Introduction: Admissions from the emergency room (ER) to an intensive care unit (ICU) depend on a number of factors that have not been well elucidated in the literature. The ER physician has to make a judgment about whether an admission is warranted based on a number of factors, some of which we can measure and some which may be physician preferences.

Purpose: We wanted to determine (1) whether ICU beds are being used where the requirement is the greatest based on acuity of care and (2) whether physician preferences influence bed utilization patterns.

Methods: To do this we developed and validated a survey that we prospectively administered to physicians who admit patients to the ICU from the ER from 30 randomly selected hospitals in Ontario, Canada with four or more ICU beds. The scenarios varied illness severity of the patient on admission, expected survival and Health related quality of life (HRQoL) post-intervention, and baseline HRQoL. Within each scenario patient age, bed availability, and presence of directives were varied. The main outcome measure was a rating of factors influencing ER physicians' decision to admit patients to the ICU. Statistical analyses included creation of a general linear model (with “scenario”, “age”, “directives”, and “bed availability” as within-subject factors, “intermediate care unit” as between-subject factor, and “years of practice” as covariate). The primary analysis focused on whether there were any main effects from within-subject factors. We also analyzed the between-subject factors, covariates and their interaction with other within-subject factors.

Results: ER Physicians were more likely to admit (1) younger patients (p<0.0001), (2) when sufficient ICU beds and staff were available (p<0.002), (3) when anticipated HRQoL post-admission was good (p<0.0001), (4) when probability of survival was good (p<0.0001), (5) when there was an immediate need for an ICU intervention (p<0.0001), and (6) when there were no advanced directives (p<0.0001).

Conclusions: Our findings support that factors related to the patient and to the ICU independently influence the decision of physicians to admit a patient to an ICU. In particular, advanced directives appropriately influence the decision, but in their absence, the decision is usually made to admit the patient to the ICU. Based on these findings, it seems reasonable to pursue the development of a framework to aid physician decision-making.