Spiraling Empiricism: When in Doubt Put on Blindfold and Shoot

By Abraham Verghese

microscopes.JPGOne of the reasons I went into infectious diseases as a sub-specialty of medicine was that I enjoyed the notion of being a sleuth at the bedside, of seeing someone with pneumonia and examining their sputum by personally carting it to the "wet lab" adjacent  to the hospital ward, then smearing a tiny sample on a slide and then, using a time-honored method of staining bacteria called the Gram stain, coming up with the exact cause of the pneumonia and thereby beginning appropriate treatment. 

A better example, and a memorable one from my training days, was seeing a patient with HIV (in the first days of AIDS when we did not know the viral cause) and watching as his fever raged on despite our efforts. One morning, I spotted some unusual skin lesions on his forehead. I scraped one with a needle and put the material on a slide. In the wet lab under the microscope, I saw tons of small budding yeast: this was cryptococcus, a fungus that affects persons with HIV.

There are so many instances in infectious diseases where a rapid and precise diagnosis can be made by directly examining sputum, wound drainage, blood smears, skin lesions and the like. Over two decades I can still remember those "Eureka" moments; it made all the difference in the care of the patient by finding precisely what was causing the problem, short circuiting what was otherwise blind therapy with many drugs, hoping one was treating the cause of the infection. Diagnosis matters--surely that's something we can all agree on.

Perhaps not. Some years ago, OSHA closed down our wet labs."These historic rooms--where interns and residents in the night or at any time could look at urine, use a Gram stain to look for bacteria in various specimens, and do other simple tests--were gone. Admittedly, they were messy rooms, and I suppose there was some potential for exposure to unsafe pathogens. But it was also the place where a generation of interns and residents learned bread and butter tricks that made them better diagnosticians. We lost that battle.

Once the wet labs closed, the choice was then to walk over to the microbiology department. Well, I got used to that, and in many ways this worked well, as the lab had more resources, better microscopes, and you could also look at the culture plates with the microbiologist and see what was brewing. Though I was willing to make the walk--microbiology is integral to my specialty-- for busy interns and residents, the walk made it less likely they would drop by the lab.

But then, many hospitals (including some where I see patients) went one further and have farmed or moved their microbiology labs off site, miles away! (I kid you not.) Say I happen to see a patient without a spleen who has come in with fever and shock, someone in whom I might make a diagnosis in minutes by quickly examining the buffy coat of blood (because such patients are prone to overwhelming infection so that their body fluids will be teeming with organisms);  now all I can do is helplessly watch as the specimen is transported away, with inevitable delay. And what is worse, more and more, no one seems to be too bothered. No one seems too worried about what exactly is causing the infection, but rather people are content instead to empirically add antibiotic after antibiotic, to "cover" the patient. And if the fever persists and the patient is worse, well then the answer is to keep "broadening" the antibiotic coverage. It's kind of like flushing out a sniper who is terrorizing your own city by bombing the city  to rubble in the hopes of silencing the sniper. Diagnosis matters.

bacteria.JPGMy friend, Victor Yu who is a wonderful clinician and Legionnaire's disease expert, has called this phenomenon "spiraling empiricism" in an editorial he wrote for the journal Chest this year. It is a wonderful term to describe the illusion that you can treat infection by just shooting in the dark. The consequences of such blind empiricism (and I see this all the time) is not just cost, but bacterial resistance, diarrhea and severe colitis related to antibiotics, not to mention the direct side effects of the drugs themselves. Indeed, as an infectious disease specialist, the number one thing I seem to be doing on hospitalized patients is stopping the cornucopia of antibiotics they seem to be on. We are great at starting empiric therapy, but terrible about stopping. And in a health care system that my colleague Alan Garber describes as a menu without prices, it is all too easy to tick off more items on the menu. That's not patient care. That's not diagnosis. 

As Congress wrestles with cost (and that is really the crux of health care reform: paying too much for things we don't need, and having perverse financial incentives to make us do things to patients that they do not need), this is the sort of nuance that does not get discussed. Indeed, the push for efficiency and "quality" has every hospital touting "pathways" and "algorithms" for the treatment of pneumonia. And with the focus on "outcomes" research we will probably be saddled with more pathways and algorithms. It is commonplace to see patients being wheeled down the "pneumonia" pathway and meeting all the quality and other metrics that measure a hospital's efficiency, only for me to disagree with the label of pneumonia. Diagnosis matters. Patients would concur, even if we seem to have forgotten.

We have overvalued the electronic medical record and its hypothetical cost savings (except that it might keep doctors and nurses so glued to the monitor that the patient will get bored and check out from the hospital) and we have underestimated the sheer waste, cost and danger of clinical error that have come about from the "business" decision to locate diagnostic microbiology labs somewhere so far away that the clinician has little chance to look at the specimen on their own patient. That is just dandy if we don't care about diagnosis.

Let's give ourselves a chance at precise diagnosis before we treat. That means good specimens, hand carried, examined by the people who care for the patient. Proxy wars never seem to work.  "Find the enemy and win the firefight" is a good philosophy for infectious diseases as it is for war. Diagnosis matters.


Photo Credit: Flickr Users US Army Africa and kaibara87

This article available online at:

http://www.theatlantic.com/technology/archive/2009/12/spiraling-empiricism-when-in-doubt-put-on-blindfold-and-shoot/32209/