How Doctors Think (review)

For fallibilism is the doctrine that our knowledge is never absolute but always swims, as it were, in a continuum of uncertainty and of indeterminacy.

—C. S. Peirce

Doctors play an iconic role in modern society. They feature prominently in popular culture, and over the past decades have assumed various personae: the humane and compassionate (think of the avuncular Marcus Welby); the harried and ground down by systems beyond their control (the long-suffering doctors of ER and St. Elsewhere); and the playful and libidinous (Grey's Anatomy). The doctor-as-thinker is not the image that first strikes one's mind, though the most recent television icon, the misanthropic Gregory House, is lauded precisely for his diagnostic acumen and capacity to unravel the most complex cases, rather than for his compassion or concern for patients as persons.

However, the doctor-as-thinker, or interest in how doctors think, seems to be an active area for exploration. In the past two years, two books with identical titles have been published, one by Kathryn Montgomery (2006), a humanities scholar at Northwestern University, and the other—the subject of this review—by Jerome Groopman, a professor at Harvard Medical School and a staff writer for the New Yorker.

Groopman's How Doctors Think is more accurately a study of when and why doctors sometimes don't think. The book is intended to inform patients of a fact of which they are probably already aware: their doctors are human and will make mistakes, and as a result, patients who partner with their physicians are more likely to get better care. This is not a new idea, but Groopman's spin is novel: doctors are most likely to err when it comes to diagnosis, and patients can lead their physicians to correct conclusions by diverting their doctors away from cognitive traps.

Groopman adopts the language of cognitive psychology when describing the types of errors to which doctors are particularly prone. One such example is the "search satisfaction error," in which a physician who has spotted an abnormality on a patient's X-ray or in a series of blood tests stops looking for other clues that might suggest another diagnosis. The author arrests his own research into physician cognition before he is really done, in order to emphasize that cognitive psychology offers just one way of thinking about thinking.

In addition to failures of the psyche, Groopman notes that doctors' errors can be traced to failures of the system. Groopman describes how managed care has imposed concerns about cash flow on physicians, thereby taxing their ability to think. Seeing more patients in less time means physicians are more likely to cut corners and produce misdiagnoses. Their ability to analyze can also be impaired—consciously or unconsciously—by the influence of industry on the delivery of health care. Groopman fears that the profession's increasing reliance on evidence-based medicine has turned diagnosis into a statistical game.

According to Groopman, being a good diagnostician requires recognition of the importance of language, mood, ego, and bias. How Doctors Think exemplifies Groopman's own ease with words and with his personal history. The writing style is light and free-flowing, and he weaves narrative from his own experiences as an intern, attending physician, and patient together with interviews, analysis, and patient cases to create a highly readable book. However, it occasionally feels too self-referential; this is not autobiography, after all. The last chapter is the most touching. In it, Groopman lauds physicians for understanding that the care of patients—who are as scared, suffering, and imperfect as their doctors—is what medicine is really about.

The epilogue picks up on that theme and reorients the reader to the book's original premise: that patients can partner with their doctors to avoid misdiagnoses. Groopman offers patients three questions that they can ask their health care providers to spur on thinking: (1) what else could it be? (2) is there anything that doesn't fit? and (3) is it possible to have...