Estimating the effects of nosocomial infections on the length of hospitalization.

Increased length of hospitalization due to nosocomial infections has generally been calculated from the arithmetic mean of the differences in hospital stay of persons with nosocomial infections and their matched control patients. This method may yield estimates inflated by a few extreme values. Data from a prospective study of nosocomial infections in patients hospitalized in surgical and orthopedic services in a large hospital in Israel are used to illustrate a modified statistical approach to this problem. The geometric mean of the ratio of length of hospitalization of 57 patients with nosocomial infections to matched control patients was determined. The increased mean hospital stay was estimated at 4.5 days for urinary tract infection, 11.9 days for surgical wound infection, and 25 days for "double infection." These figures were lower than those derived from calculation of the arithmetic mean of excess hospital stay and suggest that this modification should be incorporated in future studies of this nature.