoo away the hysteria and instead, lean on comprehensive research efforts like the work being done by Morella and Castro, to guide us forward.