Early prediction of patient deterioration using Sequential Organ Failure Assessment (SOFA) by the Rapid Response Team (RRT)

Project Background

Regardless of the immediate availability of qualified bedside clinicians, outcomes in patients hospitalized for major medical events and deteriorating conditions remain poor, resulting in unexpected transfers to higher levels of care and increased lengths of stay. Early signs of deterioration are often not monitored efficiently on general care units, where the clinician-to-patient ratio is higher than in the intensive care unit (ICU). Patient deterioration on general care units compared to critical care units has been found to be associated with a higher risk of death. To address the problem of failing to recognize patient decline in general care wards, Rapid Response Teams (RRTs) were established to assist in the management of unstable patients in general care units and to minimize further deterioration. A significant challenge facing RRT activation is triggering an RRT in the first place. Signs of decline can be subtle and easily overlooked in busy units. This gap in intervention and treatment is an area for quality and performance improvement.

The growing use of electronic health records (EHR) and digitization of data throughout the patient care process offers opportunities for the development and implementation of tools and methods to assist bedside nurses and ancillary care teams in improving patient outcomes. In this context, Sequential Organ Failure Assessment (SOFA) has been shown to assist clinical care teams, specifically RRTs, in monitoring and proactively addressing patient deterioration; however, its application outside of the ICU remains rare.

Project Purpose

The purpose of this project was to provide the RRT staff with an automatically calculated SOFA score as a potential means of improving RRT activation, RRT workflow efficiency, and RRT event outcomes.  The RRT outcomes include: (1) reducing the number of transfers to a higher levels of care, (2) reducing the number of transfers to the OR, (3) reducing the number of transfers to the ICU, and (4) reducing the number of deaths by integrating the SOFA system into Michigan Medicine's (MM) EHR to optimize patient care.  

Project Methods

This project began with identification and validation of the extent of the problem of inpatient deterioration in general, and narrowed to assessment of the issue on general care wards. The project involved a review of the peer-reviewed literature regarding the issue and a synthesis of existing evidence-based approaches to addressing the challenge of patient deterioration. This was followed by an investigation into all RRT events for a two-year time window prior to the start of this project. This analysis allowed us to look for any trends over time and allowed us to further validate the occurrence and significance of RRT events and transfers across MM. 

The intervention itself included the development and testing of the SOFA tool for use on the specifically identified general care wards at MM, a detailed plan for preparation and roll-out including onsite support, training, and education of RRT personnel, and the conduct of a user survey designed to measure opinions and perceptions of those who participated in the pilot evaluation.  During the winter term of 2018, the SOFA pilot was tested and then implemented on the identified general care wards with the goal of understanding the impact that the use of the tool might have on RRT workflow efficiency, RRT activation, and the RRT event outcomes.

Project Outcomes

Results indicate a statistically significant change in the total number of the RRT activations (p = .00); however, no significant change was found in RRT transfers following implementation of the SOFA scoring system with regard to patients not transferred at all (p = 0.43), those transferred to the ICU (p = 0.87), those transferred to telemetry/stepdown units (p = 0.35), those transferred to the OR (p = 0.31), or those who died (p = 0.32). Responses to an RRT staff survey (n=13) conducted after SOFA implementation to measure impact on workflow indicated that most respondents thought the tool was appropriate for use by bedside nurses on general care units to monitor potential patient deterioration, but that enhancement of the tool would be necessary for widespread adoption.

Project Implications

This project provided an evidence-based tool to identify patient deterioration in adult general care units with a corresponding decline in the number of RRT events. Decision-making aided by objective measures may assist RRT and general care nurses in identifying patients at risk of deterioration to encourage the implementation of interventions that will decrease patients‚unexpected transfers to the ICU and higher levels of care, OR admissions, and mortality.

While this project and pilot implementation of the SOFA tool was designed to support early intervention and facilitate decision-making for RRT members, it was not intended to replace human judgment. Therefore, this author emphasized that the SOFA score was not to be used in isolation or to overrule a patient‚ apparent need for intervention. Moreover, the SOFA score was not intended to limit an RRT to focusing solely on rounding on patients identified by a high SOFA score. Instead, in addition to enhancing care for patients on the general care unit and reacting to early signs of decline, SOFA was intended to help the RRT to prioritize and plan the rounding process. These elements were reinforced in staff training sessions.


Fadi Islim

Graduation Date

April, 2018


Sequential Organ Failure Assessments