To evaluate the usefulness of repeated prevalence surveys to determine trends in the rates of nosocomial infections and to detect changes in risk factors (e.g., use of invasive devices) associated with nosocomial infections. Ten annual prevalence surveys were conducted by trained infection control practitioners between 1985 and 1995 for acute-care patients on the medical, surgical, pediatric, and obstetric-gynecologic services at a 900-bed, tertiary-care, teaching hospital with 750 acute-care beds. The same methods of chart review and concurrent reporting from nursing, the microbiology and clinical laboratory, and the pharmacy were used each year to collect data on the prevalence of nosocomial infections, invasive-device utilization, and abnormal laboratory indicators. Although data were collected on a single day, a period-prevalence study approach was used, because charts were reviewed for any infection data occurring within the 7 days prior to the survey. The hospital census for acute care patients, as measured by the prevalence surveys, declined sharply over the 10 years, from 673 to 575 patients (P = .02). However, the medical service census increased from 150 to 188 patients (P = .01). During the same period, there was a significant decrease in the mean length of stay, from 7.3 to 6.0 days (P = .01), and a concomitant increase in the mean diagnosis related-group case-mix index, from 1.03 to 1.24 (P = .001). Overall, nosocomial infection rates remained unchanged over the study period (mean of 9.85 infections per 100 patients), but rates of nosocomial bloodstream infection increased from 0.0% in 1985 to 2.3% in 1995 (P = .05). Nosocomial infection rates were significantly higher on the medical and surgical services than on other services (P<.001). Utilization rates increased significantly for Foley catheters (9.0% to 16.0%, P = .002) and ventilators (5.0% to 8.0%, P = .05). Despite apparent increases in the severity of illness of our patients, overall rates of nosocomial infection remained stable during a decade of study. Rates of nosocomial bloodstream infection increased, in parallel with National Nosocomial Infection Surveillance System data. We found repeated prevalence surveys to be useful in following trends and rates of infection, device utilization, and abnormal laboratory values among patients at our institution. Such methodologies can be valuable and low-cost components of a comprehensive infection surveillance, prevention, and control program and other potential quality-improvement initiatives, because they enable better annual planning of departmental strategies to meet hospital needs.