Proposal for new cut-off points for vital signs used to identify at-risk patients over 65 years old in Emergency Departments

Emergency Department (ED) overcrowding and non-age-adapted triage criteria increase the risk of delayed recognition of critical conditions in older adults, who face higher morbidity, atypical presentations, and poorer short-term outcomes. We conducted a retrospective multicentre cohort study including 23,278 ED visits (patients ≥ 65 years) from 52 Spanish hospitals, drawn from the EDEN (Emergency Department and Elder Needs) cohort. Triage-recorded vital signs were analysed regarding mortality or Intensive Care Unit (ICU) admission within 72 h. Candidate optimal cut-offs were estimated using the BackAddFor algorithm within univariate logistic generalized additive models (GAMs). Thresholds were selected to maximise the area under the receiver operating characteristic curve (AUROC). External validation was performed in an independent cohort. Cut-offs were proposed for five vital signs: systolic blood pressure (SBP) at 115 and 185 mmHg, diastolic blood pressure (DBP) at 97 mmHg, respiratory rate (RR) at 15 and 21 bpm, heart rate (HR) at 45 and 105 bpm, and oxygen saturation (SatO2) at 92%. Compared with conventional values, these adapted cut-offs classified as alarm a greater proportion of patients at risk who would not have been classified as high priority according to traditional criteria: 66.1% for RR, 28.2% for SBP, 14.8% for HR, 8.6% for SatO2, and 6.6% for DBP. Findings support the identification of candidate age-adapted thresholds for older ED patients and motivate future predictive models incorporating vital sign interactions to enhance early warning systems in EDs.