When treating a patient with dementia, electronic health records fall short
By Regina Harrell and Pulse
December 23, 2013
I am a primary-care doctor who makes house calls in and around Tuscaloosa, Ala. Today my rounds start at a house located down a dirt road a few miles outside town.
... We chat about the spring garden and the rain, then we move on to Mr. Edgars’s arthritis. Earlier on in his dementia, he wandered the woods, and his wife was afraid he would get lost and die, although the entire family agreed that this was how he would want it.
... We talk about how anxious he grows whenever she’s out of his sight and how one of his children comes to sit with him so that she can run errands.
The omitted lines of this physician's encounter are poignant. Read them at the Wash. Post link above.
... When I get back to the office, I turn on the computer to write a progress note in Mr. Edgars’s electronic health record, or EHR. In addition to recording the details of our visit, I must try to meet the new federal criteria for “meaningful use,” criteria that have been adopted by my office with threats that I won’t get paid for my work if I don’t.
The "meaningful use" criteria, I point out, are an unproven experiment, decided upon by committee. They are not evidence-based. Physicians are, in effect, being threatened with nonpayment as part of the experiment. They have become experimental subjects themselves, for free, and without true informed consent.
Under “History of Present Illness” (HPI), I enter “knee pain.” Up pops a check-box menu: injury-related (surely the chronic wear on Mr. Edgars’s knees from his work as a farmer is some sort of injury, but I don’t think that’s what the computer programmer had in mind), worsening factors (I know of none that apply, since he couldn’t give his own history), relieving factors (there’s no check box for a tired, sleep-deprived wife who’s purposely keeping the dose of acetaminophen low) and so on. Nothing fits, so I exit the HPI and type in “follow-up” (f/u), for which my EHR doesn’t have a pop-up menu. It yields only a blank screen.
As a medical student, I was forbidden to use paper templates to "remind me" of what I needed to record in an H&P or progress note. The "table of contents" had to be learned and applied from memory. The continued patronization of physicians, nurses and other clinicians via templates like this, and the resultant de-skilling, time-wasting and other deleterious effects, is harmful to quality care. This physician comments later that without EHRs and "meaningful use," she could see twice as many patients.
I type the Edgars’s story in my own words, so different from the computer-speak generated by the check boxes. I move on to the Review of Systems — another pop-up menu.
Translation: handwritten or typed narrative is meaningful; computer-generated prose is largely "legible gibberish", i.e., garbage. I note that the EHR output in a recent ED visit of my own would have received a failing grade for documentation quality when I was in medical school - and I would have had serious words with a trainee who'd written such sloppy prose when I was a senior resident and attending.
I used to simply write “patient is an unreliable historian” at the beginning of this section, but the computer doesn’t understand that this statement could apply to the entire review. [Actually, the designers and programmers who believe they are "revolutionizing" the field didn't understand the real world of clinical medicine - ed.] Using a template, it generates a page of 13 sentences, one for each body system, and, under each sentence, the option “Positive: Other: unreliable historian.” [Which much then be clicked 13 times - ed.]
Sometimes I wonder if it is disrespectful to a patient to say 13 times in one progress note how unreliable a historian he or she is, but I remember that this is great data to mine for research, so I plug on.
This is not just "disrespectful", but a waste of clinician time.
Under “Physical Exam,” there is a template for geriatric patients. I pretend that the computer-speak it generates creates logical sentences, although I know better. In the check boxes, a person can be oriented to person, place and time, or not. Mr. Edgars is oriented to person and place; he knows that he is with his wife and at home, and is happy nowhere else. He no longer cares what year it is. There isn’t a check box for that.
Obviously this "feature" was not even run by medical students, who know that orientation is not "x3" or "none" as the only pertinent options.
Technically speaking, this represents inadequate and insufficiently granular data modeling ... a task I wrote years ago that requires the highest levels of clinical and biomedical informatics expertise, not computer or programming expertise. There are likely many other examples of poor data modeling in this EHR.
I remember that I must go back to “Social History” and document tobacco use. It occurs to me that if you have not tried tobacco products by your 80th birthday, you are unlikely to suddenly change your mind. Especially when you can’t remember where the store is to buy them. So I slog through the series of check boxes for “never smoker,” an extra six mouse clicks.
More wasted time.
After 15 minutes, the note is finished. And on goes my day of house calls, five in all. There aren’t enough physicians to see all the homebound patients in my area, so I try to visit as many as I can safely care for.
At day’s end, I review my meaningful use. I spent more time checking boxes than talking to patients and their families.
I could see twice as many patients if I could write their notes at the bedside while visiting with them.
In other words, in this underserved area in and around Tuscaloosa, Ala., the "meaningul use" of EHRs deprives homebound patients of care.
I would happily do this on paper or using an EHR that created a logical note within the same amount of time. But that is not an option.
Such EHRs are rare, if they exist at all. Besides, this physician is likely contractually bound to use some larger organization's choice.
The reality is that I spend more time talking to the Information Technology people about Internet connections, firewalls and box-checking than I do answering messages from concerned family members.
In other words, computerization is in the way of the best practice of medicine.
As a teaching doctor, my feedback to the residents now consists mainly of explaining how to document their visits so that we will all get paid, instead of teaching them how to take care of frail elders in their homes.
I remember my community medicine clerkship in the early 1980s in a relatively underserved region in Maine. Other classmates at Boston University School of Medicine did similar clerkships in Roxbury, an exceptionally poor and harsh section of Boston near Boston City Hospital. We were taught patient care...with nothing less than the patient's best interests at heart. Medicine today is now being financially and cybernetically deprived of its heart and soul.
Harrell is a geriatrician and assistant professor of family medicine at the College of Community Health Sciences, University of Alabama. This is an edited version of a story that originally appeared in Pulse — Voices From the Heart of Medicine, an online magazine of stories and poems from patients and health-care professionals.
I hope that patients who are not suffering dementia will increasingly take notice that these systems are depriving them of their clinicians' attention and of providing good documentation for their future care.