Niven DJ, Bastos JF, Stelfox HT.; Crit Care Med. 2014 Jan;42(1):179-87.

Objective: To determine whether critical care transition programs reduce the risk of ICU readmission or death, when compared with standard care among adults who survived their incident ICU admission.

Data Sources: MEDLINE, EMBASE, CENTRAL, CINAHL, and two clinical trial registries were searched from inception to October 2012.

Study Selection: Studies that examined the effects of critical care transition programs on the risk of ICU readmission or death among patients discharged from ICU were selected for review. A critical care transition program included any rapid response team, medical emergency team, critical care outreach team, or ICU nurse liaison program that provided follow-up for patients discharged from ICU.

Data Extraction: Two reviewers independently extracted data on study characteristics, transition program characteristics, and outcomes (number of ICU readmissions and in-hospital deaths following discharge from ICU).

Data Synthesis: From 3,120 citations, nine before-and-after studies were included. The studies examined medical-surgical populations and described transition programs that were a component of a hospital’s outreach team (n = 6) or nurse liaison program (n = 3). Meta-analysis using a fixed-effect model demonstrated a reduced risk of ICU readmission (risk ratio, 0.87 [95% CI, 0.76–0.99]; p = 0.03; I2 = 0%) but no significant reduction in hospital mortality (risk ratio, 0.84 [95% CI, 0.66–1.05]; p = 0.1; I2 = 16%) associated with a critical care transition program. The risk of ICU readmission was similar whether the transition program was included within an outreach team or a nurse liaison program and did not depend on the presence of an intensivist.

Conclusions: Critical care transition programs appear to reduce the risk of ICU readmission in patients discharged from ICU to a general hospital ward. Given methodological limitations of the included before-and-after studies, additional research should confirm these observations and explore the ideal model for these programs before recommending implementation.

Weblink here