Risk stratification to improve Pediatric Early Warning Systems: it is all about the context

Abstract

Early recognition of critically ill patients is of paramount importance to reduce pediatric mortality and morbidity. We created a risk stratification system combining vital parameters and predefined risk factors aimed at reducing the risk of unrecognized clinical deterioration compared with conventional Pediatric Early Warning Systems (PEWS). This single-center retrospective case cohort study included infants (gestational age ≥ 37 weeks) to adolescents (aged <18 years) with unplanned pediatric intensive care unit (PICU) admission between April 01, 2014, and February 28, 2018. The sensitivity in the 24 h prior to endpoint of the Pediatric Risk Evaluation and Stratification System (PRESS) was compared with that of the conventional PEWS and calculated as the proportion of study patients who received a high-risk score. Seventy-four PICU admissions were included. PRESS and PEWS sensitivities at 2 h prior to endpoint were 0.70 (95%CI 0.59 to 0.80) and 0.30 (95%CI 0.20 to 0.42) respectively (p < 0.001). Excluding patients with seizures, PRESS sensitivity increased to 0.75 (95%CI 0.64 to 0.85). Forty-nine patients (66%) scored positive on at least one high-risk factor, and "worried sign" was scored in 31 patients (42%).Conclusion: Risk stratification seems advantageous for a faster detection of clinical deterioration, providing opportunity for earlier intervention. What is Known: • Prompt detection of clinical deterioration is of essential importance to reduce morbidity and mortality. • Conventional Pediatric Early Warning Systems (PEWS) have limited sensitivity and a short window of detection of 1 to 2 h. What is New: • Risk stratification based on context factors allows earlier identification of patients at risk, well before deviation of vital signs. • Risk stratification combined with continuous monitoring of deteriorating trends in vital signs could lead to the development of next-generation warning systems achieving true patient safety.

Keywords: Early warning system (EWS); Health care quality; Patient safety; Risk stratification; Trigger tool.

Fig. 1

Visualization of PRESS in the electronic health record system. A real-time dashboard of the ward shows the PRESS status and last recorded PEWS by a colored smiley (red indicating high-risk PRESS or PEWS score above the threshold, orange indicating medium-risk PRESS or PEWS score elevated but not above the threshold, green indicating standard-risk PRESS or normal PEWS) and is fully integrated in the hospital’s electronic health record system (Epic, Systems Corp, Verona, WI). PRESS is updated manually at least once a day, while the most recent PEWS is automatically uploaded to the dashboard allowing a change in the PEWS score to be detected irrespective of the PRESS score. The dashboard displays how the PRESS risk category was determined based on the three components (1) PEWS, (2) AVPU, and (3) predefined PRESS risk factors, allowing professionals to interpret the different components easily. PEWS, Pediatric Early Warning Score; AVPU, alert-verbal-responsive to pain-unresponsive; PRESS, Pediatric Risk Evaluation and Stratification System; ICU, intensive care unit

Fig. 2

Flow diagram of inclusion. PEWS, Pediatric Early Warning Score; PRESS, Pediatric Risk Evaluation and Stratification System; PICU, pediatric intensive care unit

Fig. 3

Progression of PRESS and PEWS sensitivity over time preceding the endpoint PICU admission. The graph represents the sensitivity of PRESS and PEWS (dark and light gray respectively) at each hourly time point prior to PICU admission. 95% confidence intervals are displayed with error bars. **Difference between PRESS and PEWS at a time point significant at the 0.01 probability level. ***Difference between PRESS and PEWS at a time point significant at the 0.001 probability level