Three Remarkable Articles Last Week

As I mentioned when we launched, this blog won’t be your destination for a weekly journal update (there are plenty of sites for that). But I will keep an eye on the literature and let you know when I see something remarkable. And then I’ll try to put it in context.

Last week, there were three such studies.

First, an abstract presented at the American Heart Association’s annual conference on the efficacy of perioperative beta blockers. Remember, beta blockers were the “slam dunk” of process-outcome links – after a seminal 1999 New England Journal study found an jaw-dropping mortality benefit, periop beta blockers became one of the first surgical quality measures. In fact, the intervention seemed so good that all of the mumbo-jumbo of periop risk assessment flew out the window. Just give beta blockers to everybody, went the reasoning, and stop worrying about treadmills, persantine, echos…

But since the 1999 paper, the premise that beta blockers should be in the perioperative water has suffered a death from a thousand cuts. By now, the cuts have been deep enough that the beta blocker story is starting to resemble the name-calling extremity-less torso in Monty Python and the Holy Grail. Even before last week’s announcement, what began with “beta blockers for everybody” had gradually tapered to “make sure you don’t stop the beta blockers if the patient has been on them at home.”  

And then came last week’s POISE trial, which found that perioperative metoprolol doubled the risk of stroke and increased overall mortality risk by one-third. We’ll have to wait for the peer-reviewed paper before passing final judgment, but this may be the final blow for perioperative beta blockade, both as an intervention and a quality measure.

Several bloggers likening the beta blocker story to the much maligned four-hour door-to-antibiotic pneumonia measure (another flawed publicly-reported quality measure), but I see them as representing different problems. The pneumonia measure’s hiccups were a matter of unintended consequences – nobody realized (prospectively) that a four-hour door-to-antibiotic measure would put ED docs in a double bind: do I give antibiotics to look good on the report card, or do I wait till I’m sure the patient actually has pneumonia? (By the way, CMS and the Joint Commmission recently announced that that this measure will change from 4 to 6 hours in March, in large part because of blowback from emergency medicine physicians).

No, the perioperative beta blocker story is one of a quality measure that was promulgated after a few positive but not ironclad studies were published in prominent places. Were the studies strong enough to merit a beta blocker quality measure? I’m not sure, but the consensus back in the day was that this was a powerfully effective intervention. Luckily, researchers – recognizing that the science was imperfect – continued to pursue the question, ultimately leading to the POISE announcement last week.

To me, this is the way quality measurement is supposed to work. Some will argue that quality measures should wait for absolutely bulletproof evidence, but I do think that there are times when a quality measure can be promoted based on strong, albeit imperfect, evidence – as long as we continue to study the matter and commit to tweaking the measure as the science improves.

Next study, this one from my favorite medical journal, the New York Times, whose column last week by my friend David Leonhardt became the talk of the cocktail party circuit. David described a clever study by a young Canadian economist, who brought his deliberately sabotaged Subaru (battery cable loosened, coolant removed) to 40 service stations to see how many gave him the honest and correct skinny. The depressing result: 27 of the stations told him about the loosened battery cable, and only 11 mentioned the low coolant.

Leonhardt points out that the study illustrates the “expert service problem,” situations in which consumers aren’t sophisticated enough to make independent judgments regarding experts’ pronouncements. Would doctors perform this badly? I hope not, but it certainly throws a monkey-wrench (pun intended) into the fashionable notion of consumer-driven healthcare purchasing – which naively assumes that patients can distinguish high quality from poor care. In high-stakes and complex arenas, consumers (patients in our case) are still going to need to rely on experts, notwithstanding efforts to arm them with information. And the Subaru story argues that patients will need some “consumer protection” from unscrupulous docs – in the form of board certification, peer review, competency assessment during re-credentialing, and – gasp – the malpractice system.

Finally, my favorite study of the week, the JAMA paper that showed that a body mass index of about 28 – which happens to be mine – was associated with lower mortality than that of thin people across a whole range of illnesses. As I sit here in Starbucks sipping my mocha topped with whipped cream (who knew that it would be OK for an adult to drink something like this!), this study replaces my old favorite (which I can no longer find – it may have antedated PubMed): the one that demonstrated that, yes, jogging does improve longevity, but the extra time alive is precisely equal to the amount of time you spent jogging.

Couch potatoes of the world, rejoice!