JMIR Publications

ABSTRACT

Background:

Wireless vital signs sensors are increasingly used for monitoring surgical ward patients. Although early warning scores (EWS) are the current standard for identification of patient deterioration in a ward setting, their usefulness for continuous monitoring is unknown.

Objective:

This study investigated the predictive value of early warning scores obtained from continuous vital sign recordings for early identification of postoperative complications, and explored the performance of a sensor-based EWS alarm system in comparison to manual EWS measurements performed by nurses and threshold alarms applied to individual vital sign recordings (single-parameter alarms).

Methods:

Continuous vital signs measurements (heart rate, respiratory rate, blood oxygen saturation, axillary temperature) collected with wireless sensors in surgical ward patients were used for retrospective simulation of EWS scores (EWSsensor) for different time windows (1-120 min), adopting similar criteria as used for routine EWS measurements performed by nurses (EWSnurse). Hourly EWSsensor measurements were compared between patients with (event group; N=14) and without (control group; N=32) postoperative complications to investigate the predictive value. In addition, alarms were simulated for the EWSsensor using a wide range of alarm thresholds (1-9) and compared with alarms based on EWSnurse and single-parameter alarms simulated in hourly sensor recordings. Alarm performance was evaluated using sensitivity to predict complications within 24 hours, daily alarm rate, and false discovery rate (FDR).

Results:

The hourly EWSsensor of the event group (median: 3.4, IQR: 3.1-4.1) was significantly higher compared to the control group (median: 2.8, IQR: 2.4-3.2). Hourly EWSsensor alarm sensitivity was highest (80-67%) for thresholds of 3 to 5, which was associated with alarm rates of 2-1.2 alarms/patient/day and an FDR of 85-83% respectively. The sensitivity of EWSsensor alarms was higher than EWSnurse alarms (max: 40%) but lower than single-parameter alarms (87%) for all thresholds. Oppositely, (false) alarm rates of EWSsensor alarms were higher compared to EWSnurse alarms (max: 0.6 alarm/patient/day with FDR: 80%) but lower compared to single-parameter alarms (2 alarms/patient/day, FDR: 84%) for most thresholds. EWSsensor alarm rates increased for shorter time windows, reaching 31-70 alarms/patient/day when calculated every minute for thresholds of 3-5.

Conclusions:

EWS scores obtained using wireless vital signs sensors may contribute to early recognition of postoperative complications in a ward setting with higher alarm sensitivity compared to manual EWS measurements. Although hourly sensor-based EWS scores provide fewer alarms compared to single-parameter alarms, high (false) alarm rates can be expected when calculated over shorter time spans. Further studies are recommended to optimize care escalation criteria for continuous monitoring of vital signs in a ward setting and evaluate the effects on patient outcomes.