ABSTRACT
OBJECTIVES: Improve antibiotic therapies in a geriatrics hospital by enhancing guidelines impact. Historical background. - Study of the uses of the data-processing tool: detection of adverse drug events and inappropriate therapies; diffusion of the probable bacteria and their sensitivities to antibiotics following the infected site; creating decisional algorithms leading to suggested antibiotic therapy concept. METHOD: Referring to the conferences of consensus: drafting of local protocols of antibiotic therapies by using data processing if possible. Description of computer-systems implemented in Salt Lake City (USA) and Tubingen (Germany). With the assistance of an expert infectiologist, design of a database containing clinical details of the different situations in the pulmonary and urinary infections: once the clinical data are entered, a suggested antibiotic therapy appears on the screen, followed by a suggested prescription adapted to patient's specifications. During five years, survey of Defined Daily Dose per 1000 occupied bed-days; survey of sensitivities to antibiotics of urinary tract bacteria. RESULTS: Decrease of overall antibiotic consumption; decrease of broad-spectrum agents; modifications of E. coli sensitivities lead to variations of suggested antibiotic therapy. DISCUSSION: Seizure of the clinical determinants imposes a reducing effect; but it is necessary to apply to antibiotic therapies a hierarchicalbasis on account of bacterial resistances process and on account of "tonnage" of antibiotic prescriptions. CONCLUSION: Computer-assisted decision in antibiotic therapy can draw prescribers' attention to guidelines. The adhesion of the physicians to the computerized decision-making system is consolidated by their direct participation in the design of this system.
