Methods, systems, and computer program products using artificial intelligence for coordinated identification of patients for a clinical trial that are served by multiple providers

Patent metadata

Field	Value
Publication number	US-12562274-B2
Application number	US-202117218487-A
Country	US
Kind code	B2
Filing date	Mar 31, 2021
Priority date	Mar 31, 2021
Publication date	Feb 24, 2026
Grant date	Feb 24, 2026

Abstract

A method includes receiving patient medical record information from each of a plurality of providers, the plurality of providers being associated with a plurality of different organizational managing entities, respectively; querying the patient medical record information of each of the plurality of providers using selection criteria to identify a first subset patients having first characteristics that match first screening requirements for a clinical trial; identifying, using an artificial intelligence engine, ones of the first subset of patients whose medical record information includes second characteristics that match second screening requirements for the clinical trial as clinical trial patient candidates; and communicating identities of the clinical trial patient candidates to ones of the plurality of providers that provide healthcare services to the clinical trial patient candidates, respectively.

First claim

What is claimed is:

1 . A computer-implemented method, comprising:
obtaining, by one or more processors, a set of medical records from a set of providers; querying, by the one or more processors, the set of medical records obtained from the set of providers using one or more selection criteria to determine a patient of the set of patients has one or more characteristics that match one or more first predetermined clinical trial selection criteria, wherein the one or more selection criteria includes demographic information, one or more diagnosis codes, laboratory test values, medication names, scores for cognitive tests, medical professional observations, acute condition names, chronic condition names, or allergy names; receiving, by the one or more processors, a medical record of the set of medical records that is associated with the patient, the medical record including a first subset of data and a second subset of data, the first subset of data including structured data and the second subset of data including unstructured text; upon detecting, by the one or more processors, that the medical record includes the unstructured text:
performing, by the one or more processors and via a set of featurization layers of a neural network, feature extraction on the first subset of data to generate a first vector representation of the first subset of data to reduce the dimensionality of the first subset of data, wherein performing the feature extraction further comprises:
performing an encoding technique and an embedding technique to generate, based at least in part on the first subset of data, a categorical value information input vector and a sequence of categorical value information input vector, and
concatenating the first subset of data with the categorical value information input vector and the sequence of categorical value information input vector to generate the first vector representation;

generating, by the one or more processors and via a set of classification layers of the neural network and based at least in part on the first vector representation, a first output by correlating the first vector representation of the first subset of data with second predetermined clinical trial selection criteria, the set of classification layers being trained based at least in part on the second predetermined clinical trial selection criteria;
generating, by the one or more processors and via a content similarity engine performing a natural language processing technique, a second output by comparing a second vector representation of the second subset of data with the second predetermined clinical trial selection criteria including applying a natural language processing technique to the second subset of data to generate segments containing sequences of words, weighting the words based on importance of the words with respect to a respective segment and importance of the words with respect to all of the segments using term frequency-inverse document frequency (td-idf), and performing a weighted comparison of the second vector representation of the second subset of data with the second predetermined clinical trial selection criteria;
merging, by the one or more processors, the first output and the second output into a third output;
determining, by the one or more processors, that the patient matches the second predetermined clinical trial selection criteria based on the third output satisfying a predetermined threshold;
transmitting, by the one or more processors and via a communication network, an identity of the patient to a provider of the set of providers; and
discarding, by the one or more processors, the set of medical records so that the set of medical records is not used in determining patient candidacy in subsequent clinical trials.

2 . The computer-implemented method of claim 1 , further comprising:
identifying, by the one or more processors and using patient claim information associated with the set of providers, the provider of the set of providers having a number of patients that include one or more characteristics that satisfy provider selection criteria and that exceeds a patient number threshold; wherein the patient is one of the number of patients of the provider of the set of providers.

3 . The computer-implemented method of claim 2 , wherein identifying the provider of the set of providers comprises:
querying, by the one or more processors, the patient claim information of each of the set of providers using demographic information, one or more diagnosis codes, or pharmacy information to identify ones of the set of providers having patients with the one or more characteristics that match the one or more first predetermined clinical trial selection criteria.

4 . The computer-implemented method of claim 2 , wherein identifying the provider of the set of providers comprises:
querying, by the one or more processors, the patient claim information of each of the set of providers using de-identified patient information to identify ones of the set of providers having patients with the one or more characteristics that match the one or more first predetermined clinical trial selection criteria.

5 . The computer-implemented method of claim 1 , wherein transmitting the identity of the patient comprises:
transmitting, by the one or more processors and via the communication network, the identity of the patient to the provider of the set of providers via a networked results portal accessible by the provider.

6 . The computer-implemented method of claim 1 , further comprising:
receiving, by the one or more processors, a communication from the provider of the set of providers opting in to participating in the clinical trial.

7 . The computer-implemented method of claim 1 , wherein the second predetermined clinical trial selection criteria comprise demographic information, one or more diagnosis codes, laboratory test values, medication names, scores for cognitive tests, medical professional observations, acute condition names, chronic condition names, or allergy names.

8 . A system, comprising:
one or more processors; and one or more memories storing processor-executable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising: obtaining a set of medical records from a set of providers; querying the set of medical records obtained from the set of providers using one or more selection criteria to determine a patient of the set of patients has one or more characteristics that match one or more first predetermined clinical trial selection criteria, wherein the one or more selection criteria includes demographic information, one or more diagnosis codes, laboratory test values, medication names, scores for cognitive tests, medical professional observations, acute condition names, chronic condition names, or allergy names; receiving a medical record of the set of medical records that is associated with the patient, the medical record including a first subset of data and a second subset of data, the first subset of data including structured data and the second subset of data including unstructured text;
upon detecting that the medical record includes the unstructured text:
performing, via a set of featurization layers of a neural network, feature extraction on the first subset of data to generate a first vector representation of the first subset of data to reduce the dimensionality of the first subset of data wherein performing the feature extraction further comprises:
performing an encoding technique and an embedding technique to generate, based at least in part on the first subset of data, a categorical value information input vector and a sequence of categorical value i

Assignees

• Change Healthcare Holdings Llc

Inventors

• Morris V Scott
• Mushtaq Faisal
• Mondal Sanjoy
• Fidow Mark

Classifications

• G06N3/04

  Architecture, e.g. interconnection topology · CPC title

• G16H70/60

  relating to pathologies · CPC title

• G16H70/40

  relating to drugs, e.g. their side effects or intended usage · CPC title

• G16H10/40

  for data related to laboratory analysis, e.g. patient specimen analysis · CPC title

• G16H70/20

  relating to practices or guidelines · CPC title