Algorithms for Prediction of Clinical Deterioration on the General Wards: A Scoping Review

Area of Science:

• Medical Informatics
• Clinical Monitoring
• Predictive Analytics

Background:

• Clinical deterioration on general wards poses a significant risk to hospitalized patients.
• Current methods for detecting deterioration rely on traditional track-and-trace protocols.
• There is a need for advanced tools to improve early detection and patient outcomes.

Purpose of the Study:

• To identify and describe state-of-the-art models using vital sign monitoring for predicting clinical deterioration.
• To identify facilitators, barriers, and effects of implementing these predictive models.

Main Methods:

• A scoping review of studies published until November 2020.
• Searched PubMed, Embase, and CINAHL databases.
• Included studies comparing vital signs-based automated algorithms with track-and-trace protocols for clinical deterioration in adult general ward populations.

Main Results:

• 21 studies were included, primarily retrospective, focusing on hospitalized adults.
• Models showed variable performance (AUC 0.65-0.95) with inconsistent input variables and endpoints.
• 57 facilitators and 48 barriers to implementation were identified, with 57 positive reported effects.

Conclusions:

• Vital sign monitoring algorithms can potentially detect clinical deterioration earlier and more accurately than conventional methods.
• Standardization of input variables, predictive time horizons, and endpoint definitions is crucial for comparative research and implementation.
• Implementation of these models can lead to improved patient outcomes, including reduced mortality.