Pricing method of special insurance with disease body and related product

Abstract

The invention discloses a pricing method for a special insurance with a disease body and a related product, and can be applied to the technical field of insurance pricing, and the method comprises the steps: obtaining hospital medical data of a thyroid nodule patient; the hospital medical data comprises thyroid function examination result data, thyroid ultrasound image feature data and patient clinical feature data; predicting the cancerization risk probability of the thyroid nodule patient by using a risk probability prediction model in combination with the hospital medical data, and determining a risk level; based on the canceration risk probability and the risk level, determining the treatment cost of the thyroid nodule patient by using a medical cost estimation model; and pricing the special insurance with the disease body in combination with the canceration risk probability, the danger level and the treatment cost. Therefore, a more refined cancerization risk probability is generated through the risk probability prediction model, insurance pricing is carried out, and the pricing accuracy is improved.

Description

Technical Field

[0001] The present application relates to the technical field of insurance pricing, and in particular to a pricing method for special insurance for sick persons and related products. Background Art

[0002] With the increase of sub-healthy people, the market demand for insurance with illness is increasing. Currently, most of the insurance with illness on the market is in the form of million-dollar medical insurance or serious illness insurance, while special insurance with illness for specific diseases is less in variety due to pricing difficulties.

[0003] The pricing design of existing special insurance products for people with diseases is relatively rough. Most products still rely on expert experience and past medical research data for design and lack precise data reliance, which leads to inaccurate pricing of special insurance products for people with diseases.

[0004] Therefore, how to improve the pricing accuracy of special insurance for patients with diseases is an issue that technical personnel in this field urgently need to solve. Summary of the invention

[0005] Based on the above problems, the present application provides a pricing method and related products for special insurance for sick bodies, which generates a more refined cancer risk probability through a risk probability prediction model, and then conducts insurance pricing, thereby improving the accuracy of pricing.

[0006] In a first aspect, the present application provides a pricing method for special insurance for sick persons, characterized in that the method comprises:

[0007] Obtain hospital medical data of patients with thyroid nodules; the hospital medical data includes thyroid function test result data, thyroid ultrasound image feature data, and patient clinical feature data;

[0008] In combination with the hospital medical data, a risk probability prediction model is used to predict the cancer risk probability of the thyroid nodule patient and determine the risk level;

[0009] Based on the cancer risk probability and the risk level, determining the treatment cost of the thyroid nodule patient using a medical cost estimation model;

[0010] The special insurance for sick persons is priced based on the cancer risk probability, the risk level and the treatment costs.

[0011] Optionally, the risk probability prediction model is trained by the following method:

[0012] To obtain time series risk factor data for the evaluation of thyroid nodules;

[0013] Select key time points to analyze the time series risk factor data to determine the risk development trend information of thyroid nodules;

[0014] Training an LSTM model based on the risk development trend information and the canceration results, and obtaining an LSTM basic model;

[0015] Constructing the TIRADS basic model based on the time series risk factor data and thyroid nodule risk stratification criteria;

[0016] The LSTM basic model and the TIRADS basic model are integrated to obtain a risk probability prediction model.

[0017] Optionally, the obtaining of time series risk factor data for evaluating thyroid nodules includes:

[0018] Obtaining clinical research and medical guidelines for patients with thyroid nodules; the clinical research and medical guidelines contain medical data text corresponding to risk factors; the risk factors include clinical indicators and ultrasound features;

[0019] The medical data text is subjected to NLP and structured processing to obtain time series risk factor data for evaluating thyroid nodules.

[0020] Optionally, selecting key time points to analyze the time series risk factor data to determine the risk development trend information of thyroid nodules includes:

[0021] Select the time nodes before and after disease diagnosis, and the start and end of treatment to analyze the time series risk factor data, and determine the corresponding indicator data change trend; the change trend includes continuous rise, fall or fluctuation;

[0022] The risk development trend information of thyroid nodules is determined based on the changing trend of the indicator data.

[0023] Optionally, constructing a TIRADS basic model based on the time series risk factor data and thyroid nodule risk stratification standards includes:

[0024] Access to thyroid nodule risk stratification criteria from at least one country's thyroid imaging reporting and data system;

[0025] Combined with the time series risk factor data, TIRADS basic models corresponding to each stratification standard are constructed respectively.

[0026] Optionally, the integrating the LSTM basic model and the TIRADS basic model to obtain a risk probability prediction model includes:

[0027] In combination with the time series risk factor data, the prediction results of the LSTM basic model and the TIRADS basic model corresponding to each stratification standard are combined to form a training data set;

[0028] The meta-model is trained based on the training data set to obtain a risk probability prediction model.

[0029] Optionally, the determining the treatment cost of the thyroid nodule patient using a medical cost estimation model based on the cancer risk probability and the risk level includes:

[0030] Determining drugs and devices for treating the patient with thyroid nodules based on the cancer risk probability and the risk level;

[0031] Combined with the medicines and medical devices, and in association with the medical expense details and medical insurance settlement data, the medical expense estimation model is used to determine the amount and proportion of each expense; the amount of each expense includes the total settlement amount of the expense, the fund payment amount, the personal account payment amount, and the personal payment amount;

[0032] The treatment costs of the thyroid nodule patients are determined based on the classification of drugs in the local medical insurance catalog, the payment ratio of drugs in the local medical insurance catalog, and the amounts and proportions of the various expenses.

[0033] In a second aspect, the embodiment of the present application provides a pricing device for special insurance for sick persons, which is characterized by comprising:

[0034] An acquisition module is used to acquire hospital medical data of patients with thyroid nodules; the hospital medical data includes thyroid function test result data, thyroid ultrasound image feature data and patient clinical feature data;

[0035] A risk prediction module, used to combine the hospital medical data, use a risk probability prediction model to predict the cancer risk probability of the thyroid nodule patient, and determine the risk level;

[0036] A medical expense estimation module, used to determine the treatment cost of the thyroid nodule patient based on the cancer risk probability and the risk level using a medical expense estimation model;

[0037] The actuarial pricing module is used to price the special disease-bearing insurance based on the cancer risk probability, the risk level and the treatment cost.

[0038] In a third aspect, the embodiment of the present application provides a pricing device for special insurance for sick persons, which is characterized by comprising:

[0039] Memory for storing computer programs;

[0040] A processor is used to implement the steps of the pricing method for special insurance for sick persons as described above when executing the computer program.

[0041] In a fourth aspect, an embodiment of the present application provides a readable storage medium, characterized in that a computer program is stored on the readable storage medium, and when the computer program is executed by a processor, the steps of the pricing method for the special insurance for sick bodies as described above are implemented.

[0042] It can be seen from the above technical solutions that compared with the prior art, the present application has the following advantages:

[0043] This application first obtains the hospital medical data of patients with thyroid nodules. Among them, the hospital medical data includes thyroid function test result data, thyroid ultrasound image feature data and patient clinical feature data. Then, combined with the hospital medical data, the risk probability prediction model is used to predict the cancer risk probability of patients with thyroid nodules and determine the risk level. Based on the cancer risk probability and risk level, the medical cost estimation model is used to determine the treatment costs of patients with thyroid nodules. Finally, the special insurance for sick people is priced based on the cancer risk probability, risk level and treatment costs. In this way, a more refined cancer risk probability is generated through the risk probability prediction model, and then insurance pricing is performed, which improves the accuracy of pricing. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A flowchart of a pricing method for a special insurance policy for a sick person provided in an embodiment of the present application;

[0045] Figure 2 A flowchart of inter-module collaborative operation provided in an embodiment of the present application;

[0046] Figure 3 A flowchart for constructing a risk probability prediction module provided in an embodiment of the present application;

[0047] Figure 4 A schematic diagram of the structure of a pricing device for special insurance for sick persons provided in an embodiment of the present application. DETAILED DESCRIPTION

[0048] As mentioned above, the existing pricing of special insurance for patients with diseases is inaccurate. Specifically, the pricing of existing insurance products mostly relies on expert experience and past medical research, lacks accurate data reliance, and the target patients studied are insufficient in number, population characteristics, and diversity, which leads to inaccurate pricing of special insurance for patients with diseases.

[0049] In order to solve the above problems, the embodiment of the present application provides a pricing method for special insurance for sick people, including: first, obtaining the hospital medical data of patients with thyroid nodules. Among them, the hospital medical data includes thyroid function test result data, thyroid ultrasound image feature data and patient clinical feature data. Then, combined with the hospital medical data, the risk probability prediction model is used to predict the cancer risk probability of patients with thyroid nodules and determine the risk level. Based on the cancer risk probability and the risk level, the medical cost estimation model is used to determine the treatment cost of patients with thyroid nodules. Finally, the special insurance for sick people is priced based on the cancer risk probability, risk level and treatment cost.

[0050] In this way, a more refined cancer risk probability is generated through the risk probability prediction model, which is then used for insurance pricing, thereby improving pricing accuracy.

[0051] It should be noted that the pricing method and related products of a special insurance for a sick body provided by this application can be applied to the field of insurance pricing technology. The above is only an example and does not limit the application field of the pricing method and related products of a special insurance for a sick body provided by this application.

[0052] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0053] Figure 1 A flowchart of a pricing method for a special disease insurance provided in an embodiment of the present application. Figure 1 As shown, a pricing method for special insurance for sick persons provided in an embodiment of the present application may include:

[0054] S101: Obtain hospital medical data of patients with thyroid nodules; the hospital medical data includes thyroid function test result data, thyroid ultrasound image feature data, and patient clinical feature data.

[0055] In practical applications, the price of insurance for patients with thyroid nodules of different conditions should be different. For example, a mild condition (such as a nodule below 0.5 cm) may only need to take some Chinese herbal medicine or Chinese patent medicine to obtain a certain effect, and the price of insurance may be lower; while for more serious conditions (such as nodules above 1.5 cm), general medicines may not be effective, and more expensive drugs or even medical devices are required for auxiliary periodic treatment, so the price of insurance for such patients needs to be higher. In this way, in order to achieve reasonable pricing, the embodiment of the present application provides a pricing method for special insurance with disease, which can be priced based on the patient's condition. Specifically, it is first necessary to obtain hospital medical data including thyroid function test results (thyroid stimulating hormone TSH, free triiodothyronine FT3 and free thyroxine FT4, etc.), time series data of thyroid ultrasound imaging characteristics (position, morphology, edge, echo, structure, size, etc.) and patient clinical characteristics data (age, height, weight, BMI, family history, radiation exposure history, etc.).

[0056] S102: In combination with the hospital medical data, a risk probability prediction model is used to predict the cancer risk probability of the thyroid nodule patient and determine the risk level.

[0057] In practical applications, the trained risk probability prediction model integrates the cancer risk prediction model and the hazard level prediction model, and has higher model robustness and more accurate cancer risk probability prediction. The hospital medical data of patients with thyroid nodules is used as input and input into the trained risk probability prediction model to obtain the cancer risk probability and hazard level of the thyroid nodule patient. For example, if the indicator data of a patient A continues to rise over a period of time, the model will predict that the cancer risk probability is high, but if the specific value has been in the specified second hazard level (assuming the second level is set to 5-20%), then combined, the model may eventually output a relatively large and accurate number of 18%, 19% or 20%.

[0058] S103: Based on the cancer risk probability and the risk level, determine the treatment cost of the thyroid nodule patient using a medical cost estimation model.

[0059] In actual applications, there are different treatment methods for patients with different risk levels and whether the patients have cancer or not. Therefore, it is necessary to comprehensively consider the cancer risk probability and risk level corresponding to the above-mentioned thyroid nodule patients, and use the medical cost estimation model to determine the treatment cost of the thyroid nodule patients.

[0060] In addition, since the methods for determining treatment costs are different, the embodiments of the present application can illustrate one possible determination method.

[0061] In one case, S103: based on the cancer risk probability and the risk level, using a medical cost estimation model to determine the treatment cost of the thyroid nodule patient, specifically for:

[0062] Determining drugs and devices for treating the patient with thyroid nodules based on the cancer risk probability and the risk level;

[0063] Combined with the medicines and medical devices, and in association with the medical expense details and medical insurance settlement data, the medical expense estimation model is used to determine the amount and proportion of each expense; the amount of each expense includes the total settlement amount of the expense, the fund payment amount, the personal account payment amount, and the personal payment amount;

[0064] The treatment costs of the thyroid nodule patients are determined based on the classification of drugs in the local medical insurance catalog, the payment ratio of drugs in the local medical insurance catalog, and the amounts and proportions of the various expenses.

[0065] In practical applications, the physiotherapy cost estimation model needs to mine the drugs and medical devices (drugs and medical devices) used for treatment according to the cancer risk probability and risk level of thyroid patients, and at the same time associate the medical expense details and medical insurance settlement data to calculate the total settlement amount, fund payment amount, personal account payment amount, personal payment amount and proportion of the expenses. Then, based on the classification (Class A / Class B) and payment ratio of the local medical insurance catalog drugs, the total settlement amount, fund payment amount, personal account payment amount, personal payment amount and proportion of the above expenses are only used for verification and adjustment, and then the treatment cost for the patient with thyroid nodules is obtained.

[0066] S104: Pricing the special insurance for sick persons is determined based on the cancer risk probability, the risk level and the treatment cost.

[0067] Figure 2 A flowchart of the collaborative operation between modules provided in the embodiment of the present application. Figure 2 As shown, the design of thyroid nodule-carrying body insurance is mainly divided into three modules. The risk prediction module predicts the cancer risk probability and risk level of thyroid patients, the medical cost estimation module predicts the treatment cost of thyroid patients, and then the actuarial pricing module combines the cancer risk probability, risk level and treatment cost to price the special insurance for the sick body. Specifically, the insurance actuary can first determine the product form based on the model results of the risk probability prediction model and the pricing basis of the medical cost estimation model, so as to design the pricing of thyroid nodule-carrying body insurance products in a more refined way and store them. The pricing device can subsequently use the actuarial pricing module to combine the corresponding cancer risk probability, risk level and treatment cost of the thyroid nodule patient to provide the thyroid nodule patient with an accurately priced special insurance for the sick body.

[0068] In addition, since the methods of training risk probability prediction models are different, the embodiments of the present application can illustrate a possible training method.

[0069] In one embodiment, the risk probability prediction model is trained by the following method:

[0070] To obtain time series risk factor data for the evaluation of thyroid nodules;

[0071] Select key time points to analyze the time series risk factor data to determine the risk development trend information of thyroid nodules;

[0072] Training an LSTM model based on the risk development trend information and the canceration results, and obtaining an LSTM basic model;

[0073] Constructing the TIRADS basic model based on the time series risk factor data and thyroid nodule risk stratification criteria;

[0074] The LSTM basic model and the TIRADS basic model are integrated to obtain a risk probability prediction model.

[0075] In practical applications, unlike the traditional insurance pricing scheme that mainly relies on expert experience and diseased bodies, the embodiment of the present application takes into account the diversity of patient populations, the integrity of cohort data, and the richness of thyroid cancer risk factors, and especially uses the medical and expense settlement data of multiple hospitals to train the risk probability prediction model. Specifically, first obtain the time series risk factor data including clinical indicators and ultrasound characteristics for evaluating thyroid nodules. Then select key time points for analysis, identify potential risk change trends, and obtain the risk development trend information of thyroid nodules. Each set of time series risk factor data corresponds to a patient and corresponds to a cancer result (whether it exists). The LSTM model is trained using the risk development trend information and cancer results corresponding to a large number of patients to predict the risk probability of cancer of thyroid nodules, and the LSTM basic model is obtained. Then, based on the thyroid nodule risk stratification standards of the thyroid imaging report and data system (TIRADS) of various countries, the above-mentioned time series data is used to model the risk probability prediction for each stratification standard, that is, to construct the TIRADS basic model. Finally, the LSTM basic model and the TIRADS basic model are integrated and trained to obtain a trained risk probability prediction model.

[0076] In addition, since the methods for obtaining time series risk factor data for evaluating thyroid nodules are different, the embodiments of the present application can illustrate a possible acquisition method.

[0077] In one embodiment, the obtaining of time series risk factor data for evaluating thyroid nodules comprises:

[0078] Obtaining clinical research and medical guidelines for patients with thyroid nodules; the clinical research and medical guidelines contain medical data text corresponding to risk factors; the risk factors include clinical indicators and ultrasound features;

[0079] The medical data text is subjected to NLP and structured processing to obtain time series risk factor data for evaluating thyroid nodules.

[0080] In practical applications, we first obtain data on medical treatment and expense settlement for patients with thyroid nodules from multiple hospitals. Then, based on past clinical studies and authoritative medical guidelines, we perform NLP and structured processing on the relevant text of medical data containing key information, and extract time series risk factor data for evaluating thyroid nodules (such as age, gender, height, weight, BMI, diagnosis date, course of disease, thyroid pain, pressure, progressive enlargement symptoms, family history, radiation exposure history, and primary disease range data for patient characteristics; thyroid function data for thyroglobulin antibodies, thyroid peroxidase antibodies, thyroid stimulating hormone, free triiodothyronine, free thyroxine, triiodothyronine, and thyroxine; thyroid ultrasound data for location, orientation / morphology, edge, acoustic halo, structure, echo, echo texture, focal strong echo, posterior echo characteristics, and size data).

[0081] In addition, since the methods for determining thyroid nodule risk development trend information are different, the embodiments of the present application can illustrate a possible determination method.

[0082] In one embodiment, the step of selecting key time points to analyze the time series risk factor data and determining the risk development trend information of thyroid nodules includes:

[0083] Select the time nodes before and after disease diagnosis, and the start and end of treatment to analyze the time series risk factor data, and determine the corresponding indicator data change trend; the change trend includes continuous rise, fall or fluctuation;

[0084] The risk development trend information of thyroid nodules is determined based on the changing trend of the indicator data.

[0085] In practical applications, the above patient characteristics, seven thyroid function indicators, and thyroid ultrasound characteristics are processed to ensure that each data is generated in chronological order. After the biomarker data record is recorded, the time series data record is analyzed, and key nodes including the time nodes before disease diagnosis, the start and end of treatment are selected to identify the data changes of the indicators and obtain the change trend of the indicator data at key time points. The change trend includes continuous rise, fall, or fluctuation, and then the development trend of the patient and the nodule of the installation line is judged, and the risk development trend information data is processed.

[0086] In addition, since there are different ways to construct the TIRADS basic model, the embodiments of the present application can illustrate a possible construction method.

[0087] In one embodiment, the TIRADS basic model is constructed based on the time series risk factor data and the thyroid nodule risk stratification standard, including:

[0088] Access to thyroid nodule risk stratification criteria from at least one country's thyroid imaging reporting and data system;

[0089] Combined with the time series risk factor data, TIRADS basic models corresponding to each stratification standard are constructed respectively.

[0090] In practical applications, in order to better apply ultrasound examination to the diagnosis of thyroid nodules, many countries and societies have successively issued various standard thyroid imaging report and data systems (TIRADS), including K-TIRADS issued by the Korean Radiological Society, ACR-TIRADS formulated by the American Radiological Association, EU-TIRADS issued by the European Thyroid Association, and C-TIRADS formulated by the Chinese Medical Association, etc. Since each system has its own advantages and disadvantages, in order to make good use of hospital data and at the same time give play to the advantages of each system stratification, the embodiment of this application is based on the thyroid nodule risk stratification standard of the thyroid imaging report and data system (TIRADS) of various countries, and uses the above-mentioned processed and processed time series risk factor data to model the risk probability prediction for each layer of stratification standard, and obtain the corresponding TIRADS basic model.

[0091] In addition, since there are different ways to obtain the risk probability model, the embodiment of the present application can illustrate one possible way to obtain it.

[0092] In one embodiment, the integration of the LSTM basic model and the TIRADS basic model to obtain a risk probability prediction model includes:

[0093] In combination with the time series risk factor data, the prediction results of the LSTM basic model and the TIRADS basic model corresponding to each stratification standard are combined to form a training data set;

[0094] The meta-model is trained based on the training data set to obtain a risk probability prediction model.

[0095] Figure 3 A flowchart of constructing a risk probability prediction module provided in an embodiment of the present application. Figure 3 As shown in the figure, the time series thyroid risk factor data is processed, and then the thyroid nodule risk development trend is obtained to train the LSTM model to predict the risk probability of thyroid nodule cancer, and the LSTM basic model is obtained. In addition, the time series thyroid risk factor data is combined with the ACR-TIRADS risk stratification standard, K-TIRADS risk stratification standard, KWAK-TIRADS risk stratification standard, K-TIRADS risk stratification standard and C-TIRADS risk stratification standard to train TIRADS basic model 1, TIRADS basic model 2, TIRADS basic model 3, TIRADS basic model 4 and TIRADS basic model 5. Finally, the prediction results of the above LSTM basic model and TIRADS basic model are used as new features in combination with the time series risk factor data, and combined into a new training data set. Then use the new training data set to train a meta-model (stackingclassifier), which can be Random Forest, Gradient Boost, XGBoost, etc. The purpose of training is to enable the model to learn how to combine the prediction results of the basic model and maximize the model accuracy, so as to achieve the integration of the LSTM basic model and the TIRADS basic model and obtain a trained risk probability prediction model.

[0096] In addition, the above pricing model can also be used in other diseases with similar characteristics, such as breast nodules, lung nodules, and cardiovascular diseases. Taking lung nodules with similar characteristics as an example, in the risk prediction module, the TIRADS risk stratification standard can be replaced with the Chinese Expert Consensus on the Diagnosis and Treatment of Lung Nodules, NCCN, Fleischner Society, ACCP, Asian Consensus and other guidelines for lung nodules, and the corresponding basic model can be adjusted at the same time. In the medical cost estimation module, the analysis and statistics of drugs, equipment and costs are still applied, and only the disease type needs to be adjusted from thyroid cancer to lung cancer. Finally, based on the model results of the risk prediction module and the pricing basis of the medical cost estimation module, the product pricing of lung nodule insurance can be completed, so as to realize the promotion and reuse of the model.

[0097] In summary, this application first obtains the hospital medical data of patients with thyroid nodules. Among them, the hospital medical data includes thyroid function test result data, thyroid ultrasound image feature data and patient clinical feature data. Then, combined with the hospital medical data, the risk probability prediction model is used to predict the cancer risk probability of patients with thyroid nodules and determine the risk level. Based on the cancer risk probability and risk level, the medical cost estimation model is used to determine the treatment costs of patients with thyroid nodules. Finally, the special insurance for sick bodies is priced based on the cancer risk probability, risk level and treatment costs. In this way, a more refined cancer risk probability is generated through the risk probability prediction model, and then insurance pricing is performed, which improves the accuracy of pricing.

[0098] Figure 4 A schematic diagram of a pricing device for a special insurance policy for a sick person provided in an embodiment of the present application. Figure 4 As shown, the pricing device 400 for special insurance for sick persons provided in the embodiment of the present application includes:

[0099] An acquisition module 401 is used to acquire hospital medical data of patients with thyroid nodules; the hospital medical data includes thyroid function test result data, thyroid ultrasound image feature data, and patient clinical feature data;

[0100] The risk prediction module 402 is used to combine the hospital medical data, use the risk probability prediction model to predict the cancer risk probability of the thyroid nodule patient, and determine the risk level;

[0101] A medical expense estimation module 403 is used to determine the treatment cost of the thyroid nodule patient using a medical expense estimation model based on the cancer risk probability and the risk level;

[0102] The actuarial pricing module 404 is used to price the special disease-bearing insurance based on the cancer risk probability, the risk level and the treatment cost.

[0103] As an implementation method, regarding how to determine the treatment cost, the medical cost estimation module 403 is specifically used to:

[0104] Determining drugs and devices for treating the patient with thyroid nodules based on the cancer risk probability and the risk level;

[0105] Combined with the medicines and medical devices, and in association with the medical expense details and medical insurance settlement data, the medical expense estimation model is used to determine the amount and proportion of each expense; the amount of each expense includes the total settlement amount of the expense, the fund payment amount, the personal account payment amount, and the personal payment amount;

[0106] The treatment costs of the thyroid nodule patients are determined based on the classification of drugs in the local medical insurance catalog, the payment ratio of drugs in the local medical insurance catalog, and the amounts and proportions of the various expenses.

[0107] As an implementation method, with respect to how to train the risk probability prediction model, the above-mentioned pricing device 400 for special insurance for sick persons further includes: an acquisition submodule, a determination module, a first training module, a second training module and an integration module;

[0108] An acquisition submodule, used to obtain time series risk factor data for evaluating thyroid nodules;

[0109] A determination module is used to select key time points to analyze the time series risk factor data and determine the risk development trend information of thyroid nodules;

[0110] A first training module is used to train the LSTM model based on the risk development trend information and the canceration results, and obtain the LSTM basic model;

[0111] The second training module is used to construct a TIRADS basic model based on the time series risk factor data and thyroid nodule risk stratification standards;

[0112] The integration module is used to integrate the LSTM basic model and the TIRADS basic model to obtain a risk probability prediction model.

[0113] As an implementation method, regarding how to obtain time series risk factor data, the above acquisition submodule is specifically used to:

[0114] Obtaining clinical research and medical guidelines for patients with thyroid nodules; the clinical research and medical guidelines contain medical data text corresponding to risk factors; the risk factors include clinical indicators and ultrasound features;

[0115] The medical data text is subjected to NLP and structured processing to obtain time series risk factor data for evaluating thyroid nodules.

[0116] As an implementation method, regarding how to determine the risk development trend information, the above-mentioned determination module is specifically used to:

[0117] Select the time nodes before and after disease diagnosis, and the start and end of treatment to analyze the time series risk factor data, and determine the corresponding indicator data change trend; the change trend includes continuous rise, fall or fluctuation;

[0118] The risk development trend information of thyroid nodules is determined based on the changing trend of the indicator data.

[0119] As an implementation method, with respect to how to construct the TIRADS basic model, the second training module is specifically used for:

[0120] Access to thyroid nodule risk stratification criteria from at least one country's thyroid imaging reporting and data system;

[0121] Combined with the time series risk factor data, TIRADS basic models corresponding to each stratification standard are constructed respectively.

[0122] As an implementation method, with respect to how to integrate to obtain a risk probability prediction model, the above integration module is specifically used for:

[0123] In combination with the time series risk factor data, the prediction results of the LSTM basic model and the TIRADS basic model corresponding to each stratification standard are combined to form a training data set;

[0124] The meta-model is trained based on the training data set to obtain a risk probability prediction model.

[0125] In summary, this application first obtains the hospital medical data of patients with thyroid nodules. Among them, the hospital medical data includes thyroid function test result data, thyroid ultrasound image feature data and patient clinical feature data. Then, combined with the hospital medical data, the risk probability prediction model is used to predict the cancer risk probability of patients with thyroid nodules and determine the risk level. Based on the cancer risk probability and risk level, the medical cost estimation model is used to determine the treatment costs of patients with thyroid nodules. Finally, the special insurance for sick bodies is priced based on the cancer risk probability, risk level and treatment costs. In this way, a more refined cancer risk probability is generated through the risk probability prediction model, and then insurance pricing is performed, which improves the accuracy of pricing.

[0126] In addition, the present application also provides a pricing device for special sick body insurance, including:

[0127] Memory for storing computer programs;

[0128] A processor is used to implement the steps of the pricing method for special insurance for sick persons as described above when executing the computer program.

[0129] In addition, the present application also provides a readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the pricing method for special insurance for sick bodies as described above are implemented.

[0130] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A pricing method for special insurance for sick persons, characterized in that: The method comprises: Obtain hospital medical data of patients with thyroid nodules; the hospital medical data includes thyroid function test result data, thyroid ultrasound image feature data, and patient clinical feature data; In combination with the hospital medical data, a risk probability prediction model is used to predict the cancer risk probability of the thyroid nodule patient and determine the risk level; Based on the cancer risk probability and the risk level, determining the treatment cost of the thyroid nodule patient using a medical cost estimation model; The special insurance for sick persons is priced based on the cancer risk probability, the risk level and the treatment costs.

2. The method according to claim 1, characterized in that The risk probability prediction model is trained by the following method: To obtain time series risk factor data for the evaluation of thyroid nodules; Select key time points to analyze the time series risk factor data to determine the risk development trend information of thyroid nodules; Training an LSTM model based on the risk development trend information and the canceration results, and obtaining an LSTM basic model; Constructing the TIRADS basic model based on the time series risk factor data and thyroid nodule risk stratification criteria; The LSTM basic model and the TIRADS basic model are integrated to obtain a risk probability prediction model.

3. The method according to claim 2, characterized in that The method of obtaining time series risk factor data for evaluating thyroid nodules includes: Obtaining clinical research and medical guidelines for patients with thyroid nodules; the clinical research and medical guidelines contain medical data text corresponding to risk factors; the risk factors include clinical indicators and ultrasound features; The medical data text is subjected to NLP and structured processing to obtain time series risk factor data for evaluating thyroid nodules.

4. The method according to claim 2, characterized in that: The step of selecting key time points to analyze the time series risk factor data and determining the risk development trend information of thyroid nodules includes: Select the time nodes before and after disease diagnosis, and the start and end of treatment to analyze the time series risk factor data, and determine the corresponding indicator data change trend; the change trend includes continuous rise, fall or fluctuation; The risk development trend information of thyroid nodules is determined based on the changing trend of the indicator data.

5. The method according to claim 2, characterized in that: The constructing of the TIRADS basic model based on the time series risk factor data and the thyroid nodule risk stratification standard comprises: Access to thyroid nodule risk stratification criteria from at least one country's thyroid imaging reporting and data system; Combined with the time series risk factor data, TIRADS basic models corresponding to each stratification standard are constructed respectively.

6. The method according to claim 5, characterized in that The integrating the LSTM basic model and the TIRADS basic model to obtain a risk probability prediction model includes: In combination with the time series risk factor data, the prediction results of the LSTM basic model and the TIRADS basic model corresponding to each stratification standard are combined to form a training data set; The meta-model is trained based on the training data set to obtain a risk probability prediction model.

7. The method according to claim 1, characterized in that The method of determining the treatment cost of the thyroid nodule patient by using a medical cost estimation model based on the cancer risk probability and the risk level includes: Determining drugs and devices for treating the patient with thyroid nodules based on the cancer risk probability and the risk level; Combined with the medicines and medical devices, and in association with the medical expense details and medical insurance settlement data, the medical expense estimation model is used to determine the amount and proportion of each expense; the amount of each expense includes the total settlement amount of the expense, the fund payment amount, the personal account payment amount, and the personal payment amount; The treatment costs of the thyroid nodule patients are determined based on the classification of drugs in the local medical insurance catalog, the payment ratio of drugs in the local medical insurance catalog, and the amounts and proportions of the various expenses.

8. A pricing device for special insurance for sick persons, characterized in that: include: An acquisition module is used to obtain hospital medical data of patients with thyroid nodules; The hospital medical data includes thyroid function test result data, thyroid ultrasound image feature data and patient clinical feature data; A risk prediction module, used to combine the hospital medical data, use a risk probability prediction model to predict the cancer risk probability of the thyroid nodule patient, and determine the risk level; A medical expense estimation module, used to determine the treatment cost of the thyroid nodule patient based on the cancer risk probability and the risk level using a medical expense estimation model; The actuarial pricing module is used to price the special disease-bearing insurance based on the cancer risk probability, the risk level and the treatment cost.

9. A pricing device for special sick body insurance, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of the pricing method for special insurance for sick persons as described in any one of claims 1 to 7 when executing the computer program.

10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the pricing method for special insurance for sick persons as described in any one of claims 1 to 7.

Images