Patent Issued for System for predicting patient health conditions (USPTO 11621081): IQVIA Inc.
2023 APR 21 (NewsRx) -- By a
Patent number 11621081 is assigned to
The following quote was obtained by the news editors from the background information supplied by the inventors: “As part of the healthcare process, physicians or other medical care providers may perform clinical trials, programs, and other activities to evaluate subject safety and efficacy of a pharmaceutical drug or other medical treatment option. The use of health-related trial programs can help to identify novel treatment options for improving overall patient health and reducing health system costs. A clinical trial or program can be a single research study or multiple research studies that prospectively assigns human participants/subjects or groups of human subjects to one or more health-related interventions to evaluate the effects on health outcomes.”
In addition to the background information obtained for this patent, NewsRx journalists also obtained the inventors’ summary information for this patent: “As part of the healthcare process, physicians or other medical care providers may perform trials, programs, and other activities to evaluate the efficacy of a particular pharmaceutical drug or other medical treatment option. Conducting health-related clinical trials can help to identify medical treatment options for improving overall patient health and reducing health system costs. Clinical trials and other controlled programs are generally conducted by investigators that use a particular geographic site location(s) to interact with study subjects. In some instances, a physician for a patient can be associated with a clinical trial and the physician can refer a patient as a candidate for participation in a trial based on a diagnosed condition of the patient. An investigator, a geographic site location, or both, can form an entity that executes a program. The effectiveness of a trial program can depend on a variety of factors, such as obtaining a sufficient number of subjects that are suitable for participation in a trial, the accuracy of diagnosed conditions for each patient or subject in the program, or certain types of conditions that may prospectively affect a subject. In some cases, factors that impact the effectiveness of a clinical trial can vary depending on the treatment options being evaluated and the criteria that are associated with the trial.
“Based on the above context, this document describes a predictive computing system that uses specific computing rules or instructions (e.g., a unique algorithm) to predict or determine at least one condition that relates to the healthcare status of a patient (e.g., an undiagnosed medical condition). The condition can be a disease or medical ailment that presently affects the patient or that will prospectively affect the patient. The condition can indicate early detection of a disease or provide a basis for predicting progression of a current or future disease. In some instances, the predictive system may be configured to identify a misdiagnosis and determine a correct diagnosis with reference to a misdiagnosed condition. The predictive system trains an initial prediction model by processing sets of input data through layers of at least one neural network included at the system. The input data is uniquely structured in a sequenced format so as to enhance a feature engineering process of the predictive model. The specific instructions are learned by the predictive computing system when the neural networks process the inputs to generate the trained prediction model. The instructions can be adapted and refined for use at the trained model when new sets of input data are processed using the neural networks.
“The system can identify past healthcare providers that treated a particular set of patients (e.g., diagnosed or at-risk patients). Referencing these past providers, the system uses the specific computing rules to more efficiently determine which provider(s) to target for treating a particular predicted condition of the patient. For example, the predictive system makes this determination in response to receiving query inputs that may specify certain diagnosis codes and other relevant disease risk factors associated with the patient. The system’s trained prediction model processes the input data in accordance with the specific computing rules to generate a prediction result. The result can include an identified listing of clinical trials and site locations that execute trial programs evaluating medical options for treating the patient’s predicted condition. Further, the result can also identify one or more providers that can be targeted for participation in any relevant clinical trials as well as providers (e.g., referral physicians) that are likely to refer patients to other physicians/investigators that are involved in a clinical trial. The providers/referral physicians may be targeted as having relevance to a clinical trial based on the healthcare or medical attributes of the patients that are treated by the providers.
“One aspect of the subject matter described in this specification can be embodied in a computer-implemented method performed using a neural network comprising multiple neural network layers. The method includes: obtaining data for a set of patients that each have a particular healthcare condition; determining a first sequence of data based on the data for the set of patients, where the first sequence of data represents training data relating to healthcare transactions of the set of patients; and determining a second sequence of data based on the data for the set of patients, wherein the second sequence of data represents training data relating to healthcare providers of the set of patients.
“The method further includes, generating a scoring model at least by processing the first and second sequence of data through one or more of the layers of the neural network to train the neural network, where the scoring model is configured to determine a confidence that an individual has the particular healthcare condition; and obtaining patient scoring data for determining the confidence that the individual has the particular healthcare condition using the scoring model, wherein the patient scoring data is not included in the data for the set of patients.”
The claims supplied by the inventors are:
“1. A computer-implemented method comprising: generating a training dataset comprising a plurality of first distinct data sequences derived from information about healthcare transactions and providers for a set of patients; determining a first set of features from the plurality of first distinct data sequences in the training dataset, wherein the first set of features is determined based on commonalities among parameters in the plurality of first distinct data sequences from which the first set of features is determined; generating a predictive model by: training a first neural network to generate output vector data sets in response to processing inputs corresponding to the first set of features through a plurality of layers of the first neural network; training a second neural network to generate an indication of undiagnosed conditions of the set of patients (i) using the output vector data sets and (ii) by processing inputs corresponding to the first set of features, wherein the trained second neural network represents the predictive model; determining a second set of features from at least some of the plurality of first distinct data sequences for scoring the predictive model; in response to determining the second set of features, generating a first confidence score by providing the second set of features as input to the predictive model, wherein the first confidence score represents an accuracy of the predictive model to predict future undiagnosed health conditions of patients; comparing the first confidence score to a threshold value to evaluate the accuracy of the predictive model; in response to determining the first confidence score satisfies the threshold value, providing the predictive model for output; generating, by the predictive model, a prediction of a future undiagnosed health condition of a patient based on inferences determined from the predictive model and a third set of features determined from a plurality of second distinct data sequences, wherein the plurality of first distinct data sequences are different from the plurality of second distinct data sequences; generating, from the predictive model, an output that comprises the prediction of the future undiagnosed health condition of the patient; in response to generating the output, identifying one or more clinical trials and corresponding site locations that execute trial programs for evaluating medical options for treating the predicted future undiagnosed health condition of the patient; generating data that links medical information of the patient to information of the one or more identified clinical trials; and providing the linked data for output.
“2. The computer-implemented method of claim 1, wherein generating the prediction of the future undiagnosed health condition of the patient further comprises: determining, by the predictive model, a second confidence score that represents a confidence in an accuracy of the future undiagnosed health condition of the patient in response to processing the third set of features corresponding to the patient through one or more layers of the predictive model.
“3. The computer-implemented method of claim 2, wherein generating the predictive model comprises: providing a first data sequence and a second data sequence of the plurality of first distinct data sequences as inputs to a first layer of the first neural network, wherein the first data sequence is different from the second data sequence and each of the first data sequence and the second data sequence correspond to a respective set of features; and training the second neural network using the sets of features in response to processing the first data sequence and the second data sequence through the first neural network.
“4. The computer-implemented method of claim 3, wherein the predictive model determines the second confidence score that represents the confidence in the accuracy of the future undiagnosed health condition of the patient based on training inferences computed when the predictive model is trained using at least the first data sequence and the second data sequence of the plurality of first distinct data sequences.
“5. The computer-implemented method of claim 4, further comprising: determining a third data sequence based on the information about the healthcare transactions and the providers for the set of patients, wherein the third data sequence represents training data that describes patient specific information; and generating the predictive model at least by processing each of the first, second, and third data sequence of the plurality of first distinct data sequences through the one or more layers of the first neural network.
“6. The computer-implemented method of claim 5, wherein: the first data sequence comprises data describing procedures, products, or diagnoses and is formatted for processing through the one or more layers of the first neural network; the second data sequence comprises data describing healthcare provider specialties and is formatted for processing through the one or more layers of the first neural network; and the third data sequence comprises data describing patient demographics and is formatted for processing through the one or more layers of the first neural network.
“7. The computer-implemented method of claim 6, further comprising: encoding, as a vector, at least one of a procedure, a product, and a diagnosis in the first data sequence; and obtaining an encoding rule associated with the vector by using the second neural network to learn inferences based on analysis of data corresponding to the training dataset.
“8. The computer-implemented method of claim 1, wherein the predictive model is a trained neural network model configured to identify latent variables in patient scoring data derived from the information about the healthcare transactions and the providers for the set of patients.
“9. The computer-implemented method of claim 1, wherein: the future undiagnosed health condition of the patient reflects at least one of (i) a current healthcare condition or (ii) a prospective healthcare condition, that is not yet diagnosed for the patient; and the output: (i) identifies a particular disease that currently affects the patient, or (ii) identifies whether the patient is at risk for contracting a particular disease.
“10. The computer-implemented method of claim 9, wherein the method further comprises: identifying a healthcare provider for treating the particular disease; linking first data for the patient to second data for the healthcare provider; and providing the indication using linked data for the patient and the healthcare provider.
“11. The computer-implemented method of claim 9, wherein the method further comprises: identifying a healthcare provider for treating the particular disease.
“12. The computer-implemented method of claim 1, wherein generating the predictive model further comprises: generating, iteratively, feedback data that comprises at least some of the output vector data sets output by the first neural network; and providing the generated feedback data to be used as the training dataset for generating the predictive model.
“13. The computer-implemented method of claim 1, wherein generating the training dataset comprises: generating the training dataset from data for a subset of patients that each have the same healthcare condition.
“14. The computer-implemented method of claim 1, wherein the first and second neural networks are implemented on a hardware circuit of a predictive system that includes the predictive model and the computer-implemented method further comprises: generating, at the predictive system, the plurality of first distinct data sequences by sequencing a portion of the information from the training dataset for each of the healthcare transactions and the providers for the set of patients; generating a plurality of vector datasets from the plurality of first distinct data sequences that describes the set of patients, wherein (i) a first vector dataset describes healthcare transactions for the set of patients, (ii) a second vector dataset describes healthcare providers that engage in the healthcare transactions for treating the set of patients, and (iii) a third vector dataset describes demographics for each patient in the set of patients; and generating, by the predictive system, a set of computing rules based on the plurality of vector datasets implemented on the hardware circuit.”
There are additional claims. Please visit full patent to read further.
URL and more information on this patent, see: Alamuri, Chaitanya. System for predicting patient health conditions.
(Our reports deliver fact-based news of research and discoveries from around the world.)
Researchers from Southern Cross University Report on Findings in Life Science (A Discourse Analysis of the Representation of Nursing In the National Disability Insurance Scheme Pricing Guide and Eligibility Criteria): Life Science
Patent Issued for Delivering digital identity documents that are vetted as trusted in a secured manner (USPTO 11621955): International Business Machines Corporation
Advisor News
- Goldman Sachs, others weigh in on what’s ahead for the market in 2025
- How is consumers’ investing confidence impacting the need for advisors?
- Can a balance of withdrawals and annuities outshine the 4% rule?
- Economy ‘in a sweet spot’ but some concerns ahead
- Trump, Capitol allies likely looking at quick cryptocurrency legislation
More Advisor NewsAnnuity News
Health/Employee Benefits News
- Study Findings from Brown University Broaden Understanding of COVID-19 (Continuity of Care and Lifestyle Intervention Programs for Spanish-speaking Immigrants Without Health Insurance At a Free Clinic In Rhode Island): Coronavirus – COVID-19
- Hospital payment caps could save millions of dollars for state employee health plans: Brown University
- Burlington County defies trends by keeping health insurance rates stable
- Inside the ‘big deal’ battle for Georgia’s Medicaid contract
- Medicaid coverage to end for millions if Trump cuts funding
More Health/Employee Benefits NewsLife Insurance News