Jerome Groopman’s How Doctors Think attempts to shed light on the cognitive processes of doctors as they go about the business of diagnosing and treating patients. Groopman’s thesis is that, contra to what is taught in medical schools, the very best doctors do not follow the algorithmic, Bayesian diagnostic process. Algorithmic analyses is often adequate when diagnosing clear-cut cases. However, doctors are often faced with vague or multiple and confusing symptoms, as well as inexact or inconclusive test results:
In such cases—the kinds of cases where we most need a discerning doctor—algorithms discourage physicians from thinking independently and creatively. Instead of expanding a doctor’s thinking, they can constrain it.
Groopman cites a number of reasons for why doctors fail in their diagnosis, many relating to well known psychological biases that affect human decision making (e.g. confirmation bias, representativeness error, anchoring, etc). Rather than admonish doctors for this, he acknowledges his own failings and provides examples of his failures and how he fell prey to these same biases. Because humans are hard wired to commit these errors, doctors need to be even more aware of the limitations of Bayesian analysis.
So what do the best doctors do? First, they listen to their patients. Most of what doctors do is talk, Groopman says. The kinds of questions they ask of their patients and the way they ask them can have a big impact. For example, asking open-ended questions has the advantage of allowing the doctor to hear additional details about a patient’s symptoms and background as well as avoids structuring the kind of answer a patient may give. Second, good doctors and diagnosticians remember their mistakes and incorporate the lessons from those mistakes into their future thinking. Third, they avoid being biases by other perspectives. Rather than rely on the way another doctor has framed a patient’s illness, good doctors start fresh by interviewing the patient themselves.
There are a number of great lessons in this book for researchers of all stripes. In particular, consultants can learn a great deal by how doctors go about the process of diagnosing based on limited time and limited data. Even the most well-funded consulting engagement has a finite budget and timeline. At some point, a conclusion must be drawn and a corrective action recommended. Additionally, heuristics and schema are helpful in many situations, but we shouldn’t be afraid of uncertainty. Instead, we should value it for it’s ability to force us to think differently about a problem. Groopman writes:
Because of some puzzling, troubling, interesting phenomena, a physician expresses uncertainty, takes the time to reflect, and allows himself to be vulnerable. Then he restructures the problem. This is the key to the art of dealing with situations of uncertainty, instability, uniqueness, and value conflict.
However, I found the book uneven overall. For those that are familiar with cognitive biases and behavioral economics, much of the first few chapters will be repetitive. Each chapter feels like it could have been a separate article in a magazine, and I suspect much of the insight from the book could have been presented in just that format. I was hoping that this book would provide more detail on the diagnostic process. Groopman largely relies on brief summaries and long anecdotes when discussing process. It is certainly worth a read, but likely won’t lead to any significant revelations for most.