Systems and methods for diagnosing a health condition based on patient time series data
Disclosed systems, methods, and computer readable media can diagnose a health condition based on patient time series data. For example, a method for diagnosing a health condition based on patient time series data includes identifying a training set of health records comprising a first set of patient time series data, training a neural network using the training set of health records, and executing the trained neural network model to diagnose a health condition based on a second set of patient time series data. In further examples, the first set of patient time series data and the second set of patient time series data can each comprise electrocardiogram data and the health condition can comprise pulmonary hypertension.
1. A method comprising:
identifying, by one or more hardware processors, a training set of health records comprising a first set of patient time series data, wherein the first set of patient time series data comprises a preemptive set of patient time series data captured earlier than a predetermined amount of time before a date of a positive diagnosis for a health condition;
training, by the one or more hardware processors, a neural network model using the training set of health records, wherein training the neural network model comprises:
receiving initial parameters of the neural network model, wherein the initial parameters of the neural network model are transfer learned from an independently learned self-supervised network;
dividing the training set of health records into a training set, a validation set and a test set; and
modifying the initial parameters of the neural network model as a function of the training set, the validation set and the test set; and
executing, by the one or more hardware processors, the trained neural network model to diagnose the health condition based on a second set of patient time series data.
2. The method of claim 1 , wherein the trained neural network model receives a plurality of inputs including at least one vector representation of the second set of patient time series data and at least one spectrogram representation of the second set of patient time series data.
3. The method of claim 1 , wherein the second set of patient time series data comprises electrocardiogram data and the health condition comprises pulmonary hypertension.
4. The method of claim 1 , wherein identifying the training set of health records comprises searching a corpus of health records to identify health records associated with a cohort of patients that satisfy one or more predetermined criteria.
5. The method of claim 1 , wherein the training set of health records further comprises a discrete metric derived from the first set of patient time series data.
6. The method of claim 5 , wherein the first set of patient time series data comprises electrocardiogram data and the discrete metric comprises a QT interval.
7. The method of claim 1 , wherein the training set of health records comprises a first set of health records associated with patients diagnosed with the health condition and a second set of health records associated with patients not diagnosed with the health condition.
8. The method of claim 1 , wherein the first set of patient time series data comprises a preemptive set of patient time series data captured earlier than a predetermined amount of time before a date of a positive diagnosis for the health condition.
9. The method of claim 1 , wherein the first set of patient time series data comprises electrocardiogram data and wherein training the neural network comprises randomly shuffling a set of leads associated with the electrocardiogram data at an input of the neural network.
10. The method of claim 1 , wherein the first set of patient time series data comprises electrocardiogram data and wherein training the neural network comprises shifting a level of each lead among a set of leads associated with the electrocardiogram data by a random amount.
11. The method of claim 1 , wherein executing the trained neural network model comprises segmenting the second set of patient time series data into a plurality of time windows having a predetermined duration.
12. The method of claim 1 , wherein identifying the training set of health records comprises:
receiving a corpus of health records comprising the first set of patient time series data;
creating one or more cohorts within the corpus of health records as a function of the first set of patient time series data; and
identifying the training set of health records as a function of the corpus of health records and the one or more cohorts.
13. A system comprising:
a non-transitory memory; and
one or more hardware processors configured to read instructions from the non-transitory memory that, when executed, cause the one or more hardware processors to perform operations comprising:
identifying a training set of health records comprising a first set of patient time series data, wherein the first set of patient time series data comprises a preemptive set of patient time series data captured earlier than a predetermined amount of time before a date of a positive diagnosis for a health condition;
training a neural network model using the training set of health records, wherein training the neural network model comprises:
receiving initial parameters of the neural network model, wherein the initial parameters of the neural network model are transfer learned from an independently learned self-supervised network;
dividing the training set of health records into a training set, a validation set and a test set; and
modifying the initial parameters of the neural network model as a function of the training set, the validation set and the test set; and
executing the trained neural network model to diagnose the health condition based on a second set of patient time series data.
14. The system of claim 13 , wherein the trained neural network model receives a plurality of inputs including at least one vector representation of the second set of patient time series data and at least one spectrogram representation of the second set of patient time series data.
15. The system of claim 13 , wherein the second set of patient time series data comprises electrocardiogram data and the health condition comprises pulmonary hypertension.
16. The system of claim 13 , wherein the training set of health records further comprises a discrete metric derived from the first set of patient time series data.
17. The system of claim 16 , wherein the first set of patient time series data comprises electrocardiogram data and the discrete metric comprises a QT interval.
18. The system of claim 13 , wherein identifying the training set of health records comprises:
receiving a corpus of health records comprising the first set of patient time series data;
creating one or more cohorts within the corpus of health records as a function of the first set of patient time series data; and
identifying the training set of health records as a function of the corpus of health records and the one or more cohorts.
19. A medical instrument comprising:
an electrocardiogram monitor for capturing electrocardiogram data of a patient; and
at least one hardware processor connected to the electrocardiogram monitor, wherein the at least one hardware processor receives the electrocardiogram data from the electrocardiogram monitor and executes a trained neural network model to diagnose the patient for pulmonary hypertension based on the electrocardiogram data, wherein training the trained neural network model comprises:
identifying a training set of health records comprising a first set of patient time series data wherein the first set of patient time series data comprises a preemptive set of patient time series data captured earlier than a predetermined amount of time before a date of a positive diagnosis for a health condition; and
training, a neural network model using the training set of health records comprising:
receiving initial parameters of the neural network model, wherein the initial parameters of the neural network model are transfer learned from an independently learned self-supervised network;
dividing the training set of health records into a training set, a validation set and a test set; and
modifying the initial parameters of the neural network model as a function of the training set, the validation set and the test set.
20. The medical instrument of claim 19 , wherein the at least one hardware processor derives a QT interval from the electrocardiogram data, and wherein the QT interval and the electrocardiogram data are each provided as inputs to the trained neural network model.