To achieve great clinical care, avoid perfection!

There is an aphorism generally attributed to Voltaire which says that "perfection is the enemy of the good."  I happen to like this one a lot, because it serves as such a helpful starting point for almost all organizational improvement work. More on that another time.

Right now, however, I'm thinking about the way this phrase applies to clinical decision-making and the selection of treatments.  I was chatting recently with a very smart infectious disease specialist about my observation that every third patient on our Medicine service seems to have pneumonia and a urinary tract infection.  Not pneumonia or UTI, but both.  I confessed to her that I'm a little reluctant to use my (pediatric) reflexes to form opinions about Geriatrics, but that I was nonetheless skeptical about so many patients experiencing two, simultaneous and unrelated infections.

She said there was a simple explanation. "Most of those patients don't actually have UTI."  What they likely have, she went on, is asymptomatic bacteriuria, a common condition in which the urinary tract is colonized with bacteria in a quiescent state.  There are generally few inflammatory cells, and no symptoms.   Now asymptomatic bacteriuria is neither uncommon nor obscure.  So why would our physicians so routinely mistake it for UTI and prescribe unnecessary antibiotics?  Here's why.

Usually, infections of the lower urinary tract cause painful and frequent urination, while kidney infections cause fever, chills, abdominal or back pain and other signs of systemic illness.  Once in a great while, such infections may cause patients with dementia to be especially confused; moreover, many believe that the elderly are less likely to experience the typical symptoms of UTI. It is this confluence of facts that sets the stage for over-treatment.

In fact, we have seen this many times before.  Here's the syllogism. Condition "X" may be associated with vague symptoms. My patient has vague symptoms.  Therefore, my patient has condition "X." (This is of course a medical illustration of the famous false syllogism: A coffee table has four legs.  My dog has four legs.  Therefore, my dog is a coffee table.) In this equation, X may stand for low blood sugar, occult systemic candidiasis, Lyme disease, gluten sensitivity, various vitamin deficiencies and a host of other conditions.  Which one depends on the newest "under-diagnosed" condition, the patient's age and the physician's personal preference.

What, you ask, has this to do with Voltaire?  It is the fool's errand of seeking perfect diagnostic sensitivity that leads us astray.  The probability that each of the patients on our service diagnosed with UTI actually has the condition is not zero.  But absent any symptom but confusion in the setting of dementia, the probability of urinary tract infection is low.  Since asymptomatic bacteriuiria with or without pyuria is quite common in the elderly, the diagnostic significance of any positive result is likely small.  The question is, how does one best plan a diagnostic evaluation, and then assemble the evidence into a coherent impression?  And in doing so, how does one accommodate diagnostic uncertainty?

Some try to sidestep the problem.  "Better to over treat than under treat," they say - often with great confidence.  Sometimes that's right, as when the consequences of delayed intervention are dire.  Other times, however, the opposite is true.  No, one cannot expect to get the best clinical outcomes by ignoring the obligation to make the best decision possible under conditions of clinical uncertainty.

Osler, of course, said it best,"Even the best of men must be content with fragments, with partial glimpses, never the full fruition."