Phenotype-specific heart failure diagnosis and management

1 . A medical-device system for detecting and managing heart failure in a patient, the medical-device system comprising:

a receiver circuit configured to receive physiological and clinical information of the patient; and

a heart failure detector circuit configured to:

classify the patient into one of a plurality of phenotypes using the received physiological and clinical information, wherein each phenotype of the plurality of phenotypes is associated with a distinct false positive detection rate for heart failure detection;

modify one or more heart failure detection settings based on the false positive detection rate of the classified phenotype, the one or more heart failure detection settings including one or more of a detection parameter value or a detection algorithm, wherein the one or more heart failure detection settings are distinct from treatment parameters; and

detect a heart failure status in the patient using the received physiological and clinical information and the modified one or more heart failure detection settings.

2 . The medical-device system of claim 1 , wherein the received physiological and clinical information includes heart sound information,

wherein the heart failure detector circuit is configured to detect the heart failure status including a presence or absence of a heart failure with preserved ejection fraction (HFpEF) using the heart sound information and the modified one or more heart failure detection settings.

3 . The medical-device system of claim 1 , wherein the plurality of phenotypes are each characterized by a cluster of physiological and clinical features in a multi-dimensional feature space including at least one of:

demographic feature;

clinical or laboratory test data feature;

medical history data;

medication information;

heart failure comorbidity information; or

sensor signal features produced by one or more ambulatory physiological sensors.

4 . The medical-device system of claim 1 , wherein the heart failure detector is configured to classify the patient into one of the plurality of phenotypes using a trained machine learning (ML) or artificial intelligence model.

5 . The medical-device system of claim 1 , wherein the heart failure detector is configured to classify the patient into one of the plurality of phenotypes based on a similarity between the received physiological and clinical information of the patient and one or more of the plurality of phenotypes.

6 . The medical-device system of claim 1 , wherein the heart failure detector is further configured to:

receive information sensed from the patient by at least one phenotype-specific sensor for the classified phenotype; and

confirm or modify the classified phenotype of the patient using the received information sensed by the at least one phenotype-specific sensor.

7 . The medical-device system of claim 1 , comprising a storage device configured to store a correspondence between the plurality of phenotypes and corresponding candidate heart failure detection settings,

wherein to modify the one or more heart failure detection settings, the heart failure detector circuit is configured to select from the stored candidate heart failure detection settings based on the classified phenotype.

8 . The medical-device system of claim 1 , comprising a sensor circuit configured to selectively sense physiological signal based on the classified phenotype,

wherein the heart failure detector circuit is configured to detect the heart failure status using the selectively sensed physiologic signal.

9 . The medical-device system of claim 1 , wherein to detect the heart failure status in the patient, the heart failure detector circuit is configured to:

compute a composite signal index using the received physiological and clinical information and the modified one or more heart failure detection settings; and

detect the heart failure status in response to the composite signal index satisfying a specific condition.

10 . The medical-device system of claim 9 , wherein the heart failure detector circuit is configured to determine or adjust a detection threshold value based on the classified phenotype, and to detect the heart failure status based on a comparison between the composite signal index and the determined or adjusted detection threshold value.

11 . The medical-device system of claim 9 , wherein the heart failure detector circuit is configured to:

determine or adjust weights for one or more of a plurality of signal metrics derived from the received physiological and clinical information; and

compute the composite signal index using a weighted combination of the plurality of signal metrics.

12 . The medical-device system of claim 11 , wherein the heart failure detector circuit is configured to determine or adjust the weights for the one or more of the plurality of signal metrics further based on a severity of a symptom or a comorbid condition associated with the classified phenotype.

13 . The medical-device system of claim 1 , comprising a therapy circuit configured to generate and deliver a heart failure therapy to the patient in accordance with the detected heart failure status.

14 . The medical-device system of claim 1 , wherein to modify the one or more heart failure detection settings, the heart failure detector circuit is configured to modify a detection threshold value based on the distinct false positive detection rate of the classified phenotype.

15 . A method of detecting and managing heart failure in a patient using a medical-device system, the method comprising:

receiving physiological and clinical information of the patient using a receiving circuit of the medical-device system; and

by operation of a heart failure detector circuit implemented in the medical-device system, automatically performing operations comprising:

classifying the patient into one of a plurality of phenotypes using the received physiological and clinical information, wherein each phenotype of the plurality of phenotypes is associated with a distinct false positive detection rate for heart failure detection;

modifying one or more heart failure detection settings based on the false positive detection rate of the classified phenotype, the one or more heart failure detection settings including one or more of a detection parameter value or a detection algorithm, wherein the one or more heart failure detection settings are distinct from treatment parameters; and

detecting a heart failure status in the patient using the received physiological and clinical information and the modified one or more heart failure detection settings.

16 . The method of claim 15 , wherein the plurality of phenotypes are each characterized by a cluster of physiological and clinical features in a multi-dimensional feature space including at least one of:

demographic feature;

clinical or laboratory test data feature;

medical history data;

medication information;

heart failure comorbidity information; or

sensor signal features produced by one or more ambulatory physiological sensors.

17 . The method of claim 15 , wherein classifying the patient into one of the plurality of phenotypes is based on a similarity between the received physiological and clinical information of the patient and one or more of the plurality of phenotypes or includes using a trained machine learning (ML) or artificial intelligence model.

18 . The method of claim 15 , further comprising:

receiving information sensed from the patient by at least one phenotype-specific sensor for the classified phenotype; and

confirming or modifying the classified phenotype of the patient using the received information sensed by the at least one phenotype-specific sensor.

19 . The method of claim 15 , comprising storing a correspondence between the plurality of phenotypes and corresponding candidate heart failure detection settings,

wherein modifying the one or more heart failure detection settings includes selecting from the stored candidate heart failure detection settings based on the classified phenotype.

20 . The method of claim 15 , wherein detecting the heart failure status in the patient includes:

computing a composite signal index using a weighted combination of a plurality of signal metrics derived from the received physiological and clinical information and each scaled by an adjustable weight; and

detecting the heart failure status in response to the composite signal index satisfying a specific condition.