Intelligence-based medicine

Abstract

Despite seven hundred thousand new medical references last year, the relationship between a given set of medical features and specific pathophysiology, treatment, and criteria of improvement is often weak. Moreover, the generalization of evidences obtained in specific settings may lead to under-treat or to over-treat a significant proportion of patients. We expose an application of the cybernetic loop, based on traditional medical steps: nosology, semeiology, pathophysiology, therapy and on the four transitions between these steps. This approach leads to formulate eight basic questions evaluating the steps in terms of reproducibility and the transitions in terms of predictivity. We detail two practical applications: 1) the evaluation of a medical decision (implantation of an internal cardioverter-defibrillator) and 2) the evaluation of a specific study (EPHESUS). Using this loop allows to determine clearly when evidence is lacking and/or to which extend an evidence really increases the medical knowledge or just creates a market.