Wednesday, October 07, 2009

"If your only tool is a hammer...."

We have a new procedure at our hospital – well, new and not new. The principle has been around for – well, anecdotally, almost forever – but we have come to apply it again. It is also new to our setting.

This “new” tool is hypothermia: lowering the body temperature of a patient to slow metabolism. We’ve all heard the stories of the victim who fell through the ice on the pond. After time under water the victim is removed, cold and apparently lifeless. However, a pulse returns, and with time the victim recovers – not only physical function but mental function, too.

This has been done in hospitals in the past, and for varying conditions, but with varying results. Now, protocols have been developed, and in our hospital and many others it has become a new tool in cases of cardiac arrest when there has been some period with no pulse, and so with probability of loss of oxygen to the patient’s brain. The hope is to preserve brain. (You can see information about it here.)

Not long ago, I asked one of our cardiologists about the studies that had been done to demonstrate the value of hypothermia. I specifically asked about the results of those studies. He said that the success rate was about 10%. That is, the number of patients with good outcomes was 10% higher among those treated with hypothermia than among those who were not.

Now, in one sense, a 10% result may or may not be a big deal. If it’s the difference between respective recovery rates of 90% vs 99%, that’s important. If it’s the difference between 10% and 11%, how important is it? If it's the difference between 50% and 55%, how do we decide? And of course the answer to that question must take into account how devastating the disease (and let’s be honest: the neurological consequences of hypoxia caused by cardiac arrest can certainly be devastating), and whether it’s your loved one in the ICU bed.

That raises some corollary issues. The most important is, for which patients do we use this? The studies I’ve been able to find and look at (mostly in abstracts) are small, and address several different presenting diagnoses: cardiac arrest, certainly, but also cardiac arrhythmia, strokes, and neurotrauma. Some were more pilot studies of one sort or another, either to demonstrate enough response to be worth studying, or to look at different mechanisms for cooling and/or warming the patient. An important issue in research ethics is just how much we can generalize from one study to another – from a small study to a large one, from a procedural study to a treatment protocol, from one diagnostic patient group to another. It is also an important issue in moving from a study to a treatment protocol.

There are also issues around the duration of the protocol. You see, to commit to this protocol takes time. The patient must be gradually cooled to the appropriate temperature, maintained at that temperature for an extended period of time, and then gradually rewarmed. The point of the protocol is to protect neurologic functions, but those functions can’t be assessed until the patient has been sufficiently rewarmed. So, we’ve committed the patient and ourselves to two or more days of intensive care before we can really assess progress. If it’s successful, no one questions the value. However, if that 10% is a small margin in a situation of low expectations, one has to ask whether there has been an appropriate use of resources. There are issues of justice if we’re providing this lengthy and expensive protocol (in all resources, but money symbolizes them all) with little expectation of meaningful recovery.

We’ve also committed the family to two or more days of watching and waiting. During that time, there is little we can tell them. After all, the patient has been put on the protocol because of a devastating event. However, while the protocol is in process, we can’t really assess the consequences of the event, nor advise the family on what to expect. We speak often in health care about the difference between extending life and prolonging dying. If we can’t assess, how do we know which we’re doing, and what can we tell the family about that? We don’t want to give false hope – indeed, in my hospital that is explicitly stated in orientation to the protocol – but to engage the protocol at all is to offer hope before we can know whether it’s reasonable or false. I will acknowledge that this is true of all emergency medicine, and much intensive care; but that doesn’t absolve us of addressing this concern.

What this really presses us to do is ask when it’s appropriate to use the protocol. That really is a question of which patient is an appropriate participant: which diagnoses suggest it, and which related factors exclude it. The thing is that the studies don’t really clarify that. Once again, the few studies address a number of different presenting diagnoses, and those differences make a difference. In addition, any study also has exclusion criteria. A good study is looking at a specific, relatively narrow question, and not every patient would help answer that question. So, there are criteria for including a patient in a study group, and criteria for excluding. However, different studies have different exclusion criteria. That also makes the results hard to generalize. Narrow criteria for inclusion and variable criteria for exclusion add to the difficulty deciding which patients might benefit from this protocol for treatment. Can the patient be too young or too old? Are there other physical conditions that might affect the patient’s survival independent of brain injury, such as liver failure or metastatic disease? It’s possible to look online at hypothermia protocols from various institutions, and see that there are differences between their exclusion criteria (for example, here or here). So, once again, have we invested resources in patients who won’t benefit for reasons independent of this event? Have we offered hope to a family in a situation that is medically futile for reasons independent of this event?

So, how is it that we offer this protocol to our patients? The real answer is that the events are indeed as devastating as I’ve suggested. Loss of blood flow and of oxygen to the brain is devastating, whatever the cause. Our colleagues specializing in stroke have for some time encouraged rapid response to stroke symptoms with the phrase, “Time is brain;” and that’s just as true when it’s caused by heart attack or a fatal heart rhythm. So, in a situation where before we felt there was nothing we could do, now we have something we can try.

Which brings me back to the title of this post. Most readers will recognize the saying: “If your only tool is a hammer, every problem begins to look like a nail.” So it happens in medicine. As hard as we try to measure the effectiveness of treatments, and as hard as we try to be careful and scientific about it, we are also prone to the temptation to “do something, even if it’s wrong.” Or more clearly, we’re more likely to do something because we don’t know whether it’s wrong. I think of it as a special application of the ethical principle of the technological imperative: "we can do something, therefore we must." However, it’s also an application of my own comment on ethics: critical decisions are usually made emotionally, and justified rationally afterward.

The thing is, the technological imperative is a fallacy. It does not follow that because we can, we must. In this instance, we must consider how to use this tool, to optimize the benefits it can offer in light of the costs both in resources and in the emotions of patients and families. Indeed, we must consider these questions with any such tool in health care – something we do reasonably well with drugs but not so well with new procedures or devices. And I would argue it is even more important when the circumstances are devastating; for that is when we are most likely to invest our time, resources, and emotional energies in our actions.

So in my own work I watch and wait with families to see if this will help; and I celebrate with staff and families when it does. That also means I grieve with staff and families when it doesn’t; but that’s the nature of my vocation. And I do ask questions, trying to help focus on what can help, and on what extends life rather than simply prolonging dying. And I especially ask this question: is this problem really a nail, or is it just that all we have is a hammer?

1 comment:

Kirkepiscatoid said...

You're now touching upon something in our "brave new medical world" that makes me nervous--the increasing use of "algorithmic medicine" in certain diagnoses and its tie-in with "performance based medicine."

We have so many more "protocols" than when I started. We are doing them, not just because of scientific evidence based stuff, but b/c you do them right enough times, your hospital gets a designation as a "stroke center," "High risk birthing center," etc.

The flip side, though, is if something is missed in the protocol, and the patient outcome is bad (no matter if it would or would not have affected patient outcome), families now have a new "handle" for lawsuits. It injects a new "false hope" for families if the chance of improvement is low.

It's problematic, at the very least, when we are talking all except for the protocols which show relatively good data in change in outcome.

But let's be real; when people see you using a "cookbook," they're expecting a "cake" at the end of it. They may not understand that this protocol may not add as much hope as they think.

I remember back to a time as a 4th year medical student, when I got the tough task (because I was lowest in the pecking order, of course) to tell an elderly man his wife had succumbed to her cardiac arrest in the ER. My heart broke as he clutched her pill bottle and sobbed, "But...but...she still has pills left."