Patent Issued for Disease development risk prediction system, disease development risk prediction method, and disease development risk prediction program (USPTO 11437146): NEC Corporation

2022 SEP 27 (NewsRx) -- By a News Reporter-Staff News Editor at Insurance Daily News -- NEC Corporation (Tokyo, Japan) has been issued patent number 11437146, according to news reporting originating out of Alexandria, Virginia, by NewsRx editors.

The patent’s inventor is Fukunishi, Hiroaki (Tokyo, Japan).

This patent was filed on August 9, 2017 and was published online on September 6, 2022.

From the background information supplied by the inventors, news correspondents obtained the following quote: “BACKGROUND ART

“In the national health insurance program operated by local governments and the health insurance program operated by health insurance societies established by corporations, the formulation of measures for reducing medical care costs on insured persons using the programs is promoted. When formulating such measures, for example, the health conditions of insured persons are analyzed using health care data such as receipts.

“Health conditions are analyzed, for example, by artificial intelligence (AI). If future deterioration in the health conditions of insured persons is predicted, local governments or health insurance societies can plan to take measures such as activities to prevent deterioration in the health conditions of insured persons.

“One cause of deterioration in health conditions is population aging. An increasing number of patients due to population aging is regarded as a major social problem currently in Japan. One reason why an increasing number of patients is regarded as a major social problem is that patient care requires enormous costs.

“Costs for patient care include medical care cost, nursing care cost, and informal care cost, and the like. That is, with an increase in the number of patients, social costs borne by the national and local governments.

“Informal care means not support provided on the basis of formal programs by local governments, specialized agencies, etc., but unofficial support provided by families, friends, local residents, volunteers, and the like and not on the basis of formal programs. Informal care is also referred to as informal service.

“If the risk of insured persons developing a predetermined disease in the future can be predicted early, each local government or each health insurance society may be able to suppress the development of the predetermined disease by taking preventive measures. As a result of suppressing the development of the predetermined disease by insured persons, the foregoing social costs can be reduced.

“Patent Literature (PTL) 1 and PTL 2 describe techniques of predicting, for example, the risk of developing a predetermined disease. For example, PTL 1 describes a method of predicting the risk of developing Alzheimer’s disease.

“The method described in PTL 1 predicts the risk of developing Alzheimer’s disease by determining human lipocalin type prostaglandin D synthetase (b-trace) losing chaperone activity of amyloid b peptide existing in biological fluid collected from a human. Alternatively, the method described in PTL 1 predicts the risk of developing Alzheimer’s disease by measuring chaperone activity of amyloid b peptide in a biological fluid collected from a human.

“PTL 2 describes a method of, for example in treatment of primary breast cancer, predicting axillary lymph node (AxLN) metastasis (AxLN metastasis) using a prediction model formed by an alternative decision tree (AD tree). For example, a learning device for learning the prediction model described in PTL 2 uses clinical data obtained backward by tracing back to the past, as training data.

“PTL 3 describes a medical data analysis system for predicting medical care cost reduction effect by health guidance, by generating and visualizing, on the basis of medical check-up information and receipt information, a graphical model having each item of the medical check-up information and the receipt information as a random variable.”

Supplementing the background information on this patent, NewsRx reporters also obtained the inventor’s summary information for this patent: “Technical Problem

“When performing the method described in PTL 1, special testing is required in order to predict the risk of developing Alzheimer’s disease, as mentioned above. That is, the method described in PTL 1 does not assume predicting the risk of developing Alzheimer’s disease using existing information which is available without special testing.

“The learning device described in PTL 2 uses, as training data, clinical data which is existing information, as mentioned above. However, the learning device described in PTL 2 does not assume using data other than clinical data as training data.

“The medical data analysis system described in PTL 3 generates and visualizes a graphical model by combining a plurality of sets of data from different sources, as mentioned above. However, the medical data analysis system described in PTL 3 does not assume concealing information for identifying individual persons when combining a plurality of sets of data.

“Object of Invention”

The claims supplied by the inventors are:

“1. A disease development risk prediction system comprising: a processor; a memory storing instructions executable by the processor to: generate combination data by combining at least two different types of receipt data using a combination key, wherein the receipt data includes an insured person number for an insured person which was converted using a predetermined method and birth year of the insured person, and wherein the combination key combines the converted insured person number and the birth year; generate a machine learning prediction model predicting a risk of the insured person of developing a predetermined disease using generated combination data; and apply the machine learning prediction model to the generated combination data for the insured person to predict the risk of the insured person of developing the predetermined disease, wherein the receipt data is any of: medical receipt data indicating a receipt for a medical act; and dispensing receipt data indicating a receipt for a dispensing act, the processor generates the combination data using at least the medical receipt data and the dispensing receipt data, the processor excludes, from the generated combination data, data of the insured person who developed the predetermined disease before or in a predetermined year, the processor generates the machine learning prediction model for predicting the risk of the insured person developing the predetermined disease for a first time, using the combination data from which the data of the insured person has been excluded, the processor adds, to the generated combination data, an attribute indicating whether the insured person developed the predetermined disease in or after a year following the predetermined year, the processor generates the machine learning prediction model, using the added attribute as an objective variable and information from the predetermined year backward included in the combination data as an explanatory variable, and the processor predicts the risk of the insured person developing the predetermined disease for the first time, using the generated machine learning prediction model.

“2. The disease development risk prediction system according to claim 1, wherein the receipt data includes the birth year and month of the insured person, and wherein the processor combines the at least two different types of receipt data using the combination key, and wherein the combination key combines the converted insured person number and the birth year and the month.

“3. The disease development risk prediction system according to claim wherein the receipt data includes nursing care insurance data indicating a receipt for a nursing care service.

“4. The disease development risk prediction system according to claim 2, wherein the receipt data includes gender of the insured person, and wherein the processor combines the at least two different types of receipt data using the combination key, and wherein the combination key combines the converted insured person number, the birth year and the month, and the gender.

“5. The disease development risk prediction system according to claim 1, wherein the processor combines the at least two different types of receipt data using the combination key including age of the insured person.

“6. The disease development risk prediction system according to claim 1, wherein the processor combines the at least two different types of receipt data using the combination key including the insured person number subjected to hashing.

“7. The disease development risk prediction system according to claim 1, wherein the processor combines the at least two different types of receipt data using the combination key including the insured person number subjected to encryption.

“8. The disease development risk prediction system according to claim 1, wherein the processor determines the insured person for which a number of times an injury/disease code corresponding to the predetermined disease is included in the medical receipt data from the predetermined year backward is more than or equal to a designated number, as the insured person who developed the predetermined disease before or in the predetermined year.

“9. The disease development risk prediction system according to claim 1, wherein the processor uses a code of middle classification corresponding to an ICD-10 code included in the medical receipt data, as the explanatory variable.

“10. The disease development risk prediction system according to claim 1, wherein the processor uses a number representing drug efficacy in a national health insurance drug list included in the dispensing receipt data, as the explanatory variable.

“11. The disease development risk prediction system according to claim 1, wherein the instructions are executable by the processor to further: predict the insured person having a potential to become a patient of the predetermined disease, using the generated machine learning prediction model.

“12. A disease development risk prediction method comprising: generating combination data by combining at least two different types of receipt data using a combination key, wherein the receipt data includes an insured person number for an insured person which was converted using a predetermined method and birth year of the insured person, and wherein the combination key combines the converted insured person number and the birth year; generating a machine learning prediction model predicting a risk of the insured person of developing a predetermined disease using generated combination data; applying the machine learning prediction model to the generated combination data for the insured person to predict the risk of the insured person of developing the predetermined disease, wherein the receipt data is any of: medical receipt data indicating a receipt for a medical act; and dispensing receipt data indicating a receipt for a dispensing act, further comprising: generating the combination data using at least the medical receipt data and the dispensing receipt data; excluding, from the generated combination data, data of the insured person who developed the predetermined disease before or in a predetermined year; generating the machine learning prediction model for predicting the risk of the insured person developing the predetermined disease for a first time, using the combination data from which the data of the insured person has been excluded; adding, to the generated combination data, an attribute indicating whether the insured person developed the predetermined disease in or after a year following the predetermined year; generating the machine learning prediction model, using the added attribute as an objective variable and information from the predetermined year backward included in the combination data as an explanatory variable; and predicting the risk of the insured person developing the predetermined disease for the first time, using the generated machine learning prediction model.

“13. The disease development risk prediction method according to claim 12, wherein the receipt data includes the birth year and month of the insured person, and wherein the at least two different types of receipt data are combined using the combination key, and wherein the combination key combines the converted insured person number and the birth year and the month.

“14. A non-transitory computer-readable capturing medium having captured therein a disease development risk prediction program for causing a computer to execute: a first generation process of generating combination data by combining at least two different types of receipt data using a combination key, wherein the receipt data includes an insured person number for an insured person which was converted using a predetermined method and birth year of the insured person, and wherein the combination key combines the converted insured person number and the birth year; a second generation process of generating a machine learning prediction model predicting a risk of the insured person of developing a predetermined disease using generated combination data; an application process of applying the machine learning prediction model to the generated combination data for the insured person to predict the risk of the insured person of developing the predetermined disease, wherein the receipt data is any of: medical receipt data indicating a receipt for a medical act; and dispensing receipt data indicating a receipt for a dispensing act, the computer is caused, in the first generation process, to generate the combination data using at least the medical receipt data and the dispensing receipt data, exclude, from the generated combination data, data of the insured person who developed the predetermined disease before or in a predetermined year, and add, to the generated combination data, an attribute indicating whether the insured person developed the predetermined disease in or after a year following the predetermined year, the computer is caused, in the second generation process, to generate the machine learning prediction model for predicting the risk of the insured person developing the predetermined disease for a first time, using the combination data from which the data of the insured person has been excluded, the added attribute as an objective variable and information from the predetermined year backward included in the combination data as an explanatory variable, and the computer is caused, in the application process, to predict the risk of the insured person developing the predetermined disease for the first time, using the generated machine learning prediction model.”

There are additional claims. Please visit full patent to read further.

For the URL and additional information on this patent, see: Fukunishi, Hiroaki. Disease development risk prediction system, disease development risk prediction method, and disease development risk prediction program. U.S. Patent Number 11437146, filed August 9, 2017, and published online on September 6, 2022. Patent URL: http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PALL&p=1&u=%2Fnetahtml%2FPTO%2Fsrchnum.htm&r=1&f=G&l=50&s1=11437146.PN.&OS=PN/11437146RS=PN/11437146

(Our reports deliver fact-based news of research and discoveries from around the world.)