Abstract 2534: Impact of Rapid Response Systems Implementation on Hospitalized Patient Outcomes: A Systematic Review and Meta-analysis
Background: The Institute for Healthcare Improvement recommends use of rapid response systems (RRS) to prevent cardiopulmonary arrests and mortality among hospitalized patients. However, studies have reported conflicting results, and the data underlying this recommendation should be critically evaluated. We aimed to examine the impact of the introduction of RRS on hospitalized patient mortality and cardiopulmonary arrests.
Methods: We searched MEDLINE (1950 –2009), the Cochrane library, Scopus, EMBASE, and reference lists for randomized controlled trials and observational studies including primarily adults where RRS were implemented in acute care hospitals and in-hospital mortality and cardiopulmonary arrests were compared with a control group without RRS. Of 198 identified abstracts, 19 studies met inclusion criteria. Two reviewers independently extracted information using a pre-specified protocol. Study investigators were contacted for missing data. We used random effects models to estimate relative risks (RR).
Results: A total of 19 studies (2 randomized trials, 17 cohort studies) representing more than 900,000 hospitalized patients were included. Overall, studies were at high risk of bias and yielded heterogeneous results. There was no interaction between study design and each outcome, so results from trials and observational studies were pooled. RRS were associated with a reduction in in-hospital mortality (16 studies reporting, RR 0.89, 95% CI 0.83– 0.96, I2=82%), hospital-wide cardiopulmonary arrests (13 studies, RR 0.71, 95% CI 0.57– 0.89, I2= 88%), and non-intensive care unit cardiopulmonary arrests (11 studies, RR 0.72, 95% CI 0.57– 0.93, I2=71%). A greater reduction in hospital-wide cardiopulmonary arrests was observed among studies where RRS was implemented for at least 6 months prior to recording of arrests compared with those with RRS in place <6 months (RR 0.52, 95% CI 0.39 – 0.70 vs. RR 0.86, 95% CI 0.66 –1.12, p value for interaction=0.012).
Conclusion: Available evidence at high risk of bias suggests RRS may be associated with reductions in mortality and cardiopulmonary arrests among hospitalized adults.