To improve the efficiency of
nosocomial infection detection, a highly structured system combining initial reporting by the bedside
night nurse of symptoms possibly related to
infection with follow-up by the infection control nurse (ICN) was developed: The Infection Control Sentinel Sheet System (ICSSS). Between July 1, 1987 and February 28, 1988, a prospective comparison of results obtained through ICSSS and daily bedside observation/chart review by a full-time trained intensivist was undertaken in the pediatric intensive care unit (PICU). Ratios of
nosocomial infections and nosocomially-infected patients were 15.8 and 7.0 respectively among 685 admissions; included are seven
infections identified only through the ICSSS so that the "gold standard" became an amalgamation of the two systems. The sensitivity for detection of nosocomially-infected patients by bedside observation/chart review and ICSSS was 100% and 87% respectively. The sensitivity for detection of standard
infections (blood,
wound and urine) was 88% and 85% respectively. The sensitivity for detection of
nosocomial infections at all sites was 94% and 72% respectively. Missed
infections were minor (e.g., drain, skin, eye), required physician diagnosis (e.g.,
pneumonia), were not requested on the sentinel sheet (SS) (e.g.,
otitis media), related to follow-up of deceased patients or were minor misclassifications or failures to associate with device (e.g., central-line related). Daily PICU surveillance by the ICN required only 20 minutes a day. The ICSSS appears highly promising and has many unmeasured benefits.