A Foolish Consistency: Policy, Practice, and Rationality in Medical Decision Making

"A foolish consistency is the hobgoblin of little minds, adored by little statesmen and philosophers and divines" - Ralph Waldo Emerson

As I was making coffee this morning, I grabbed a filter and realized that two may be stuck together. I faced a common dilemma: do I bother to separate them so as to not waste a filter? A trivial example, I will use it as an analogy for decisions we face every day in life and medicine.

Folks approach this kind of decision (or the decision to brush one's teeth, or wear a seatbelt) in various ways. Paul may have a "Paulicy" of never wasting anything, and separates them; doesn't-give-a-damn Sam would never bother; Sometimes Suzie separates them when she has time, and doesn't when she's in a hurry.

The decision hinges on the value of a filter in terms of cost, waste, etc. The value of a single filter is a rounding error, but clearly a lifetime of wasting filters can add up to a lot of filters magnified by the number of people wasting them. Why not go to great lengths to save every filter possible?

In this kind of situation, a rational agent may decide to institute a policy of doing things in a certain way, but accept deviation from that policy under special circumstances. My favorite example of this situation is in medicine with central lines. A left-sided line is not supposed to come at a 45-degree angle to the wall of the SVC; rare cases are reported where, over time, it can erode through the great veins and cause fatal hemorrhagic mediastinum. But what is the probability of that? Something on the order of one instance per 10,000 lines. Supposing you have just placed a left internal jugular hemodialysis catheter and the tip is at the proscribed 45-degree angle. You can't retract that catheter, and you can't advance it either. Do you replace it with a longer one? What if your right internal jugular 20 cm catheter is "too deep" into the right atrium; do you retract it and use the little clamp, knowing that often the line slides in it, increasing risk for infection? What if the upper great veins are suboptimal targets, but you've read the Parienti paper in NEJM and know that the risk of infectious and thrombotic complications doubles (but increases only a percent or so, absolute) with femoral vein catheters? Should the femoral vein be verboten?

In cases such as this and many more analogous ones, a rational agent may employ a general policy of always trying a priori to avoid the 45-degree angle; select the appropriate length catheter; and favor the upper access sites, but to allow exceptions whenever the probability of an untoward outcome in an individual case is low, and the costs of avoiding it are higher. In the case of the 45-degree dialysis catheter, the risks of infectious and mechanical complications of repositioning or replacing it exceed the benefits, especially if it is going to be in place for a short period of time. Reluctance to use the femoral site which has but a 1% increase in adverse events compared to the jugular can lead to far greater complications from forcing patients to lie flat for prolonged periods, or from failed attempts under more difficult operating conditions with non-collaborator patients (delirious, thrashing). The policy to default to a preferred standard becomes irrational when, in a specific case the preferred standard doesn't deserve its status.

The same situation is true with many policies. Studies have shown that a policy of routine use of right heart catheters in sepsis or ARDS does not benefit patients (it also does little harm). So, we have policy of NOT doing routine Swan-Ganz catheters. But that hardly means that in specific circumstances we should not do them. The same holds true for so much of what we do in medicine. Part of the problem is that physicians memorize the policy and the direction of harm/benefit ("there's an increased risk of DVT with femoral lines!"), but they don't know the magnitude of the effect.

So, to be a rational agent, you should establish a policy of what is best, under optimal conditions, for all-comers. But when you have more information a posteriori (because the line went into the wrong place; or you forgot your helmet; or your kid fell asleep in the car seat), default to the policy only if a cost-benefit analysis justifies it. Also, be on the lookout for situations where the all-comer policy does not apply. Like the 90-year-old patient with a small (250cc) complicated parapneumonic effusion. Policy says drain. Rational practice says hold tight; it's a small target, risk of complication or failed drainage higher, benefits of avoiding a small fibrothorax (losing 250cc of vital capacity) in someone with reduced life expectancy and limited mobility are small.

I was not in a hurry, so I separated the filters. Had I been running late or the kids were screaming, I may not have.

Edit to add later in the day: it occurred to me that I do a lot of anal-retentive things in my home, many concerned for the environment or recycling. Example: we don't use paper towels, or disposable plates/silverware. 

The consequence is when we have a get-together, guests can be out of sorts with my "environmental policies." Even though our family, as a routine policy, is (I hope) happy to adapt to these well-intentioned standards, the social costs of this routine policy are off-putting to my guests. So, I keep things in perspective: this is just one gathering, a roll of paper towels here and there is no big deal to avoid the off-putting social costs. So we use disposable stuff on occasion when the convenience in a specific instance overrides the general routine policy goals. Same would be true if you had a general policy of "no shoes in the house" that you let lapse when you invited guests into your home.

Many doctors would do well to make such exceptions to their general policies, in the name of building (or not spending) "social capital."

The Number Needed Not To Treat To Harm (NNNTTTH): A Heuristic for Evaluating Trade-offs in Medical Decisions

A frequent conundrum of decision making that arises in medicine is when there is a generally indicated therapy, say, anticoagulation for atrial fibrillation, that poses unique risks in a particular patient.  CHADS2 and HAS-BLED scores are calculated, but don't quiet the hemming and hawing or quell the hand-wringing.  What is usually a simple dichotomous decision is now one laden with probabilities, risks and benefits, and compromise between competing objectives.  (See:  The Therapeutic Paradox:  What's Right for the Population May Not Be Right for the Patient.)  In order to restore nuance to the decision, we need to try to estimate the numerical values of the risks and benefits to determine if the net utility of anticoagulation is positive or negative, something the aforementioned calculators are intended to do in a semi-quantitative way.  But what if you opine that your patient has a specially enhanced risk of side effects and you're worried about falls or bleeding but ambivalent because of a concurrent fear of denying him of the benefit of stroke prophylaxis?  What if you think that he would have never been included in a trial of stroke prophylaxis and the results of those trials may have limited generalizability to him?  What if you think he has only a year to live?

The number needed not to treat to harm (NNNTTTH) is the number of patients whom you have to not treat with something beneficial in order to cause one harm from your omission.  It is numerically equivalent to the number needed to treat (NNT), but it reframes the decision from action to omission and from benefit to harm.  Ignoring bleeding altogether (because making relative utilities for bleeding and stroke is a fraught endeavor), you could ask yourself "how many patients can I withhold stroke prophylaxis from for one year before I statistically cause (or allow to happen, if you are prone to omission bias) a stroke?"  For most patients, withholding stroke prophylaxis has a NNNTTTH of about 25-30 per year (check the corresponding NNT from CHADS2 for a more "precise" estimate).  Reframing the question into "how much am I asking the patient to pay, in terms of statistical likelihood of stroke, to avoid anticoagulation and the particular side effects that cause me concern in his case?" can often provide some reassurance for the clinician and the patient alike.

Don't Judge a Brain By It's Scan: The Importance of Actionable Information in Medical Decisions

A common and vexing problem in ICUs everywhere is posed by the patient with diffuse encephalopathy or coma - patients who are colloquially said to be "out to lunch" or "the lights are on, but nobody's home."  The underlying cause is usually a toxic, anoxic, or metabolic malady and the only available treatment is to remove the underlying insult (toxin, hypoxia, or metabolic derangement, e.g, uremia) and wait patiently to see if the neuronal injury improves or resolves.  That is perhaps the hardest part - waiting patiently, with all the attendant uncertainty and resulting anxiety, for days, or even weeks.  This, despite my knowledge that the vast majority of these patients, save for those with severe anoxic brain injury (ABI) will mostly or completely recover with supportive care and "tincture of time."  It is very difficult for me, as a physician, to watch, mostly knowing that the patient has been "out to lunch" for the past 72 or 96 hours because of drug effects, which is very common.  "Shouldn't the drugs be cleared by now?", I keep asking myself.  "Maybe I'm missing something," comes the id's reply.  This anxious uncertainty serves as an impetus for action - but what action is there to take, and is it helpful on balance?

The urge usually is to get a variety of neuroimaging tests, CT (the "donut of truth"), MRI, MRA, and probably an EEG.  But I resist this urge mightily.  These tests rarely yield actionable information - that is, data that I can act upon to change the course of care, as an alternative to waiting.  Say for example, as is often the case, the CT scan and the MRI scan are normal.  How has this helped me?  I still have a comatose patient, and I'm still stuck waiting.  (Also, without meticulous care in explaining the test's purpose and results to her family, confusion ensues.  "If the MRI is normal, why is she in a coma?" is a commonly uttered frustration.

Alternatively, the MRI can look horrible and the patient can still do very well.  I recall a case about a decade ago where we sent a patient with hepatic/toxic/metabolic/anoxic encephalopathy to "the magnet" (MRI scanner) and the neuroradiologist soon after called:  "Is this patient still alive?" he demanded.  Yes, indeed she was.  "Well she won't live long, her whole cortex is necrotic! [rotten]," he forcefully propounded.  Two days later the patient awakened.  I do not believe the radiologist's read of the MRI was in error, only that he erred in confidently inferring that function follows form.  "Don't judge a brain by it's scan" might be a useful heuristic here.  Telling this anecdote to patients' families led me to stumble upon a communication pearl:  My not getting the scan is in essence giving the patient the benefit of the doubt - because I don't want to find something that looks bad and ruins my optimism.  This is generally a very positively received explanation for families who may themselves be demanding scans, action, do something, Doc!

There are indeed cases in which radiological imaging does reveal important and undeniable information, such as stroke, hemorrhage, herniation, and swelling.  I do not impugn all scans, many times they are very useful.  I only wish to cast skeptical doubt on that large fraction of scans which are done as a pat response to non-specific encephalopathy where they are unlikely to yield actionable information.

Beyond the costs, radiation, and risks of transportation associated with the routine use of these scans, which are often not counterbalanced by the yield of actionable information, there is a psychosocial cost - in my observation, these scans increase families' anxiety.  Firstly, when you say you're going to order an MRI, false hope that we can learn what cannot yet be learned (until enough time has passed for the patient to awaken) is engendered.  The family inevitably wants to know when the MRI will be done.  The nurse says 4 PM, there are invariable delays, it gets done at 10 PM.  Anxiety pervades all of those hours.  When will it be read?  Nobody knows, but we offer "in an hour" as a guess.  More anxiety, until 1AM when the report becomes available - but it's inscrutible.  It talks of diffusion weighted this and that, and the neurologist is not available to put it into context.  Anxiety mounts.  In the morning, the neurologist attempts to explain all these little areas of this and that, or a normal scan or whatever.  But none of it tells the patient's family what they want to know, namely "Will my mom awaken?  When?"  Those questions remain unanswerable, but the family has been put on a psychic roller coaster for the past 24 hours thinking the MRI will answer them.  When it is finally done, the reason they can't divine the meaning of the scan is because there is no meaning of the scan - it is devoid of actionable information and should not have been ordered in the first place.

An Opportunity Lost is an Opportunity Cost: Doubling Down with Your Final Days

Experience and study teach that decision making in and about life is more akin to chess than it is to checkers.  A good decision maker will think several steps into the future and will consider multiple alternatives, not just the obvious first order choices.

 In medicine, we are often perilously mired in first order choices, to the detriment of patients.  We act as though there are just simple binary choices to make, such as treatment and life versus no treatment and death.  Would that it were so simple.

Someone I knew, a decade ago, made a courageous choice.  Faced with the grim prognosis of an aggressive metastatic cancer, he elected to forego any treatment and take his chances.  He left the hospital and got one month of relative freedom from medical burdens.  He got his affairs in order.  He selected his own grave site.  He visited with friends and family and doubtless did countless other things that he could not have done had he elected to receive chemotherapy and or radiation or debulking surgery or basically any medical intervention given the desperate nature of his case.  His last month could have been characterized by painful procedures and  repeated scans, nausea, vomiting, and anorexia induced by chemotherapy, cumbersome trips to radiotherapy - you get the picture.  In another parallel universe his doppelganger, selecting "treatment" for this runaway cancer, would have traded away the last month of his life - and probably would not have benefited from the trade.  An opportunity lost is an opportunity cost.

Let's look at another case.  An 82 year old frail man on dialysis for several months presents with increasing fluid in the chest around the lungs   Initial (non-invasive) testing suggests malignancy/cancer.  The patient can be steered in several directions ranging from hospice (given age, functional status, and co-morbidities, the prognosis is poor), to straightforward diagnosis via needle biopsy sampling (of fluid or superficial lymph nodes) followed by consideration of treatment options, or to aggressive video assisted thoracic surgery under general anesthesia to take control of the fluid, scar down the lung to prevent recurrence, and get a definitive diagnosis by surgical biopsy.

Note that as the aggressiveness of the diagnostic and treatment approach increases, so does the likelihood of never leaving the hospital as a result of complications.  An opportunity lost is an opportunity cost.  And the patient who selects or is steered towards the most aggressive treatment option may well pay with their final days for the marginal chance of improving their outcome - measured, of course, in the number of days that they can live unencumbered at the end of their lives.  In this case, regretfully, an aggressive approach was taken, and the patient died in the hospital, on life support.

We have to be very cognizant of these costly lost opportunities when we present options to patients, lest they double down with their final days.