Usefulness of Severity-of-Illness Scores Based on Admission Data Only in Nosocomial Infection Surveillance Systems

Abstract

Surveillance of nosocomial infection (NI) and the use of reference data for comparison is recommended to improve the quality of patient care. In addition to standardization according to device use, another stratification of reference data according to patients' severity-of-illness scores is often required for benchmarking in intensive care units (ICUs). To determine whether severity-of-illness scores on admission to the ICU are sufficient data for predicting the development of NI. This study was performed in an interdisciplinary ICU at a teaching hospital. Two scores were studied: the Acute Physiology and Chronic Health Evaluation (APACHE) II scoring system and the Therapeutic Intervention Scoring System (TISS). The patient's clinical condition was evaluated on admission and reevaluated daily during the period before the development of NI. In addition, we recorded the number of intubations for every patient-day, the age and sex of the patients, and their history of operations. The Fisher exact test and the stepwise multiple logistic regression model were applied to identify significant predictors of NI. During a 12-month period, 270 patients with ICU stays of more than 24 hours were included in the study. Sixty-nine NIs were identified (incidence, 25.6 cases per 100 patients [95% confidence interval, 19.9-32.3]). A mean APACHE II score and a mean TISS score above the median for these scores, duration of ventilation above the median in the period before the development of NI, and patient age were significantly associated with the development of NI; the score data on admission provided a clearly poorer prediction. The APACHE II and TISS scores on admission are not useful predictors for NI in ICUs.