US 12,476,000 B2
Machine learning to manage sensor use for patient monitoring
Keegan Duane Dsouza, Ocala, FL (US); Kedar Mangesh Kadam, Halifax (CA); and Adhiraj Ganpat Prajapati, St. Paul, MN (US)
Assigned to MATRIXCARE, INC., Bloomington, MN (US)
Filed by MatrixCare, Inc., Bloomington, MN (US)
Filed on Oct. 1, 2021, as Appl. No. 17/449,751.
Prior Publication US 2023/0107394 A1, Apr. 6, 2023
Int. Cl. G16H 40/63 (2018.01)
CPC G16H 40/63 (2018.01) 20 Claims
OG exemplary drawing
 
1. A method, comprising:
generating a first elopement likelihood indicating a probability of elopement for a patient in a care setting by processing, using a trained machine learning model, first patient data comprising (i) one or more caregiver assessments for the patient and (ii) non-patient-specific extrinsic data;
determining whether the first elopement likelihood exceeds a threshold;
in response to determining that the first elopement likelihood exceeds the threshold:
enabling, by a monitoring system, one or more sensor devices comprising one or more wearable biometric sensors, comprising transmitting, by the monitoring system and to the one or more sensor devices, an instruction causing the one or more sensor devices to begin transmitting sensor data to the monitoring system; and
subsequent to enabling the one or more sensor devices, receiving biometric data for the patient from the enabled one or more sensor devices, wherein the biometric data comprises real-time data;
selecting an intervention for the patient based on the biometric data; and
enacting the intervention.