Alarm management for monitored physiological variables

1 . A device for monitoring a patient, comprising:

a sensor interface that operates to communicate with a motion sensor positioned on a patient support system and at least one physiological sensor;

a display configured to display data acquired from the at least one physiological sensor;

a speaker;

at least one processing device communicatively connected to the sensor interface, the display, and the speaker; and

a memory device storing instructions which, when executed by the at least one processing device, cause the at least one processing device to:

determine a set of predicted physiological variable values using a model trained using physiological variable measurements collected from the patient;

receive via the sensor interface a set of physiological variable measurements from the at least one physiological sensor and a motion measurement from the motion sensor;

compare the set of physiological variable measurements to the set of predicted physiological variable values to determine a residual vector;

classify the residual vector using a vector motion error based on the motion measurement to indicate a likelihood that a change in a condition of the patient is a true event, wherein the vector motion error represents a predicted influence that the motion measurement has on the physiological variable measurements;

perform an action based on the classification of the residual vector, the action including triggering an alarm on the device by displaying a visual alarm on the display or causing the speaker to output an audible alarm, pausing the alarm, or increasing a delay for triggering the alarm; and

transmit an alarm notification to a remote device when the classification of the residual vector indicates the true event such that the change in the condition of the patient is detected before a conventional alarm is triggered while mitigating false alarms.

2 . The device of claim 1 , wherein the set of physiological variable measurements includes a heart rate, a respiration rate, an end tidal carbon dioxide, and a blood oxygen saturation.

3 . The device of claim 1 , wherein the vector motion error is learned from a motion profile that includes a sequence of activities performed by the patient.

4 . The device of claim 1 , wherein the model is trained without motion artifacts.

5 . The device of claim 1 , wherein the instructions, when executed by the at least one processing device, further cause the at least one processing device to:

monitor a respiration rate of the patient;

determine whether the respiration rate is outside of a threshold range;

when the respiration rate is outside of the threshold range, collect data detected from a sensor in an area where the patient is located;

process the data to determine whether the patient is talking; and

suppress the alarm when it is determined that the patient is talking.

6 . The device of claim 5 , wherein the data includes sounds detected from a microphone inside the area where the patient is located.

7 . The device of claim 5 , wherein the sensor is positioned proximate the patient's larynx, and the data includes vibrations detected from the sensor.

8 . The device of claim 1 , wherein the instructions, when executed by the at least one processing device, further cause the at least one processing device to:

initiate a measurement protocol for measuring the physiological variable measurements;

apply a first delay after a noise artifact is detected;

when the noise artifact does not stop after the first delay, communicate an instruction to the patient to stop an activity relevant to the noise artifact;

measure the physiological variable measurements after a second delay; and

perform the action based on whether the patient stops the activity and the physiological variable measurements are within one or more threshold ranges.

9 . The device of claim 8 , wherein the noise artifact is movement by the patient, and the instruction alerts the patient to stop moving.

10 . The device of claim 8 , wherein the noise artifact is talking by the patient, and the instruction alerts the patient to stop talking.

11 . The device of claim 8 , wherein the action includes triggering the alarm when the patient does not stop the activity or the physiological variable measurements are measured outside of the one or more threshold ranges after the patient stops the activity.

12 . The device of claim 8 , wherein the action includes suppressing the alarm when the patient stops the activity and the physiological variable measurements are within the one or more threshold ranges.

13 . The device of claim 8 , wherein the action includes increasing at least one of the first delay and the second delay.

14 . A non-transitory computer readable storage medium storing instructions, which when executed by at least one processing device, cause the at least one processing device to:

determine a set of predicted physiological variable values using a model trained using physiological variable measurements collected from a patient;

receive via a sensor interface a set of physiological variable measurements from at least one physiological sensor and a motion measurement from a motion sensor;

compare the set of physiological variable measurements to the set of predicted physiological variable values to determine a residual vector;

classify the residual vector using a vector motion error based on the motion measurement to indicate a likelihood that a change in a condition of the patient is a true event, wherein the vector motion error represents a predicted influence that the motion measurement has on the physiological variable measurements;

perform an action based on the classification of the residual vector, the action including triggering an alarm by displaying a visual alarm on a display or causing a speaker to output an audible alarm, pausing the alarm, or increasing a delay for triggering the alarm; and

transmit an alarm notification to a remote device when the classification of the residual vector indicates the true event such that the change in the condition of the patient is detected before a conventional alarm is triggered while mitigating false alarms.

15 . The non-transitory computer readable storage medium of claim 14 , further comprising instructions that cause the at least one processing device to:

monitor a respiration rate of the patient;

determine whether the respiration rate is outside of a threshold range;

when the respiration rate is outside of the threshold range, collect data detected from a sensor in an area where the patient is located;

process the data to determine whether the patient is talking; and

suppress the alarm when the patient is talking.

16 . The non-transitory computer readable storage medium of claim 14 , further comprising instructions that cause the at least one processing device to:

initiate a measurement protocol for measuring the physiological variable measurements;

apply a first delay after a noise artifact is detected;

when the noise artifact does not stop after the first delay, communicate an instruction to the patient to stop an activity relevant to the noise artifact;

measure the physiological variable measurements after a second delay; and

perform the action based on whether the patient stops the activity and the physiological variable measurements are within a threshold range.

17 . A method for monitoring a patient, comprising:

determining a set of predicted physiological variable values using a model trained using physiological variable measurements collected from the patient;

receiving via a sensor interface a set of physiological variable measurements from at least one physiological sensor and a motion measurement from a motion sensor;

comparing the set of physiological variable measurements to the set of predicted physiological variable values to determine a residual vector;

classifying the residual vector using a vector motion error based on the motion measurement to indicate a likelihood that a change in a condition of the patient is a true event, wherein the vector motion error represents a predicted influence that the motion measurement has on the physiological variable measurements;

performing an action based on the classification of the residual vector, the action including triggering an alarm on a monitoring device by displaying a visual alarm on a display or causing a speaker to output an audible alarm, pausing the alarm, or increasing a delay for triggering the alarm; and

transmitting an alarm notification to a remote device when the classification of the residual vector indicates the true event such that the change in the condition of the patient is detected before a conventional alarm is triggered while mitigating false alarms.

18 . The method of claim 17 , further comprising:

monitoring a respiration rate of the patient;

determining whether the respiration rate is outside of a threshold range;

collecting data detected from a sensor in an area where the patient is located when the respiration rate is outside of the threshold range;

processing the data to determine whether the patient is talking; and

suppressing the alarm when the patient is talking.

19 . The method of claim 17 , further comprising:

initiating a measurement protocol for measuring the physiological variable measurements;

applying a first delay after a noise artifact is detected;

communicating an instruction to the patient to stop an activity relevant to the noise artifact when the noise artifact does not stop after the first delay;

measuring the physiological variable measurements after a second delay; and

performing the action based on whether the patient stops the activity and the physiological variable measurements are within a threshold range.