I have to give a presentation to the residents on “Clinical Decision Making” and this caused me to reflect on  “How Doctors Think” (Jerome Groopman’s book) and, more importantly to me, “How Should Primary Care Doctors Think?”. In an interview on NPR, Dr Groopman follows the story of a patient who lived with severe nausea, cramps, and weight loss ( mis-diagnosed  as anorexia for 15 years when in fact she suffered from a gluten intolerance) and who has seen approximately 30 physicians, none of whom are very helpful. The patient is saved when a physician (self-referred) sits down with her, elicits her entire story, then does the appropriate diagnostic test.

Dr Groopman feels that we in medicine are letting our patients down . He bemoans the fact that our training has become less apprentice like where we learn at the feet of the great clinicians. He suggests that the use of clinical algorithms  has led us to place patients in clinical “boxes” which benefit insurance companies, and pharmaceutical companies, but not patients. He feels (and I agree to an extent) that the solution isn’t following evidence based algorithms but lies in listening to the patients narrative. Where he and I differ is that where he thinks we need to focus more on critical clinical thinking in medical education and less on algorithms, I feel we need to teach folks when to rethink and how to put systems in place which limit the consequences of poor clinical thinking. The patient whose case he uses seems to me to have been let down by a system which encourages sloppy thinking, includes limited quality assurance, and rewards procedural efficiency.

First, the advantage primary care physicians have is that of time. We have done a poor job of teaching physicians how to utilize time as an aspect of disease management. In a separate interview, Dr Goopman identifies “anchoring” (when physicians latch onto a piece of information and do not change despite evidence to the contrary) as a problem which leads to missed diagnoses. If a patient such as this one is mislabeled as having a certain illness, multiple visits should offer the clinician a clue something else might need investigation. For example, abdominal cramps and intense nausea are not the diagnostic criteria for anorexia (see below) and in this patients care should have led to further investigation

Criteria for anorexia

  • Body weight < 85% of expected weight
  • Intense fear of gaining weight
  • Undue emphasis on body shape or weight
  • Amenorrhea (in girls and in women after menarche) for three consecutive months

    Secondly, in Advanced Primary Care involving the use of a high end electronic health record, algorithms can be used not to limit thinking but to confirm diagnostic labels. For example, if a diagnosis of diabetes is added to a patient’s medical record, the diagnositic criteria could be placed in front of the clinician to get confirmation that this is what was really meant and avoid mis-labeling. From a quality assurance standpoint, a diagnosis should be confirmed and the management should then be optimized based on accepted guidelines. Guidelines should be used to guide testing and therapy, not to limit thinking.

    Thirdly, we have lost critical thinking in all of medicine but it has been especially missed among subspecialists. Dr Groopman suggests that primary care docs, with only 12 minutes per visit, merely get a sketchy complaint from the patient and then route them to the appropriate “subspecialist”. In truth, the 12 minutes is a very loose average. The average primary care doc sees approximately 25 patients in an 8 hour day, resulting in approximately 20 useable minutes per patient. An ear infection takes approximately 3 minutes. We can (and do) use this extra 17 minutes to listen to and work with complex patients over the better part of an hour. What we (and patients) would benefit from in the way of subspecialty care are physicians who will listen to the patient’s story again, and work with us to help make a correct diagnosis and determine the appropriate treatment rather than calculating how to extract money from the patient’s insurance via invansive procedures. I was taught at Tulane by George Burch, C. Thorpe Ray, and others who prided themselves on being the good kind of consultant. What I try to encourage my learners to do is to find those types of consultants and latch onto them so that their patients will get complete care.