A pilot model for predicting the success of prehospital endotracheal intubation

Abstract

Objectives: We sought to evaluate the success of prehospital, non-drug-assisted endotracheal intubation (ETI) performed by Virginia prehospital care providers and to develop a model designed to predict the probability of success of ETI.

Methods: We conducted a retrospective observational study on prehospital, non-drug-assisted ETI (N=4002) performed by Virginia prehospital care providers, from January 1, 2012, to December 31, 2012. Using descriptive statistics, we quantified patient, provider, and system characteristics. Success rates were calculated by provider certification level and number of ETI attempts. Procedure complications were evaluated for the entire cohort. Variables were recoded for modeling purposes. Univariate analyses using χ2 tests were performed to identify candidate parameters to be included in the model. We performed a backward stepwise logistic regression to predict ETI success.

Results: An overall success rate of 69.9% was found. Binary logistic regression revealed the following covariates associated with ETI success: community type, provider certification level, gender, age group, myocardial infarction, and ethnicity which were all significant (P<0.05) with a -2 log-likelihood value of 3705.574. This was the most parsimonious model evaluated and demonstrated good fit (Hosmer-Lemeshow test P=.646) but poor discrimination (area under the receiver operating characteristic curve=0.595).

Conclusion: This study characterized prehospital ETI success using retrospective state data and found a low overall success rate. Binary logistic regression was performed to create a model and equation identifying a set of factors associated with ETI success.