Health quality


In 1966, a guy named Avedis Donabedian wrote the article that defined the metrics for healthcare quality.  In his article, Donabedian defined healthcare quality as a function of the structure, process and outcomes in healthcare.  For the next approximately 35 years, these three pillars of quality framed the discussion of what was quality care.  Quality gurus like Dartmouth’s Gerry O’Connor led a revolution in approaches to care by tracking therapies and outcomes, and figuring out how the best doctors got the best results.  Gerry, who was a consultant on a project I led, was once hired by a hospital to prove to NIH that their heart surgery program was great.  He found that it was, in fact, below average.  However, he pointed out that the process by which he determined that they were sub-standard could be evolved into one for tracking and evaluating quality.  At the start of the project, risk-weighted mortality varied by a factor of 5 between the best and worst heart surgeons.  After 10 years, the average doctor was better than the best one was at the start of the process.  Gerry is why patients are given aspirin after heart surgery.  Gerry turned his sights to cystic fibrosis, and, thanks in part to him, pediatricians who specialize in cystic fibrosis are following patients into their 60’s.  His efforts were celebrated in Atul Gawande’s New Yorker article, The Bell Curve. [N.B.: the phenomenon Gawande describes is not gaussian, i.e., not a bell-shaped curve, but that’s for another post.]

In 2001, the world was changing fast.  Genomics and proteomics were revolutionizing medicine.  The Institute of Medicine issued a report called, “Crossing the Quality Chasm.”  In it, IOM reported that our healthcare system is not prepared to advance into the 21st century: most physicians practice medicine as it was when they were trained.  New best practices do not get adopted.  In 2003, RAND published a study where they found that the average patient gets only 55% of recommended care.  This rocked the healthcare world: I personally have heard scientists, administrators, and politicians cite this statistic.  I’ve cited it myself, using a slide based on one presented by Rhode Island’s Patrick Kennedy.

Based on this, CMS issued a roadmap for quality care, which is leading to pay for performance standards.  Quality care is based on delivering evidence-based care.  All of it.  To receive full reimbursement, physicians should deliver to their patients the therapies that are agreed to represent the state of the art care.  Everyone is on board with this.  The AMA puts together committees to define the therapies required for quality care.  CMS adopts these standards and pays physicians more who deliver them.  Other insurers follow CMS’s lead.  It’s a great system that results in better care.

Or does it?

This system fails to adopt Donabedian’s method.  It is a system of incentives (structure) for delivering recommended care (process), but outcomes are never measured.  Peter Drucker, the guru of manufacturing quality, is known for saying, “What is measured improves.”  What is measured — what is rewarded — is delivering care, and in this case, more care.  Achieving results is not incentivized and it’s not even measured.  Physicians are, in general, payed for delivering services.  Based on the RAND study, the CMS approach to quality rewards them for delivering more care.

Is this a problem?  Maybe.

What defines the “recommended care” that is driving pay for performance?  Not clinical outcomes.  Real-world clinical outcomes are not measured as part of pay for performance.  Recommended care is determined in research.  The research that is driving quality indicators is the kind of great work that appears in major medical journals, like, for example, The New England Journal of Medicine.  But wait a second: Marcia Angell, an editor of NEJM for two decades, has said,  “It is simply no longer possible to believe much of the clinical research that is published, or to rely on the judgment of trusted physicians or authoritative medical guidelines. I take no pleasure in this conclusion, which I reached slowly and reluctantly over my two decades as an editor of The New England Journal of Medicine.” (The link is to where I first saw the quote.)

The very research cited by Dr. Angell, the same authoritative medical guidelines, are the ones that are defining quality care.

When I developed a quality program, our work was designed from the start to track outcomes versus therapy, defining success as delivering to patients better health.  We need to think hard about whether CMS’s approach, focusing on process but not outcomes, is the right way for America to deal with our unsustainable healthcare bubble.

5 thoughts on “Health quality

  1. btw, a suggestion – given the statistical illiteracy you hint at, it would probably help most readers if you changed “is not gaussian” to “is not gaussian, i.e. not a bell-shaped curve.”

    The approach I’ve found effective in my talks is to assume the majority of my audience needs a quick (ahem) “reminder” on just about everything.

  2. Thanks, Dave! I’ve done some work with government and found that many of the departments are overworked and get saddled with new priorities driven by politics not common sense. For example, a contact of mine was a senior person at HHS who argued against the doughnut hole in Part D and was effectively fired for it. The doughnut hole was a terrible idea, but it satisfied the key political stakeholders.

    CMS is definitely interested in outcomes, but how to measure? They accept as “outcomes” the results achieved in the clinical trials that establish a therapy. The problem with measuring outcomes is that, for much of medicine, it is the physician that dominates the outcomes, not the therapies. The only thing worse than having CMS propose a system of rating physicians would be see the results of those ratings. There’s enough data that we know that if CMS tracked outcomes, they would identify things politicians don’t want to know: rural and inner-city populations get terrible care.

  3. Another problem with “evidence-based medicine” and quality guidelines is that nobody anywhere along the line — not the committees composed of physicians, not AHRQ, not the IOM or CMS — is ultimately responsible for cost-benefit analysis. In fact, many actors are explicitly told that they may not consider cost. For example, one cancer doctor I know was recently on the physician panel charged with making a recommendation about the use of Provenge. Provenge had been shown in clinical trials to add an average of four months to the lives of men with prostate cancer. Provenge does not affect the severity of symptoms or the underlying cancer, and costs more than $90,000 for a single course of treatment — a treatment that may or may not do anything for that particular patient. (This is not like chemo and radiation treatments, which are assessed on an ongoing basis and can be altered depending on how the patient responds.) Most of the patients who take Provenge have a limited life expectancy and poor quality of life. Yet when my friend raised the cost-benefit issue on her panel, she was told that it wasn’t relevant to the task the panel was charged with (recommending whether this treatment should be part of the standard of care for prostate cancer patients) and it isn’t physicians’ job to make cost-benefit judgments. The problem is that in our system, it doesn’t seem to be ANYONE’S job to think about the costs as well as benefits of treatments when constructing quality of care guidelines.

Leave a Reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload the CAPTCHA.