Abnormal vital signs are associated with an increased risk for critical events in US veteran inpatients

Abstract

Aim: Establish the frequency of abnormal vital signs in medical and surgical ward patients; study their association with "critical events," which for the purposes of this study, were mortality, cardiac arrests and unplanned ICU transfers.

Design and methods: Four-month prospective, observational cohort study; University-affiliated US Veteran's hospital. Vital signs from all regular ward medical and surgical inpatients were recorded over the study period and compared with records of cardiac arrests, mortality and ICU admissions.

Results: Using the Hospital's Medical Emergency Team criteria to define normal/abnormal thresholds for vital signs, abnormal vital signs (VS(MET)) were found in 16% of patients; of these; 35% experienced a critical event vs. 2.5% in the patients with normal vital signs (OR 21, 95% CI 12-35, p<0.001). The sensitivity of VS(MET) to predict a critical event was 0.72 and the positive predictive value was 0.35; sensitivity decreased to 0.28 and positive predictive value increased to 0.78 for patients that had two different VS(MET). Survival was significantly lower in both medical and surgical patients with VS(MET) at both 30 days and at 1 year following discharge (p<0.02). Both medical and surgical patients with VS(MET) had twice the length of stay of patients with normal vitals (3 vs. 7 days; p<0.001).

Conclusions: Even single recordings of VS(MET) signaled increased risk for critical events in hospital ward patients. Use of vital signs as criteria for additional patient assessment and possible ICU admission appears justified. Development of abnormal vitals during hospitalization may signify impaired physiologic reserve that places a patient at higher risk for mortality after discharge.