To measure the prevalence of interventions used to circumvent intensive care access block and to estimate the attributable mortality and additional hospital bed-days associated with them.

Retrospective observational study of 11 adult public hospital intensive care units (ICUs) in Melbourne, Victoria, July 2004 – June 2006.

Prevalence of five interventions in response to access block; attributable fatalities and/or increased length of stay associated with each.

21 896 ICU admissions and 3039 inhospital deaths (13.9%) were screened. All hospitals reported ICU access block. There were 6787 interventions for access block (mean, 9.3/day) — 4070 (18.6% of admissions) instances of after-hours step-down from an ICU to a low-acuity ward; 1115 (5.1%) delays in an emergency department > 8 hours; 895 (4.1%) postponed major surgeries; 487 (2.2%) interhospital transfers; and 220 (1.0%) instances of premature cessation of intensive care. Based on published risk estimates, these interventions may have resulted in 91.1 (95% CI, 34.7–147.2) attributable deaths and 4368 (95% CI, 333–10 050) additional hospital bed-days each year.

Intensive care access block is frequent, and measures to circumvent it increase mortality and length of stay. Further study of the health and financial implications of access block are warranted.