The Once and Future EHR
On the Occasion of the 100th Birthday of Morrie Collen, MD
Biomedical Informatics Pioneer
In 1968 Morrie and colleagues published “Computer-Stored Medical Record” in Computers and Biomedical Research. The abstract from that article follows.
“In order to satisfy all the input and output requirements and process all the forms of essential medical information for all office and hospital visits for the lifetime of a patient, it has become necessary to develop a more complex programming system for a variable-length and variable-format computer-stored medical record. This programming system is designed to control and coordinate general and fundamental routines which include the handling of direct access storage, as well as remote terminal input and output devices.
“An encoded computer-stored patient medical record, organized basically as a tree-structure record, is partitioned into patients’ visits which are subdivided into parts containing medical history, examination findings, diagnoses, laboratory and X-ray reports, etc. Active patient records are permanently stored on a large mass storage (Data Cell) device, but temporarily are moved to high-speed random-access disc drives while the patient is under actual care in the hospital or outpatient clinics. Although the computer record can store straight English text, in order to conserve storage space, the system utilizes an English word dictionary to encode diagnoses and an item catalog to identify standard tests and procedures, standard medical questions, etc.
“To permit convenient revision of input and output forms, the system includes an input language compiler, a medical report generator, and an information retrieval compiler. The multiprogramming system allows multiple medical functions to operate simultaneously and permit access to the same patient’s records. Such a system requires extensive communications equipment to handle input and output requirements for multiple facilities and at least some limited backup capabilities in the event of equipment failure.”
- Morrie and his colleagues wrote: “…essential medical information for all office and hospital visits for the lifetime of a patient….” Morrie was among the first to recognize that information was not just important for delivering care, one delivers care by managing information. Or as Don Berwick has said: Information IS care.
- Morrie and his colleagues recognized the centrality of the data dictionary as the heart and soul of the EHR and that it had to make clinical sense. This was because Morrie was fundamentally a clinician, not a computer geek. Hence the systems and programs they created had to serve clinicians and their patients, and they did.
- These basic tenets of Biomedical Informatics were reflected in the landmark 1997 IOM report, The Computer-Based Patient Record: An Essential Technology for Health Care. Many of us participated on that IOM Committee and its meetings. This was the only in-depth personal interactions that I had with Morrie, a kind and gentle but strong force of leadership on the Committee. This IOM report led to the HITECH Act and the wide implementation of EHRs we are experiencing.
So where do we go now, building on the leadership of Morrie and his generation of Titans?
- Interoperability: We pay a lot of lip service to having health information follow patients wherever they go, but right now, that’s just an aspirational goal. A family friend called me last week from NYC and asked me to help her find a general internist. She eventually rejected both of the people I referred her to because her ophthalmologist, whom she liked a lot, wanted her to see a general internist in the same system so she could see their electronic records. We need to be able to exchange data easily between systems, and to do so we need a sustainable business model for health information exchanges.
- Documentation: EHRs are still mainly electronic versions of paper records, with lots of unnecessary duplication, too much dependency on physician documentation to the point where they feel like “data entry clerks.” There’s also too much reliance on free text which can’t be searched easily and leads to dangerous short cuts like cutting and pasting. In my opinion, writing a note is two separate activities—documentation and reporting—we and need to separate them. All those participating in a patient encounter, including the patients and their families, need to record the information that they control. Patients should record symptoms, history, family/social history. Clerks: demographics, socioeconomic information. Nursing assistants: vital signs. Pharmacists: mediation reconciliation and adherence. Nurses: patient self-management and education. Lab and pharmacy: lab and drug info. Physicians should add additional history, physical exam findings, comment on findings of others, and develop an assessment and plan in concert with nurses and pharmacists. Team-based care requires team-based documentation.
- Reporting: The information entered during a clinical encounter can be formatted for a visit note, billing record, consultation letter, quality improvement report, etc. IMO, there doesn’t have to be an actual note in the physical sense. All the information is there.
- Data analysis: We have all been guilty of spending much more time getting data into EHRs than getting information out. There is much to learn about the practice of health care and prevention, about what works and what doesn’t, about who should be treated how, about cost-effectiveness. But the basic science of clinical database epidemiology is in its infancy. The data are messy: half the diagnoses of heart failure are wrong. All definitions of conditions have sensitivity and specificity tradeoffs. Yet like ancient Roman mosaics, with a sufficient number of pieces, even messy pieces, a cogent picture of a patient can emerge that is sufficient for informed clinical decision-making.
- Clinical sociology of EHRs: Introducing computers into exam rooms fundamentally changes the clinical encounter, and often for the worse. A Regenstrief colleague of mine conducted a video study of what happened when Kaiser-Permanente introduced computers into the exam room. Too often, the doctor had his back turned to the patient, looking at the computer, or both the patient and doctor were staring at the computer, not at each other. During this study, patients’ satisfaction with their doctors decreased. We need to learn much more about how to use EHRs as tools and avoid their being a disruptive third person in the clinical encounter.
- Patient involvement and control of their information: Like it or not, patients will not only contribute to their EHRs, fair information practices will dictate that patients will have some control over who sees what’s in them. Like the privacy discussions, we can go nuts and require patients to say who can see each piece of data in their EHRs. That’s crazy, but doing nothing is not an option either. Rather than fighting this, we—the academic descendants of Morrie and his colleagues—need to lead this discussion and find the sane middle ground.
- Innovation: Most academic medical centers with strong Biomedical Informatics Departments have ditched their home-grown EHRs for commercial vendors. So where will innovation take place? What is the role of academia? Is it possible to have a public-private partnership with the big EHR vendors around innovation? A recent EHR Summit we had at Regenstrief that was attended by 22 vendors, including most of the big ones (Cerner, Epic, etc.) suggests that they are open to this.
Morrie Collen and his pioneering colleagues launched our field. It’s up to us to carry the torch forward, building upon their founding principles, and maintaining the centrality of the clinicians and their patients and the high quality care they expect and deserve.