Frequency of early warning score assessment and clinical deterioration in hospitalized patients: A randomized trial.

Abstract

To explore whether early warning score (EWS) measurements at 8h intervals is associated with better outcomes than 12h intervals. We hypothesized that the proportion of patients that deteriorated to a higher EWS at 24h after hospital admission would be lower with 8h interval than with 12h interval. This was a pragmatic, ward-level randomized, non-blinded, controlled trial at an urban University hospital. During two six weeks periods acutely admitted surgical and medical patients, with an initial EWS of 0 or 1, were monitored either every 8th hour or every 12th hour. The primary outcome was clinical deterioration 24h post-admission, estimated by the proportion of patients with an EWS≥2 at 24h after the initial EWS on admission. Of 3185 patients screened for eligibility, 1346 patients were included to the trial. Forty-nine percent were allocated to the 8h group and 51% to the 12h group; of these, 23% and 20% had an elevated EWS≥2 at 24h, respectively (p=0.456), OR 1.17 (0.78-1.76); 3.4% and 2.2%, respectively had an EWS≥5 (p=0.391), and one patient in each group had an EWS≥7 at 24h (p=1.0). Multiple logistic regression analysis showed no significant interactions for the primary outcome and the predefined variables: age, gender, ward type, and inclusion period, with an adjusted OR 1.20 (0.79-1.82). There were no significant differences in regard to the secondary outcomes: cardiac arrests, ICU admissions, review by medical emergency team (MET), length of hospital stay, or elevated EWS at 48 h. Thirty-day mortality was 1.1% vs. 1.8% (p=0.357) in the 8h group and the 12h-group, respectively (OR=0.60 (0.23-1.50), p=0.279). We found no significant reduction in the proportion of clinical deterioration with monitoring frequencies of 3 vs. 2 times daily among patients acutely admitted to a surgical or medical ward and an initial EWS of 0-1.