Computerized surveillance in clinical microbiology with time series analysis.

An automatic surveillance system to detect changes in the incidences of microorganisms diagnosed in the department of clinical microbiology has been developed. The program is incorporated into the laboratory computer system and gives a weekly list of microorganisms whose isolation rates compared with those of a previous period exceed a chosen limit. The system uses time series analysis with moving weighted averages, and the detection limit is based on the distribution of the residuals. Output from the system included information about potential outbreaks of gastroenteritis, nosocomial infection with Corynebacterium jeikeium, and a seasonal epidemic of respiratory syncytial virus. The system also listed organisms not commonly isolated in the laboratory and detected incorrect reports. We conclude that continuous surveillance of laboratory data with time series analysis is a valuable tool for epidemiologic surveillance and quality control. Large quantities of data may be screened.