Prospective evaluation of a clinical decision tool to reduce childhood admissions to PICUs for poisoning: ReCAP²

Abstract

Aim: To prospectively validate a pediatric clinical prediction model to identify children at low risk of clinically significant ingestions to prevent unnecessary pediatric intensive care unit (PICU) admissions.

Methods: Calls received by the Georgia Poison Center about children for acute ingestions between May 25, 2017 and May 17, 2018 were scored in real time using the full, age-stratified, and simplified clinical scoring tool to reduce childhood admissions to PICUs for poisoning (RECAP²). Clinically significant ingestions with a poison center recommendation of PICU admission are defined in the simple RECAP² model as ingestion of clonidine, ethanol, an oral anti-hyperglycemic agent, or exposure to carbon monoxide, as well as the presence of symptoms occurring within 2 h for an immediate release, or 4 h for an extended release, medication exposure. Model statistics and percent reduction in PICU admissions were computed.

Results: There were 886 children admitted after ingestions, of which 454 (51.2%) children were admitted to intensive care. At the time of the initial poison center call to report the ingestion, 44 cases (5%) were incomplete using the full, age-stratified model compared to the complete scoring using the simple scoring model. Seventy-two children (8.1%) required monitoring or interventions performed only in a PICU. Real-time application of the full model compared with the simple model would have reduced PICU admissions by 33.3 and 31.7%, respectively.

Conclusions: The simple RECAP² clinical scoring model is a sensitive prediction tool to identify children at very low risk for clinically significant ingestions for whom PICU admission can be avoided. Clinical implementation of the simple RECAP² model and recommendation for admission to an inpatient unit versus PICU should be further evaluated, to reduce unnecessary PICU admissions following acute ingestions.

Keywords:

• Poisoning
• ingestion
• intoxication
• prediction model
• pediatric intensive care unit

Disclosure statement

No potential conflict of interest was reported by the authors.