The medical and economic consequences of automation in bacteriology: A case study in a French University Hospital

In a bacteriology laboratory where automated and non-automated procedures co-existed during the study period (1 year), patients were randomly assigned to each type of procedure and we observed the physicians behaviour as well as patients well-being in a surgical service using the results from the laboratory. Contrary to our expectations, the reduction in the time delay necessary to obtain information did not alter either the prescribing behaviour of physicians nor the welfare of patients. Besides, the gain in time was significantly lower than expected. We also discuss in detail the meaning and relevance of the results.