Medical Premonitory Event Estimation

Abstract

A system and method for medical premonitory event estimation includes one or more processors to perform operations comprising: acquiring a first set of physiological information of a subject, and a second set of physiological information of the subject received during a second period of time; calculating first and second risk scores associated with estimating a risk of a potential cardiac arrhythmia event for the subject based on applying the first and second sets of physiological information to one or more machine learning classifier models, providing at least the first and second risk scores associated with the potential cardiac arrhythmia event as a time changing series of risk scores, and classifying the first and second risk scores associated with estimating the risk of the potential cardiac arrhythmia event for the subject based on the one or more thresholds.

What is claimed is: 
     
         1 . A medical premonitory event estimation system, comprising:
 a non-transitory computer-readable storage medium in communication with one or more processors and having instructions stored thereon which, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
 for a plurality of time periods, calculating an event estimation of risk score associated with a potential medical event for a subject occurring within the associated time period based at least partly on physiological parameter data of the subject. 
   
     
     
         2 . The system of  claim 1 , wherein the physiological parameter data comprises ECG data. 
     
     
         3 . The system of  claim 1 , wherein the physiological parameter data comprises at least one of blood pressure data, heart rate data, thoracic impedance data, pulse oxygen level data, respiration rate data, heart sound data, lung sound data, and activity level data. 
     
     
         4 . The system of  claim 1 , wherein the potential medical event comprises a cardiac event. 
     
     
         5 . The system of  claim 4 , wherein the cardiac event comprises at least one of an ectopic beat, a run of ectopic beats, a ventricular tachycardia, a bradycardia, asystole, and a T-wave abnormality. 
     
     
         6 . The system of  claim 1 , wherein the potential medical event comprises at least one of a plurality of medical events, an increase in a rate of medical events, and/or an increase in an intensity of medical events. 
     
     
         7 . The system of  claim 1 , wherein the potential medical event is defined in a multidimensional parameter space comprising the physiological parameter data and at least one other type of physiological parameter data and/or demographic data of the subject. 
     
     
         8 . The system of  claim 1 , wherein the one or more processors perform operations comprising:
 calculating a plurality of different event estimation of risk scores associated with the potential medical event for the subject within the associated time period based at least partly on the physiological parameter data.   
     
     
         9 . The system of  claim 1 , wherein the one or more processors perform operations comprising:
 calculating a plurality of different event estimation of risk scores associated with a plurality of different potential medical events for the subject within the associated time period based at least partly on the physiological parameter data.   
     
     
         10 . The system of  claim 1 , wherein the calculating the event estimation of risk score comprises calculating a criticality score indicating a significance of the potential medical event with respect to at least one other potential medical event. 
     
     
         11 . The system of  claim 1 , wherein the calculating the event estimation of risk score comprises calculating a confidence score including a probability that the potential medical event occurs within the associated time period. 
     
     
         12 . The system of  claim 1 , wherein the one or more processors perform operations comprising:
 determining that the event estimation of risk score satisfies one or more event estimation of risk thresholds for the associated time period; and   determining a response to the potential medical event based at least partly on the one or more event estimation of risk thresholds determined to be satisfied.   
     
     
         13 . The system of  claim 12 , wherein the determined response to the potential medical event includes at least one of informing the subject of advanced diagnostics, advising the subject against removal of equipment, advising the subject of a behavior modification, alerting a medical professional, and preparing a device for treatment. 
     
     
         14 . The system of  claim 12 , wherein each of the one or more event estimation of risk thresholds include at least one confidence threshold including a required probability that the potential medical event occurs within the associated time period and at least one criticality threshold including a required significance of the potential medical event with respect to at least one other potential medical event. 
     
     
         15 . The system of  claim 12 , wherein the one or more event estimation of risk thresholds comprise a plurality of different event estimation of risk thresholds for the associated time period. 
     
     
         16 . The system of  claim 12 , wherein the one or more event estimation of risk thresholds for a first time period are different than the one or more event estimation of risk thresholds for a second time period. 
     
     
         17 . The system of  claim 12 , wherein the determined response to the potential medical event for the subject occurring within a first associated time period is different that the determined response to the potential medical event for the subject occurring within a second associated time period. 
     
     
         18 . The system of  claim 1 , wherein the one or more processors perform operations comprising:
 determining that the event estimation of risk score fails to satisfies at least one event estimation of risk threshold for the associated time period;   receiving additional data of the subject; and   calculating an enhanced event estimation of risk score associated with the potential medical event for the subject occurring within the associated time period based at least partly on the physiological parameter data and the additional data.   
     
     
         19 . The system of  claim 18 , wherein the additional data comprises at least one of image data of the subject, audio data including the voice of the subject, and data based on a galvanic skin response of the subject. 
     
     
         20 . The system of  claim 12 , wherein the one or more processors perform operations comprising:
 setting the one or more event estimation of risk thresholds based at least partly on historical patient data collected from a plurality of patients.   
     
     
         21 . The system of  claim 12 , wherein the one or more processors perform operations comprising:
 setting the one or more event estimation of risk thresholds based at least partly on input from a user.   
     
     
         22 . The system of  claim 1 , wherein the one or more processors perform operations comprising:
 calculating the event estimation of risk score at periodic time intervals.   
     
     
         23 . The system of  claim 1 , wherein the one or more processors perform operations comprising:
 calculating the event estimation of risk score is at dynamic time intervals, wherein a duration of the dynamic time intervals is based at least partly on the event estimation of risk score.   
     
     
         24 . The system of  claim 1 , wherein the one or more processors perform operations comprising:
 continuously calculating the event estimation of risk score is continuously calculated.   
     
     
         25 . The system of  claim 1 , wherein the calculating the event estimation of risk score comprises applying a logistic regression model to the physiological parameter data to determine the event estimation of risk score. 
     
     
         26 . The system of  claim 1 , wherein the calculating the event estimation of risk score comprises:
 generating at least two generally orthogonal vectors based at least partly on the physiological parameter data;   processing the at least two generally orthogonal vectors to determine a loop trajectory of the physiological parameter data; and   identifying a trajectory bifurcation by:
 characterizing a group of control loop trajectories that includes one or more loop trajectories obtained during a first time period; 
 characterizing a group of test loop trajectories that includes one or more loop trajectories obtained during a second time period that is subsequent to the first time period; 
 comparing the characterization of the group of control loop trajectories to the characterization of the group of test loop trajectories; 
 measuring a degree of trajectory bifurcation between the group of control loop trajectories and the group of test loop trajectories; and 
 calculating the event estimation of risk score based at least in part on the measure of the degree of trajectory bifurcation. 
   
     
     
         27 . The system of  claim 1 , wherein the calculating the event estimation of risk score comprises:
 calculating a first event estimation of risk score including a first criticality score for a first potential medical event; and   calculating a second event estimation of risk score including a second criticality score for a second potential medical event based at least partly on the first event estimation of risk score, wherein the first criticality score indicates that a significance of the first potential medical event is different than a significance of the second potential medical event.   
     
     
         28 . The system of  claim 1 , wherein the calculating the event estimation of risk score comprises:
 calculating a first event estimation of risk score associated with the potential medical event for the subject occurring within of the associated time period based on a first shockable rhythm detection algorithm; and   calculating a second event estimation of risk score associated with the potential medical event for the subject occurring within of the associated time period based on a second rhythm detection algorithm, wherein the second rhythm detection algorithm is tuned for a higher sensitivity on the physiological data than the first rhythm detection algorithm.   
     
     
         29 . The system of  claim 1 , wherein the calculating the event estimation of risk score comprises:
 applying at least two different rhythm detection algorithms to different time segments of the physiological data.   
     
     
         30 . The system of  claim 1 , wherein the calculating the event estimation of risk score comprises:
 receiving data indicating a viability of a patient; and   determining a response to the potential medical event based at least partly on the viability of the patient.   
     
     
         31 . The system of  claim 1 , comprising:
 a medical device comprising one or more sensors configured to sense the physiological parameter data of the subject.   
     
     
         32 . The system of  claim 31 , wherein the medical devices comprises a wearable medical device, wherein the one or more sensors comprise a plurality of ECG sensors, wherein the physiological parameter data of the subject comprises ECG data, and wherein the potential medical event comprises a cardiac event. 
     
     
         33 . The system of  claim 31 , comprising:
 a communications network configured to communicate at least one of the physiological parameter data and the event estimation of risk score from the medical device to an another computing device.   
     
     
         34 . The system of  claim 1 , comprising:
 a display for displaying a time-based visual indicator of the event estimation of risk score for the plurality of time periods.   
     
     
         35 . The system of  claim 1 , wherein the one or more processors perform operations comprising:
 determining a response to the potential medical event based at least partly on the event estimation of risk score.   
     
     
         36 . The system of  claim 35 , wherein the determined response to the potential medical event includes providing an instruction to the subject to contact a medical professional. 
     
     
         37 . The system of  claim 35 , comprising:
 a wearable medical device, wherein the determined response to the potential medical event includes providing an instruction to the subject to check a battery of the wearable medical device.   
     
     
         38 . The system of  claim 35 , comprising:
 a wearable medical device, wherein the determined response to the potential medical event includes charging a shocking mechanism of the wearable medical device.   
     
     
         39 . The system of  claim 35 , wherein the one or more processors perform operations comprising:
 determining the response to the potential medical event based at least partly on a sensitivity and a specificity of the event estimation of risk score.   
     
     
         40 . The system of  claim 39 , wherein the determined response based on a first sensitivity and a first specificity is different than the determined response based on a second different sensitivity and second different specificity. 
     
     
         41 . The system of  claim 40 , wherein the determined response to the potential medical event includes at least one of informing the subject of advanced diagnostics, advising the subject against removal of equipment, advising the subject of a behavior modification, alerting a medical professional, and preparing a device for treatment. 
     
     
         42 . The system of  claim 35 , wherein the one or more processors perform operations comprising:
 modifying a sensitivity of an algorithm for determining the event estimation of risk score based on a risk level of the subject.   
     
     
         43 . The system of  claim 1 , wherein the plurality of time period comprise at least one time period of less than about ten minutes, at least one time period of less than about one hour, at least one time period of less than about three hours, at least one time period of less than about one day, at least one time period of less than about one week, and at least one time period of less than about one month. 
     
     
         44 . The system of  claim 1 , wherein the calculating the event estimation of risk score comprises calculating a confidence band of the event estimation of risk score. 
     
     
         45 . The system of  claim 1 , wherein the calculating the event estimation of risk score comprises calculating an error band of the event estimation of risk score. 
     
     
         46 . The system of  claim 1 , wherein the plurality of time periods comprise a plurality of time periods of less than four hours. 
     
     
         47 . A method for medical premonitory event estimation, comprising:
 receiving, by one or more processors, physiological parameter data of a subject; and   for a plurality of time periods, calculating, by the one or more processors, an event estimation of risk score associated with a potential medical event for the subject occurring within the associated time period based at least partly on the physiological parameter data of the subject.   
     
     
         48 . The method of  claim 47 , wherein the physiological parameter data comprises ECG data. 
     
     
         49 . The method of  claim 47 , wherein the physiological parameter data comprises at least one of blood pressure data, heart rate data, thoracic impedance data, pulse oxygen level data, respiration rate data, heart sound data, lung sound data, and activity level data. 
     
     
         50 . The method of  claim 47 , wherein the potential medical event comprises a cardiac event. 
     
     
         51 . The method of  claim 50 , wherein the cardiac event comprises at least one of an ectopic beat, a run of ectopic beats, a ventricular tachycardia, a bradycardia, asystole, and a T-wave abnormality. 
     
     
         52 . The method of  claim 47 , wherein the potential medical event comprises at least one of a plurality of medical events, an increase in a rate of medical events, and/or an increase in an intensity of medical events. 
     
     
         53 . The method of  claim 47 , wherein the potential medical event is defined in a multidimensional parameter space comprising the physiological parameter data and at least one other type of physiological parameter data and/or demographic data of the subject. 
     
     
         54 . The method of  claim 47 , comprising:
 calculating, by the one or more processors, a plurality of different event estimation of risk scores associated with the potential medical event for the subject within the associated time period based at least partly on the physiological parameter data.   
     
     
         55 . The method of  claim 47 , comprising:
 calculating, by the one or more processors, a plurality of different event estimation of risk scores associated with a plurality of different potential medical events for the subject within the associated time period based at least partly on the physiological parameter data.   
     
     
         56 . The method of  claim 47 , wherein the calculating the event estimation of risk score comprises calculating a criticality score indicating a significance of the potential medical event with respect to at least one other potential medical event. 
     
     
         57 . The method of  claim 47 , wherein the calculating the event estimation of risk score comprises calculating a confidence score including a probability that the potential medical event occurs within the associated time period. 
     
     
         58 . The method of  claim 47 , comprising:
 determining, by the one or more processors, that the event estimation of risk score satisfies one or more event estimation of risk thresholds for the associated time period; and   determining, by the one or more processors, a response to the potential medical event based at least partly on the one or more event estimation of risk thresholds determined to be satisfied.   
     
     
         59 . The method of  claim 58 , wherein the determined response to the potential medical event includes at least one of informing the subject of advanced diagnostics, advising the subject against removal of equipment, advising the subject of a behavior modification, alerting a medical professional, and preparing a device for treatment. 
     
     
         60 . The method of  claim 58 , wherein each of the one or more event estimation of risk thresholds include at least one confidence threshold including a required probability that the potential medical event occurs within the associated time period and at least one criticality threshold including a required significance of the potential medical event with respect to at least one other potential medical event. 
     
     
         61 . The method of  claim 58 , wherein the one or more event estimation of risk thresholds comprise a plurality of different event estimation of risk thresholds for the associated time period. 
     
     
         62 . The method of  claim 58 , wherein the one or more event estimation of risk thresholds for a first time period are different than the one or more event estimation of risk thresholds for a second time period. 
     
     
         63 . The method of  claim 58 , wherein the determined response to the potential medical event for the subject occurring within a first associated time period is different that the determined response to the potential medical event for the subject occurring within a second associated time period. 
     
     
         64 . The method of  claim 47 , comprising:
 determining, by the one or more processors, that the event estimation of risk score fails to satisfies at least one event estimation of risk threshold for the associated time period;   receiving, by the one or more processors, additional data of the subject; and   calculating, by the one or more processors, an enhanced event estimation of risk score associated with the potential medical event for the subject occurring within the associated time period based at least partly on the physiological parameter data and the additional data.   
     
     
         65 . The method of  claim 64 , wherein the additional data comprises at least one of image data of the subject, audio data including the voice of the subject, and data based on a galvanic skin response of the subject. 
     
     
         66 . The method of  claim 58 , comprising:
 setting, by the one or more processors, the one or more event estimation of risk thresholds based at least partly on historical patient data collected from a plurality of patients.   
     
     
         67 . The method of  claim 58 , comprising:
 setting, by the one or more processors, the one or more event estimation of risk thresholds based at least partly on input from a user.   
     
     
         68 . The method of  claim 47 , comprising:
 calculating, by the one or more processors, the event estimation of risk score at periodic time intervals.   
     
     
         69 . The method of  claim 47 , comprising:
 calculating, by the one or more processors, the event estimation of risk score is at dynamic time intervals, wherein a duration of the dynamic time intervals is based at least partly on the event estimation of risk score.   
     
     
         70 . The method of  claim 47 , comprising:
 continuously calculating, by the one or more processors, the event estimation of risk score is continuously calculated.   
     
     
         71 . The method of  claim 47 , wherein the calculating the event estimation of risk score comprises applying a logistic regression model to the physiological parameter data to determine the event estimation of risk score. 
     
     
         72 . The method of  claim 47 , wherein the calculating the event estimation of risk score comprises:
 generating at least two generally orthogonal vectors based at least partly on the physiological parameter data;   processing the at least two generally orthogonal vectors to determine a loop trajectory of the physiological parameter data; and   identifying a trajectory bifurcation by:
 characterizing a group of control loop trajectories that includes one or more loop trajectories obtained during a first time period; 
 characterizing a group of test loop trajectories that includes one or more loop trajectories obtained during a second time period that is subsequent to the first time period; 
 comparing the characterization of the group of control loop trajectories to the characterization of the group of test loop trajectories; 
 measuring a degree of trajectory bifurcation between the group of control loop trajectories and the group of test loop trajectories; and 
 calculating the event estimation of risk score based at least in part on the measure of the degree of trajectory bifurcation. 
   
     
     
         73 . The method of  claim 47 , wherein the calculating the event estimation of risk score comprises:
 calculating a first event estimation of risk score including a first criticality score for a first potential medical event; and   calculating a second event estimation of risk score including a second criticality score for a second potential medical event based at least partly on the first event estimation of risk score, wherein the first criticality score indicates that a significance of the first potential medical event is different than a significance of the second potential medical event.   
     
     
         74 . The method of  claim 47 , wherein the calculating the event estimation of risk score comprises:
 calculating a first event estimation of risk score associated with the potential medical event for the subject occurring within of the associated time period based on a first shockable rhythm detection algorithm; and   calculating a second event estimation of risk score associated with the potential medical event for the subject occurring within of the associated time period based on a second rhythm detection algorithm, wherein the second rhythm detection algorithm is tuned for a higher sensitivity on the physiological data than the first rhythm detection algorithm.   
     
     
         75 . The method of  claim 47 , wherein the calculating the event estimation of risk score comprises:
 applying at least two different rhythm detection algorithms to different time segments of the physiological data.   
     
     
         76 . The method of  claim 47 , wherein the calculating the event estimation of risk score comprises:
 receiving data indicating a viability of a patient; and   determining a response to the potential medical event based at least partly on the viability of the patient.   
     
     
         77 . The method of  claim 47 , comprising:
 sensing, by a medical device comprising one or more sensors, the physiological parameter data of the subject.   
     
     
         78 . The method of  claim 47 , wherein the medical devices comprises a wearable medical device, wherein the one or more sensors comprise a plurality of ECG sensors, wherein the physiological parameter data of the subject comprises ECG data, and wherein the potential medical event comprises a cardiac event. 
     
     
         79 . The method of  claim 47 , comprising:
 communicating, by a communications network, at least one of the physiological parameter data and the event estimation of risk score from the medical device to an another computing device.   
     
     
         80 . The method of  claim 47 , comprising:
 displaying, by a display controlled by the one or more processors, a time-based visual indicator of the event estimation of risk score for the plurality of time periods.   
     
     
         81 . The method of  claim 47 , wherein the plurality of time periods comprise a plurality of time periods of less than four hours. 
     
     
         82 . The method of  claim 47 , wherein the plurality of time period comprise at least one time period of less than about ten minutes, at least one time period of less than about one hour, at least one time period of less than about three hours, at least one time period of less than about one day, at least one time period of less than about one week, and at least one time period of less than about one month. 
     
     
         83 . The method of  claim 47 , wherein the calculating the event estimation of risk score comprises calculating a confidence band of the event estimation of risk score. 
     
     
         84 . The method of  claim 47 , wherein the calculating the event estimation of risk score comprises calculating an error band of the event estimation of risk score. 
     
     
         85 . The method of  claim 47 , comprising:
 determining a response to the potential medical event based at least partly on the event estimation of risk score.   
     
     
         86 . The method of  claim 85 , wherein the determined response to the potential medical event includes providing an instruction to the subject to contact a medical professional. 
     
     
         87 . The system of  claim 85 , wherein the determined response to the potential medical event includes providing an instruction to the subject to check a battery of a wearable medical device. 
     
     
         88 . The system of  claim 85 , wherein the determined response to the potential medical event includes charging a shocking mechanism of a wearable medical device. 
     
     
         89 . The system of  claim 85 , comprising:
 determining the response to the potential medical event based at least partly on a sensitivity and a specificity of the event estimation of risk score.   
     
     
         90 . The system of  claim 89 , wherein the determined response based on a first sensitivity and a first specificity is different than the determined response based on a second different sensitivity and second different specificity. 
     
     
         91 . The system of  claim 90 , wherein the determined response to the potential medical event includes at least one of informing the subject of advanced diagnostics, advising the subject against removal of equipment, advising the subject of a behavior modification, alerting a medical professional, and preparing a device for treatment. 
     
     
         92 . The system of  claim 85 , comprising:
 modifying a sensitivity of an algorithm for determining the event estimation of risk score based on a risk level of the subject.   
     
     
         93 . A medical event estimation system, comprising:
 at least one non-transitory computer-readable storage medium in communication with at least one processor and having instructions stored thereon which, when executed by the at least one processor, cause the at least one processor to perform operations comprising:
 receiving at least one signal comprising physiological parameter data of a subject; and 
 calculating estimation of time to potential events associated with at least two different potential medical events for the subject based at least partly on the physiological parameter data. 
   
     
     
         94 . A medical event estimation system, comprising:
 at least one non-transitory computer-readable storage medium in communication with at least one processor and having instructions stored thereon which, when executed by the at least one processor, cause the at least one processor to perform operations comprising:
 receiving at least one signal comprising physiological parameter data of a subject; and 
 calculating risk scores based upon estimation of time to potential events associated with at least two different potential medical events for the subject based at least partly on the physiological parameter data. 
   
     
     
         95 . A medical event estimation system, comprising:
 a medical device configured to measure physiological parameter data of a subject, transmit a signal comprising the physiological parameter data of the subject, and perform a plurality of different actions in response to a plurality of different medical events; and   at least one non-transitory computer-readable storage medium in communication with at least one processor and having instructions stored thereon which, when executed by the at least one processor, cause the at least one processor to perform operations comprising:
 receiving the at least one signal comprising the physiological parameter data of the subject; 
 calculating risk scores based upon estimation of time to potential events associated with at least two different potential medical events for the subject based at least partly on the physiological parameter data; and 
 controlling the medical device to perform at least one action of the plurality of different actions based at least partly on at least one of the calculated risk scores. 
   
     
     
         96 . A method for medical event estimation, comprising:
 measuring, by a medical device, physiological parameter data of a subject;   transmitting, by the medical device, a signal comprising the physiological parameter data of the subject;   receiving, by one or more processors, the at least one signal comprising the physiological parameter data of the subject;   calculating, by the one or more processors, risk scores based upon estimation of time to potential events associated with at least two different potential medical events for the subject based at least partly on the physiological parameter data; and   performing, by the medical device, at least one action of a plurality of different actions based at least partly on at least one of the calculated risk scores.   
     
     
         97 . A medical premonitory event estimation system, comprising:
 a non-transitory computer-readable storage medium in communication with one or more processors and having instructions stored thereon which, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
 calculating an event estimation of risk score associated with a potential medical event for a subject occurring within an associated time period based at least partly on a physiological parameter signal of the subject, wherein the calculating the event estimation of risk score comprises:
 extracting a plurality of physiological measurements from the physiological parameter signal of the subject; 
 combining at least a portion of the plurality of physiological measurements in a multivariate parameter signal; 
 applying a change point analysis to the multivariate parameter signal to determine at least one change point; 
 determining a distance of the at least one change point from a baseline; 
 applying an anomaly detection to the multivariate parameter signal to determine at least one anomaly score; and 
 determining the event estimation of risk score based at least partly on the at least one distance and the at least one anomaly score. 
 
   
     
     
         98 . The system of  claim 97 , wherein the physiological parameter signal comprises an ECG signal of the subject. 
     
     
         99 . The system of  claim 98 , wherein the plurality of physiological measurements extracted from the ECG signal include at least one of heart rate, heart rate variability, PVC burden or counts, activity, noise quantifications, atrial fibrillation, momentary pauses, heart rate turbulence, QRS height, QRS width, changes in the size or shape of the morphology, cosine R-T, artificial pacing, corrected QT interval, QT variability, T wave width, T wave alternans, T-wave variability, ST segment changes, early repolarization, late potentials, fractionated QRS/HF content, and fractionated T wave/HF content. 
     
     
         100 . The system of  claim 98 , wherein the physiological parameter signal comprises an accelerometer signal. 
     
     
         101 . The system of  claim 100 , wherein the calculating the event estimation of risk scores comprises:
 removing one or more ECG signal artifacts from the ECG signal based at least partly on the accelerometer signal.   
     
     
         102 . The system of  claim 97 , wherein the calculating the event estimation of risk score comprises:
 calculating a statistical value of the multivariate parameter signal for at least one interval of the multivariate parameter signal associated with the at least one change point; and   determining a difference between the at least one statistical value and a corresponding statistical value of the baseline.   
     
     
         103 . The system of  claim 97 , wherein the applying the anomaly detection comprises determining a plurality of raw single-parameter outputs corresponding to the plurality of physiological parameter measurements. 
     
     
         104 . The system of  claim 97 , wherein the anomaly detection comprises a neural network, and wherein the neural network is trained on the baseline of the multivariate parameter signal. 
     
     
         105 . The system of  claim 97 , wherein the determining the event estimation of risk score based at least partly on the at least one distance and the at least one anomaly score comprises classifying the at least one distance and the at least one anomaly score using one or more machine learning processes. 
     
     
         106 . The system of  claim 97 , wherein the calculating the event estimation of risk score comprises:
 applying a change point analysis to the multivariate parameter signal to determine a plurality of change points;   determining a plurality of distances of the plurality of change points from a baseline;   applying an anomaly detection to the multivariate parameter signal to determine a plurality of anomaly scores corresponding to the plurality of change points; and   determining the event estimation of risk score for the plurality of change points based the plurality of distances and the plurality of anomaly scores.   
     
     
         107 . The system of  claim 97 , comprising a wearable medical device. 
     
     
         108 . A method for medical premonitory event estimation, the method comprising:
 calculating, by one or more processors, an event estimation of risk score associated with a potential medical event for a subject occurring within an associated time period based at least partly on a physiological parameter signal of the subject, wherein the calculating the event estimation of risk score comprises:
 extracting a plurality of physiological measurements from the physiological parameter signal of the subject; 
 combining at least a portion of the plurality of physiological measurements in a multivariate parameter signal; 
 applying a change point analysis to the multivariate parameter signal to determine at least one change point; 
 determining a distance of the at least one change point from a baseline; 
 applying an anomaly detection to the multivariate parameter signal to determine at least one anomaly score; and 
 determining the event estimation of risk score based at least partly on the at least one distance and the at least one anomaly score. 
   
     
     
         109 . The method of  claim 108 , wherein the physiological parameter signal comprises an ECG signal of the subject. 
     
     
         110 . The method of  claim 109 , wherein the plurality of physiological measurements extracted from the ECG signal include at least one of heart rate, heart rate variability, PVC burden or counts, activity, noise quantifications, atrial fibrillation, momentary pauses, heart rate turbulence, QRS height, QRS width, changes in the size or shape of the morphology, cosine R-T, artificial pacing, corrected QT interval, QT variability, T wave width, T wave alternans, T-wave variability, ST segment changes, early repolarization, late potentials, fractionated QRS/HF content, and fractionated T wave/HF content. 
     
     
         111 . The method of  claim 109 , wherein the physiological parameter signal comprises an accelerometer signal. 
     
     
         112 . The method of  claim 111 , wherein the calculating the event estimation of risk scores comprises:
 removing one or more ECG signal artifacts from the ECG signal based at least partly on the accelerometer signal.   
     
     
         113 . The method of  claim 108 , wherein the calculating the event estimation of risk score comprises:
 calculating a statistical value of the multivariate parameter signal for at least one interval of the multivariate parameter signal associated with the at least one change point; and   determining a difference between the at least one statistical value and a corresponding statistical value of the baseline.   
     
     
         114 . The method of  claim 108 , wherein the applying the anomaly detection comprises determining a plurality of raw single-parameter outputs corresponding to the plurality of physiological parameter measurements. 
     
     
         115 . The method of  claim 114 , wherein the anomaly detection comprises a neural network, and wherein the neural network is trained on the baseline of the multivariate parameter signal. 
     
     
         116 . The method of  claim 108 , wherein the determining the event estimation of risk score based at least partly on the at least one distance and the at least one anomaly score comprises classifying the at least one distance and the at least one anomaly score using one or more machine learning processes. 
     
     
         117 . The method of  claim 108 , wherein the calculating the event estimation of risk score comprises:
 applying a change point analysis to the multivariate parameter signal to determine a plurality of change points;   determining a plurality of distances of the plurality of change points from a baseline;   applying an anomaly detection to the multivariate parameter signal to determine a plurality of anomaly scores corresponding to the plurality of change points; and   determining the event estimation of risk score for the plurality of change points based the plurality of distances and the plurality of anomaly scores.   
     
     
         118 . A wearable medical device configured to perform the method of  claim 108 . 
     
     
         119 . A medical premonitory event estimation system, comprising:
 a non-transitory computer-readable storage medium in communication with one or more processors and having instructions stored thereon which, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
 acquiring a first set of physiological information of a subject received during a first period of time and based at least in part on a first ECG signal of the subject, and a second set of physiological information of the subject received during a second period of time; 
 calculating first and second risk scores associated with estimating a risk of a potential cardiac arrhythmia event for the subject based on applying the first and second sets of physiological information to one or more machine learning classifier models trained on training metrics comprising at least one of i) cardiac electrophysiology metrics of a first plurality of subjects, and ii) at least one of demographic metrics and medical history metrics of the first plurality of subjects,
 wherein the one or more machine learning classifier models is validated on validation metrics of a second plurality of subjects, and wherein one or more thresholds of the one or more machine learning classifier models is set based on the validation; 
 
 providing at least the first and second risk scores associated with the potential cardiac arrhythmia event as a time changing series of risk scores; and 
 classifying the first and second risk scores associated with estimating the risk of the potential cardiac arrhythmia event for the subject based on the one or more thresholds. 
   
     
     
         120 . The system of  claim 119 , wherein the one or more thresholds comprise at least an elevated risk threshold and an immediate risk threshold, and wherein the classifying comprises,
 for each of the first and second risk scores,
 classifying the risk of the potential cardiac arrhythmia event for the subject as an elevated risk based on the first or second risk score transgressing the elevated risk threshold; and 
 classifying the risk of the potential cardiac arrhythmia event for the subject as an immediate risk based on the first or second risk score transgressing the immediate risk threshold. 
   
     
     
         121 . The system of  claim 119 , wherein the classifying comprises a time changing classification of the risk of the potential cardiac arrhythmia event for the subject based on the time changing series of risk scores. 
     
     
         122 . The system of  claim 119 , wherein the classifying comprises adjusting an underlying specificity of the one or more machine learning classifier models to reduce false positives in the underlying classification of the risk of the potential cardiac arrhythmia event for the subject. 
     
     
         123 . The system of  claim 119 , wherein the validation metrics comprises a plurality of one or more of cardiac electrophysiology metrics, demographic metrics, and medical history metrics of the second plurality of subjects. 
     
     
         124 . The system of  claim 119 , wherein the one or more processors perform operations comprising:
 updating the validation metrics by at least one of 1) adjusting one or more of the metrics in the validation metrics, and 2) expanding the validation metrics based on appending additional one or more subjects to the second plurality of subjects; and   refining the one or more thresholds based on the updated validation metrics.   
     
     
         125 . The system of  claim 119 , wherein the one or more processors perform operations comprising:
 updating the training metrics by at least one of 1) adjusting one or more of the metrics in the training metrics, and 2) expanding the training metrics based on appending additional one or more subjects to the first plurality of subjects; and   retraining the one or more machine learning classifier models based on the updated training metrics.   
     
     
         126 . The system of  claim 119 , wherein the validation metrics of the second plurality of subjects is independent from the training metrics of the first plurality of subjects. 
     
     
         127 . The system of  claim 119 , wherein the one or more machine learning classifier models are validated on the validation metrics, and wherein the validation metrics comprises an indication of the presence or absence of ectopic beats in an underlying one or more of the validation metrics. 
     
     
         128 . The system of  claim 119 , wherein the one or more processors perform operations comprising:
 discriminating between normal and ectopic beats in the first set of physiological information of the subject received during the first period of time and based at least in part on the first ECG signal.   
     
     
         129 . The system of  claim 119 , wherein at least one of the training metrics and the validation metrics comprises metrics based on at least one of heart rate, heart rate variability, non-sustained ventricular tachycardia (VT) episodes count, and premature ventricular contraction (PVC) count. 
     
     
         130 . The system of  claim 119 , wherein at least one of the training metrics and the validation metrics comprises metrics based on heart rate variability, and wherein the metrics comprise a standard deviation over time of normal-to-normal intervals. 
     
     
         131 . The system of  claim 119 , wherein at least one of the training metrics and the validation metrics comprises metrics based on at least one of QRS width, QRS height, single lead QRS morphology, and dual lead QRS morphology. 
     
     
         132 . The system of  claim 119 , wherein at least one of the training metrics and the validation metrics comprises metrics based on single lead QRS morphology, and wherein the metrics comprise an average over time of similarity scores respectively on side-to-side (SS) and front-to-back (FB) channels. 
     
     
         133 . The system of  claim 119 , wherein at least one of the training metrics and the validation metrics comprises metrics based on QRS width, and wherein the metrics comprise at least one of a standard deviation over time of an estimated width of QRS complexes and a mean over time of the estimated width of QRS complexes. 
     
     
         134 . The system of  claim 119 , wherein at least one of the training metrics and the validation metrics comprises metrics based on QRS height, and wherein the metrics comprise a standard deviation over time of an estimated height of QRS complexes. 
     
     
         135 . The system of  claim 119 , wherein at least one of the training metrics and the validation metrics comprises metrics based on at least one of QT variability, ST depression, elevation, and/or slope, T-wave alternant, T-wave variability, and dual lead T-wave morphology. 
     
     
         136 . The system of  claim 119 , wherein at least one of the training metrics and the validation metrics comprises heart sounds metrics. 
     
     
         137 . The system of  claim 136 , wherein the heart sounds metrics comprise S3 and S4 heart sound metrics. 
     
     
         138 . The system of  claim 119 , wherein at least one of the training metrics and the validation metrics comprises electromechanical activation time metrics describing an interval from a first predetermined fiducial timepoint in the electrocardiograph (ECG) to a second predetermined fiducial timepoint in a subsequent mechanical activity of the heart. 
     
     
         139 . The system of  claim 138 , wherein the first predetermined fiducial timepoint in the ECG comprises an onset of P-wave and QRS complexes, wherein the onset of the P-wave and QRS complexes comprises timepoints relating to at least one of a) a P-wave, b) a Q-wave, c) an R-wave, and d) an S-wave. 
     
     
         140 . The system of  claim 138 , wherein the subsequent mechanical activity of the heart comprises left ventricular wall motion. 
     
     
         141 . The system of  claim 140 , wherein the second predetermined fiducial timepoint in the subsequent mechanical activity of the heart comprises at least one of a) a timepoint of maximal left ventricular wall motion, and b) a state of a relaxation of the left ventricular wall motion. 
     
     
         142 . The system of  claim 138 , wherein the second predetermined fiducial timepoint in the subsequent mechanical activity of the heart comprises a timepoint of peak intensity of the 51 heart sound. 
     
     
         143 . The system of  claim 138 , wherein the second predetermined fiducial timepoint in the subsequent mechanical activity of the heart is based on ultrasound measurements of the heart. 
     
     
         144 . The system of  claim 138 , wherein the electromechanical activation time metrics comprises a percent electromechanical activation time metric. 
     
     
         145 . The system of  claim 119 , wherein the classifying comprises calculating at least one of an area under a plotted curve of the time changing series of risk scores and a mean of the time changing series of risk scores. 
     
     
         146 . The system of  claim 119 , wherein the first and second risk scores are classified based on an amount that the first and second risk scores transgress the one or more thresholds. 
     
     
         147 . The system of  claim 119 , wherein the first and second risk scores are classified based on a number of times that the time changing series of risk scores transgress the one or more thresholds. 
     
     
         148 . The system of  claim 119 , wherein the one or more processors perform operations comprising:
 notifying at least one of the subject and a third party based on the classification of the first and second risk scores.   
     
     
         149 . The system of  claim 148 , wherein the classification of the first and second risk scores indicates at least one of an elevated risk and an immediate risk. 
     
     
         150 . The system of  claim 148 , wherein the notifying comprises sending a notification to at least one member of a medical team of the subject. 
     
     
         151 . The system of  claim 119 , wherein the one or more processors perform operations comprising:
 adjusting a time interval between detection of a cardiac event and a treatment for the cardiac event based on the classification of the first and second risk scores.   
     
     
         152 . An external medical device comprising the medical premonitory event estimation system of  claim 1 , wherein the external medical device is configured to monitor a cardiac condition of the subject. 
     
     
         153 . The external medical device of  claim 152 , wherein the external medical device comprises a wearable medical device. 
     
     
         154 . The external medical device of  claim 152 , wherein the one or more processors perform operations comprising:
 modifying one or more functions or features of a user interface of the external medical device based on the classification of the first and second risk scores.   
     
     
         155 . A medical premonitory event estimation system, comprising:
 a non-transitory computer-readable storage medium in communication with one or more processors and having instructions stored thereon which, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
 calculating an event estimation of risk score associated with a potential medical event for a subject, wherein the event estimation of risk score is calculated based on a machine learning classifier, wherein the machine learning classifier is trained on training metrics comprising at least one of i) cardiac electrophysiology metrics of a plurality of subjects and ii) at least one of demographic metrics and medical history metrics of the plurality of subjects. 
   
     
     
         156 . The system of  claim 155 , wherein the machine learning classifier is validated based on a validation metrics independent from the training metrics, and wherein at least one threshold of the machine learning classifier is set based on the validation. 
     
     
         157 . The system of  claim 156 , wherein the one or more processors perform operations comprising:
 classifying the event estimation of risk score based on at least one threshold.   
     
     
         158 . The system of  claim 156 , wherein the at least one threshold is substantially continuously refined based on a substantially continuously updated registry of metrics. 
     
     
         159 . The system of  claim 155 , wherein the one or more processors perform operations comprising:
 calculating a plurality of event estimation of risk scores associated with the potential medical event for the subject over a time series; and   analyzing the plurality of event estimation of risk scores to determine at least one trend associated with the potential medical event for the subject over the time series.   
     
     
         160 . The system of  claim 156 , wherein at least one of the training metrics and the validation metrics comprises heart sounds metrics. 
     
     
         161 . The system of  claim 160 , wherein the heart sounds metrics comprise S3 and S4 heart sound metrics. 
     
     
         162 . The system of  claim 156 , wherein at least one of the training metrics and the validation metrics comprises electromechanical activation time metrics describing an interval from a first predetermined fiducial timepoint in the electrocardiograph (ECG) to a second predetermined fiducial timepoint in a subsequent mechanical activity of the heart. 
     
     
         163 . The system of  claim 162 , wherein the first predetermined fiducial timepoint in the ECG comprises an onset of P-wave and QRS complexes, wherein the onset of the P-wave and QRS complexes comprises timepoints relating to at least one of a) a P-wave, b) a Q-wave, c) an R-wave, and d) an S-wave. 
     
     
         164 . The system of  claim 162 , wherein the subsequent mechanical activity of the heart comprises left ventricular wall motion. 
     
     
         165 . The system of  claim 164 , wherein the second predetermined fiducial timepoint in the subsequent mechanical activity of the heart comprises at least one of a) a timepoint of maximal left ventricular wall motion, and b) a state of a relaxation of the left ventricular wall motion. 
     
     
         166 . The system of  claim 162 , wherein the second predetermined fiducial timepoint in the subsequent mechanical activity of the heart comprises a timepoint of peak intensity of the 51 heart sound. 
     
     
         167 . The system of  claim 162 , wherein the second predetermined fiducial timepoint in the subsequent mechanical activity of the heart is based on ultrasound measurements of the heart. 
     
     
         168 . The system of  claim 162 , wherein the electromechanical activation time metrics comprises a percent electromechanical activation time metric. 
     
     
         169 . A medical premonitory event estimation method, comprising:
 acquiring a first set of physiological information of a subject received during a first period of time and based at least in part on a first ECG signal of the subject, and a second set of physiological information of the subject received during a second period of time;   calculating first and second risk scores associated with estimating a risk of a potential cardiac arrhythmia event for the subject based on applying the first and second sets of physiological information to one or more machine learning classifier models trained on training metrics comprising at least one of i) cardiac electrophysiology metrics of a first plurality of subjects, and ii) at least one of demographic metrics and medical history metrics of the first plurality of subjects,
 wherein the one or more machine learning classifier models is validated on a validation metrics of a second plurality of subjects, and wherein one or more thresholds of the one or more machine learning classifier models is set based on the validation; 
   providing at least the first and second risk scores associated with the potential cardiac arrhythmia event as a time changing series of risk scores; and   classifying the first and second risk scores associated with estimating the risk of the potential cardiac arrhythmia event for the subject based on the one or more thresholds.   
     
     
         170 . The method of  claim 169 , wherein the one or more thresholds comprise at least an elevated risk threshold and an immediate risk threshold, and wherein the classifying comprises,
 for each of the first and second risk scores,
 classifying the risk of the potential cardiac arrhythmia event for the subject as an elevated risk based on the first or second risk score transgressing the elevated risk threshold; and 
 classifying the risk of the potential cardiac arrhythmia event for the subject as an immediate risk based on the first or second risk score transgressing the immediate risk threshold. 
   
     
     
         171 . The method of  claim 169 , wherein the classifying comprises a time changing classification of the risk of the potential cardiac arrhythmia event for the subject based on the time changing series of risk scores. 
     
     
         172 . The method of  claim 169 , wherein the classifying comprises adjusting an underlying specificity of the one or more machine learning classifier models to reduce type 1 errors (false positives) in the underlying classification of the risk of the potential cardiac arrhythmia event for the subject. 
     
     
         173 . The method of  claim 169 , wherein the validation metrics comprises a plurality of one or more of cardiac electrophysiology metrics, demographic metrics, and medical history metrics of the second plurality of subjects. 
     
     
         174 . The method of  claim 169 , wherein at least one of the training metrics and the validation metrics comprises heart sounds metrics. 
     
     
         175 . The method of  claim 169 , wherein the heart sounds metrics comprise S3 and S4 heart sound metrics.