Personalized digital health assessment method based on digital twin and generative ai

Abstract

The application discloses a personalized digital health assessment method based on digital twinning and generative AI, and relates to the technical field of intelligent health management. Through the collection of multi-modal data related to the health of the user to be evaluated, the multi-source heterogeneous data obtained is integrated by using data processing technology to construct a health data set of the user to be evaluated, a digital twinning model for reflecting the health status of the user to be evaluated in real time and simulating the development process of the target disease of the user to be evaluated is constructed based on the integrated multi-modal data, the generative AI model is used to learn the integrated multi-modal data, the learning result including the data distribution state and the logical relationship between variables is obtained, the synthetic data of the multi-modal data is created, the synthetic data of the multi-modal data is imported into the digital twinning model, the current health status evaluation report of the user to be evaluated is generated, and the incidence risk and risk level of the target disease under the current health status are obtained.

Description

Technical Field

[0001] This invention relates to the field of intelligent health management technology, and in particular to a personalized digital health assessment method based on digital twins and generative AI. Background Technology

[0002] The core of digital twin technology is the interactive mapping between digital models in virtual space and physical entities. In recent years, digital twin technology has been widely applied in various fields, such as aerospace, engineering construction, and intelligent manufacturing. In the healthcare field, digital twin technology can be used to construct a patient's medical digital twin. By integrating information such as the patient's health records, medical history, and monitoring data from smart wearable devices, it simulates human function in the cloud, enabling real-time monitoring, predictive analysis, and precise medical diagnosis of the patient's health status. Generative AI is an important branch of artificial intelligence. By learning from large amounts of data and patterns, it can generate new content, such as text, images, and audio. In the field of health management, generative AI can be used to process complex multimodal data, such as electrocardiograms, pulse waves, and facial spectra, thereby achieving rapid assessment and prediction of health indicators.

[0003] With increasing health awareness and strained medical resources, the demand for personalized digital health assessments is growing. Traditional health assessment methods often rely on doctors' experience and limited testing methods, making it difficult to meet people's needs for accurate, efficient, and personalized health assessments. While existing digital health assessment tools have improved efficiency and accuracy to some extent, they still fall short in terms of personalization and precision. Therefore, a new approach is needed that fully leverages the advantages of digital twins and generative AI technologies to achieve comprehensive, accurate, and real-time assessments of individual health status, providing personalized health recommendations and intervention plans. The development of digital twins and generative AI technologies offers new possibilities for personalized digital health assessments. By combining these technologies, it is hoped that the limitations of traditional health assessment methods can be overcome, providing people with more accurate, efficient, and personalized health assessment services. Therefore, this paper proposes a personalized digital health assessment method based on digital twins and generative AI. Summary of the Invention

[0004] The main objective of this invention is to provide a personalized digital health assessment method based on digital twins and generative AI, which can effectively solve the problems in the background art.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: Personalized digital health assessment methods based on digital twins and generative AI include: Step 1: Collect multimodal data related to the health of the user to be evaluated. The multimodal data includes clinical information, real-time physiological data, lifestyle data, health records, medical history, medication history, and monitoring data obtained using smart wearable devices. Data processing technology is used to integrate the acquired multi-source heterogeneous data to construct a health dataset of the user to be evaluated. Step 2: Based on the integrated multimodal data, a digital twin model is constructed using mathematical modeling and computer simulation techniques to reflect the health status of the user to be assessed in real time. The digital twin model is also used to simulate the development process of the target disease for the user to be assessed. Step 3: Use a generative AI model to learn from the integrated multimodal data, obtain learning results including the data distribution state and the logical relationships between variables, and based on the data learning results, sample data from the data distribution and variable relationships of the constructed digital twin model to create synthetic multimodal data; Step 4: Import the synthesized data of the multimodal data into the digital twin model to generate a current health status assessment report for the user to be evaluated, as well as the risk and risk level of the target disease based on the current health status.

[0006] The method further includes: Step 5: Set the data update cycle to T, update the multimodal data in the health dataset according to the update cycle T, and dynamically update the digital twin model and the synthesized data of the multimodal data according to the update results of the health dataset.

[0007] Furthermore, in step two, the assessment process for the user's health status includes the following steps: A health status assessment index system is constructed, and health status levels are defined. The health status levels are defined as four categories: healthy status, basic healthy status, sub-healthy status, and unhealthy status. Construct a membership function to describe the degree of membership of any of the health status assessment indicators under different health status levels. Based on the collected multimodal data of the user to be assessed, determine the health status assessment indicator data of the user in the current state, and perform preprocessing including data cleaning and standardization. Based on the constructed membership function and the obtained health status assessment index data, the membership degree of each evaluation index under different health status levels is calculated. The membership degrees of all evaluation indicators are weighted to obtain the health status assessment results of the user to be evaluated, and the health status level is determined based on the health status assessment results.

[0008] Furthermore, the assessment indicators include one or more of the following: physical health indicators, mental health indicators, social health indicators, lifestyle indicators, user behavior indicators, environmental factor indicators, and self-assessment indicators.

[0009] Furthermore, in step two, simulating the specific process of the target disease development for the user being evaluated using a digital twin model includes the following steps: Determine the data types required for simulating the development process of the target disease, and extract the required data from the collected multimodal data of the users to be evaluated based on the determination results; Historical case data on the development process of the target disease are obtained as learning corpus. The digital twin model is trained using the learning corpus so that it can learn the development patterns and rules of the disease. By comparing with actual case data, the relevant parameters of the digital twin model are adjusted until its accuracy and reliability meet the set expected values. Using the user's current health status as the initial condition, the initial parameters of the digital twin model are set according to the initial condition. By adjusting the parameters in the model, the progression of the disease under different conditions is simulated. The multimodal data in the health dataset is updated, and the digital twin model is dynamically updated based on the update results to ensure that the accuracy and reliability of the digital twin model simulation results are not lower than the set expected value. Establish a feedback mechanism to continuously optimize the constructed digital twin model based on the actual health changes and feedback of the users to be evaluated.

[0010] Furthermore, the physical health indicators include at least one of physiological indicators, disease state, physical function, and health risk behaviors; The mental health indicators include at least one of emotional state, psychological resilience, and cognitive function. The social health indicators include at least one of social support, social participation, and social adaptability; The lifestyle indicators include at least one of dietary habits, exercise habits, and sleep quality; The user behavior indicators include at least one of preventive healthcare behaviors and self-management behaviors; The environmental factor indicators include at least one of the living environment and the working environment; The self-assessment indicators include at least one of self-health assessment and quality of life assessment.

[0011] The present invention has the following beneficial effects: Compared with existing technologies, this method collects multimodal data related to the health of the user to be assessed, integrates the acquired multi-source heterogeneous data using data processing technology to construct a health dataset for the user to be assessed, and uses mathematical modeling and computer simulation technology to construct a digital twin model to reflect the health status of the user to be assessed in real time. The digital twin model is also used to simulate the development process of the target disease for the user to be assessed. A generative AI model is used to learn from the integrated multimodal data to obtain learning results including the data distribution status and the logical relationships between variables. Based on the data learning results, data is sampled from the data distribution and variable relationships of the constructed digital twin model to create synthetic multimodal data. The synthetic multimodal data is imported into the digital twin model to generate a current health status assessment report for the user to be assessed, as well as the risk and risk level of the target disease based on the current health status. This method can fully utilize the advantages of digital twin and generative AI technologies, and overcome the limitations of traditional health assessment methods by combining these technologies, thus achieving a comprehensive, accurate, and real-time assessment of individual health status. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the overall structure of the personalized digital health assessment method based on digital twins and generative AI of the present invention. Figure 2 This is a schematic diagram of the assessment process for the health status of the user to be assessed in an embodiment of the present invention; Figure 3 This is a schematic diagram of a simulation process of the development of a target disease by a user to be evaluated in an embodiment of the present invention. Detailed Implementation

[0013] The present invention will be further described below with reference to specific embodiments. The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the present invention. In order to better illustrate the specific embodiments of the present invention, some parts in the drawings may be omitted, enlarged or reduced, and do not represent the actual product size.

[0014] The specific implementation process of the technical solution of this invention includes the following steps: Step 1: Collect multimodal data related to the health of the user to be evaluated. This multimodal data includes clinical information, real-time physiological data, lifestyle data, health records, medical history, medication history, and monitoring data obtained using smart wearable devices. Data processing techniques are used to integrate the acquired multi-source heterogeneous data to construct a health dataset for the user to be evaluated. The specific process of multimodal data integration can be divided into the following key steps, covering the entire process from data collection to fusion model design and optimization: Step S11: Data Collection and Preprocessing Data acquisition: Collecting data of different modalities from multiple channels, such as images, text, audio, sensor data, etc.

[0015] Data cleaning: operations such as removing noise, filling in missing values, and standardizing timestamps to ensure data quality.

[0016] Standardization processing: Converting data in different formats into a unified standard format to facilitate subsequent processing.

[0017] Step S12: Feature Extraction Image feature extraction: Image data is processed using methods such as convolutional neural networks (CNN).

[0018] Text feature extraction: Use pre-trained language models (such as BERT, GPT) or word embedding methods to process text data.

[0019] Other modal feature extraction: Select appropriate feature extraction methods based on data type, such as Long Short-Term Memory (LSTM) networks for processing sequence data.

[0020] Step S13: Data Alignment Time and space alignment: Aligning data from different sources in time and space to ensure consistency of the merged data.

[0021] Step S14: Data Fusion Early Fusion: This involves directly fusing raw data or low-level features from different modalities at the input layer. For example, concatenating image pixels and text word vectors before inputting them into the model.

[0022] Middle-level fusion (or representation-level fusion): Each modality is encoded separately first, and then interacts through attention mechanisms or graph networks. For example, cross-modal attention (such as CLIP) in visual language models.

[0023] Late Fusion: This involves fusing the scores (decision decisions) from classifiers trained on different modalities.

[0024] Step S15: Fusion Model Design and Optimization Choose a fusion model: Select a suitable fusion model based on the application scenario and data characteristics, such as weighted average, neural network, support vector machine, etc.

[0025] Optimize fusion strategies: Employ methods such as shared attention mechanisms to reduce computational load and improve the model's adaptability to weak modal data.

[0026] Step S16: Application and Evaluation Application scenario development: Apply the fused data to specific tasks, such as autonomous driving, medical diagnosis, and sentiment analysis.

[0027] Dataset evaluation: Evaluate the quality, diversity, and representativeness of the dataset to ensure it meets practical needs.

[0028] Through the above process, multimodal data can be effectively integrated to provide richer information and more accurate decision support for various application scenarios.

[0029] Step 2: Based on the integrated multimodal data, a digital twin model is constructed using mathematical modeling and computer simulation techniques to reflect the health status of the user to be assessed in real time. The digital twin model is also used to simulate the development process of the target disease for the user to be assessed. The assessment process for the health status of the user to be evaluated includes the following steps: Step S211: Construct a health status assessment index system and define health status levels, wherein the health status levels are defined as four categories in sequence: healthy status, basic healthy status, sub-healthy status and unhealthy status. Step S212: Construct a membership function to describe the degree of membership of any health status assessment index under different health status levels. Based on the collected multimodal data of the user to be assessed, determine the data of each health status assessment index in the current state, and perform preprocessing including data cleaning and standardization. Step S213: Calculate the membership degree of each evaluation index under different health status levels based on the constructed membership function and the obtained health status assessment index data; Step S214: Weight the membership degrees of all evaluation indicators to obtain the health status assessment results of the user to be evaluated, and determine the health status level based on the health status assessment results.

[0030] The specific process of simulating the progression of a target disease for a user being evaluated includes the following steps: Step S221: Determine the data types required for simulating the development process of the target disease, and extract the required data from the collected multimodal data of the users to be evaluated based on the determination results; Step S222: Obtain historical actual case data of the development process of the target disease as learning corpus, use the learning corpus to train the digital twin model so that it can learn the development pattern and law of the disease, and adjust the relevant parameters of the digital twin model to meet the set expected values ​​in terms of accuracy and reliability by comparing it with actual case data. Step S223: Using the current health status of the user to be evaluated as the initial condition, set the initial parameters of the digital twin model according to the initial condition, and simulate the progression of the disease under different conditions by adjusting the parameters in the model. Step S224: Update the multimodal data in the health dataset, and dynamically update the digital twin model based on the update results of the health dataset, so that the accuracy and reliability of the simulation results of the digital twin model are not lower than the set expected value; Step S225: Establish a feedback mechanism to continuously optimize the constructed digital twin model based on the actual health changes and feedback of the users to be evaluated.

[0031] It should be noted that health status assessment indicators typically encompass multiple dimensions to comprehensively reflect an individual's health status. Below are some common health status assessment indicators, referenced from the latest research and official guidelines: Physical health Physiological indicators include blood pressure, blood sugar levels, blood lipid levels (such as total cholesterol, HDL cholesterol, LDL cholesterol, and blood triglycerides), and heart rate.

[0032] Disease status: such as cancer stage, control status of chronic diseases (such as diabetes, cardiovascular disease).

[0033] Physical function: such as body mass index (BMI), physical activity level, muscle strength, joint flexibility, etc.

[0034] Health risk behaviors: such as smoking, drinking alcohol, and drug abuse.

[0035] mental health Emotional state: such as depression, anxiety, mood swings, etc.

[0036] Psychological resilience: such as the ability to cope with stress, life satisfaction, etc.

[0037] Cognitive functions: such as memory, attention, and mental agility.

[0038] Social health Social support: such as the quality of family relationships, social networks, and social support systems.

[0039] Social participation: such as whether or not one participates in community activities, volunteer activities, etc.

[0040] Social adaptability: such as job satisfaction, life adaptability, and fulfillment of social roles.

[0041] lifestyle Dietary habits: such as nutritional balance, whether meals are eaten regularly, and whether there are specific dietary restrictions.

[0042] Exercise habits: such as weekly exercise frequency, type of exercise, and exercise intensity.

[0043] Sleep quality: such as sleep duration, sleep quality, and sleep disorders.

[0044] User behavior Preventive healthcare behaviors: such as regular physical examinations, vaccinations, and cancer screenings.

[0045] Self-management behaviors: such as the ability of patients with chronic diseases to manage their own diseases and medication adherence.

[0046] Environmental factors Living environment: such as air quality, safety of the living environment, noise level, etc.

[0047] Work environment: such as work pressure, occupational exposure risks, etc.

[0048] self assessment Self-health assessment: such as an individual's overall assessment of their own health status.

[0049] Quality of life assessment: such as life satisfaction, degree of activity limitation, etc.

[0050] These indicators can be adjusted and selected according to different evaluation purposes and objects to ensure the accuracy and comprehensiveness of the evaluation results.

[0051] Step 3: Utilize a generative AI model to learn from the integrated multimodal data, obtaining learning results including data distribution and logical relationships between variables. Based on these learning results, sample data from the data distribution and variable relationships of the constructed digital twin model to create synthetic multimodal data. The synthetic data can be generated through the following process, with the specific steps as follows: Step S31: Data Understanding and Preprocessing Data understanding: First, fully understand the real dataset, including the distribution of the data, the relationships between variables, missing data elements, and extreme values.

[0052] Data preprocessing: Remove or fill in missing values, correct errors, and standardize data format. Simultaneously, remove or encrypt any personally identifiable information (PII) to ensure privacy.

[0053] Step S32: Select synthetic data generation technology Deep learning models: Generative Adversarial Networks (GANs): GANs consist of a generator and a discriminator. The generator is responsible for creating synthetic data, while the discriminator is responsible for distinguishing between synthetic and real data. Through adversarial training, the generator can generate synthetic data that is similar to the distribution of real data.

[0054] Variational Autoencoders (VAEs): VAEs are unsupervised machine learning models where the encoder compresses and integrates real-world data, while the decoder analyzes this data to generate a representation of the real-world data. A key advantage of VAEs is ensuring that the input and output data remain highly similar.

[0055] Generative Pretrained Transformer (GPT): GPT is a generative model based on the Transformer architecture that can generate high-quality text data. In the medical field, it can be used to generate text descriptions related to patient health.

[0056] Data augmentation: Although data augmentation is not synthetic data, it can improve the generalization ability of a model by adding new data to an existing dataset.

[0057] Step S33: Synthetic Data Creation and Validation Synthetic data creation: New data points are created by sampling from the modeled distributions and relationships, ensuring that the synthetic data retains the statistical properties of the original data.

[0058] Post-processing: The synthetic data is altered by adding subtle differences and avoiding any indication of a deliberately created consistent pattern. The synthetic data is then examined to confirm that it does not contain any re-identifiable personal information.

[0059] Validation and usability evaluation: Validate the quality of the synthetic data to ensure that it is statistically similar to real data, and evaluate its effectiveness in practical applications.

[0060] Step S34: Optimize the digital twin model using synthetic data Filling data gaps: Synthetic data can generate data and long-tail data in sensitive or high-security areas, thereby filling gaps in real data and improving the accuracy of twin models.

[0061] Enhanced model training: Using synthetic data for model training can improve the model's generalization ability and adaptability, enabling it to better simulate the patient's health status and the progression of the target disease.

[0062] Dynamic updates and optimizations: As new data is generated and the model is trained, the digital twin model is continuously updated to achieve more accurate health predictions and personalized intervention plans.

[0063] Using the methods described above, generative AI can generate synthetic data for digital twin models used to optimize patient health, providing strong support for research and practice in the medical field.

[0064] Step 4: Import the synthesized multimodal data into the digital twin model to generate a current health status assessment report for the user to be evaluated, as well as the risk and risk level of the target disease based on the current health status. The method for determining the risk and risk level is the same as in steps S211-S214, and will not be repeated here.

[0065] Step 5: Set the data update cycle to T, update the multimodal data in the health dataset at the update cycle T, and dynamically update the digital twin model and the synthetic data of the multimodal data based on the update results of the health dataset.

[0066] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A personalized digital health assessment method based on digital twins and generative AI, characterized in that, include: Step 1: Collect multimodal data related to the health of the user to be evaluated. The multimodal data includes clinical information, real-time physiological data, lifestyle data, health records, medical history, medication history, and monitoring data obtained using smart wearable devices. Data processing technology is used to integrate the acquired multi-source heterogeneous data to construct a health dataset of the user to be evaluated. Step 2: Based on the integrated multimodal data, a digital twin model is constructed using mathematical modeling and computer simulation techniques to reflect the health status of the user to be assessed in real time. The digital twin model is also used to simulate the development process of the target disease for the user to be assessed. Step 3: Use a generative AI model to learn from the integrated multimodal data, obtain learning results including the data distribution state and the logical relationships between variables, and based on the data learning results, sample data from the data distribution and variable relationships of the constructed digital twin model to create synthetic multimodal data; Step 4: Import the synthesized data of the multimodal data into the digital twin model to generate a current health status assessment report for the user to be evaluated, as well as the risk and risk level of the target disease based on the current health status.

2. The personalized digital health assessment method based on digital twins and generative AI according to claim 1, characterized in that, The method further includes: Step 5: Set the data update cycle to T, update the multimodal data in the health dataset according to the update cycle T, and dynamically update the digital twin model and the synthesized data of the multimodal data according to the update results of the health dataset.

3. The personalized digital health assessment method based on digital twins and generative AI according to claim 1, characterized in that, In step two, the assessment process for the user's health status includes the following steps: A health status assessment index system is constructed, and health status levels are defined. The health status levels are defined as four categories: healthy status, basic healthy status, sub-healthy status, and unhealthy status. Construct a membership function to describe the degree of membership of any of the health status assessment indicators under different health status levels. Based on the collected multimodal data of the user to be assessed, determine the health status assessment indicator data of the user in the current state, and perform preprocessing including data cleaning and standardization. Based on the constructed membership function and the obtained health status assessment index data, the membership degree of each evaluation index under different health status levels is calculated. The membership degrees of all evaluation indicators are weighted to obtain the health status assessment results of the user to be evaluated, and the health status level is determined based on the health status assessment results.

4. The personalized digital health assessment method based on digital twins and generative AI according to claim 1, characterized in that, The assessment indicators include one or more of the following: physical health indicators, mental health indicators, social health indicators, lifestyle indicators, user behavior indicators, environmental factor indicators, and self-assessment indicators.

5. The personalized digital health assessment method based on digital twins and generative AI according to claim 1, characterized in that, In step two, the specific process of simulating the development of the target disease for the user being evaluated using a digital twin model includes the following steps: Determine the data types required for simulating the development process of the target disease, and extract the required data from the collected multimodal data of the users to be evaluated based on the determination results; Historical case data on the development process of the target disease are obtained as learning corpus. The digital twin model is trained using the learning corpus so that it can learn the development patterns and rules of the disease. By comparing with actual case data, the relevant parameters of the digital twin model are adjusted until its accuracy and reliability meet the set expected values. Using the user's current health status as the initial condition, the initial parameters of the digital twin model are set according to the initial condition. By adjusting the parameters in the model, the progression of the disease under different conditions is simulated. The multimodal data in the health dataset is updated, and the digital twin model is dynamically updated based on the update results to ensure that the accuracy and reliability of the digital twin model simulation results are not lower than the set expected value. Establish a feedback mechanism to continuously optimize the constructed digital twin model based on the actual health changes and feedback of the users to be evaluated.

6. The personalized digital health assessment method based on digital twins and generative AI according to claim 4, characterized in that, The physical health indicators include at least one of physiological indicators, disease state, physical function, and health risk behaviors; The mental health indicators include at least one of emotional state, psychological resilience, and cognitive function. The social health indicators include at least one of social support, social participation, and social adaptability; The lifestyle indicators include at least one of dietary habits, exercise habits, and sleep quality; The user behavior indicators include at least one of preventive healthcare behaviors and self-management behaviors; The environmental factor indicators include at least one of the living environment and the working environment; The self-assessment indicators include at least one of self-health assessment and quality of life assessment.

Images