Assistive technology for operating nursing homes and other health care facilities
A facility for coordinating the safety care of a person is described. The facility receives sensor output indicating that the person has departed a place of repose. In response to this receiving, the facility establishes an alarm identifying the person, accesses a set of on-duty caregivers, and applies one or more precedence rules to establish a precedence among at least a portion of the set of on-duty caregivers. Until a caregiver has accepted the alarm, for each caregiver in the established precedence, the facility: causes a mobile device carried by the caregiver to render a message notifying the caregiver of the alarm and soliciting the caregiver's acceptance of the alarm; and allows the caregiver an interval of time of a first predetermined length in which to accept the alarm before proceeding to the next caregiver in the established precedence.
1. One or more computer memories collectively having contents configured to cause a computing system to perform a method, the method comprising:
receiving input specifying one or more criteria customized for each particular resident of a plurality of identified residents for triggering a first type of safety notification regarding a target behavior for each particular resident;
storing a representation of the one or more criteria specified by the received input;
obtaining output of sensors positioned in the residential facility that correlate with the first type of safety notification;
using at least a portion of the output of sensors that has been produced proximate in time with an occurrence of at least one instance of the first type of safety notification, generating one or more machine learning-based inference rules regarding the target behavior; and
triggering a second type of safety notification regarding the target behavior for each particular resident of the plurality of identified residents based, at least in part, on the one or more machine learning-based inference rules.
2. The computer memories of claim 1 , the method further comprising:
for a resident of the residential facility among the identified residents, receiving output of at least one sensor positioned proximate to the resident in the residential facility; and
based upon the stored representation, applying to the received sensor output the one or more criteria to determine whether to trigger an instance of the first type of safety notification for the resident.
3. The computer memories of claim 2 , the method further comprising:
in response to determining to trigger an instance of the first type of safety notification for the resident:
triggering an instance of the first type of safety notification for the resident; and
communicating the instance of the first type of safety notification triggered for the resident to at least one staff member of the residential facility.
4. A method in a computing system for managing safety notifications to care providers on behalf of a resident of a residential facility, comprising:
accessing at least one user-defined criterion customized for triggering a first type of safety notification regarding a target behavior for the resident;
obtaining output of sensors in the residential facility that correlate with the first type of safety notification;
using at least a portion of the output of sensors that has been produced proximate in time with an occurrence of at least one instance of the first type of safety notification, generating one or more machine learning-based inference rules regarding the target behavior; and
triggering a second type of safety notification regarding occurrence of the target behavior for the resident based, at least in part, on the one or more machine learning-based inference rules.
5. The method of claim 4 wherein the at least one user-defined criterion includes departure of the resident from a resting place.
6. The method of claim 4 , further comprising:
using output of the sensors to train a statistical model that predicts that near-future output of the sensors will satisfy the at least one user-defined criterion based upon preceding output of the sensors,
and wherein the generating the one or more machine learning-based inference rules comprises applying the trained statistical model to contemporaneous output of the sensors.
7. The method of claim 4 , further comprising:
at a time before the at least one user-defined criterion is satisfied, in response to triggering the second type of safety notification, causing a verbal cue to be audibly presented to the resident.
8. One or more computer memories collectively storing a personalized notification criteria data structure, the data structure comprising a plurality of entries, each entry corresponding to a person and comprising:
information identifying the person to whom the entry corresponds; and
information specifying one or more user-defined criteria for triggering a first type of safety notification for the person,
wherein the contents of the entry corresponding to the person are accessible to (a) evaluate the one or more user-defined criteria for triggering one or more instances of the first type of safety notification, and (b) in accordance with output of sensors that has been produced proximate in time with the occurrence of the one or more instances of the first type of safety notification, generate one or more inference rules; as at least a partial basis for triggering a second type of safety notification for the person.
9. The computer memories of claim 8 wherein, for a distinguished one of the entries, the information specifying one or more user-defined criteria specifies a monitoring mode for the person to whom the distinguished entry corresponds that is a fall monitoring monitoring mode, an activity monitoring monitoring mode, or a combined fall and activity monitoring monitoring mode.
10. The computer memories of claim 8 wherein, for a distinguished one of the entries, the information specifying user-defined one or more criteria specifies that an instance of the first type of safety notification is triggered if the entire weight of the person to whom the distinguished entry corresponds departs a resting place.
11. The computer memories of claim 8 wherein, for a distinguished one of the entries, the information specifying one or more user-defined criteria specifies that an instance of the first type of safety notification is triggered if less than the entire weight of the person to whom the distinguished entry corresponds departs a resting place.
12. The computer memories of claim 8 wherein, for a distinguished one of the entries, the information specifying one or more user-defined criteria specifies that an instance of the first type of safety notification is triggered if the weight of the person to whom the distinguished entry corresponds fails to depart a resting place for at least a threshold amount of time.
13. The computer memories of claim 8 wherein, for a distinguished one of the entries, the information specifying one or more user-defined criteria specifies that an instance of the first type of safety notification is triggered if the weight of the person to whom the distinguished entry corresponds fails to arrive at a second resting place within a threshold amount of time of departing a first resting place.
14. The computer memories of claim 8 wherein, for a distinguished one of the entries, the information specifying one or more user-defined criteria specifies that an instance of the first type of safety notification is triggered if the weight of the person departs a room or departs a portion of a room.