Monday, April 01, 2013
Collective Medical Learning (and its future)
This needs some more thought and writing to pull it into a sensible form with a cogent argument; I hope to make that transformation soon. The objectives, consequential questions, and sources I wanted to include are here. But this is not really a note yet, this is just notes.--Maurice
Over the past few years, an accumulation of observations suggests the need ask whether medical society is learning its lessons from its experience, particularly from its failures. Or whether it could and should be learning more, and more readily. The set of observations accumulated by this observer will be presented in summary fashion below, with brief notes as to why they are included.
Medical SocietyThe collective "medical society" is perhaps not a usual term, so should be explained. It is deliberately chosen for its vague inclusiveness. The subject at hand is what Russ Ackoff aptly called a mess, a complex system with multiple stakeholders. In our "medical society" we have doctors and patients, but also patients' families, hospitals and their owners and administrators, pharmaceutical and other laboratories, medical equipment makers, and insurance companies and other healthcare mutualisers; and perhaps others. It is vast, with boundaries often unclear yet interests often very narrow. In such a mess, where should the responsibility for collective learning reside? Shouldn't it be shared, with each stakeholder contributing according to his abilities? Or should it be the business of regulatory agencies, as in one of the exhibits below?
"Contemporary medical society should be learning more from its experience. It remains to determine how could that greater learning should be achieved?"
That is the topic suggested by the observations.
The Accumulated CluesThe exhibits leading to the asserted need include:
- Exhibit demonstrating that practitioners learn what they want to learn, and may not collaborate (across regions and specialities) to arrive at a (global) consensus on best practices. The citation intended here observed regional differences in treatments and effectiveness, was in an article by Atul Gawande (to find). This shows that medical society has trouble learning from its experience. Perhaps it also does not want to drive back surgeons out of business even if they aren't providing the best treatment: what else are they going to do? Probably have to skip this exhibit (or spend a lot of time reviewing Gawande's extensive corpus).
An exhibit illustrating two problems: the difficulty of getting the questions (framing of the monitoring) right, and the difficulty of assessing a treatment comparatively when the benchmarks for comparison are constantly evolving. This is from the Harvard Magazine piece A Cardiac Conundrum (review of the book Broken Hearts: The Tangled History of Cardiac Care), review page titled "The passion for procedures to fix ailing arteries and hearts may be misguided." Essential passage:
“Doctors generate better knowledge of efficacy than of risk, and this skews decision making,” he says. “They design treatments to do something specific, and design studies to see if those treatments achieved those outcomes; and so accumulate lots of data on whether treatments produce the desired effects. Capturing good knowledge of side effects, especially the unanticipated ones that are so common, is both less interesting and more difficult. Whenever doctors have more thorough knowledge of the possible benefits of a treatment than they do of its potential risks, patients and doctors will lean towards intervention.”This is another instance of medical society framing the monitoring in such a way that collective learning is not as great as it could be.
- An example of collective learning that appears to have arrived at a global recommendation, showing it perhaps can be done; the advisory may not be universally accepted, it may have missed other lessons available from experience, but who can say? It does seem to have asked the right question. It also relies on actual double-blind studies on a large scale (both people and years), not just monitoring. The example is the change in emphasis on treatment of high blood pressure: systolic, mainly, not so much concern with diastolic : NHLBI, Clinical Advisory Statement: Importance of Systolic Blood Pressure in Older Americans [[http://www.nhlbi.nih.gov/health/prof/heart/hbp/hbpstmt/hbpstmt.htm][Clinical Advisory Statement]]. By the way, this advisory statement was issued in 2000, but that is not at all apparent from the web page, where only the current date is shown.
- Exhibit reviewing the current ignorance as to whether diagnoses (and treatments) were correct. If not checked (audited), how can medical society learn to get it right (or righter) next time? More autopsies. Certified causes of death inaccurate in fifth of cases, study suggests Cancer and heart disease death rates may be higher than thought and lung deaths lower, according to ONS findings The Guardian, December, 2012, concerning Great Britain or England.
- Exhibit reviewing death investigation in America (ProPublica, December, 2011 ) : Without Autopsies, Hospitals Bury Their Mistakes.
... even sudden unexpected deaths do not trigger postmortem reviews. Hospitals are not required to offer or perform autopsies. Insurers don’t pay for them. Some facilities and doctors shy away from them, fearing they may reveal malpractice. The downward trend is well-known — it’s been studied for years.
What has not been appreciated, pathologists and public health officials say, are the far-reaching consequences for U.S. health care of minuscule autopsy rates.
Diagnostic errors, which studies show are common, go undiscovered, allowing physicians to practice on other patients with a false sense of security. Opportunities are lost to learn about the effectiveness of medical treatments and the progression of diseases. Inaccurate information winds up on death certificates, undermining the reliability of crucial health statistics.
ImplicationsGiven those exhibits, this observer arrives at these points and derived questions:
- More autopsies, rather than fewer, would probably be worth doing.
- But how much would they be worth, what would their value be? How can it be assessed?
- Is there a strategy (or policy) that can optimize selection of those to perform?
- How should more autopsies be paid for?
Cadavers are still needed for medical students.N.B.: That was from another exhibit, missing above!
- What is the relationship of cadavers for medical school to those autopsied? Do autopsies performed in anatomy labs of med schools have any research value, or do they serve only to provide concrete experience to the students?
- If one (such as I) considers his organs too worn to be worth transplanting, and wishes to either give his body to a medical school or have it autopsied, how should he choose? Or should he leave the ultimate choice to his heirs, who might prefer autopsy if the cause of death is likely to be interesting (surprising)?
- The question of who is to blame for the decline in autopsies is interesting, but unlikely to be the most expedient way to reverse the trend. Is it economics? Religion? Big pharma in collusion with the HMO and insurance companies? Who benefits? Interesting, but more important is what society loses and finding ways to restore effective learning practices to the medical sector (to the sector collectively, not to each individual in the sector).This is just a restatement of the futility of trying to untangle a mess: there will always be other candidates for blame.
Topics for Further StudyIt could be interesting and useful to frame this mess in Argyris and Schoen's learning types model. Most likely, individual stakeholders are predominantly Type I. Is it possible to get the regulatory agencies to Type II (double loop learning) and a role as facilitator and mentor? Or will professional clans fight hard and cling to sectarian power, whether or not it is in the common interest?
Leftovers --please ignore
The MisledeA joke heard many years ago starts by asking "How many social scientists does it take to change a light bulb?" (Moreover, a clever recent short essay addresses the question "How many historians ..."). The old answer was "Social scientists don't change light bulbs, they try to explain why one is burnt out."
That may not be the best introduction for the question at hand but the part about figuring out why there had been a failure brought it to mind.
While somewhat fun and catchy, that joke did not provide an analogy useful making sense of what was to follow.
Stuff that somehow was transformed ...from sentences to attributes of some weird tag, in the course of html parsing and checking and style rewriting. Probably hopeless (and useless) to try to put back together again.
--="" a="" about="" alone="" and="" be="" because="" begin="" but="" by="" coroners="" could="" cure="" die.="" diseased="" do="" doctors="" does="" effectively="" enough="" explain="" hard="" introduction="" is="" it="" line="" make="" many="" medical="" misleding.="" more="" nbsp="" not="" obvious.="" one="" ow="" patient="" people="" perhaps="" punch="" question="" reply="" seriously="" society="" suitable="" take="" that="" the="" then="" they="" to="" too="" tries="" unfair="" way="" we="" whether="" why="" would="">