Please note that this e-mail is not secure (encrypted). If you do not wish to continue communication over unencrypted e-mail, please notify the sender of this message immediately. Continuing to send or respond to e-mail after receiving this message means you understand and accept this risk and wish to continue to communicate over unencrypted e-mail.
Agree that novice diagnosticians will not be inhibited from relying on heuristics to generate a differential. However, I don't agree that the process of diagnosis follows the scientific method so closely as is argued. In the scientific method, a single hypothesis is generated, tested usually with repeat experimentation, and then either rejected or accepted on statistical grounds or reproducible results from identical experimental conditions – even a lone human does not present identical experimental conditions. In the diagnostic process, diagnosticians are not typically pursuing a verifying a single disease one test at a time but rather pursuing multiple avenues until enough testing provides sufficient resemblance to a leading/working diagnosis to proceed with treatment. Consider the a child presenting with sore throat. That is too broad a complaint to start with testing for just strep pharyngitis with too many other competing equally or more concerning conditions must also be explored. Through iterative questioning, hypothesis driven exam one develops a refined narrower set of possibilities where representativeness certainly will drive subsequent testing decisions. Base rate neglect, zebra retreat, and search satisfying may all lead a clinician to falsely pursue one diagnosis over others. "Ah hah! The rapid antigen detection test is + for GAS – this child has strep!" (not if they're under 3 y/o – they're only colonized but the search was satisfied and his RPA was missed). "Ah hah, the radiologist says the fully vaccinated patient who got one racemic and is now happily coloring has epiglottitis – I shall put her in a helicopter and send her to the referal peds hospital" (and risk her falling out of the sky due to base rate neglect). "This teen has trismus, pain with turning his neck, high fevers and SOB – alas it could be Lemierre syndrome but I've never seen it, it's rare compared to mono and a monospot is unreliable so I'll just treat symptomatically." (and the patient returns with MODS from septic emboli).
I think Geoff Norman and Laura Zwaan and colleagues have done good work showing that heuristics are not all bad and probably very good when well calibrated with feedback. But if it were easy to study reliance on heuristics in real-world scenarios with practicing clinicians (rather than vignettes on trainees), then we'd have seen these studies start to come out – maybe we don't yet have the tech to find them in action? Some of Gary Klein's work on naturalistic decision-making seems to have a component of heuristic reliance and NDM often leads to correct conclusions without deliberate conscious hypothesis testing and verification
Nonetheless, absence of evidence is not evidence of absence. We theorized about the existence of gravitational waves but not until about the last 5 years did the technology exist to measure them. For someone to have argued that since we have no direct evidence they exist they must not exist is folly.
Joe Grubenhoff, MD, MSCS (he/him/his) | Associate Professor of Pediatrics
Section of Emergency Medicine | University of Colorado
Medical Director – Diagnostic Safety Program
Children's Hospital Colorado
13123 East 16th Avenue, Box 251 | Anschutz Medical Campus | Aurora, CO 80045 | Phone: (303) 724-2581 | Fax: (720) 777-7317
Connect with Children's Hospital Colorado on Facebook and Twitter
For a child's sake...
We are a caring community called to honor the sacred trust of our patients, families and each other through
humble expertise, generous service and boundless creativity.
...This is the moment.