As often happens, a Susannah Fox post has led to lingering questions. This time I’ve figured out what I want to say.:-)
Last week, in a side note on her World AIDS Day post, she inserted this:
… Let’s review the basic math of health services delivery in the U.S., beautifully, devastatingly summarized by Peter Margolis of Cincinnati Children’s and the C3N Project:
.5 x .5 = .25
As he recently explained:
The dominant paradigm of chronic care delivery in the US is characterized by health “services” “delivered” to relatively passive patients or “consumers” in clinician-mediated encounters using knowledge that is produced by researchers and clinical experts. In this current system, physicians provide only 50% of indicated care and patients adhere to about 50% of what is recommended. The result is a system where only about 25% of patients achieve optimal care, despite spending that far exceeds other industrialized countries. What would outcomes be like if doctors and patients got what they needed 90% of the time?
Let’s break apart the first part of that quote:
The dominant paradigm of chronic care delivery in the US is characterized by
- health “services” “delivered”
- to relatively passive patients or “consumers”
- in clinician-mediated encounters
- using knowledge that is produced by researchers and clinical experts.
And:
In this current system,
- physicians provide only 50% of indicated care
- and patients adhere to about 50% of what is recommended.
The result is a system where only about 25% of patients achieve optimal care, despite spending that far exceeds other industrialized countries. [Italics added]
Lots going on here.
- The first sentence is about today’s paradigm about healthcare: “all knowledge arises in the profession and is dispensed onto patients.” (“services delivered,” mediated by clinicians, to passive consumers)
- On this blog (and in our Society) we know otherwise. But most people running hospitals and doctor offices still presume their job is to “do things to us,” including giving us instructions.
- The second is about the often-observed reality that neither providers nor patients stick to plans.
- One might guess that this is a human failing.
- And note that he phrased the result in terms of patient impact: “only 25% of patients achieve optimal care.”
I wonder: what else does he envision, when he first talks about our current model, and then talks about “.5 x .5 = .25”? Does he assume that part 2 is a consequence of part 1?
I don’t think so – I think they’re two different problems. But I do think that part 1 causes both parties to feel hopeless about part 2.
Or is part 2 a consequence of part 1? Do we perform more poorly if our assumptions about who has the power are fundamentally flawed?
Dave,
I’m curious why you think that part 1 might not be related to part 2. My own take on this is that he does assume that part 1 is a significant causal component of the observations in part 2.
Let’s deconstruct further:
The first 50% attenuation of knowledge (in the paternal model) is caused in large part by variation across doctors and institutions. This variation stems from differences in knowledge, lack of standardization of practice for common cases, lack of methodology for uncommon cases, etc. Projects like the ImproveCareNow network of which Peter is member can have a significant impact on this multiplier, and on outcomes, without making significant changes in paternalism; although I’ll argue that Pediatric practice in the ICN is probably closer to the kind of system we would all like to see than your average practice (there are patient and parent advisory councils, etc)
The second 50% term in the equation has to do with patients failing to follow the recommendations given by their treating physician. The assumption that being treated as a passive consumer of wisdom from on high has the effect of exacerbating the human tendency to adhere to pre-existing habits.
The thinking behind the C3N project that Margolis leads is to remove paternalism and make families and patients equal partners. This, they believe, will have substantial effects on both sides of the equation. In such a world, one of the key questions they are asking is whether they can apply the same process improvement techniques they used to improve the first term in the equation to help optimize the second? If patients are equal participants in the system, then they have a role and a stake in governance and outcomes.
I’m most excited by the idea that these new kinds of encounters and interactions can lead to new sources of knowledge that impact care ahead of the rigorous mechanisms of clinical discovery and that is clearly on the radar for this team.
For example, if a family uses a treatment for which there is no clinically-derived knowledge, the C3N project brings in a doctor and a statistician to work with that family to try that treatment out to find out for sure if it is having an effect (e.g. eating granola bars reduces pain/flares, or probiotics). This is a responsible response and treats patients and caregivers as responsible actors, takes their beliefs seriously, while still bringing the discipline of their training to answering the patient’s question. This one intervention alone is a huge paradigm shift from the typical doctor-patient encounter.
Any errors or confusions produced here I attribute to the 101 degree core temperature I’m enjoying today.
Ian and all, I’ll be replying – love your stuff here! – but I’m wrapping up an intensive week, no time to THINK about this stuff yet.
And this is deep. Wickedly powerful, as we try to troubleshoot healthcare and engineer some interventions. Need to know how we miss the mark today.
So, I quibble with the math here. If the average physician delivers 50% of recommended care and the average patient adheres to 50% of care delivered, the math is that the average patient gets 25% of recommended care, not 25% of patients get optimal care. Based on the way populations tend to work, I’d expect that zero patients get recommended care. [I’m restraining myself from providing a dissertation drawing on stochastic mechanics and statistical physics… so tempted…]
Even worse, those docs who are delivering 50% of recommended care are likely also delivering just as much deprecated (unrecommended) care — the problem isn’t that they skimp on care, it’s that they aren’t giving the right care. Even more insidious, commonly the care that is given but not beneficial is stuff that sounds good. Read the story of the physician who tried to introduce hand washing into 19th century surgeries and this really comes across. A modern example is glutithione in Parkinson’s disease: in PD the brain is short of glutithione, but expensive injections don’t help because it doesn’t cross the blood-brain barrier.
Oh, and by the way, I’m leading a project to address just these issues: http://www.parkinson.org/Improving-Care/Research/Quality-Improvement-Initiative
We’re looking at not just what care is recommended to patients, but what they are actually doing and the results they are achieving. When we see that one doc is recommending exercise, for example, we look at what, a year later, that doc’s patients did based on the recommendation. We’re attacking both of the 50%’s.
I may have done Peter a disservice by posting only a short excerpt of what was a great speech at the National Meeting on Collaborative Improvement Networks in Children’s Healthcare last month.
I love the discussion happening, but let’s not “deconstruct Margolis” too much since it’s my fault that we have only a piece of the puzzle he describes.
Ian, post more when you’re feverish – it’s lucid & awesome.
Pete, thanks so much for the link – exploring the site now.