Dr. Jeff Gold: EHRs, Education, and Patient Care
Dr. Jeff Gold is a pulmonary critical care physician-scientist with an interest in EHR research. He serves as the program director for the Oregon Health and Science University’s (OHSU) Pulmonary Critical Care and Medical Critical Care Fellowships. In addition, he is the associate director of OHSU’s Adult Cystic Fibrosis Center.
What first inspired you to engage in EHR research? What aspects of EHR research do you think have been overlooked?
When I was director of the critical care services at Bellevue Hospital in New York, I would look at the electronic data of every patient that suffered an unexpected cardiac arrest in the ICU. I was looking for data that predicted this would happen, and based on that, I realized much of the information that was available to us was not being acted upon. When I joined OHSU to take over their fellowship programs, I saw the same issues with the EHR there. Critical information was not being processed appropriately. Safety issues were being overlooked because of all the clutter and inability to use the system.
I wanted to create a training opportunity for my fellows to teach them to better use the EHR, while still recognizing imminent patient safety issues. I wouldn’t say that EHR research itself has been overlooked, but rather the access of information as a source of research has been overlooked. Once we got EHRs, the influx of data made things even more confusing. Now, there’s no excuse not to have information; whether you can find it is a different story. That’s what our research is about — how can we find the tools to improve this.
You told us an anecdote about a situation you encountered early in your career where your care team came out of a weekly M&M without reporting any mistakes. However, in reviewing the most complex patient of the week, you discovered there were fifteen to eighteen unnoticed mistakes. How could the care team have spotted them earlier? How could we have stopped them from happening?
This was an eye-opening “aha!” moment for me, and it comes down to training. In order to prevent errors and get better at recognizing them, you have to practice with real-life patient scenarios built into EHRs. Bottom line, no one wants to make a mistake, but if you constantly make a mistake and no one points it out, you may not realize or improve your behavior. Simulation as an exercise changes this. In order to make training as realistic as possible and reproduce possible blind spots, we have to build cases that reproduce the complexity of our sickest patients.
It’s important to train doctors and students on electronic tools like any other crucial medical tool. At OHSU, we integrate the EHR in all aspects of the medical school curriculum from day one. We should be training our students on the care of patients with these tools, not the use of the tool by itself. EHR training is too generic; it is formed based around the tool, rather than the outcome of the patient. I believe training needs to be a patient-centered paradigm.
How do you hope your research will impact the way medical students and doctors are trained going forward?
As more people understand my research, I hope the EHR will be integrated into all forms of patient care training activities. For example, as rapid response and cardiac arrest teams undergo code training, I hope that EHRs are involved in the process. As we evolve medical training to train in all aspects in patient care, we should use EHR training in a similar robust manner. Eventually, there will be a library of high fidelity use cases that test whether people can both efficiently and accurately process EHR information. We want individuals to be able to overcome initial cognitive biases about a case to make the appropriate decision.
Have you implemented any best practices at OHSU that other organizations could also learn from?
We’ve defined a series of EHR competencies that individuals should be able to use. I also believe we need to integrate the EHR into graduate medical education training and physician and nursing training. However, I think the hardest thing that medical centers struggle with is the lack of resources. We were lucky that OHSU was able to supply these resources, but integration and design of these training activities is not easy.
Having worked with many tools over the years, I find that people struggle with the same issues in the parent interfaces and the new, novelty interfaces. In our high tech world, we value user satisfaction and efficiency. Safety and efficacy is determined by straightforward use cases. The device, whether EHR or interface, needs to be tested with the rigor of simulation test cases because they may be creating the same cognitive blind spots. We have to understand the blind spots of the tools while focusing on the care of the patient.
What has surprised you in your research? What results have validated what you already knew anecdotally?
What surprised me the most is how reliant we’ve become on electronic tools. The use of tools or the appropriate screens has become an end unto itself as opposed to a means to an end. Instead of saying “let me go see what’s happened to the blood pressure over the past few days,” they say “this is the screen where I see the blood pressure.” Then, they hope trends become apparent to them. Secondly, as EHRs become more widely used, groups other than physicians are using these technologies. For example, physician use of medication screens has gone down as pharmacists have taken over this responsibility. We’ve shifted the errors from one user group to another. Even though we’ve consolidated all the information in one place, we’ve siloed the use of the info among multiple user groups. The third thing that surprised me is the over-reliance on and the use of macro and import tools. Macros have huge blind spots, yet they’ve become the easiest way to find information.
What are other initiatives are you currently working on? What is the impact you’re hoping to achieve?
We recently received a grant from the Agency for Healthcare Research and Quality (AHRQ) to better understand how medical scribes interface with the EHR. Another unintended consequence of EHRs is their contribution to physician burnout. Because of this, scribes are used by 20% of physicians across the U.S. These are normally students without long-term medical training. Now that we’ve taken documentation and EHR functionality and given it to this “untrained group,” we want to see what they’re doing and how effectively they are doing it. We’re going to look at this over five years and collect a library of testing use cases so institutions will be able to access the accuracy of scribe-generated activity for further training tools.
