By Robert D. Lafsky, MD
Robert D. Lafsky is a gastroenterologist in Virginia.
It’s been a year since Dr. Lawrence Weed passed on at 93. He got a mention in HIStalk and a longer obituary in the New York Times, where he’s credited as a major innovator in the organization and computerization of medical records.
He was. But reading his later work, one has to doubt that Weed would have been happy with the Times statement about his Problem Oriented Medical Record, that “two of its features have become nearly universal in health care: the compiling of problem lists and the SOAP system for writing out notes in a patient chart.”
Oh sure, you can look at the average EMR chart these days and see a “problem list” and SOAP designations on at least some of the progress notes. But do the problem lists reflect the sort of organized rigorous thinking and aggressive pruning he advocated as necessary to keeping them useful?
Not very often. Especially after several admissions, a hospital patient’s list has a long string of overlapping, duplicative, or clearly unnecessary subordinate “problems.” It’s of little use to anybody. And are the SOAP notes problem-specific and do they clearly divide up the information as intended and point toward action?
Don’t get me started. Let’s just say that if we revived Dr. Weed and turned him loose with a current day EMR and gave him an hour at the lectern, his critical dissection and ridicule of the clinical work therein would be strikingly similar to what he does in his famous video from 1971.
The key concept underlying the Weed scheme was that one doctor brain couldn’t hold enough information to organize information and make good decisions about a patient. The process needs a more systematic and documented approach.
As obvious as that may sound to readers here, medical giants walked the Earth 50 years ago, and a Big Ego telling other Big Egos — especially specialists — that their egos were too big didn’t always go over so well. But there was a more fundamental problem and Weed had to deal with it.
The original Weed system made sense dealing with the management of known and established problems, but the “unknown unknown,” the diagnosis problem, was the flaw in the scheme. I saw this myself when I started practice as a specialist in the early 1980s in a then-hotbed of Weed methodology — a small hospital with young family practitioners trained on the POMR concept.
What I saw repeatedly as a consultant was that no amount of dogged problem list maintenance could get you to see that problems 1, 3, and 5 were actually components of a single syndromic diagnosis. You just had to know that. And in those days, without sufficient training in the field in question, the light bulb never went off over your head.
Actually that light bulb method is still what we’re doing, but Weed spent his later career working on a computerized improvement. A trained interviewer (not necessarily a physician) would work with the patient and the records to input extremely granular information in a neutral fashion, avoiding the leading questions that the current heuristic system requires. The computer would then go to work applying a series of “knowledge couplers,” what I believe would be considered an expert system in current terminology, to generate a complete list of diagnostic possibilities. Only after that would a physician start dealing with the case and sort out the problems in light of that information.
How did that work out? You can read Weed’s book for a very full discussion of his later views. But the business of the ensuing business enterprise is a checkered story. Here’s an article from 2002 about his system and its fate. Suffice it to say that this was not a system that took over the world.
Is Weed doomed to be an obscure historical figure in medical history at best and a minimally successful software developer at worst? I’ve had conversations with very highly-placed medical people who had never even heard of him. But it’s hard to look at the current morale problem in medicine and not see him as a prophetic visionary. Every week or so, I see two or three “burnout” articles or videos, mainly focusing on the current EMR experience. Everybody complains that they now have two jobs, data entry and actual thinking, or at least trying to make the light bulb turn on.
Will a Weed-like diagnostic system take over eventually and automate the light bulb? The problem is it’s going to take a lot more time and disruption to get something like that working and working well.
But in smaller but still important ways, Weed’s legacy can and should come into play right now. The Weed argument would be that this burnout crisis was foreseeable, a result of medicine never controlling the data design process in the first place. And we never developed an ethos that requires that everyone have the individual discipline to actually contribute value to structured data with rigorous truthfulness, regardless of specialty orientation. And then to rely on what’s in there, or if necessary, correct it.
In particular, although specialists are necessary, they have special duty in a shared hospital EMR environment to pay attention to what’s in the data tables and not just churn out unstructured and often contradictory text reports. Detail management is hard but critical, and although details span a range of importance, failure (say, to get a fresh and confirmed cancer diagnosis on the problem list before discharge) should be considered somewhat above the misdemeanor level.
The burnout crisis reflects a pervasive sense that medicine has lost its autonomy to business and IT interests. But a key Weed-based insight is that we can’t start to get it back without taking more responsibility for what’s gone wrong.
Weed can be seen to offer a tough but fair path off of the beachhead we seem to be stuck on. The profession as a whole can regain autonomy, but the individuals in it have to give up some of individual ego-tripping many in it have enjoyed for too long.
There’s great potential for a better software environment in the future. Weed’s legacy will be more clear to everyone in the future. Right now, we have to pitch in now to work better with what we have.