Systems and methods for machine learning from medical records

Patent metadata

Field	Value
Publication number	US-12562283-B2
Application number	US-202217732322-A
Country	US
Kind code	B2
Filing date	Apr 28, 2022
Priority date	Apr 28, 2021
Publication date	Feb 24, 2026
Grant date	Feb 24, 2026

Abstract

Systems and methods for machine learning of medical records are provided. The system can execute multiple machine learning models on the medical records in parallel using multi-threaded approach wherein each machine learning model executes using its own, dedicated computational thread in order to significantly speed up the time with which relevant information can be identified from documents by the system. The multi-threaded machine learning models can include, but are not limited to, sentence classification models, comorbidity models, ICD models, body parts models, prescription models, and provider name models. The system can also utilize combined convolutional neural networks and long short-term models (CNN+LSTMs) as well as ensemble machine learning models to categorize sentences in medical records. The system can also extract service provider, medical specializations, and dates of service information from medical records.

First claim

What is claimed is:

1 . A machine learning system for automatically extracting information from medical records, comprising:
a memory storing a plurality of medical records; and a processor in communication with the memory, the processor programmed to perform the steps of: retrieving the plurality of medical records from the memory; identifying a plurality of machine learning models suitable for processing the plurality of medical records to extract desired information from the plurality of medical records; executing each of the plurality of machine learning models on the plurality of medical records such that each of the plurality of machine learning models executes in a dedicated computational thread dedicated solely to a single one of said plurality of machine learning models, each computational thread including a timeout parameter specifying a time at which model results are collected in the event that model execution is not completed before the timeout parameter; processing outputs of the plurality of machine learning models to identify the extracted desired information; and returning the extracted desired information.

2 . The system of claim 1 , wherein the plurality of medical records include nurse summaries, doctor summaries, medical claims data, or insurance claims data.

3 . The system of claim 1 , wherein the plurality of machine learning models include at least one of a sentence classification model, a comorbidity model, an ICD model, a body parts model, a prescription model, a provider name model, a medical specialization extraction model, or a date of service extraction model.

4 . The system of claim 1 , wherein the dedicated computational threads are executed in parallel.

5 . The system of claim 1 , wherein the plurality of machine learning models include one or more of a Word2Vec model, a convolutional neural network (CNN), a long short-term model (LSTM), or an ensemble machine learning model.

6 . The system of claim 1 , wherein the plurality of medical records are retrieved from a medical records computer system.

7 . The system of claim 1 , wherein the plurality of machine learning models include a plurality a sentence relevance model and a plurality of sentence category models.

8 . The system of claim 7 , wherein the plurality of sentence category models include an assessment model, a recommendation model, and a procedure model.

9 . The system of claim 8 , wherein the sentence relevance model outputs predicted in-summary sentences for processing by the assessment model.

10 . The system of claim 9 , wherein the assessment model processes the predicted in-summary sentences and outputs predicted assessment and non-assessment sentences.

11 . The system of claim 10 , wherein the recommendation model processes the predicted non-assessment sentences and outputs predicted recommendation and non-recommendation sentences.

12 . The system of claim 11 , wherein the procedure model processes the predicted non-recommendation sentences and outputs predicted procedure and non-procedure sentences.

13 . The system of claim 1 , wherein the extracted desired information is accessible via an application programming interface (API).

14 . A machine learning method for automatically extracting information from medical records, comprising the steps of:
retrieving by a processor a plurality of medical records from a memory in communication with the processor; identifying a plurality of machine learning models suitable for processing the plurality of medical records to extract desired information from the plurality of medical records; executing each of the plurality of machine learning models on the plurality of medical records such that each of the plurality of machine learning models executes in a dedicated computational thread dedicated solely to a single one of said plurality of machine learning models, each computational thread including a timeout parameter specifying a time at which model results are collected in the event that model execution is not completed before the timeout parameter; processing outputs of the plurality of machine learning models to identify the extracted desired information; and returning the extracted desired information.

15 . The method of claim 14 , wherein the plurality of medical records include nurse summaries, doctor summaries, medical claims data, or insurance claims data.

16 . The method of claim 14 , wherein the plurality of machine learning models include at least one of a sentence classification model, a comorbidity model, an ICD model, a body parts model, a prescription model, a provider name model, a medical specialization extraction model, or a date of service extraction model.

17 . The method of claim 14 , wherein the dedicated computational threads are executed in parallel.

18 . The method of claim 14 , wherein the plurality of machine learning models include one or more of a Word2Vec model, a convolutional neural network (CNN), a long short-term model (LSTM), or an ensemble machine learning model.

19 . The method of claim 14 , wherein the plurality of medical records are retrieved from a medical records computer system.

20 . The method of claim 14 , wherein the plurality of machine learning models include a plurality a sentence relevance model and a plurality of sentence category models.

21 . The method of claim 14 , wherein the plurality of sentence category models include an assessment model, a recommendation model, and a procedure model.

22 . The method of claim 21 , wherein the sentence relevance model outputs predicted in-summary sentences for processing by the assessment model.

23 . The method of claim 21 , wherein the assessment model processes the predicted in-summary sentences and outputs predicted assessment and non-assessment sentences.

24 . The method of claim 23 , wherein the recommendation model processes the predicted non-assessment sentences and outputs predicted recommendation and non-recommendation sentences.

25 . The method of claim 24 , wherein the procedure model processes the predicted non-recommendation sentences and outputs predicted procedure and non-procedure sentences.

26 . The method of claim 14 , wherein the extracted desired information is accessible via an application programming interface (API).

Assignees

• Insurance Services Office Inc

Inventors

• Godbole Sneha
• Riordan Kate M
• Lall Abhishek
• Lung Pei-Yau
• Kalra Shubham
• Kunapuli Gautam
• Bordia Shikha
• Jain Tushar
• Ahmed Syed Danish

Classifications

• G16H50/20

  for computer-aided diagnosis, e.g. based on medical expert systems · CPC title

• G16H10/60

  for patient-specific data, e.g. for electronic patient records · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title

• G06N3/08

  Learning methods · CPC title