It seems CMS and I differ in our definitions of quality when it comes to the impact of EHRs on patient care.
A recent e-mail from CMS boldly announced that a "New Study Links Higher Care Quality to EHRs." They go on to describe the key findings published in the Journal of General Internal Medicine (JGIM).
"The study looked at the association between EHRs and ambulatory quality in a community-based setting, finding that EHR use led to:
• Increased appropriate hemoglobin A1C testing for patients with diabetes;
• Greater provider success in meeting quality measures for breast cancer screening;
• More providers meeting quality measures for Chlamydia screening; and
• Increased number of providers meeting quality measures for colorectal cancer screening."
This definition of quality emphasizes the procedural aspect of healthcare - performing the "proper" procedures equals quality. A healthcare organization could score well on these criteria even if it employed no physicians. Any nurse or PA could perform these procedures and score high on quality. This kind of quality is important but it is tangential to either what physicians can contribute or to the use of an EHR. A checklist and a routine that does not allow a visit to be ended until the items have been performed is all that is required. If a megabuck EHR can help achieve this kind of quality, that's nice, but not all EHRs help. Furthermore, this goal can be accomplished without an EHR - if the organization is committed to it.
This concept of quality would not have helped the patients described in this September's article from Rx for Prevention (from the L.A. County Public Health Department) entitled "Botulism: A Commonly Unrecognized Disease."
It describes a cluster of two botulism cases in which the diagnosis and treatment were delayed by (in my opinion) a combination physician ignorance, failure to document the history and findings in detail, failure to recognize the existence of the shared living arrangement of the two cases, and the inability of the patient to carry the records of their prior care to the next facility they visited.
To summarize: The two individuals consumed some rancid-tasting food that had been left unrefrigerated. Both awoke with ptosis, diplopia, dry mouth, and dysphagia - potential signs of cranial nerve dysfunction. The first patient was "diagnosed" with a throat infection and given antibiotics. The second patient, seen at different facilities, was also given antibiotics in spite of her additional complaints (by that time) of upper extremity weakness. After multiple trips to different facilities both were eventually admitted, diagnosed correctly, and treated successfully.
Perhaps the most valuable contribution made by Larry Weed when he described the problem-oriented medical record was to emphasize that if a physician is ignorant about something that might be affecting the patient, that ignorance constituted a significant, independent problem and it should appear on the problem list. Before addressing the ignorance problem, perhaps with a referral or discussion with a colleague, it behooves the physician to develop the "knowledge base" (Weed's term) as completely as possible by performing and documenting a thorough history and physical and perhaps some initial studies.
These cases highlights the other (in my opinion more important) kind of quality - that the physician does a uniformly excellent job of exploring and documenting the patient's problem and either arrives at a (provably) correct diagnosis or seeks assistance. Prescribing antibiotics for every URI is not quality. Labeling something about which the physician has no clue as epiglottitis or pharyngitis in order to rationalize a prescription for antibiotics is not quality.
Could a computer system help with some aspects of this? Sure - but only if physicians do their work and document it. Jumping to an incorrect conclusion and documenting only those things that support it is not quality. No computer system could possibly help unless it is supplied with data. Narrative notes are not data. Legibility and retrievability are important but you don't need an EHR to store and retrieve typed narrative. The computer could also help if it allowed people who are not patients to be added to the database which would make it possible to group individuals into family groups, but billing-account-oriented EHRs rarely allow this. Providing each patient with a copy of the visit documentation could help if it was complete and if subsequent practitioners would take the time to read and believe it. Again, a computer might help, but so would pressing the "print" button or making a photocopy. Doing this requires a commitment to quality medicine - it does not require an EHR.
CMS likes their definition of quality because it is easy to tabulate. To me, quality means having a doctor who is a good history-taker, knows enough to avoid significant blunders or delays in diagnosis and treatment and is meticulous.