Read on Twitter
When I was in training, I did a small study with 6 breast cancer specialists. I presented a patient with newly diagnosed breast cancer to each (in private), & asked if they would give adjuvant chemo, adjuvant hormonal therapy, both, or neither. #medtwitter #medstudenttwitter
1/
1/
2 specialists said adjuvant chemo, 2 picked hormonal therapy, 1 recommended both, and 1 neither.
I had also asked them to estimate the 20 year risk of recurrence. Their responses varied widely. This study was before the era of computerized decision aids and such.
I had also asked them to estimate the 20 year risk of recurrence. Their responses varied widely. This study was before the era of computerized decision aids and such.
The truth on what the best therapy is for this patient is not in the middle. One of the 4 choices is the best. But sadly not only will the patient have no idea on who is right, but the treatment she receives will depend on who she was randomly assigned to see.
I did the study because as a trainee I would feel great about learning the absolute perfect treatment strategy from one specialist. And then for the next similar patient I would confidently present the “correct” answer only to be told by the next specialist that I was wrong.
It was very disheartening that I could never seem to get it right. Now I tell this story to my trainees to help them understand that major variations in interpretation of data occurs often among specialists, and not to get frustrated but learn to critically evaluate evidence.
Knowledge gap: Variation in practice recommendations may not be due to differences in interpreting data. Rather it may be simply due to not knowing relevant data that’s available on that question. If estimates of recurrence are wrong, recommendation for treatment will be affected
In any case, this particular situation has been mitigated by computerized decision aids to accurately estimate risk of recurrence taking into account multiple variables, and more optimal choice of intervention.
But many cancers (and diseases) have not reached this level.
But many cancers (and diseases) have not reached this level.
Making the right call for our patients is what we all strive to do. Learning critical appraisal is important. So is acquiring adequate in depth knowledge in your field. Given the complexity of medicine, I think we will need technology to assist us in both these aspects.
Sorry due to typos I had to delete and repost the last 3 tweets in this thread.
