How can we make a best guess with incomplete information?

I have spent three days at the Society to Improve Diagnosis in Medicine. Whenever I come to this meeting, I have insights from listening to talks and many conversations with leaders in the field.

When one considers diagnostic errors, one must consider two important factors: physician factors and system factors. We have a major system factor that can cause problems. Most hospitals in the U.S. require a diagnosis for admission. I believe this rule increases diagnostic errors.

Currently, I am developing a talk based on the lessons we can learn from fictional detectives. One example that I will likely use came from a relatively minor TV show: the Glades. In one episode, the detective is walking around the crime scene making observations. Someone comes to him and asks what he thinks happened. He turns and tells the person that he is still collecting data and does not have enough to develop a hypothesis (or something like that).

Patients get admitted, either from the emergency department or from an outpatient office because the patient is clearly not well. But too often the admitting physician really does not have enough data to make a correct diagnosis. When the system requires an admitting diagnosis, they give a best guess diagnosis.

But how can we make a best guess with incomplete information? When the patient comes to the receiving physician (often a hospitalist), the patient already has a diagnostic label.

This sequence, which occurs every day in most hospitals, often results in premature closure.

Back in the day (talking about the 1970s), we received patients with a list of symptoms or abnormal laboratory tests. Sometimes the admitting resident would call and tell us that the patient is sick.

Could we decrease diagnostic errors through a simple process of delaying writing down a diagnosis until we had sufficient data? Some physicians ignore admitting diagnoses and start over with each patient. But too often we see insufficient questioning of premature diagnosis.

So please comment on your experience. Am I just a jaded old guy or does my hypothesis have merit? Should we challenge the current status quo and ask for this “requirement” to change? Am I thinking to hard about this issue, or is it actually very important?

Robert Centor is an internal medicine physician who blogs at DB’s Medical Rants.

Image credit:

View 1 Comments >

Get KevinMD's 5 most popular stories.