To explore the ability of the Pediatric Status Epilepticus Severity Score (STEPSS) for assessment of children in terms of mortality and negative outcomes for those admitted to Pediatric Intensive Care Unit (PICU) with status epilepticus. Children admitted to PICU with status epilepticus were included in the study. Descriptive and clinical details, STEPSS, and Pediatric Overall Performance Category Scale (POPC) scores were computed. We examined the performance of STEPSS in predicting adverse outcomes and death. 51 children were involved; most of the children were infants (median = 24 months), and 27.5% had a CNS infection. We found that STEPSS > 2 had poor estimating death rates, with a level of sensitivity of 59%, specificity of 37.9%, a positive predictive ability of 41.9%, a negative predictive ability of 55%, positive likelihood ratio of 0.95, and a negative likelihood ratio of 1.08. The ROC curve (area under the curve: 0.51; 95% confidence interval: 0.36–0.65). STEPSS has a sensitivity of 67.6%, a specificity of 52.9%, a PPV of 74.2%, and an NPV of 45% for poor outcomes (POPC score ≥ 3). The ROC curve (area under the curve: 0.53; 95% confidence interval: 0.38–0.67). The STEPSS showed poor discriminative ability in our PICU-only cohort, with an AUC close to 0.5 reflecting chance-level performance and non-informative likelihood ratios in the prediction of death and unsatisfactory functioning consequences.
Abuelela et al. (Thu,) studied this question.