Abstract: A system for interpreting an electrocardiogram identifies a segment of the electrocardiogram having an abnormality using a machine learning algorithm. The system assigns an interpretation output to the segment of the electrocardiogram having the abnormality. The interpretation output describes a morphology of the abnormality.

Abstract: This disclosure is directed towards detecting artifacts in an ECG signal. An ECG system may include multiple sensors which can sense an ECG signal when attached to a patient. Bipolar leads connect the sensors, and provide the ECG signal from the sensors to a computing device. The computing device receives respective signals from the bipolar leads, where the respective signals are indicative of the ECG signal. The computing device identifies, based on the respective signals, a potential artifact corresponding to a subset of the plurality of bipolar leads. The computing device determines that each lead of the subset of the plurality of bipolar leads is connected to a common sensor. The computing device may use signals originating from a remainder of the bipolar leads (e.g., the bipolar leads that are not connected to the sensor(s) where the artifact is detected) to detect a condition of the patient.

Abstract: This disclosure is directed towards a patient management system for selectively generating, storing, and/or sharing patient activity data. A patient management system may receive, at an activity device, patient activity data indicative of activities of a patient. The patient management system may determine a raw score for the patient based on the activity data. The patient management system may combine component scores associated with activities performed by the patient to determine the raw score. Further, the patient management system may determine that the raw score is greater than a score threshold, and generate an activity score based on determining that the raw score is greater than the score threshold. The patient management system may output the activity score to an electronic device, such as a clinician device and/or a device associated with the patient, determine trends associated with the activity score, and/or compare the activity score to a mobility protocol.

Abstract: Systems, methods, and devices relate to monitoring biological conditions (e.g., vital signs) using patient-specific ranges. An example method includes receiving, from a sensor, first measurements of a vital sign of an individual determined automatically during a first time period and determining a baseline of the vital sign based on the first measurements by averaging the first measurements. A threshold of the vital sign can be a predetermined percentage of the baseline. Second measurements of the vital sign can be received from the sensor. In response to determining that at least one of the second measurements is outside of the range, an alarm may be triggered.

Abstract: A sepsis monitoring system determines an initial sepsis risk assessment score, and automatically and continuously updates the sepsis risk assessment score using a sepsis risk assessment model that receives vital signs data, electronic medical records data, and admissions data to continuously update the sepsis risk assessment score. The system determines whether the patient is likely to develop sepsis based on the updated sepsis risk assessment score, and in response to determining that the patient is likely to develop sepsis, generates a notification to drive an early intervention by one or more caregivers.

Abstract: This disclosure is directed towards detecting artifacts in an ECG signal. An ECG system may include multiple sensors which can sense an ECG signal when attached to a patient. Bipolar leads connect the sensors, and provide the ECG signal from the sensors to a computing device. The computing device receives respective signals from the bipolar leads, where the respective signals are indicative of the ECG signal. The computing device identifies, based on the respective signals, a potential artifact corresponding to a subset of the plurality of bipolar leads. The computing device determines that each lead of the subset of the plurality of bipolar leads is connected to a common sensor. The computing device may use signals originating from a remainder of the bipolar leads (e.g., the bipolar leads that are not connected to the sensor(s) where the artifact is detected) to detect a condition of the patient.