Related Experiment Video
Updated: Sep 20, 2026
Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
Applications of the Early Warning Score as an outcome measure: a systematic review of methodologies
Katherine Griffin1,2, Jessica M Ryan3,4, Jan Sorensen5
1School of Postgraduate Studies, Royal College of Surgeons in Ireland, Dublin, Ireland katherinegriffin23@rcsi.ie.
Objectives:
To evaluate how Early Warning Scores (EWS) are used and analysed as research outcome measures in interventional and comparative studies and to assess the consistency and methodological rigour of these approaches.
Methods:
We conducted a systematic review in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (INPLASY 202540054), searching OVID Medline (MEDLINE, Embase, CENTRAL) (1946 to 6 January 2026) for interventional or comparative studies using an EWS as a research outcome measure. Three reviewers independently screened articles, extracted data and performed a narrative synthesis of methodological approaches regarding the use and analysis of EWS as a research outcome.
Results:
EWS were most commonly derived from intermittent ward observations (92/127, 72.0%) rather than continuous monitoring (2/127, 1.6%). Outcomes were typically defined as clinically interpretable summary measures, including index-time values, maximum scores within defined windows or threshold-based categories. Temporal analysis was limited, with most studies not incorporating time explicitly (91/127, 71.7%). Where time was considered (32/127, 25.2%), approaches focused on short-horizon prediction or fixed time point comparisons rather than modelling longitudinal trajectories; explicit assessment of change in EWS was uncommon (14/127, 11.0%). Reporting prioritised discrimination (98/127, 77.2%) over calibration (38/127, 29.9%), with variable consideration of repeated measures, missing data and validation.
Conclusions:
EWS are predominantly used as simple, clinically interpretable severity endpoints rather than as longitudinal measures of physiological change. Inconsistent handling of temporality and limited methodological standardisation restrict comparability and interpretation despite the availability of repeated time-stamped data. Development of reporting and analytic guidance is needed to support more robust use of EWS as research outcomes.
