Substantially-median-based determination of long-term heart rates from ECG data of wearable cardioverter defibrillator (WCD) system
A wearable medical monitoring (WMM) system may be worn for a long time. Some embodiments of WMM systems are wearable cardioverter defibrillator (WCD) systems. In such systems, ECG electrodes sense an ECG signal of the patient, and store it over the long-term. The stored ECG signal can be analyzed for helping long-term heart rate monitoring of the patient. The heart rate monitoring can be assisted a) by special filtering techniques that remove short-term variations inherent in patients' short-term heart rate determinations, and b) by indication techniques that indicate when conditions hampered sensing of the ECG signal too much for a reliable heart rate determination.
1 . A computer system for helping monitor a patient, the patient having worn a wearable medical monitor (WMM) system for at least one hour during which the WMM system included a plurality of Electrocardiogram (ECG) electrodes, a discharge circuit to store electrical charge, a support structure worn by the patient so as to maintain the plurality of ECG electrodes on a body of the patient, the plurality of ECG electrodes thus defining two or more channels and sensing two or more versions of an ECG signal of the patient across the two or more channels, a WMM processor analyzing short segments of the sensed ECG signal, and a memory storing WMM system data about the sensed ECG signal, the stored WMM system data being generated from the at least one hour of the sensed ECG signal, the computer system comprising:
one or more computer processors distinct from the WMM processor; and
a non-transitory computer-readable storage medium storing instructions which, when executed by the one or more computer processors, result in operations comprising:
receiving the stored WMM system data;
inputting computed raw heart rate (HR) values for respective ones of the short ECG signal segments;
aggregating groups of the computed raw HR values into respective first time bins that are arranged in a time sequence;
discarding, from their respective first time bins, computed raw HR values that meet an error condition, wherein a computed raw HR value in a certain one of the first time bins meets the error condition when the computed raw HR value differs from another raw HR value aggregated into the certain one of the first time bins by at least an error HR threshold, and no two other raw HR values aggregated into the certain one of the first time bins differ from each other by as much as the error HR threshold;
deriving first HR values for respective ones of the first time bins, a first HR value of a certain one of the first time bins being derived from the computed raw HR values aggregated and remaining into the certain one of the first time bins after the computed raw HR values that meet the error condition are discarded, wherein the first HR value is derived using a median-based selection of the remaining HR values;
storing at least some of the first HR values, the stored first HR values having been thus derived from the stored WMM system data that is generated from the at least one hour of the sensed ECG signal; and
utilizing the stored first HR values to perform long-term monitoring of heart rate of the patient, wherein the long-term monitoring of the heart rate of the patient comprises detecting a medical condition of the patient, and wherein the medical condition comprises a shockable arrhythmia;
wherein responsive to detecting the medical condition and determining that a shock criterion is met, the WMM system is configured to cause at least some of the electrical charge stored in the discharge circuit to be discharged through the patient while the support structure is worn by the patient, to deliver a shock to the patient.
2 . The computer system of claim 1 , wherein:
the stored WMM system data encodes amplitude values of the sensed ECG signal, and
when the instructions are executed by the one or more computer processors, the resulting operations further comprise:
computing, from the amplitude values, the raw HR values that are subsequently inputted.
3 . The computer system of claim 1 , wherein:
the WMM processor further computes the raw HR values, and
the stored WMM system data includes the raw HR values that are subsequently inputted.
4 . The computer system of claim 1 , wherein the error condition includes that one of the plurality of ECG electrodes is detected to be off.
5 . The computer system of claim 1 , wherein the error condition includes that, in at least one of the two or more versions, a detected noise exceeds a noise threshold.
6 . The computer system of claim 1 , wherein:
when the instructions are executed by the one or more computer processors, the resulting operations further comprise:
after deriving and before storing, replacing a certain one of the first HR values of a certain one of the first time bins by an adjusted first HR value that is derived from the first HR value of the first time bin that is within a filter range of the certain one of the first time bins in the time sequence, and is not derived from the certain one of the first HR values.
7 . The computer system of claim 1 , further comprising:
a screen, and
wherein when the instructions are executed by the one or more computer processors, the resulting operations further comprise:
marking as error-prone at least some of the first HR values derived for the first time bins from which the computed raw HR values were discarded; and
displaying the stored first HR values, and displaying error indicia in relation to the displayed first HR values that are marked as error-prone.
8 . The computer system of claim 1 , wherein:
when the instructions are executed by the one or more computer processors, the resulting operations further comprise:
aggregating groups of the first time bins into respective second time bins; and
deriving second HR values for respective ones of the second time bins, the second HR value of a certain one of the second time bins being derived from at least some of the first HR values of the first time bins that are aggregated into the certain one of the second time bins, and the second HR values are stored instead of the first HR values.
9 . The computer system of claim 8 , further comprising:
a screen, and
wherein when the instructions are executed by the one or more computer processors, the resulting operations further comprise:
marking as error-prone at least some of the second HR values derived for second time bins that include aggregated first time bins from which the computed raw HR values were discarded;
displaying the stored second HR values; and
displaying error indicia in relation to the displayed second HR values that are marked as error-prone.
10 . A method for helping monitor a patient, the patient having worn a wearable medical monitor (WMM) system for at least one hour during which the WMM system included a plurality of Electrocardiogram (ECG) electrodes, a discharge circuit to store electrical charge, a support structure worn by the patient so as to maintain the plurality of ECG electrodes on a body of the patient, the plurality of ECG electrodes thus defining two or more channels and sensing two or more versions of an ECG signal of the patient across the two or more channels, a WMM processor analyzing short segments of the sensed ECG signal, and a memory storing WMM system data about the sensed ECG signal, the stored WMM system data being generated from the at least one hour of the sensed ECG signal, the method being performed by one or more computer processors distinct from the WMM processor, and the method comprising:
receiving the stored WMM system data;
inputting computed raw heart rate (HR) values for respective ones of the short ECG signal segments;
aggregating groups of the computed raw HR values into respective first time bins that are arranged in a time sequence;
discarding, from their respective first time bins, the computed raw HR values that meet an error condition, wherein a computed raw HR value in a certain one of the first time bins meets the error condition when the computed raw HR value differs from another raw HR value aggregated into the certain one of the first time bins by at least an error HR threshold, and no two other raw HR values aggregated into the certain one of the first time bins differ from each other by as much as the error HR threshold;
deriving first HR values for respective ones of the first time bins, a first HR value of a certain one of the first time bins being derived from the computed raw HR values aggregated and remaining into the certain one of the first time bins after the computed raw HR values that meet the error condition are discarded, wherein the first HR value is derived using a median-based selection of the remaining HR values;
storing at least some of the first HR values, the stored first HR values having been thus derived from the stored WMM system data that is generated from the at least one hour of the sensed ECG signal; and
utilizing the stored first HR values to perform long-term monitoring of heart rate of the patient, wherein the long-term monitoring of the heart rate of the patient comprises detecting a medical condition of the patient, and wherein the medical condition comprises a shockable arrhythmia;
wherein responsive to detecting the medical condition and determining that a shock criterion is met, the WMM system is configured to cause at least some of the electrical charge stored in the discharge circuit to be discharged through the patient while the support structure is worn by the patient, to deliver a shock to the patient.
11 . The method of claim 10 , wherein the stored WMM system data encodes amplitude values of the sensed ECG signal, and the method further comprising:
computing, from the amplitude values, the raw HR values that are subsequently inputted.
12 . The method of claim 10 , wherein:
the WMM processor further computes the raw HR values, and
the stored WMM system data includes the raw HR values that are subsequently inputted.
13 . The method of claim 10 , wherein the error condition includes that one of the plurality of ECG electrodes is detected to be off.
14 . The method of claim 10 , wherein the error condition includes that, in at least one of the two or more versions, a detected noise exceeds a noise threshold.
15 . The method of claim 10 , further comprising:
after deriving and before storing, replacing a certain one of the first HR values of a certain one of the first time bins by an adjusted first HR value that is derived from the first HR value of the first time bin that is within a filter range of the certain one of the first time bins in the time sequence, and is not derived from the certain one of the first HR values.
16 . The method of claim 10 , further comprising:
marking as error-prone at least some of the first HR values derived for first time bins from which the computed raw HR values were discarded; and
displaying the stored first HR values, and displaying error indicia in relation to the displayed first HR values that are marked as error-prone.
17 . The method of claim 10 , further comprising:
aggregating groups of the first time bins into respective second time bins; and
deriving second HR values for respective ones of the second time bins, the second HR value of a certain one of the second time bins being derived from at least some of the first HR values of the first time bins that are aggregated into the certain one of the second time bins, and
the second HR values are stored instead of the first HR values.
18 . The method of claim 17 , further comprising:
marking as error-prone at least some of the second HR values derived for second time bins that include aggregated first time bins from which the computed raw HR values were discarded;
displaying the stored second HR values; and
displaying error indicia in relation to the displayed second HR values that are marked as error-prone.