11.07.2009

Medical Reasoning

Lately I’ve been hearing one tale after another about the problems people are having with their medical care—can’t get an appointment, duplicate billing, failure to return calls, in some cases, calls that require immediate attention and finally perhaps the most frequent, incorrect or delayed diagnosis. Who has not heard such tales?

In the November 2009 New York Review of Books, Jerome Groopman, author of How Doctors Think, gives a thought-provoking account of why patients sometimes receive such poor care. He begins by describing a clinical conference he conducted for interns and residents at the Massachusetts General Hospital.

The conference focused on how doctors arrive at a diagnosis of their patient’s ills. “Some 10 to 15 percent of all patients either suffer from a delay in making the correct diagnosis or die before the correct diagnosis is made.”

At once I was struck by Groopman’s methodological approach to this problem. Unlike the usual one of discussing medical successes, Groopman begins by discussing failures. He writes, “The most instructive moments are when you are proven wrong, and realize that you believed you know more than you did, wrongly dismissing a key bit of information that contradicted your presumed diagnosis…”

Science or any empirical discipline (or individual for that matter) doesn’t move forward by pointing to its successes. If that were the case, it would scarcely ever change. Rather we learn far more from the mistakes that have made, from those cases that disprove our conjectures.

Groopman points out that the most common sources of diagnostic errors are the cognitive biases that physician’s make in trying to understand a patient’s condition. (These errors are not confined to physicians. Rather they are errors that anyone is prone to make in making a decision under conditions of uncertainty). He identifies three of the most common biases:

anchoring where a person overvalues the first data he encounters …; availability where recent or dramatic cases quickly come to mind and color judgment about the situation at hand; and attribution where stereotypes can prejudice thinking so conclusions arise not from data but from such preconceptions.

The second notable methodological point in Groopman’s account is his emphasis on the limits of empirical generalizations in any particular case. He points out that subjects in clinical trial investigations (upon which these generalizations are based) are often highly selective as those who have multiple conditions or are taking other medications or do not fit into a narrow age range (usually too old or too young) are excluded from the study

Groopman comments, “Yet these excluded patients are the very people who heavily populate doctor’s clinics and seek their care.”

The other major sources of physician error stems from the heavy patient load they are now asked to carry. One physician “said she spends less and less time conversing with her patients. Instead she felt glued to a computer screen, checking off boxes on an electronic medical record…”

Another pointed out that “…work rounds were frequently conducted in a closed conference room with a computer rather than at the patient’s bedside.” And finally in describing the case of a seriously ill cancer patient, Groopman reports that “… no one attending to her had sat down in a chair at her [hospital] bedside and conversed at eye level, asking questions and probing her thoughts and feelings about what was being done to combat her cancer and how much more treatment she was willing to undergo.”

This may be hard to believe for anyone familiar with the days when doctors routinely came to your home if you were will or told you to come right over to his or her office if you felt poorly, or indeed, called you at the end of the day to see if you were feeling any better.

In the end Groopman makes clear that the solution to these problems will come about “…only by dogged thinking that requires the kind of time and inquiry that is absent in much of modern medical care." Dream on Dr. Groopman