Using Technical Filtering and Clinical Criteria to Reduce Vital Sign Alarms Using Continuous Monitoring in the Ward

The purpose of this study is to describe the extent of vital sign threshold alarm reduction in relation to several simple multilevel minimum duration criteria and artefact removal. We hypothesize that the amount off false alarms will be decreased using different filters.

Introduction: Early identification of patients who are at risk of severe vital sign deterioration is essential in reducing mortality in critically ill patients. Abnormal physiology often precedes adverse outcomes like unexpected cardiac arrest, unexpected admission to intensive care units (ICU) and unexpected death among hospitalized patients. The system, which is currently being used at Danish hospitals for monitoring patient's vital signs, is the EWS system (Early Warning Score) carried out primarily by nurses. The EWS monitors 7 parameters: Blood pressure (BP), heart rate (HR), blood saturation (SpO2), respiratory rate (RR), temperature, level of alertness and oxygen supplementation. Depending on the measured values, the EWS system determines the intervals with which nurses must measure vital signs. In case of aberrant signs medical staff is informed. While simple continuous monitoring is rarely used outside the PACU or ICU, it presents the possibility of detecting early signs of deteriorations (ESODs), but has shown to induce alarm fatigue among hospital staff, thus increasing the risk of mortality. In addition the system is based on a "better-safe-than-sorry-logic", where a large amount of false alarms are accepted, rather than risking missing one correct alarm. Aim: The aim of this study is to describe the extent of vital sign threshold alarm reduction in relation to several simple multilevel minimum duration criteria and artefact removal. We hypothesize that the amount off false alarms will be decreased using different filters. Methods: The study is a prospective, descriptive cohort study, where patients were included from prospective WARD studies. Data will be analysed using different filters and a duration criteria. First data is analysed without artefact removal, where alarms are collected as in the ward, where alarms are activated if any vitals sign is below or above the threshold for a minimum of 15 seconds. Next data is filtered and analysed with artefact removal and finally data is filtered and analysed with a duration criteria. Furthermore, the available data are without staff interventions and may thus represent cases where the first alert would result in effective treatment and that subsequent alerts were averted. To assess this the data will be analysed for number of alerts if the first alert removed 100%, 75%, 50%, 25% and 0% of the following same alerts within the first hour, the following 3, 6, 12 and 24 hours. Results: The primary analysis will be number of vital sign threshold alarms/patients/day for the alarms with artefact removal but without a duration criteria compared to the alarms with artefact removal and a duration criteria. The two filters will be compared using either a paired Student's t-test or a paired Mann-Whitney U test. The secondary and tertiary analysis of the secondary and tertiary outcomes will be performed using either a paired Student's t-test or a paired Mann-Whitney U test.