Identification of predictors of outcome at the scene of drowning events could guide prevention, care and resource utilization. This review aimed to describe where, what and how scene predictive factors have been evaluated in drowning outcome studies.

We reviewed studies reporting scene drowning predictors published between 2003 and 2019. Data extraction included study populations, data sources, predictor factors (victim, incident, rescue, resuscitation and hospital-related), outcome measures and type of analyses.

Of 49 studies, 87.6% were from high-income countries, 57.1% used data from only one source (92.9% of these from either hospital or EMS), 73.5% included cases who received medical care and 53.1% defined outcomes as survival or death. A total of 78 different factors were studied; the most commonly studied group of factors described victim demographics, included in 42 studies (85.7%), followed by resuscitation factors, included in 30 studies (61.2%). Few studies described rescue (6.1%). The most frequent statistically significant single predictors of outcome known at the scene were submersion duration (evaluated in 19, predictor in 14) and age (evaluated in 31, predictor in 16). Only 38.7% of studies employed multivariable methods.

Gaps to be addressed in drowning outcomes research include data from low- and middle-income countries, standardized definition of factors to allow evaluation across studies, inclusive study populations that can be generalized beyond those receiving medical care, study rescue and resuscitation factors, use of more meaningful outcomes (survival with good neurologic status) and advanced analyses to identify which factors are true predictors versus confounding variables.