Validation of the Signs of Inflammation in Children that can Kill (SICK) score for assessment of illness severity

Abstract

Background The Signs of Inflammation in Children that can Kill
(SICK) score has been developed in the search for a practical
triage tool in settings with limited resources for immediate, noninvasive
assessment of illness severity. Its seven parameters are
heart rate, respiratory rate, systolic blood pressure, temperature,
blood oxygen saturation (Sp02), capillary refill time (CRT), and
level of consciousness. The SICK score also takes into account
the age of the child.
Objective To assess the validity of SICK scores for differentiating
between high and low probabilities of death in children.
Methods We performed a prospective evaluation of all children
aged between one month to twelve years admitted to the Pediatric
Emergency Care Unit at Prof. Dr. R.D. Kandou Hospital, Manado
between October 2011 to January 2012. We calculated SICK
scores at the time of presentation and assessed their correlation
with subsequent in-hospital mortality using logis tic regression
analysis.
Results During the study period, we observed 230 patients,
of whom 199 survived and 31 died. There were 134 males, of
whom 117 survived and 17 died. The remaining 96 subjects were
female, of whom 82 survived and 14 died. Logistic regression
analysis revealed a significant relationship between SICK score
and mortality (P< 0.001). With a probability of 0.5, we attained
a cut off score of 4.74 points, with 96.8% sensitivity and 99.5%
specificity for the prediction of mortality.
Conclusion T he high SICK score is associated with higher
probability of death. A cut off score of 4. 74 has high sensitivity and
specificity for predicting the probability of death. The SICK score
may be useful as a triage tool at the patient's initial presentation,
particularly in settings with limited resources.