Aim: The effects of using handheld devices in combination with filtering and delaying alarms were investigated. Effects on the number of alarms, patient safety, and nurses' experience were evaluated. Methods: Alarm and physiological trend data were collected over two periods of three months for a control (n = 54) and intervention (n = 47) group. During the intervention period, an adapted alarm architecture, filtering and delaying alarms, was implemented, and the number of alarms, critical cardiorespiratory events, and episodes of decreased oxygen saturation and heart rate were compared to the contemporary alarm architecture. Nurses filled out a survey on their experiences. Results: The adapted alarm architecture reduced the number of alarms by 84%. This reduction did not result in significant differences in the number of critical events. Additionally, the duration and depth of the patient's episodes of mildly decreased oxygen saturation and heart rate were unaffected. Nurses reported that they continue to receive too many alarms and occasionally miss alarms. Conclusion: Alarms can be filtered and delayed, reducing the number of alarms and preventing alarm fatigue. Patient safety is not at risk since the number of critical events and the decreases in oxygen saturation and heart rate do not differ significantly between the groups.