Whether or not more regulatory oversight of Health IT is needed, I suspect many of us have experienced instances where health information about us is found to be in error. I recall when my mother was hospitalized for chest pain that doctors were treating her as though she had been a life-long smoker. In fact she had never, ever been a smoker. At some point in time, information about smoking history had been erroneously entered into the electronic record. Now, the doctors treating her for chest pain were making decisions about the likelihood of heart disease based in part on that information about smoking history. In my own medical records I have also found, and had to correct, occasional errors in medication history, allergies, and immunizations over the years.
Despite this, I would tend to put the blame not on the computer or the software. It is not generally these systems making the errors, but rather the people using them. Sometimes the wrong information has been entered into the system, as in the case of my mother. Sometimes, errors are made because the information being displayed is in the wrong chronological order or is buried in a user interface that is out of synch with real-world, clinical workflow. In both instances, the problem is with people—those who designed the software and those who use it, but not with the software itself or the machines running it. How can we improve on this situation? Here are four ideas:
Involve the Patient Right from the Start
In gathering the information that becomes the foundation of our medical records, we are putting too much burden on caregivers. How much of the complete medical history or SOAP note is information that comes directly from the patient? Chief complaint, history of present illness, past medical history, social, family and occupational history, medications, allergies, review of systems? All of this information is retrieved by “interviewing” the patient. Perhaps it would be more efficient and more accurate if the patient himself entered all that information into a kiosk, or some other kind of fully automated, information intake solution. Surely with today’s technology we could design systems that would do a more consistent and comprehensive patient interview and subsequent documentation of information without taking even a minute of clinical staff time. Patients could then review the information captured about them for accuracy before it was officially entered into their record.
Ease the Documentation Burden on Clinicians
We need to ease up on documentation requirements for clinical staff. The patient-centered machine capture solution mentioned above would help remove a lot of the documentation burden. The remaining documentation of the exam, differential diagnosis, and treatment plan could be better facilitated by free text, medical dictation solutions with natural language processing and coding technology on the back end. Nothing is more important that freeing our clinicians of the time currently being spent doing data entry.
Prohibit Templates, Cut and Paste
Templates simply don’t work because it is impossible to template the “patient story” and all of the other nuances of a good clinical exam. Likewise, cut and paste solutions are probably responsible for more medical misinformation and errors than anything else. EHRs should ban “cut and paste” capabilities altogether.
Share Information with Patients
At the end of the day, I believe all information in the medical record should be shared with the patient. The patient is an extra set of eyes, an extra check point if you will, against medical errors. Giving patients complete and full access to the information about them is not only a better way to engage patients in their care, but also a way to help make sure everyone is on the same page about their care. As eHealth advocates proclaim, “Nothing about me, without me!” I think this is sage advice for preventing misinformation and the introduction of errors in our medical records.
I would also be the first to admit that many, if not most of today’s electronic health record solutions are still too hard to use. They have been poorly designed in our attempt to replicate a clinical workflow previously based on paper records. As I have stated many times before, there is a unique opportunity to design solutions that really take full advantage of today’s technology. That includes incorporating ways to give clinicians greater choice and flexibility in the data capture methods they use. In addition to text, we should be making better use of audio and video in our medical records. We also need consensus and greater standardization of the clinical user interface itself. What a nightmare it is for clinicians who practice in a number of different settings, each with a different EHR. This is not only a challenge for clinicians, but yet another way bad information gets replicated in our clinical systems due to data entry errors or omissions made by people who are unfamiliar with the solutions they are forced to use.
