Thursday, January 12, 2006

Clausewitz, defensive medicine, and teaching the residents

Karl von Clausewitz was a famous 19th Century Prussian military strategist, whose writings on War (On War) are broadly applicable to all situations in which strategic planning is required. One of my favorite of Clausewitz's observations is on the subject of "friction" in the chain of command. He notes that "countless minor incidents - the kind you can never really foresee - combine to lower the general level of performance, so that one always falls short of the intended goal. Iron will-power can overcome this friction; it pulverizes every obstacle, but of course it wears down the machine as well."

Recently, with one of my residents, I was performing what was expected to be a fairly simple laparoscopy on a young woman on the residents' service (ie, a patient in whom the residents are expected to have the primary role in management). Due to massive intrapelvic adhesions between the uterus and pretty much everything else in the abdomen and pelvis, the procedure took over three hours. When we closed there was some disseminated capillary oozing from the raw surfaces we had separated, but nothing that in my best judgement would put her at risk.

However, what I thought to be insignificant oozing might actually turn out to be worse, so I asked the resident to admit this woman overnight & obtain both an immediate blood count and a repeat count the next morning, in order to demonstrate objectively a stable blood count.

Normally, I take the "iron will-power" approach described by Clausewitz, and I am very strict with the residents, incurring the pros and cons as noted by old man Karl - the job gets done my way, obstacles be damned, but at the cost of wearing down the machine: the House Staff, who have to do it my way despite not being me, and me as well, as I continually have the same conversations about how I don't really care if they are "busy," the tests still need to be checked, the patients still need to be seen, etc.).

In this case, I tried a different approach: give the Chief some latitude - I was confident this situation wasn't risky, and I had been pretty explicit about my orders (how could they screw that up?). It isn't July, the chiefs have three years and six months of on-the-job experience and in just over another five months they'll be unleashed on the general public as doctors.

So I restrained my iron willpower and didn't micromanage the team. But I had operated on this woman and I am morally obligated to continue to care for her (not to mention legally obligated). Late the next day I checked in with the chief resident, and here's how it went:

Me: "So I assume Our Patient was fine and was discharged without trouble?"
CR: "Yep! She had some pain but felt better and went home."

I notice the CR didn't mention the blood count. I'm compulsive, remember? Can't let that slide -

Me: "Great!....and, uh...I take it her blood count was the same?"
CR: "...Actually it dropped from 37 yesterday to 30 this morning."
Me: (pause)
CR: "...uh...but her urine output was fine and her abdomen was soft, and the drop was consistent with the estimated blood loss at surgery."

I would have repeated the test again, given a falling blood count in a patient I was already monitoring for blood loss, even with the reassuring signs mentioned, but I suppose someone with more confidence in their clinical judgment could explain that drop as noted. But I didn't get angry. I just tried to make that teaching point.

Me: " what you're saying is, in your clinical judgment she was fine and the drop was explainable?"
CR: "Right."
Me: "Ok...(walking away down the hall)...but I still would have repeated it at 12:00 before I discharged her." (The blood test is generally taken at 6-7 am).

And that's when the icy hand of Clausewitz threw some Friction our way.

CR: "...Ah...Well...I actually didn't get the result of the first test until 12."

Now this is unacceptable. The test should be drawn on morning rounds, first thing, and be back by 10 am at the latest. Iron willpower wakes up:

Me: "I would have been compulsively checking for the results since about 10 am and then finding out why they weren't in the computer by 11."
CR: "...I did...but the specimen was 'lost' on the way to the lab...the patient even had a band-aid on so I know it was drawn this morning..."
Me: "...ok...I guess that explains your timing, but still, I would have kept her and repeated it at 3 pm just to be 100% safe."

And that was that. The patient turned out to be fine, as expected, but in my attempt to prove this I ran into precisely the kind of "countless minor incidents - the kind you can never really foresee - [that] combine to lower the general level of performance, so that one always falls short of the intended goal" as described by KvonC:
  • The specimen was drawn in the morning but lost
  • Friction in the system is so bad that the CR is used to drawing inferences like "she has a band-aid, the blood must have been drawn," because we have to expect our orders will not be followed. This means we spend half the day just finding out if the things we want done are actually getting done.
  • The specimen was then redrawn but the results were not available until late in the day
  • The combination of the time delay and the low-risk nature of the clinical situation led the CR not to fulfill her orders, as she judged the cost of time and effort outweighed the clinical benefit.
So big deal, right? The patient was fine - why even care about this?

Two reasons. First, it's all too possible for this situation to end with "the patient was not in fact fine and had been bleeding internally all morning." That happens. That's why I wanted her to be checked.

Second, there's the issue of Defensive Medicine. Clinically speaking, the CR was probably right not to bother with the test again - we were just looking for confirmation of the myriad signs that the patient was ok and not really expecting contrary information. But my desire for an "extra" blood count, and "extra" hours of hospitalization for this healthy young woman was DEFENSIVE, intended in part to prove with explict evidence what I "knew," that when I let her go home she was fine.

Why? Because a list of signs and suppositions a mile long can easily be cast into doubt by a malpractice lawyer who after the fact will say, "OBVIOUSLY your judgment was wrong, since she came back 12 hours later bleeding severely!" (Even if it weren't obvious, that's what they'll say). And a stable blood test is my trump card: "The blood count was 30, and 12 hours later it was 30. And that, in addition to everything else, was why I let her go home."

So Clausewitz's observations are a daily experience for me. Even being the iron will can't fix these problems, unless everyone else is also an iron will, but what truly rankles is that the everyday friction is exacerbated by the omnipresent hanging fear of litigation, which requires us to fulfill a standard often not supported by the clinical requirements of our patients, but by the emotional and evidentiary standards of a "jury of our peers."


This page is powered by Blogger. Isn't yours?

Weblog Commenting and Trackback by