"Clinical Judgment" revisited: the distraction of quantitative models.

More than 25 years ago, in a book called Clinical Judgment, each act of patient care was described as having an experimental structure. The "experiments" needed substantial scientific improvement, however, in quality of basic data, taxonomic classification of phenomena, and specifications of clinical reasoning. During the past 2 decades, these improvements have not occurred as extensively as expected because many investigators working in clinical forms of clinical research have not addressed these basic scientific challenges in data, taxonomy, and reasoning. Instead, the investigators have applied quantitative "models," derived from non-clinical domains, that focus on hard data, randomized trials, Bayes theorem, quantitative decision analysis, and psychometric strategies for clinimetric measurement. Consequently, the main challenges of clinical judgment still remain generally available for basic scientific research by investigative clinicians.