A full-function electronic medical record can’t always tell the whole story.
|
| Betty Rabinowitz, M.D. |
I recently saw a 56-year-old woman whom I have followed for hypertension for more than 15 years. Usually bubbly and very positive, she seemed quiet and subdued, and her blood pressure was higher than it had been in a while.
With her electronic medical record open on the screen in front of us, we reviewed her medications, renewing her prescriptions electronically. The system offered no drug interaction alerts. Following an on-screen reminder, I alerted her that her mammogram was soon due. Then, looking at the electronic flow sheet of her blood pressure, I commented that her blood pressure seemed a little high. Her eyes welled up as she told me that she was “not herself.” In the last week, her husband had lost his job. We spoke about her fears of losing their home and her disappointment in the company that had employed him for many years.
Finally, she talked about her hope that he would find a new job, and about the friends and family that had reached out to them with kindness and offers of generosity. She seemed to brighten, and there was a glimmer of her usual demeanor when she said, “I feel a little better. Could you recheck my blood pressure now?” It had normalized.
Automation’s Mixed Bag
In 2005, my organization, the University of Rochester Medical Center, implemented an ambulatory electronic medical record for approximately 500 physician users and close to 1,500 general users. As EMR implementations go, this was an intense, highly successful implementation with strong physician adoption, gains in safety and efficiency as well as cost reduction across the enterprise.
In medical literature, thought leaders in the EMR field often refer to the change brought on by implementation of an EMR as “automating medicine.” Automation means controlling a process by automatic means and reducing human intervention to a minimum. In many ways, this short definition reveals all the benefits and drawbacks of current EMR systems.
The benefits are easy to see: Many processes in medicine require a level of detail, accuracy and timing that are not optimally handled by human intervention. For example, automated alerts on high-risk drug interactions are a powerfully useful tool offered by full-function EMRs because the human brain is ill-equipped to remember and cite all the possible interactions among thousands of available medications. Providing reminders to physicians about indicated testing or an intervention is another example. There would have been no way for a clinician caring for more than 2,000 patients to consistently remember when a patient’s mammogram was due.
On the other hand, the drawback of automation is a direct result of attempting to reduce human intervention. There is no better example of this than in the notes produced using many EMR note modules. Many rely heavily on text templates or templates using discrete codified data. This method of documentation is efficient, and it creates documents that are searchable and accessible to data mining tools. Unfortunately, the notes produced tend to have a uniform look and feel with very little variation. For my patient, a template “hypertension note” would have done my interaction with her a grave injustice because the patient’s narrative was central to this encounter. Understanding her story dictated the medical decision not to increase her blood pressure medication in spite of the unusually high reading.
To counteract the drawbacks of automation, we have worked very hard at the University of Rochester Medical Center to find a balanced approach to medical documentation within our EMR. In our primary care network, we have adopted a documentation workflow that combines speech recognition software with the EMR’s note module. The voice recognition software is used for the narrative sections of the note that can’t be easily automated, and we rely on the EMR note module to generate those parts of the documentation that lend themselves to the use of automated templates. This hybrid form of documentation is the best of both worlds.
Unique and Efficient
Prior to implementing the EMR, the URMC primary care network relied on the medical center’s transcription vendor for most of its transcription needs. It is estimated that the primary care network alone, with around 115 providers, was spending approximately $1 million a year on transcription. With the implementation of speech recognition along with the EMR, the primary care transcription budget was reduced to practically zero.
As more physicians and health care systems implement EMRs, there is a growing risk that we will be lured into extreme automation of the medical encounter. Although we will do so in our continuous and commendable search for increased safety and efficiency in health care, we risk losing our patients’ voices along with our own. It is important that we create and disseminate electronic workflows that preserve what is unique about each clinical encounter—the patient’s story in his or her own words—along with capturing the physician’s voice sharing his or her diagnostic and therapeutic thoughts regarding each individual encounter. By combining a full-function EMR with integrated speech recognition, the most important parts of the patient encounter are preserved.
Betty Rabinowitz, M.D., is associate professor of clinical medicine, University of Rochester School of Medicine, New York.
GIVE US YOUR COMMENTS!
HHNMostWired welcomes your comment on this article. E-mail your comments to hhn@healthforum.com, fax them to Most Wired Magazine Editor at (312) 422-4500, or mail them to Editor, Most Wired Magazine, Health Forum, One North Franklin, Chicago, IL 60606.