PREDICTORS OF PATIENT DETERIORATION OUTSIDE CRITICAL CARE UNITS AND OUTCOMES OF RAPID RESPONSE TEAM IMPLEMENTATION IN A TERTIARY CENTER IN A DEVELOPING THIRD WORLD COUNTRY

Abstract

Rapid Response Teams (RRTs) are in constant development and implemented across different hospitals with varying practices. RRTs have been implemented by many hospitals to provide urgent critical care expertise when activated. It is important to accurately identify patients with increased risk of mortality during their hospital stay, as this could lead to better prioritization for admission to critical care units and timely reexaminations of their care objectives. There is a gap in knowledge regarding the understanding of predictors that contribute to in-hospital mortality, cardiac arrests, and transfers to critical care units among patients who require rapid response team activations.

The primary aim of this quality improvement project is to identify standardized and well-defined predictors for patient deterioration (defined as a composite outcome of subsequent cardiac arrest, Intensive Care Unit [ICU] admission, or unexpected death within 24 hours of the activation of an RRT) as well as mortality before discharge in the index admission in adult Medical Surgical units. The secondary aim is to calculate quality indicators for RRT that will be used as a point reference for evaluating RRT efficiency at a tertiary medical center in Lebanon.

This quality improvement project was a retrospective chart review that included all RRT activations between January 2021 to June 2021. Cases were included if the patient was an adult (18 years and above) admitted to an inpatient Medical Surgical unit and an RRT activation was triggered. A list of possible predictors for deterioration in patients who had an RRT activation were collected and compared statistically with the patients’ outcome after RRT activation, for the first aim. A list of literature generated quality indicators was used to report RRT quality indicators for the second aim. Using statistical tests, a bivariate analysis for subsequent cardiac arrest, ICU admission, or unexpected death were done using contingency tables, independent t test, or analysis of variance test, appropriately. Transfer to critical care and mortality before discharge, the primary outcomes, were used as the outcome variables to create two logistic regression models using significant variables from the bivariate analyses.

During January 2021 to June 2021, RRT was activated 204 times; 5 of these events were changed to code activation or cancelled RRT activations. A review of 199 cases revealed 56 (37.6%) patients were transferred to critical care units and 47 (31.5%) patients had mortality before discharge. Of the 199 activations, 3 (2%), patients had subsequent cardiac arrests and 2 (1.3%) patients had mortality within 24 hours of RRT activation. After adjusting for covariates, as lactic acid increased the likelihood of critical care transfers increased (adjusted odds ratio 1.824; 95% confidence interval 1.195 – 2.786; p = .005), as Charlson comorbidity index increased the likelihood of mortality before discharge increased (adjusted interval odds ratio .838; 95% confidence interval .720 - .976; p = .023). The dummy variable cardiac in the reason for RRT activation was significant in comparison to sepsis (reference group) (adjusted odds ratio 3.443; 95% confidence interval 1.403 – 8.448; p = .007). Those with a cardiac reason for RRT activation were 3.443 times less likely to have mortality before discharge than those with a sepsis reason for RRT activation. In-hospital patients with RRT activations had readmissions at 30-days (44 [29.5%]), 60-days (29 [19.5%]), and 90-days (18 [12.1%]).

RRT activation among non-ICU patients identifies a population at high risk of transfers to critical care units and mortality at discharge. We identified several predictors for patient deterioration and outcomes of RRTs, which provide opportunities for future quality improvement and patient safety opportunities.