The intervention consists of the implementation of a silent alarm system for the clinical environment. The silent alarm system consists of a self-locating interface device, which is interposed between an alarm-generating device and its audio output. This device contains communication, locating, motion, logic, and relay chips, which enable the interface device to identify its location and staff responsible for that location, as well as to detect audio output from the alarm-generating device, and to control the audible state of that output. The Interface device can communicate with separate bone conduction headsets worn by staff, which contain sensors that confirm staff presence at the headset, and buttons for response to an alarm announcement. The interface device then delivers alarms silently to those staff specifically responsible for its location when a responsible staff member can be identified, confirmed to be present, and accepts responsibility for the alarm through button action

In the control arm of the study, interface devices (as described in the experimental arm) will remain in place for data collection only, but will allow all alarms to be audible, and will not attempt silent alarms.

The intervention consists of the implementation of a silent alarm system in the clinical environment. The silent alarm system consists of a self-locating interface device, which is interposed between an alarm-generating device and its audio output. This device contains communication, locating, motion, logic, and relay chips, which enable the interface device to identify its location and staff responsible for that location, as well as to detect audio output from the alarm-generating device, and to control the audible state of that output. The Interface device can communicate with separate bone conduction headsets worn by staff, which contain sensors that confirm staff presence at the headset, and buttons for response to an alarm announcement. The interface device then delivers alarms silently to those staff specifically responsible for its location when a responsible staff member can be identified, confirmed to be present, and accepts responsibility for the alarm through button action.

The CAM-S Long form is a structured assessment to evaluate the presence and severity of delirium. It consists of a structured clinical assessment that has proven to be reproducible and an accurate measure of delirium presence and severity. The CAM-S has a scoring range of 0-19 with higher scores indicating a greater severity of delirium.

Unit noise levels will be recorded using an integrating Decibel monitor measuring 10 second periods for A-weighted equivalent sound pressure levels (LA-EQ), and also C-weighted peak sound pressure (LC-Peak) for those same 10 second periods. Both LA-EQ and LC-Peak will be averaged, and reported on the Decibel scale, a logarithmic scale of sound pressure levels from 0 Db(least sound pressure) to 194 Db (maximum measurable sound pressure).

Alarm fatigue is an involuntary learned response among staff alarm responders that occurs over time. When presented with a high frequency of alarms which either lack specificity of the condition to which they intend to call attention, or which have a low frequency of required staff intervention ("nuisance alarms"), staff response times become extended, and in some cases, responses do not occur at all. This trial will track staff response time as the time from alarm initiation to staff button response, or alarm audio termination.

Alarm fatigue, assessed as the response time from alarm initiation to staff response (headset button response) or alarm audio termination, measured in seconds on every alarm occurring on every participating device during the study and control periods, each lasting four weeks. Time points for recording responses will be determined by alarm occurrence. Response times for the study and control periods will be averaged and compared statistically.

Nursing Work - Pedometer Measurements. All nurses in each participating unit will wear a pedometer during work shifts. Pedometers will record the number of steps taken during each shift. At the end of each shift, nurses will record the number of steps taken, and the length of the shift (hours). At the end of each four-week period as either study or control, the total number of steps recorded for nursing shifts in that unit will be divided by the total number of shift hours recorded in that unit over the four-week period, to determine the average hourly step count for that unit during the four-week period of either study or control unit.

The total and proportional alarm time that equipment spent in audible and silent states will be recorded during the intervention period.