• Charles Weissman, Peter Vernon van Heerden, and Charles L Sprung.
• Department of Anesthesiology, Critical Care Medicine and Pain Management, Israel; Hospital Administration, Hadassah - Hebrew University Medical Center, Hebrew University - Hadassah School of Medicine, Kiryat Hadassah, POB 12000, Jerusalem 91120, Israel. Electronic address: vernon@hadassah.org.il.
• J Crit Care. 2020 Oct 1; 59: 130-135.

PurposeDemonstrate the practical range of information that can be obtained about ICU mortality/survival from limited administrative data.Materials And MethodsProspectively collected administrative data (length-of stay, survival/mortality, referring service) from a university medical center's General ICU was subjected to retrospective analysis to demonstrate ways of presenting and analyzing mortality/survival information.Results16,022 patients (87,624 patient-days) admitted over 23 years were included. 28% of all deaths occurred on ICU day 1. When considering all admissions, mortality on ICU day 1 was 2%, while the overall crude mortality rate revealed that the chances of dying during an ICU stay was 8.6%. Mortality rates in the overall population steadily increased over ICU days 1-5, plateaued during days 6 to 50, decreasing after day 50. The general surgery subgroup had a similar pattern. This contrasted with the internal medicine subgroup where mortality steadily increased over the initial 14 ICU days then plateauing at rates of 40-50%.InterpretationSimple calculations using the few variables found in administrative database enhanced information provided by the crude mortality rate and demonstrated that temporal patterns of mortality change as stay lengthens. These results highlight the limitations of just using overall crude mortality rates.Copyright © 2020 Elsevier Inc. All rights reserved.

16 keywords...

hide…