The Bell Curve

Anesthesiologists overtreat, and it is the right thing to do. If I had known this statement prior to residency, I would have been shocked. In my past blogs, I have written a lot about the problems of doing too much in preventive health screening and the cost of overtreatment. Being too zealous - such as doing unnecessary colonoscopies, prostate cancer screening, or mammograms - can lead to waste, false positives, unnecessary testing, and side effects. I imagine that if I were a primary care practitioner, I'd be prudent in trying to minimize overtreatment.

But anesthesiology is the opposite, and I am only slowly being convinced of it. In anesthesia, we cannot treat to the "average" (mean, median, mode) patient. Say the bell curve above shows the distribution of a patient's response to a muscle relaxant. Often, a patient is paralyzed for the surgeons to work. After a couple hours, the muscle relaxant will have worn off for 95% of patients. The "average" patient will not need reversal of the paralytic. The "average" patient would do fine if we woke them up. But 5% of patients may still have residual paralysis, and waking them up would be unpleasant and dangerous for them. In this situation, I give the reversal agent; I'm not treating the average patient, but rather the tail ends of the normal distribution. In 95% of patients, I have given something unnecessary, overtreated. But I do so to avoid adverse consequences to the last 5% of patients.

I've been struggling with this notion. Why is it that ordering screening mammograms on healthy 40 year olds is inappropriate, that it will lead to unnecessary anxiety, biopsies, cost, and waste? But that giving an antiemetic during anesthesia is acceptable, even though it may be unnecessary for many patients? Is this difference in approach a result of a different culture in each specialty or a different threshold in accepting needless cost or something else? This is one of the facets of clinical decision making that fascinates me.

Image shown under Creative Commons Attribution License.