Experienced doctors, friend or foe?

I read an editorial from the New England Journal of Medicine by AM Stuebe in which she talks about the notion of Level IV Evidence. Doctors and scientists have many ways of classifying studies and one divides them into three basic levels. Level I is the best quality randomized double blind, placebo controlled trial. Level II is anything that is retrospective, unrandomized, and so on. Level III is "expert opinion", or what would your peers recommend. She suggests the importance of Level IV evidence, or the adverse anecdote.

During residency we referred to this phenomenon as IGBO or MAGBO (I got burned once, or my attending got burned once). The general idea is that when caring for a patient, if something goes wrong, you will always, and strongly remember the circumstances and hopefully carry that experience with you during future patient encounters and avoid a repeat performance. Of course, most of the time, these are chance events and may or may not be avoidable.

Logic would seem to suggest that being an older, or should I say, more experienced doctor would allow you to collect more Level IV evidence and thus be a better doctor. Indeed, this is what Dr. Stuebe suggests in her editorial. On the other hand, what if you are the type of person who does not learn from your experiences?

A study published in the American Journal of Medicine in July looked at this question. Southern et al. approached it using evidence from 59 doctors over 2 years and a total of 6,572 hospital admissions. They divided the doctors by how many years of experience they had after finishing their training (ie: residency), under 5, 6-10, 11-20, and over 20 years. When adjusting for differences in variables, and comparing under 5 years of experience to over 20 years of experience, they found that the older doctors on average kept patients in the hospital about a half-day longer and the rate of death while in the hospital was 70% higher!

The limits of the data they used cannot really give us any more information as to why or how this occurs, but it does raise some interesting questions. The authors primary conclusion is that doctors, after finishing their training may continue to gain Level IV evidence but have little time and expertise to absorb Level I evidence. But like I said, we do not really know if this is due to older doctors using outmoded practices or is it due to something else? Are older doctors more likely to rely on their instincts and less likely to second guess themselves? Are they more likely to move forward without asking a consulting doctor for a second opinion?