Some of my comments and lessons from the book.
Doctors, like all of us, are subject to many of the 'fast thinking' pattern recognition (System 1), to use Kahneman's phrase as all of us. According to one study cited by Groopman, some 80 percent of misdiagnoses could be attributed to a cascade of cognitive errors, not lack of medical knowledge.
Groopman walks through a large number of examples from a range of medical fields to illustrate some of the more common cognitive errors:
- Attribution errors, particularly when patients fit negative stereotypes;
- Affective errors, when we follow our wishes, or treat someone we like;
- Availability, based upon the ease which relevant examples come to mind, e.g., 'When you hear hoofbeats, think about horses, not zebras';
- Confirmation bias and anchoring, where we latch on to a single cause;
- Diagnosis momentum, as doctors build upon the earlier diagnosis and don't easily go back to first principles (decision-tree effect);
- Commission bias, or tendency towards action rather than inaction;
- Satisfaction of search, or stop when you have found something that fits, rather than finding everything; and,
- Influence of emotion on cognition and that first impressions are more influential than one might think.
Good practices and questions for doctors, to check their first impressions from 'fast thinking' and move to 'slow thinking':
- Step back from first impressions and be aware of emotional reactions (positive or negative);
-Appropriate use of checklists and/or thoroughness to make sure that one is not missing something (but with the risk that in some cases, checklists will also miss things), to avoid the risk of search satisfaction;
-Be aware of the risks of electronic templates and health records for what might be missed;
Ask:
- 'What might I be missing in this case?'
- 'And what would be the worst thing that could be missed?'
- 'What else could this be?'
- 'Tell me the story again as if I'D never heard it - what you felt, how it happened, when it happened;'
- Recognize that 'there are aspects to human biology and human physiology that you just can't predict;'
- Recognize uncertainty, not disregard it, and the natural tendency to be over-confident and focus more on positive than negative data;
- 'Don't just do something, stand there,' to address commission bias;
- Send the patient to another doctor for a fresh perspective in one doesn't know what the problem is or what to do.
Some good questions patients can use to slow down the thinking of doctors:
Questions:
- What's the worst thing this can be?';
- 'What body parts are near where I am having my symptom?'
- 'What else could it be?'
- 'Is there anything that doesn't fit?'
- 'Is it possible I have more than one problem?'
- 'I am worried about x being something more serious?'
Should doctors recommend a procedure or treatment, ask:
- 'What are the potential complications and their frequency?'
- 'What will be the pain level compared to having a tooth pulled under Novocaine?'
- 'Why, what might be found, with what probability, and how much difference it will make to further treatment decisions?'
- 'Why, if this particular treatment is not working, should we continue with this treatment?'
As well as some approaches and things to keep in mind:
- Acknowledge positive bias (e.g., if one has a close relation with one's doctor) and negative bias (e.g., if one's appearance or behaviour may suggest a certain diagnosis);
- Use humour to suggest why other doctors may not have taken complaints seriously (one patient said 'I may be perceived as kooky but ....');
- If unclear, ask for the explanation in non-technical language;
- Recognize that the perfect is the enemy of the good, miracles without side effects are unlikely, and that many treatments may not completely return one to the pre-disease or operation state; and
- Focus on the long (or medium) term goal with respect to the disease, not the short-term fear of treatment.
The book also has some good insights into some of the perverse incentives, either from drug companies or from fee-for-service that may cloud physician judgement. Overall, a good, interesting and helpful read.
Favourite quote and advice:
"Informed choice means, in part, learning how different doctors think about a particular medical problem and how science, tradition, financial incentives and personal bias mold that thinking. There is no single source for all of this information, so a patient and family should ask the doctor whether a proposed treatment is standard or whether different specialists recommend different approaches, and why. Laypeople also should inquire about how time-tested a new treatment is."