Thursday, July 29, 2010

Dr. Gawande, Hospice, The New Yorker and the Curious Absence of Some Inconvenient Truths: Here's the Rest of the Story

Media Golden Boy and eponymous web site star Atul Gawande has done it again with a compelling essay about death in the August 2 issue of The New Yorker. Pulling from a grab bag of instructive anecdotes, he describes America's perfect storm of imperfect dying. It's all thanks to the twin ingredients of a culture unprepared to think about mortality combined with a curiously passive health system willing to offer up any treatment, no matter how distant the odds. Steering clear of the "shoals" of blaming dysfunctional economic incentives, Dr. Gawande reminds us about some handy solutions: living wills, which introduce the living to the ars moriendi and hospice, which helps the dying meet a good end.

Unfortunately, expert essayist Dr. Gawande has skipped over some inconvenient truths. The DMCB brings you the rest of the story:

When confronted with a critically ill cancer patient (of the kind compellingly described by Dr. Gawande), popular culture would have you think the physicians can predict the likelihood of not making it out of the ICU alive and can therefore treat accordingly. The problem is that the prediction is far from perfect with an ROC, according to this study, of about 0.8 (where 1 is perfect). In other words, there are enough false positives to give physicians pause before recommending pulling the plug. By the way, for non-cancer patients, the prognostic tools are even worse.

Here's a DMCB anecdote: a patient with wildly metastatic colon cancer showed up in its clinic after a yucky ICU stay with a huge grin and said "Hey doc! Surprised to see me?"

Speaking of yucky ICU stays, common sense tells us that given a choice between a "florescent" ICU death with emotionally distant nurses versus a sunset-lit bedroom death with family, everyone would chose the latter. Suppose, however, if you asked ICU survivors about the wisdom of rolling the dice for that third option of surviving? There is research that says up to 70% would do it all over again and that, among the elderly, a majority are often satisfied with their quality of life afterwards.

Another DMCB anecdote: The DMCB accepted the transfer of an elderly patient out of the ICU onto its inpatient service. When asked about being at the business end of all those plastic tubes for all those weeks, the patient's response was "I don't remember much."

Unfortunately, neither does DMCB and its physician colleagues. Armed with state-of-the-art adult learning theories aimed at the promotion of a patently obvious societal need, the SUPPORT investigators set out to help physicians solicit patient preferences about end-of-life care. They failed. Dr. Gawande's New Yorker article makes for a great read, but it's a recycling of nostrums that's unlikely to change anything.

Which begs the question: why the interest in changing things? Could it be that the 12 months prior to death accounts for a quarter of Medicare costs? As a result, government and commercial insurers have a considerable economic incentive to steer patients with a poor prognosis away from costly health care services. The DMCB recalls another anecdote from when it was a medical director: a physician with special interest in end-of-life care asked the DMCB's managed care plan for financial support to promote living wills and surrogate decision making among our enrol lees. Among the many right reasons for doing this, he also suggested the savings from reduced claims expense would be huge and we'd make a killing. We agreed and took a pass on the creepy opportunity. Years later, this same conflict of interest was the key ingredient in the traction behind the otherwise despicable "death panels" controversy

Readers may also recall that hospitals are generally reimbursed with a fixed payment "DRG." Since the inpatient payment (with many exceptions) per patient is typically capped, patients with prolonged stays in the ICU often turn out to have costs that exceed the payment. Accordingly, steering patients toward palliative care can have a tidy return on investment. Now that is yucky.

Last but not least, it's ironic, given our national manic compulsion for "evidence-based" science to guide treatment decisions, that the research supporting the benefits of hospice is decidedly shaky. Dr. Gawande only quoted some studies that happened to support his point of view. That is the luxury of writing for The New Yorker. Unfortunately, it doesn't inform policy and one can only hope the White House will not repeat its past mistake and not pay attention.

So, what could be done? Writing in Population Health Management, Dr. Kash and colleagues point out that disease management for oncology can facilitate informed decision making about end-of-life options because it's typically external to the risk bearing sponsor and has less of a conflict of interest. The DMCB performed unpublished research that demonstrated that nurses can fill in for docs and are quite capable of doing the heavy lifting on promotion of living wills.

In addition, measurable outcomes in end-of-life cancer care are well within reach, according to Don Fetterolf, also writing in Population Health Management. They include the number of times cancer drugs are used "off label," (particularly as second or third line treatments) as well as the percent of patients that received chemotherapy within 14 days of death. These measures are not supposed to be zero, but Dr. Gawande points out that the number seems to be unreasonably high.


7 comments:

Roger Collier said...

"As a result, government and commercial insurers have a considerable economic incentive to steer patients with a poor prognosis away from costly health care services." ... Well yes, Jaan, but since it's pretty much all of us who are paying taxes and/or insurance premiums, this is an economic incentive in which we all have an interest. So, how should we control end-of-life costs?

Jaan Sidorov said...

It's ironic that with the government's foray into going beyond just PAYING for health care ("usual and customary") to promoting QUALITY that there was such political blowback over EOL care. Like it or not, health care consumers will never trust Medicare on the issue, because it now shares in the same conflict of interest.

What would I do?

First off, the percentage of costs occupied by EOL care has remained stable over the years: if we want to reduce that cost, we have to reduce TOTAL health care costs, i.e., bend the curve.

Secondly, I'd forget about any big band solutions: a bill, a regulation, a speech. This is a journey that will take at least a decade, and it's small steps, including the promotion of ground level quality initiatives that promote the discussion between provider and patient; for example, many disease management have been routinely doing this because it is a long established feature of care planning.

AND as mentioned in my blog, I'd start looking at the extremes of one side of the tail of variation with credible measures, such as the percent of patients getting aggressive treatment within "X" days of death. If it's high, that's a problem. It should be in a medium range. There is probably good benchmark data out there: let's find out what it is and make sure it fits with consumer expectations.

Last but not least, this may be were Berwick's "Confessions of an Extremist" could help in real world terms. Patients have an absolute right to make the calls. Using the right kind of Shared Decision Making could make a difference.

skierpage said...

Unlike Atul Gawande's piece,to a bystander your post is unclear and poorly argued (maybe understanding the links help); your inconvenient truths seem murky or unrelated. In particular, whether the people who survived ICU would do it all over again is a weak argument. I want the best quality of life overall, including all the futures in which I die.

Your cynical first answer to "why the interest in changing things?" is an insult. Millions of people die every year receiving medical care they didn't ask for and don't want. Everyone thinking about that, including you, is a hero. (And it doesn't BEG the question, it RAISES it.)

Phil 314 said...

The "problem" with advanced directives is they primarily direct when one has no hope for meaningful recovery. The harder discussion is how to express one's health choices and values prior to those situations that might put you in a permanent vegetative state (such as complex open heart surgery in someone who already has poor lung and kidney function0

This book has, in my experience, one of the best approaches to such a discussion. It doesn't require a doctor or a governmental bureaucrat or an insurance nurse. And here's and example of the form.

eric said...

The issue is deeper than a doctor's inability to predict a person's last illness or a person's inability to choose what they might want done in every possible situation. It is a spiritual matter of deciding at some point that I won't be greedy for more life and will accept the end no matter how it comes. "Is there a duty to die?" asked John Hardwig in The Hastings Center Report, volume 27, 1997. Arthur Kleinman exhorts us to avoid medicalizing a normal process and to recognize that people near the end of life "are experiencing serious suffering, a crucial
moral condition: one that doctors of past generations
recognized and dealt with as fundamental to
living and dying with meaning...When what is often needed is
a conversation opener so that people at the end of
life can express and explore the most tellingly
human of fears and concerns."

eric said...

source for Arthur Kleinman quote was Mount Sinai Journal of Medicine, volume 73, 2006.

Hank said...

The survey of survivors says to me the opposite of what you intend with it. 30% of people who win the medical lottery and survive their stay in the ICU wouldn't do it over if they could go back again! That's shockingly high to me. These are the fortunate ones!

I also don't think it's reasonable to describe hospice as "pulling the plug", its just prioritizing quality of life over quantity of life. For every patient with a very poor prognosis who makes it, several, possibly dozens or hundreds, will have their health destroyed prematurely by traumatic low odds treatments. We're not talking about taking away people's options but rather being responsible enough to inform them of the honest costs and benefits of their choices.

Frequently patients will feel that lower cost courses of treatment offer a higher value set of contingencies than more expensive ones. Do people really need to make themselves miserable in the last months or weeks of life just to squeeze every last nickel out of those greedy insurers?