Issues in imaging, screening, and information processing discussed in a special issue of this Journal (Henry Ford Hosp Med J 1985;33:65-148) have implications for the decision-making algorithms of all clinicians who use those imaging and screening techniques. Epidemiologic and psychological research show that clinicians, like other professionals, do not obey the laws of conditional probability in their judgments of risk or outcomes under uncertainty. Although physicians cannot be expected to make complex probabilistic calculations every time they receive the result of a screening test, teachers of medicine should present algorithms that would allow the physician-in-training to make efficient use of information from all sources, including screening tests. In the long run, more formal training of both clinicians and their teachers in basic epidemiology and biostatistics, especially Bayesian probability, might yield more efficient use of information from the complex screening systems now available.
Lane, John C.
"Imaging, Screening, Artificial Intelligence, and the Diagnostic Dilemma: An Epidemiologist's Response,"
Henry Ford Hospital Medical Journal
: Vol. 34
Available at: https://scholarlycommons.henryford.com/hfhmedjournal/vol34/iss2/20