Use of EHRs in Hospital Systems Carries Patient Safety Concerns

A recent study published in the Journal of American Medical Informatics Association, identified safety concerns related to the adoption of electronic health records (EHR) in hospitals, as hospitals and other large medical institutions are pushing the adoption of EHRs in the U.S. and worldwide. This has resulted in a recent and dramatic shift in how are collect, store, and distribute electronic health data. It has also resulted in the creation of a new set of errors, even as the use of EHR reduces errors in other areas.

I have been an early adopter and "power user" of our facilities' EHR for a number of years now. Our private practice also uses a fully implemented EHR that is tablet-based, so things are advancing quite nicely to the dream of a fully portable medical record.

There are things that I really like about large scale EHRs, such as the ones used in hospitals, and things that I think are problematic.

I have my own thoughts on the four areas identified as problematic in the study:

1. System to system interface. This is what I consider the holy grail of the EHR. The promise of the EHR was to create a "portable" patient EHR that follows the patient independent of where it was created or what type of software was used to create it. This is the goal, and I can report that we have a long way in which to go to achieve that goal. Even within our own hospital's regional facility. A digital CT created at one hospital can't be viewed digitally on the EHR of another hospital. We are forced the load and view the study from a DVD as there is no system-wide solution for labs, diagnostics, patient notes, or other things, even though we are one organization. Multiply this by the thousands of healthcare systems in the U.S., and you get a good idea of the scope of the problem. We have come a long way in that patients, typically on the initiative of hospital systems, HMOs, etc., have better access to their health data electronically via patient portals on websites, but this doesn't do much for the busy medical staffs of facilities trying to track down patient information when it is collected outside of their immediate ecosystem.

2. Unmet display needs. This refers to what the researchers refer to as a mismatch between information needs and content display. A prime example of this is the typical user interface of a hospital-based EHR. The available content for display on the EHR user interface is staggering to say the least. Available data includes patient notes, nursing notes, notes from PT, OT, dietary and other services, labs, diagnostics, patient orders both completed and pending, just to name a few. While the user interface of our system allows significant customization and user organization of the displayed data, it is still easy to see how critical data can be missed or inappropriately ignored, based on the shear volume of data available to the medical staff in an inpatient facility. And the user customization aspects of our EHR are challenging and largely ignored by our medical staff as they don't have the time to experiment with the system.

3. Software modifications. Our EHR vendor is required to fix bugs, comply with changes in regulation, improve operability, and install upgrades to our EHR continuously. All in all, this is a good thing and the system continues to improve in accuracy and performance. On the other hand, many times, fixing one problem exposes an issue in another area. I experienced this first hand this past month when a medical staff patient-education module was implemented. It worked great and was customizable. However, the IT folks didn't tie it into the discharge process, so unless you printed it out yourself and delivered it to the patient, the patient and the discharging nurse never saw it.

4. Hidden dependencies in distributed system. Anything as complex as an EHR in a hospital environment is impossible to perfect and make error free. I'm reminded of the quote, "You can't get there from here." Many times, I and others on the medical staff are trying to do simple things, like simultaneously order and hold units of blood to be given at a future time, alert nursing staff to a time sensitive order, and provide continuity of orders and medications as patients are transferred between units, etc. There still seems to be a significant disconnect between the folks writing the software that we rely on, and the actual use of that software to get the job done - caring for patients at the trench level. In defense of the IT folks, there is most often a way in which to do it and our medical team is more than willing to help them identify that solution, but this costs the medical staff precious time. These dependencies are doing nothing but expanding and becoming more complex, lending a heightened sense of frustration to all users of the system.

I'm a firm believer in the EHR, and believe that when the end users have a good sense of the limitations of the EHR, and an understanding as to how errors occur, that it is overall safer for the patient and improves information gathering and decision support. However, we clearly have a long way to go to realize the promise of the EHR in the U.S., and need to keep studying the widespread implementation of the EHR from every angle.

This blog was provided in partnership with the American Academy of Physician Assistants.