O2103: The Accuracy of a Novel Auto-launch Program for a Combined Neonatal and Pediatric Critical Care Transport Team in a Hosted Model

Purpose/Objectives: Hosted models in inter-facility transport, in which a critical care transport team partners with a third party emergency medical services (EMS) agency for patient transport, require the development of operative work flow models to minimize the time from call receipt to team launch. As part of a continuous quality improvement program, our team identified EMS-related delays as the primary contributor for failure to meet benchmarks for timely departure for emergent neonatal and pediatric cases. The aim of this study was to evaluate the performance of a novel auto-launch process developed to streamline transport requests and rapidly facilitate team launch for emergent neonatal and pediatric transports identified a priori as part of the development process for the tool. Additionally, we aimed to identify failure process points, including human error and specific call types that contributed to inaccurate triage.

Design/Methods: This is a retrospective review of prospectively gathered quality improvement data. We included all transport runs for which the auto-launch process was initiated and excluded runs with incomplete data. High acuity patients were defined based on a pre-established deployment matrix following standard descriptions of acuity. We provide descriptive statistics and performed a sensitivity analysis to evaluate the ability of the process to accurately identify tier one (emergent) calls and dispatch appropriate resources.

Results: In 2021, the EMS and critical care transport teams transported 2,920 children; 2,620 pediatric (94%), of which 441(17%) were emergent; and 300 (6%) neonatal, of which 72 (24%) were emergent. 2,115 runs were evaluated for eligibility for auto-launch based on the program start-date, of which 283 (13%) met auto-launch criteria for tier 1 emergent transfers. The auto-launch process appropriately triaged 2,013 patients (95%) and inaccurately triaged 102 patients (5%). Of the 102 inaccurately identified runs, 44 (43%) were over-triaged, including calls for ‘major trauma’ (26%) and ‘respiratory distress’ (8%). 58 runs (56%) were under-triaged. Under-triage results from call center agent error (40%), for call types DKA (21%) and drug over-dose (4%). The process performed with a sensitivity of 80.5%, specificity of 98%, PPV of 84%, and NPV of 97%.

Conclusion/Discussion: Our novel and iterative auto-launch process accurately triages pediatric and neonatal patients who require critical care resources during inter-facility transport. Continuous quality improvement processes to identify process pitfalls may enhance program performance.