Abstract Body: Introduction
Previous work suggested a benefit of laryngeal tube (LT) insertion over endotracheal intubation (ETI) in adults with out-of-hospital cardiac arrest (OHCA). But, it remains unclear whether the effectiveness of LT insertion differs across patient phenotypes, given the heterogeneous nature of OHCA.

Research Question
Does the benefit of LT insertion vs. ETI for adults with OHCA differ based on patient phenotypes?

Methods
We conducted a retrospective analysis of adults who underwent successful first attempt LT insertion or ETI in the Pragmatic Airway Resuscitation Trial (PART), a clinical trial comparing the effectiveness of LT insertion vs. ETI for adult OHCAs. The outcome was ROSC at ED arrival. We developed a causal forest-based machine learning (ML) model to predict the individualized treatment effects (ITE) of allocation to LT insertion vs. ETI. A quarter of the data were used to construct the tree structure, another quarter of the data was used to make predictions, and the remaining data were used to test the developing model. We estimated the association between the assigned airway and the outcome across the levels of age, sex, witness status, bystander CPR, initial rhythms, ROSC and emesis before airway attempt, and interval from dispatch to airway attempt. Finally, we developed a decision tree model (i.e., policytree) to guide the optimal airway management based on the predicted ITE.

Results
We identified 1,594 patients who underwent first attempt LT insertion or ETI success. 1,010 (63%) were assigned to LT insertion. The ML model revealed that LT insertion was more beneficial in 456 (29%) patients, whereas ETI was more beneficial in 1,138 (71%). The heterogeneous treatment effects (HTE) of LT insertion vs. ETI were derived largely from the time to first airway attempt, and earlier airway attempts favored LT insertion (Fig. 1). The HTE was consistent regardless of patients’ attributes in Fig. 2. The policytree showed LT insertion was preferable to ETI for patients who underwent the initial airway attempt within 18 minutes from dispatch call and did not receive bystander CPR (absolute increase in the probability of ROSC with LT insertion [95%CI], 7.9% [–3.1 to 19.0%], Fig. 3); otherwise, LT insertion’s benefit was limited (–6.7% [–12.1 to –1.3%]).

Conclusion
We identified the HTE of LT insertion vs. ETI on ROSC at ED arrival. Although an external validation is warranted, individualized airway management may optimize prehospital care for OHCA.