Great quote by USC cardiologist Leslie Saxon (a reporter reached her on her cell phone as Leslie was shopping) on this week’s NEJM study on delayed defibrillation: “You’re better off having your arrest [here] at Nordstrom [than in a hospital]… because there are 15 people around me.”
You’ve probably seen the study, a detailed analysis of cardiopulmonary resuscitation results of nearly 7,000 patients who coded with “shockable” rhythms (ventricular tachycardia or fibrillation) at 369 hospitals participating in a large CPR registry. The study found that the overall hospital survival rate for these patients was one-in-three. Note that this rate is substantially higher than the usually-quoted hospital CPR survival rate of about 15% because of the exclusion of patients with “non-shockable” rhythms like asystole.
[I can’t resist a little riff on estimated CPR survival rates: Several years ago, Pittsburgh and Duke investigators asked real people to estimate the CPR survival rate in hospitalized patients. Their answer was 70% (ie, a five-fold overestimate), which turns out to precisely match the survival rate of patients undergoing CPR on TV shows! (Just in case you were wondering where patients get their medical information.) By the way, the prognosis of patients undergoing television CPR markedly improved if one of the actors declared, “it’d be a miracle if s/he makes it” before the arrest.]
But I digress. Returning to this week’s CPR study, of the 34% of patients who survived to hospital discharge, 57% left with no neurological deficits and 33% with no functional deficits. Taken together, we see that even a hospitalized patient with an initially shockable rhythm has a surprisingly low chance of returning home both neurologically and functionally intact.
Perhaps, the authors wondered, we could do better. Well, yes. It seems that about 30% of the time, patients did not receive their initial attempt at defibrillation for more than 2 minutes after the arrhythmia was discovered. And – I hope you’re sitting down (the Code Blue team surely was) – 11% didn’t receive their first shock for more than six minutes after the nasty rhythm was discovered. He-llo?
The remainder of the study focused on the risk factors for delayed defibrillation, and the outcomes as a function of these delays. Patients whose shocks were delayed more than 2 minutes were more likely African-American (raising disparity concerns), in smaller hospitals, on med-surg units (rather than monitored units or ICUs), to have medical diagnoses (as opposed to cardiac ones), and to have had the temerity to code on the night shift.
And delayed defibrillation was bad: patients with late shocks had about a 50% lower odds of leaving the hospital alive, and 25% lower odds of avoiding significant neurological or functional impairments.
In an editorial accompanying the New England Journal study, Dr. Saxon (a former UCSF colleague of mine) focused mostly on technological fixes. She’s no doubt partly right. If your heart stops at the baseball stadium, the airport, or the casino, there is likely to be a goof-proof Automatic External Defibrillator (AED) nearby, which can be operated correctly by anybody, including kids and octogenarians. In a hospital, our defibrillators are the fancy, “what do you think that rhythm is, Sue?” types. Why? Partly tradition, I suspect, and partly hubris. It rubs us the wrong way to think that we’re using the same machine they sell at the AED Superstore, and there are no doubt cases in which we need a sophisticated rhythm strip to do the right thing. So how about keeping some of the fancy ones in the ICU, telemetry unit, and the cath lab, while peppering the rest of the hospital with AEDs (Leslie argues that there should be one in every high risk patient’s room, which may be over(un)kill, but probably isn’t too far off).
Dr. Saxon also emphasizes the need for monitoring, arguing for markedly increased use of Telemetry-Lite: new-fangled telemetry systems that don’t require a tele-nurse to review every rhythm, but which apply sophisticated computerized algorithms to look for early hints of arrhythmia. There is even now a bed covering that can detect and monitor your heart rate and respiratory rate (through your PJs), without any electrodes [Truth-in-advertising: I’m a paid member of the Scientific Advisory Board of the company that makes such a product. Still, it is pretty cool and might be a part of the solution].
But Leslie doesn’t touch on the human factors that almost certainly help determine the timeliness and effectiveness of CPR response. A 2005 University of Chicago study found that a university-based Code Blue team got it wrong much of the time, and a recent Northwestern study found that CPR technique can be markedly improved by using simulation and mock codes. It is shameful that we don’t (and aren’t required to) do this. The Fire Marshall makes us perform fire drills – I’d wager that the chances of a hospitalized patient dying because of poorly performed or delayed CPR are 1000 times higher than they are of a patient dying from a dysfunctional response to a fire.
Finally, the study once again highlights the perpetual neglect of the med-surg patient. Even though such patients make up the biggest cohort in virtually every hospital, the fact that they are heterogeneous, often scattered around the building, bring in relatively low reimbursements, are in a comparatively low tech environment, and lack powerful advocates (namely, surgeons and proceduralists) means that it is easy to forget them at every level: resources, training, equipment, and organizational focus. One of the advantages of the hospitalist movement should be the creation of a group of physicians who, working with their colleagues, can advocate for these patients and help improve the quality of their care.
So, if you’re a hospitalist, I hope that you and your colleagues will take a hard look at the way your hospital organizes its CPR response, particularly on the floor – who is on the team, what is their training, do you use simulation and mock codes, do you have the right equipment? – and resolve to make it better. You’ll save some lives.
And after you’re feeling all warm and fuzzy about that, go treat yourself to a trip to Nordstrom.
Many folks have been asking about the shutdown of the Pronovost Michigan checklist study, recently profiled so beautifully by Atul Gawande in the New Yorker, because of issues concerning informed consent. There have been some excellent articles and blogs (here and here) on this problematic example of regulatory overreach. I’m just gathering my facts and thoughts (and speaking to the principals), and will write on it in the next few days.
Hope everybody had a terrific holiday.