Thursday, December 2, 2010

Healthcare quality: 5 ways to stop the insanity

It was Einstein, I think, who defined insanity as doing the same thing over and over again and expecting a different result. You could say that perhaps we are living this definition in the healthcare system today.

I talk a lot on this blog about quality and harms associated with healthcare. My take all along has been that we are continuing to miss the mark. I have known this because I follow and contribute to the emerging literature indicating that we continue to cause healthcare-associated infections and other complications at an alarming rate, even after the massive issue of avoidable hospital deaths was uncovered and popularized by the landmark Institute of Medicine 1999 report "To Err Is Human". And sure enough, over a decade after defining the problem, we now have not one but two studies coming our in tandem to indicate that we have not advanced an inch. At virtually the same time we are also learning that screening heavy smokers with CT scans can reduce lung cancer mortality by a whopping 20% (which in reality turns out to be only 0.3%, alas), with an accompanying risk of a false positive of 25%, as well as the fact that nearly 1/3 of all end-stage cancer patients die in our already overcrowded ICUs. Amid all of this bustle, there are breathless reports of eliminating nosocomial infections through simple checklists and hand washing, yet why are we not seeing any improvement in what really matters -- whether a patient who is meant to leave the hospital alive does in fact do so?

Well, to me this is what defines madness in our healthcare system. The news of the studies coming out of the Office of the Inspector General of the Department of Health and Human Services and from Harvard were not really new, as I have already indicated. More than any other specialty, the ICU community is well aware of the ongoing issues that arise from an onslaught of ever escalating numbers of patients with increasingly complex burdens of illnesses, a spiraling cognitive load of checklists and "evidence-based" quality indicators, and crushing documentation burdens in the face of overwhelming personnel short-falls and diminishing bedside time. Yet the Dartmouth researchers continue to show what they have been showing for decades: we are generously extending limited resources (both of interventions and human cognitive capital) to all who care to partake, without any limitation in the name of appropriateness or humanity. Death panels indeed! And, as the NLST underscores, we continue to look for salvation at the margins, where it is not only financially costly, but, because of the risk of adverse outcomes following invasive work-ups in patients with false positive CT findings, is likely to create an additional cadre of the chronically ill out of people who might not otherwise need to risk exposure to our already overwhelmed healthcare system.

At the same time, because patient turnover is what drives the bottom line, clinicians are compelled by the business of medicine to care for more and more patients. This discourages the time consumption of a thoughtful clinical encounter in favor of quick reactions usually involving multiple expensive tests, many of which may be avoidable with more time and attention to the specific patient at hand. Alas, clinicians who are willing to spend this kind of time do so at the peril to themselves, their families, and their sanity, putting themselves at a high risk for burnout, as their work days bleed into any semblance of personal time they might have hoped for. Because these are the very clinicians highly sought after by many in their communities, they end up giving up sleep in order to serve, and, well you know the story of how sleep deprivation affects judgment adversely, blah blah blah... And then they up and quit medicine.

Is this enough to diagnose insanity yet? Well, if not, then let's go to "evidence" in the relentless juggernaut of evidence-based practice guidelines and policies and reimbursement and quality metrics and and and... We have talked ad nauseam about evidence -- incomplete, invalid in some respects, non-individualizable. Yet, in the buzzing beehive of today's healthcare, it is this very evidence that must replace physician's thoughtfulness about any specific patient. Given its unavailability and inadequacy, coupled with the rush of a typical encounter, is it any wonder we are failing to fix our little quality quagmire? In fact are we not likely to make it even worse by continuing in this vein and rewarding behaviors that ostensibly impact the outcomes but in fact may represent nothing but noise?

So, here are my five potential solutions to the problem. They are not easy fixes, and they will not fit easily in a fortune inside a cookie or on a bumper sticker. All of them require broad educational efforts and social and scientific changes. Yet, we need to consider them seriously if we want to get back to first doing no harm:

1. Empower clinicians to provide only care that is likely to produce a benefit that outweighs risks, be they physical or emotional.
2. Reward the signal and not the noise. I wrote about this here and here.
3. Reward clinicians with more time rather than money. Although I am not aware of any data to back up this hypothesis, my intuition is that slowing down the appointment may result not only in reduction of harm by cutting out unnecessary interventions, but also in overall lowering of healthcare expenditures. It is also sure to improve the crumbling therapeutic relationship.
4. We need to re-engineer our research enterprise for the most important stakeholder in healthcare: the clinician-patient dyad. We need to make the data that are currently manufactured and consumed for large scale policy decisions more friendly at the individual level. And as a corollary, we need to re-think how we help information diffuse into practice and adopt some of the methods of the social sciences.
5. Let's get back to the tried and true methods of public health, where an ounce of prevention continues to be worth a pound of cure. Yes, let's strive for reducing cancer mortality, but let us invest appropriately in stuffing that tobacco horse back into its barn -- getting people to stop smoking will reduce lung cancer mortality by 85% rather than 0.3%, and at a much lower cost with no complications or false positives. Same goes for our national nutrition and physical activity struggles. Our social policies must support these well-recognized and efficient population interventions.

This may be the watershed moment to stop the treadmill of insanity that the business of medicine has created and continues to fuel. We must do better. And to do better, we must change course.           


  1. Marya,

    On a somewhat related note, yesterday I attended the NEHI & Mass Technology Collaborative meeting to announce publication of their new report, Critical Care, Critical Choices: The Case for Tele-ICUs in Intensive Care. The full report is available at

    The report chronicles the experience of implementing a VISICU system (now Philips VISICU) at UMass Memorial Medical Center--and extrapolates quality improvement and cost saving results to all US hospitals. I’ve only skimmed the full report, but here are a couple of points I found interesting during the panel discussion:

    --Academic medical centers that were asked to participate in pilot weren't interested and typically responded that they were already practicing "perfect" medicine.

    --Predictive modeling is used to provide advanced warnings when alarms will likely go off, which helps prepare a faster and better coordinated response.

    --Better outcomes were largely attributed to both a faster response and having a "second set of eyes".

    I asked about the analytics and benefits of tracking data from the system. Dr. Craig Lilly said he finds the data compiled by the system incredibly helpful & in fact found that at least one guideline coded into the system provided advice that was harmful and it was subsequently modified.

    One other benefit for patients (apart from shorter stays in the ICU and improved outcomes): reduced need to be transferred to medical centers farther from home. Other key outcomes of the study are summarized in the exec summary.

    There clearly are cultural hurdles to overcome in changing our current healthcare practices. But, I think the mounting evidence of the benefits of collaborative care can be used to persuade doctors and patients of the need for changes.

    Final comment: I was just speaking with a friend who knows a nurse who uses the VISICU system and her experience has been very positive—especially with respect to the responsiveness of the VISICU support team.

  2. Janice, thanks so much for this synthesis! I have spoken with Craig pretty extensively about the UMass experience with the VISICU, and his thoughts generally echo what you have written here. What never ceases to amaze me is the attitude that can generally be described as "I am not the problem but my neighbor is". Time and time again, when docs are asked whether they are practicing best medicine, they respond in the affirmative, but when asked the same question about their colleagues they are much less positive. So, it is always the other guy. And is not the first step to solving any problem to admit that one has it?

    Thanks again for your thoughtful reply.

  3. It's going to take a lot more than technology to improve medicine, and the doctors who practice medicine, at UMass Memorial Medical Center.