A method for predicting multi-drug treatment response based on changes in blood viscosity

Abstract

The present invention provides a method for predicting multi-drug treatment response based on blood viscosity changes, comprising: dynamically detecting blood rheology parameters through a fully automatic blood viscosity monitor, evaluating blood viscosity changes, and providing a basis for formulating a multi-drug treatment plan; constructing a knowledge map of interactions between multiple drugs, collecting drug-drug, drug-food, and drug-disease interaction relationship data, and providing a reference for the selection of multi-drug combination medications; collecting patient compliance and lifestyle information during multi-drug treatment, establishing a model to evaluate the correlation between these factors and medication efficacy and adverse reactions, and providing health management advice; designing an intelligent voice assistant for medication reminders and monitoring, performing disease tracking, consulting services, etc., to help patients use multiple drugs correctly; re-detecting the patient's hemostasis parameters, evaluating the blood viscosity changes before and after multi-drug treatment, establishing a database to collect effective cases, and providing support for new drug research and development.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a method for predicting multi-drug treatment response based on changes in blood viscosity. Background Art

[0002] At present, multi-drug therapy is widely used in clinical practice, but it faces a series of challenges. Traditional clinical methods mainly formulate treatment plans based on the symptoms of the disease and the physiological indicators of the patient, ignoring the complex interrelationships between individual differences and multiple drugs. This leads to uncertainty in the treatment effect and adverse reactions of patients during treatment. Under the current technical framework, the introduction of blood viscosity monitors provides a new direction for multi-drug therapy. However, traditional blood parameter monitoring methods are often static and cannot reflect dynamic changes in blood rheology. In addition, there are still contradictions in the comprehensive consideration of individual differences and multi-drug interactions in existing technologies. The formulation of treatment plans requires more precise guidance to ensure that patients can obtain the best efficacy and minimal adverse reactions during treatment. On the other hand, the current treatment plans lack personalized adjustments for different individuals. The existing efficacy prediction models do not fully consider factors such as the patient's age, gender, and underlying diseases, resulting in uncertainty in the treatment effect. In addition, the monitoring of vital signs changes and patient lifestyles in multi-drug therapy has not yet achieved real-time and comprehensive coverage. In the context of the complex application of multi-drugs, the establishment of the existing multi-drug compatibility system and interaction knowledge graph is still not intelligent enough. The lack of comprehensive consideration of individual parameters, severity of the disease and other factors makes it difficult to provide accurate predictions of efficacy and risks. In addition, the assessment of patient compliance and lifestyle is also lacking, which limits the comprehensive understanding of various factors in the treatment process. In summary, the existing technology has contradictions and shortcomings in the individualization, dynamicization and intelligence of multi-drug therapy. Therefore, it is necessary to introduce an innovative method to provide more accurate and personalized prediction and decision support for multi-drug therapy through comprehensive, dynamic and intelligent monitoring and analysis. Summary of the invention

[0003] In view of this, the present invention provides a method for predicting multi-drug treatment response based on blood viscosity changes, which mainly includes the following steps:

[0004] The blood rheology parameters are dynamically detected by the fully automatic blood viscosity monitor to evaluate the changes in blood viscosity and provide a basis for formulating multi-drug treatment plans;

[0005] Build a large sample database of patients of different ages, genders, and underlying diseases, collect clinical data such as treatment response and adverse reactions during multi-drug treatment, use deep learning technology to analyze the characteristics and outcomes of these patients, and establish an accurate prognosis prediction model;

[0006] Design an intelligent multi-drug compatibility system that uses machine learning algorithms to predict the efficacy and risks of different medication patterns based on individual patient parameters, disease severity, and other factors, and provide decision support for multi-drug treatment plans;

[0007] Wearable sensors are used to monitor the patient's vital signs in real time during multi-drug treatment, evaluate the safety of different dosing regimens, and adjust the medication plan as needed in conjunction with the above-mentioned blood viscosity monitoring results;

[0008] Construct a knowledge graph of drug-drug interactions, collect drug-drug, drug-food, and drug-disease interaction data, and provide a reference for the selection of drug combinations for multi-drug combination therapy;

[0009] Collect information on patients' compliance and lifestyle during multi-drug treatment, establish models to evaluate the correlation between these factors and medication efficacy and adverse reactions, and provide health management advice;

[0010] Design intelligent voice assistants for medication reminders and monitoring, to provide condition tracking, consulting services, etc., to help patients use multiple medications correctly;

[0011] The patient's hemostasis parameters were tested again, the changes in blood viscosity before and after multi-drug treatment were evaluated, and a database was established to collect effective cases to provide support for new drug development.

[0012] Furthermore, the blood rheology parameters are dynamically detected by the fully automatic blood viscosity monitor to evaluate the changes in blood viscosity, providing a basis for formulating a multi-drug treatment plan, including:

[0013] The blood samples of patients are tested by automatic blood viscosity analyzer to collect data and evaluate blood fluidity; blood rheology parameters are analyzed and compared to determine the differences between samples and set the standard range of abnormal values; if the blood viscosity is detected to deviate from the standard range, the support vector machine algorithm is used to predict the possible clinical diseases causing the abnormality; based on the disease prediction results, the medical drug database is accessed to obtain information on targeted therapeutic drugs; if the prediction results show that there are multiple treatment options, the decision tree algorithm is used to analyze the effectiveness and interaction between various drugs; the algorithm is used to analyze the various drug compatibility options that are distinguished, and their effects on blood viscosity are calculated to obtain the optimal combination; the optimal drug combination will be compared and analyzed with historical case data, and the final personalized treatment plan will be determined with the assistance of the artificial intelligence system; the screened treatment plan will be entered into the drug management system, and the system will automatically generate a personalized drug use plan and a corresponding early warning mechanism; after the drug management plan is implemented, the patient's blood viscosity will be monitored again using the automatic blood viscosity analyzer to verify the continuity and stability of the treatment effect.

[0014] Furthermore, the construction includes a large sample database of patients of different ages, genders, and underlying diseases, collects their clinical data such as treatment response and adverse reactions during multi-drug treatment, uses deep learning technology to analyze the characteristics and outcomes of these patients, and establishes an accurate prognosis prediction model, including:

[0015] Through the construction of a professional database, the patient's age, gender and existing underlying diseases are recorded as an important basis for primary screening of patients; the collected clinical data are processed using data cleaning technology, the key is to extract information such as the patient's treatment drugs, reactions to these drugs, and possible adverse reactions; the cleaned data are classified, and based on the accepted treatment plan, multiple sub-databases can be created to analyze the reactions caused by different drug treatments; through text mining technology, key narratives describing the treatment effect and detailed information on complications are deeply mined and sorted out in the clinical data; feature selection algorithms are used, combined with in-depth analysis of patient characteristics, to screen out those closely related to the treatment effect. The deep learning model is used to identify the characteristic factors that are closely related and have great influence; these screened patient characteristics and treatment effect data are input into the deep learning model, which is used as the basis for the first training of the model and the prediction of the patient's prognosis; in order to ensure the stability and accuracy of the prognosis prediction model, the cross-validation method is used for evaluation, which can ensure that the model meets the statistical requirements of the study; after the model evaluation meets expectations, it is further optimized and adjusted, and the incremental learning technology is used here to improve the prediction accuracy of the model while continuously introducing new clinical data; the optimized prognosis prediction model is deployed into the clinical decision support system, so that the accurate prognosis information provided by the model can be used to help formulate individualized treatment plans for individual patients.

[0016] Furthermore, the intelligent multi-drug compatibility system is designed to use machine learning algorithms to predict the efficacy and risks of different medication modes based on individual patient parameters, disease severity and other factors, and provide decision support for multi-drug treatment plans, including:

[0017] Collect individual parameters such as the patient's medical history and vital signs, and integrate this information to ensure the accuracy of each data; use convolutional neural network algorithms to analyze historical medical data, including treatment effects on similar cases; based on these data, build a primary drug compatibility model, and define an initial weight score for each drug based on the patient's individual parameters and relevant medical literature and research. This score reflects the possible effectiveness and risk level of the drug. By further adjusting these initial weights, the selection and compatibility of drugs can be optimized until a combination suitable for a specific patient is found; conduct a risk assessment based on the severity of the patient's current condition to help determine whether medication is urgently needed and prevent possible side effects in advance; apply convolutional neural network algorithms to different Predict and analyze the potential efficacy under different medication patterns, and propose multiple treatment plans; evaluate the long-term and short-term risks of these plans, conduct in-depth analysis and comparison to identify the risks that patients may face; to ensure the safety of treatment, use the drug interaction database to check the drug compatibility in these plans to avoid potential adverse drug reactions; based on the feedback of real-time medical monitoring data, continuously analyze the changes in physical signs during treatment, and adjust the drug plan and dosage accordingly; use the expert system as a decision support tool to combine the expert knowledge and the plan generated by the algorithm to provide a tailored multi-drug treatment plan; design a feedback mechanism to continuously track the patient's treatment effect, optimize the convolutional neural network algorithm through actual results, and improve the accuracy and safety of the treatment plan.

[0018] Furthermore, the use of wearable sensors to monitor the patient's vital signs in real time during multi-drug treatment, evaluate the safety of different medication regimens, and adjust the medication plan as needed in conjunction with the above-mentioned blood viscosity monitoring results, includes:

[0019] Wearable sensor technology is used to monitor the patient's vital signs and collect real-time data. The data is processed by the intelligent data analysis system to obtain detailed physiological parameters and viscosity monitoring results. The collected data is transmitted in real time to the data storage and management module, which is responsible for recording the patient's health history information and using this information as an auxiliary basis for the clinical decision support system. The personalized medical strategy module analyzes the monitoring data according to the patient's specific situation and determines whether the current medication regimen needs to be adjusted to ensure the best treatment effect. When the monitoring data shows that the patient has potential symptoms or risks of maladaptation, the risk warning system can promptly issue an alarm to the doctor or patient. The safety assessment module continuously evaluates the patient's real-time vital sign data to determine whether there are adverse reactions caused by drugs. If the evaluation results show potential risks, the medication plan adjustment module is activated to re-formulate the treatment plan, and the multi-drug treatment analysis module is used to evaluate the interaction between different drugs to optimize the treatment strategy. The treatment effect monitoring module continuously tracks the new treatment plan to evaluate the efficacy of the plan and ensure that the treatment goals are achieved. If the efficacy monitoring module reports that the efficacy is poor, the medication plan adjustment module will be activated again and continuously optimized until the most suitable treatment plan for the patient is found.

[0020] Furthermore, the construction of the multi-drug interaction knowledge graph collects drug-drug, drug-food, and drug-disease interaction data to provide a reference for the selection of multi-drug combination medications, including:

[0021] Use text mining technology to analyze medical literature and drug instructions to extract information on drug properties and mechanisms of action; use information from food ingredient databases, combined with relevant knowledge of nutrition, to determine the impact that specific food ingredients may have on drug efficacy; conduct in-depth analysis of disease-related data to identify the correlation between drug mechanisms of action and disease characteristics; by identifying these correlations, you can know which drugs may have therapeutic or side effect risks for specific diseases; through drug metabolism analysis tools, you can better understand the metabolic process of drugs in the body, especially the interaction between metabolites and other drugs or nutrients; by analyzing medical guidelines and treatment manuals, you can provide medical professionals with recommendations for the combined use of multiple drugs to help develop the most reasonable combination treatment plan; by tracking Drug concentration monitoring data can observe in real time the impact of combined medication on drug plasma concentration and analyze the changing trend of drug concentration under various conditions; in analyzing side effect data, disease characteristics will be taken into account to determine the risk of side effects that different patient groups may encounter when using specific drugs; based on the above collected data and previous research results, a drug efficacy impact assessment model is constructed, which aims to predict the interactions of different drug combinations and evaluate how they affect the treatment effect; integrating the data and evaluation results obtained in all the above steps, and deploying a dynamic knowledge graph update mechanism using pattern recognition and machine learning algorithms; this mechanism can automatically incorporate new research results and real-time data, thereby continuously improving the knowledge graph and providing medical workers with scientific decision-making support on drug combination therapy.

[0022] Furthermore, the patient's compliance and lifestyle information during multi-drug treatment is collected, a model is established to evaluate the correlation between these factors and drug efficacy and adverse reactions, and health management suggestions are provided, including:

[0023] Collect detailed information about patients' medication treatment, including the name of the drug, dosage, frequency of medication, and specific time period of medication, to establish a basic database for subsequent analysis; collect data on patients' compliance with medical treatment, including whether patients take medications on time according to doctor's orders, monitor and mark incidents of missed or wrong medications, which provides direct evidence for evaluating the actual implementation of medication treatment; conduct a comprehensive survey of patients' lifestyles, including evaluation and recording of diet, exercise, sleep, and ways of coping with stress, to form a lifestyle data set related to health status; combine patients' basic health information, such as medical history and family history, to build a comprehensive database that comprehensively records the overall health status of individuals and provides the necessary background information for further analysis; use data mining technology to conduct a comprehensive survey of medication treatment An in-depth analysis of compliance and its relationship with patient health is aimed at identifying key factors that may affect the effectiveness of drug treatment; statistical methods are used to study the relationship between lifestyle and drug efficacy and adverse reactions, and to explore the potential connection between lifestyle habits and drug safety; based on the above analysis, a prediction model is established, which can evaluate the possibility of different lifestyles affecting patients' drug responses, and then the model is calibrated and verified to ensure its accuracy; based on the results predicted by the model, personalized health management suggestions are provided to patients to help them make adjustments in drug use and encourage them to improve their lifestyle habits; dynamic monitoring of patients' drug treatment compliance and lifestyle changes is achieved, and these data are reviewed regularly to continuously optimize the model's prediction accuracy and adjust health management measures accordingly.

[0024] Furthermore, the intelligent voice assistant designed for medication reminder and monitoring can provide condition tracking, consulting services, etc., to help patients use a variety of drugs correctly, including:

[0025] The disease tracking system collects patients' health and drug use records to monitor and analyze the patient's disease progression and drug use patterns. By comparing the health data stored in the patient's medical records with the drug characteristics in the drug information database, the system automatically checks the potential interactions of current medications to prevent unsafe drug combinations. Based on the results of the drug interaction analysis, the system uses a compound pharmacology algorithm to calculate safe and effective drug dosages and adjusts personalized medication reminder plans for patients to ensure the accuracy of medication timing and dosage. The intelligent voice interaction function uses voice recognition technology to allow patients to verbally query their medication records and doctor's orders for quick self-management. When patients have questions about medications or diseases, they can automatically Connect to relevant medical professionals to provide instant professional consulting services; extract appropriate information from the medication guidance knowledge base to educate patients on the correct use of drugs and enhance their ability to follow medical advice; use the disease management statistics function to regularly generate disease progress reports containing patient disease progression information, and adjust the patient's treatment plan based on these reports to help doctors evaluate the effectiveness of treatment; synchronize patient feedback and health data in real time to ensure that all relevant medical information is kept up to date, so that medical staff can effectively monitor and intervene in a timely manner; continuously monitor system performance, evaluate potential risks or existing problems, and regularly optimize system functions based on collected data and user feedback to continuously improve service accuracy and user experience.

[0026] Furthermore, the patient's hemostasis parameters are re-tested, the changes in blood viscosity before and after multi-drug treatment are evaluated, and a database is established to collect effective cases to provide support for new drug research and development, including:

[0027] In order to accurately monitor the changes in patients' hemostatic parameters, we first deployed an automated blood analysis system to track patients' hemostatic parameters to ensure that detailed data before and after treatment were collected. We used the collected hemostatic parameters as the basis for research and used hemodynamic simulation software to perform cortical analysis of blood status before and after multi-drug treatment in order to understand the specific changes in blood viscosity. To improve data management efficiency, we built an effective database architecture to integrate hemostatic parameters and blood viscosity change data to ensure the integrity and consistency of these key information. We used data inspiration and data mining algorithms for optimization to provide real-time and effective decision support for ongoing new drug research and development. We defined and completed Improve the efficacy evaluation criteria, and use them to analyze and verify the actual effects of the treatment plans recorded in the database; use the defined efficacy evaluation criteria combined with statistical methods to further analyze the correlation between hemostasis parameters and blood viscosity before and after treatment to identify significant differences; make necessary adjustments to the existing multi-drug treatment plans based on the data trends obtained from the analysis to improve the accuracy and pertinence of the treatment plans; implement a dynamic update mechanism to ensure that the latest research findings, treatment results and related parameters can be reflected in the database in real time; establish a regular reporting system to automatically generate a comprehensive treatment effect report that includes treatment plan adjustment suggestions and changes in patient hemostasis parameters.

[0028] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0029] The present invention discloses a method for predicting the response of multi-drug treatment based on changes in blood viscosity. By introducing a fully automatic blood viscosity monitor, the problem of static monitoring of traditional blood parameters is solved, and dynamic detection of blood rheology parameters is realized, providing a more accurate basis for formulating multi-drug treatment plans. Secondly, by constructing a large sample database and applying deep learning technology, the problem of insufficient comprehensive consideration of individual differences and multi-drug relationships in the existing efficacy prediction model is solved, and a more accurate prognosis prediction model is established. Most importantly, the present invention designs an intelligent multi-drug compatibility system, which, combined with a machine learning algorithm, solves the problem of insufficient intelligence of the current compatibility system, provides patients with personalized treatment plans, and solves the safety and patient compliance problems during the treatment process through real-time monitoring of vital signs and intelligent voice assistants. Overall, the present invention effectively solves the problems existing in the individualization, dynamism, and intelligence of multi-drug treatment, and provides a comprehensive solution for improving treatment effects and patient quality of life. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 The present invention is a flowchart of the method for predicting multi-drug treatment response based on blood viscosity changes.

[0031] Figure 2It is a schematic diagram of the method for predicting multi-drug treatment response based on blood viscosity changes of the present invention.

[0032] Figure 3 This is another schematic diagram of the method for predicting multi-drug treatment response based on blood viscosity changes of the present invention. DETAILED DESCRIPTION

[0033] In order to allow those skilled in the art to understand the present invention more clearly and intuitively, the present invention will be further described below in conjunction with the accompanying drawings.

[0034] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this specification.

[0035] In this embodiment, a method for predicting multi-drug treatment response based on blood viscosity change may specifically include:

[0036] Step S101, dynamically detect blood rheology parameters through a fully automatic blood viscosity monitor, evaluate changes in blood viscosity, and provide a basis for formulating a multi-drug treatment plan.

[0037] Patient blood samples are tested using an automated blood viscosity analyzer to collect data and assess blood fluidity. Blood rheology parameters are analyzed and compared to determine differences between samples and set standard ranges for abnormal values. If blood viscosity is detected to be out of the standard range, a support vector machine algorithm is used to predict possible clinical diseases that cause the abnormality. Based on the disease prediction results, the medical drug database is accessed to obtain information on targeted therapeutic drugs. If the prediction results show that there are multiple treatment options, a decision tree algorithm is used to analyze the effectiveness and interaction between various drugs. The algorithm is used to analyze the various drug compatibility options that are distinguished, and their effects on blood viscosity are calculated to obtain the optimal combination. The optimal drug combination will be compared and analyzed with historical case data, and the final personalized treatment plan will be determined with the assistance of an artificial intelligence system. The selected treatment plan is entered into the drug management system, and the system will automatically generate a personalized drug use plan and a corresponding early warning mechanism. After the drug management plan is implemented, the patient's blood viscosity is monitored again using an automated blood viscosity analyzer to verify the continuity and stability of the treatment effect.

[0038] For example, in medical practice, it is critical to accurately measure the rheological parameters of a patient's blood through automated equipment. For example, using an automatic blood viscosity analyzer, the viscosity of a blood sample can be measured to be 3.5 cP (millipascal seconds), and the normal range is 2.5-3.5 cP. When a patient's blood viscosity is measured to be 4.2 cP, that is, beyond the normal range, a support vector machine (SVM) may be used to compare the data point with the pathological data in the training set to predict possible disease factors, such as whether the patient suffers from hypertension or polycythemia. On this basis, the artificial intelligence system can be linked to a medical drug database to collect drug information that matches the predicted results. For example, if hypertension is predicted, drug information such as ACE inhibitors and diuretics may be extracted. Subsequently, a decision tree algorithm is used to analyze the effectiveness and interactions of different drugs. Taking diuretics and ACE inhibitors as examples, a decision tree is constructed, whose nodes represent drug selection and edges represent parameters that affect the decision, such as efficacy, side effects, etc. The algorithm calculates the final score of each path based on the utility function and the patient's specific conditions (age, weight, comorbidities, etc.) to make the best choice. Next, the selected drug combination is analyzed in depth. Quantitative pharmacological models, such as the Hill equation, are used to estimate the effect of drug combinations on blood viscosity. For example, for ACE inhibitors, their effect on blood viscosity may be described by the Hill equation V = Vmax*^n / (Kd^n+^n), where V is the effect of the drug, Vmax is the maximum potential effect of the drug, is the drug concentration, Kd is the half-maximal effect drug concentration, and n is the slope of the drug response curve. According to the drug response data of the subjects, the parameters in the Hill equation are adjusted to predict the efficacy of the optimal drug combination. Thereafter, by comparing with historical case data, machine learning techniques such as the k-nearest neighbor (k-NN) algorithm can be used to refine the treatment plan. The system analyzes the treatment responses in historical cases that are similar to the predicted situation and optimizes personalized treatment plans. The final optimized treatment plan will be incorporated into the drug management system. The system will automatically generate a drug use plan for the patient and establish an early warning mechanism, such as setting time points for regular blood pressure or blood viscosity testing. This monitoring can rely on the patient's biomarkers. For example, if the blood viscosity is higher than 4.0cP for three consecutive days, the system will automatically send a reminder to visit the doctor. After the implementation of the drug management plan, the continuity and stability of the treatment effect should be tracked. Real-time monitoring data will be collected by an automatic blood viscosity analyzer, and the algorithm analysis will determine the treatment effect. The patient's blood viscosity data reflected in the results will be compared with the data before and after treatment. For example, after treatment, the patient's blood viscosity value is stable at around 3.0cP, which proves the effectiveness and safety of the treatment plan.

[0039] Step S102, construct a large sample database containing patients of different ages, genders, and underlying diseases, collect their clinical data such as treatment response and adverse reactions during multi-drug treatment, use deep learning technology to analyze the characteristics and outcomes of these patients, and establish an accurate prognosis prediction model.

[0040] Through the construction of a professional database, the patient's age, gender and existing underlying diseases are recorded as an important basis for primary screening of patients. The key to processing the collected clinical data using data cleaning technology is to extract information such as the patient's treatment drugs, reactions to these drugs, and possible adverse reactions. The cleaned data is classified, and multiple sub-databases can be created based on the accepted treatment plan to analyze the reactions caused by different drug treatments. Through text mining technology, key narratives describing the treatment effect and detailed information about complications are deeply mined and sorted out in the clinical data. Using feature selection algorithms, combined with in-depth analysis of patient characteristics, feature factors that are closely related to the treatment effect and have a great influence are screened out. These screened patient characteristics and treatment effect data are input into the deep learning model, which is used as the basis for the first training of the model and predicting the patient's prognosis. In order to ensure the stability and accuracy of the prognosis prediction model, the cross-validation method is used for evaluation, which can ensure that the model meets the statistical requirements of the study. After the model evaluation meets expectations, it is further optimized and adjusted. Here, incremental learning technology should be used to improve the prediction accuracy of the model while continuously introducing new clinical data. The optimized prognostic prediction model is deployed in the clinical decision support system, so that the accurate prognostic information provided by the model can be used to help formulate personalized treatment plans for individual patients.

[0041] Exemplarily, when constructing a patient database, SQL language is used for structured query, and a table is created to store information such as the patient's age, gender, and medical history, for example, using "CREATE TABLE Patients(PatientIDINT, AgeINT, GenderCHAR(1), Disease History VARCHAR(255));". Subsequently, the collected clinical data are cleaned, and missing values ​​and outliers are removed using Python's pandas library, for example, using the `df.dropna()` function to delete missing data. The valid information after data cleaning, such as the patient's response to a certain drug A, is extracted and classified into a sub-database, such as "DataFrame == 'DrugA']". Next, text mining tools in NLP technology are applied, such as using the NLTK library or the spaCy library, to analyze the free text in the clinical data, extract key words that describe the treatment effect, such as "significant efficacy", and encode this information into a data format that can be used for subsequent calculations. In the feature selection phase, the feature importance evaluation method in machine learning algorithms such as logistic regression or decision trees is used, such as using sklearn's "SelectFromModel" class with "LogisticRegression" to score and screen the importance of features. Then load the selected feature data into a deep learning framework such as TensorFlow or PyTorch to build a neural network model. For example, design a multi-layer perceptron model with three hidden layers, set a reasonable number of neurons such as 64-128-64, and use the ReLU activation function for nonlinear transformation. Perform the first training with the prepared dataset, and use the mean square error (MSE) as the loss function to measure the prediction accuracy, and continuously optimize the model weights through the back propagation algorithm. After that, introduce cross-validation technology to verify the model, and you may use K-fold cross-validation to randomly divide the dataset into K subsets, use one of the subsets as the validation set, and the rest as the training set to evaluate the generalization ability of the model. As the evaluation results accumulate, you can calculate the overall performance indicators of the model, such as accuracy, recall, and F1 score, to ensure that they have reached an acceptable level. In order to achieve continuous optimization of the model, new data can be gradually introduced using online learning or batch learning methods to adjust model parameters. For example, when new patient data is added, the model can be adapted to the latest data features by adjusting hyperparameters such as the learning rate, such as "optimizer = tf.train.AdamOptimizer (learning_rate = 0.001).minimize (loss)". Here, the Adam optimizer is used to adjust the learning rate. Finally, the trained and optimized deep learning model is embedded in the clinical decision support system, which receives real-time patient information through the system interface and uses the model to predict prognosis.The system will assist doctors in developing more personalized and accurate treatment plans based on the probability output predicted by the model, such as predicting that a patient's 5-year survival rate under drug A treatment is 75%.

[0042] Step S103, design an intelligent multi-drug compatibility system, use machine learning algorithms to predict the efficacy and risks of different medication modes based on patient individual parameters, disease severity and other factors, and provide decision support for multi-drug treatment plans.

[0043] Collect individual parameters such as the patient's medical history and vital signs, and integrate this information to ensure the accuracy of each data. Use convolutional neural network algorithms to analyze historical medical data, including treatment effects for similar cases. Based on this data, build a primary drug compatibility model, define an initial weight score for each drug based on the patient's individual parameters and relevant medical literature and research, and this score reflects the possible effectiveness and risk level of the drug. By further adjusting these initial weights, the selection and compatibility of drugs can be optimized until a combination suitable for a specific patient is found. Perform risk assessment based on the severity of the patient's current condition to help determine whether medication is urgently needed and prevent possible side effects in advance. Apply convolutional neural network algorithms to predict and analyze the potential efficacy under different medication modes, and propose multiple treatment options. Evaluate the long-term and short-term risks of these options, and conduct in-depth analysis and comparison to identify the risks that patients may face. To ensure treatment safety, use the drug interaction database to check the drug compatibility in these options to avoid potential adverse drug reactions. Based on the feedback from real-time medical monitoring data, continuously analyze the changes in physical signs during treatment, and adjust the drug regimen and dosage accordingly. Using expert systems as decision support tools, we combine expert knowledge with algorithm-generated solutions to provide a tailored multi-drug treatment plan. We also design a feedback mechanism to continuously track the treatment effects of patients, optimize the convolutional neural network algorithm through actual results, and improve the accuracy and safety of the treatment plan.

[0044] For example, in the process of building a personalized treatment plan, the patient's medical history is first collected in detail. For example, the historical blood pressure record of patient J can be represented by a numerical list: [120, 118, 122, 119, 124] mmHg. At the same time, J's vital signs are monitored, and the heart rate is recorded as [78, 82, 81, 77, 80] bpm. A deep learning algorithm such as a convolutional neural network (CNN) is used to train the collected data and historical medical data of similar patients to extract features for predicting treatment effects. When conducting a quantitative risk assessment of J's condition, the cardiac risk score formula can be used: (R = W\timesH), where W is the body weight, standardized to a certain standard body weight of 1, and H is the systolic high blood pressure. After standardization, the normal systolic blood pressure is 1. The deep learning algorithm is used to predict the efficacy of different treatment plans and score the treatment effects under different drug combinations. The higher the score, the better the efficacy. For example, the combination of drugs A and B may get an effect score of 0.9, and the combination of drugs A and C is 0.85. This score can be generated by a model trained with historical data. Monte Carlo simulation is further used to evaluate the long-term and short-term risks of the treatment plan. Considering that long-term use of drug A may increase the probability of renal impairment, 10,000 patients can be simulated for drug A treatment. The simulation results show that 2% of patients have renal impairment. In terms of drug compatibility, the drug interaction database can be used to query the interaction index between different drugs, such as the CYP450 enzyme impact index. If the index is greater than a certain threshold, such as 1.5, it is considered that there may be a significant interaction. For J's treatment plan, if the CYP450 enzyme impact index of drugs A and B is 1.2, and A and C is 1.6, then the combination of drugs A and B is selected to reduce the risk of interaction. As the treatment progresses, the real-time medical monitoring data reflects that J's heart rate has increased to 88bpm. According to the preset decision tree model, if the heart rate continues to be higher than 85bpm, the dose of drug A is reduced to 80% of the original. Drug combinations can be further adjusted through decision support from expert systems, combined with the opinions of clinical experts and model predictions. If an expert system recommends that the use of drug D can reduce the probability of cardiac events in patients at high risk for heart disease, this information should be considered in the treatment plan even if it is not directly related to the treatment effect score. During the treatment process, the designed feedback mechanism can fine-tune the deep learning algorithm based on the treatment effect. If a deviation between the drug combination effect score and the actual response is observed during the treatment of J, the algorithm can be retrained based on J's response data to optimize the prediction model.

[0045] Step S104, using wearable sensors to monitor the patient's vital signs changes in real time during multi-drug treatment, evaluate the safety of different medication regimens, and adjust the medication plan as needed in conjunction with the above-mentioned blood viscosity monitoring results.

[0046] Wearable sensor technology is used to monitor the patient's vital signs and collect real-time data. The data is processed by the intelligent data analysis system to obtain detailed physiological parameters and viscosity monitoring results. The collected data is transmitted to the data storage and management module in real time, which is responsible for recording the patient's health history information and using this information as an auxiliary basis for the clinical decision support system. The personalized medical strategy module analyzes the monitoring data according to the patient's specific situation and determines whether the current medication regimen needs to be adjusted to ensure the best treatment effect. When the monitoring data shows that the patient has potential symptoms or risks of maladaptation, the risk warning system can promptly issue an alarm to the doctor or patient. The safety assessment module continuously evaluates the patient's real-time vital signs data to determine whether there are adverse reactions caused by drugs. If the evaluation results show potential risks, the medication plan adjustment module is activated to re-formulate the treatment plan, and the multi-drug treatment analysis module is used to evaluate the interaction between different drugs to optimize the treatment strategy. The treatment effect monitoring module continuously tracks the new treatment plan to evaluate the efficacy of the plan and ensure that the treatment goals are achieved. If the efficacy monitoring module reports that the efficacy is poor, the medication plan adjustment module will be activated again and continuously optimized until the most suitable treatment plan for the patient is found.

[0047] Exemplarily, when monitoring vital signs, the patient's heart rate data recorded by the wearable sensor is a series of time series values, and the average heart rate per minute is calculated using a sliding window algorithm. By summing the number of heartbeats in the past 60 seconds and dividing it by 60, the heart rate value per minute is obtained, and the result is 72 times / minute at this moment. Next, the intelligent data analysis system will compare this result with the patient's heart rate baseline. If the deviation exceeds the set threshold, such as 5 times / minute, the data will be marked as abnormal and a risk warning will be triggered. If the patient needs to adjust the dosing regimen, the personalized medical strategy module will apply a drug dosage adjustment algorithm based on the patient's age, gender, weight and other information, as well as the pharmacokinetic properties of the drug, such as the Cockcroft-Gaultequation to calculate the renal clearance rate, and then deduce the individualized drug dosage. This module may recommend reducing the dose of a drug from 5 mg once a day to 4 mg. The safety assessment module continuously monitors vital sign data, such as matching heart rate with known data in the drug side effect database, and applies machine learning algorithms such as support vector machines (SVM) to determine whether there are signs of adverse drug reactions. If the predicted probability exceeds an acceptable threshold, such as 20%, it prompts to adjust the medication plan. After receiving the adjustment signal, the medication plan adjustment module will use the decision tree algorithm to evaluate the current prescription and combine the drug interaction data provided by the multi-drug treatment analysis module, such as using the drug interaction checker, to ensure that the adjusted plan can avoid potential drug interactions. This may involve staggering the administration of two drugs that may interact with each other to reduce interaction, such as adjusting the administration time of drug A from morning to evening. In order to continuously evaluate the efficacy of the new treatment plan, the treatment effect monitoring module will regularly check key physiological parameters and use logistic regression analysis to predict the treatment response to determine whether the expected treatment goal is achieved. For example, for patients with hypertension, based on the average systolic blood pressure readings for three consecutive days, if the decline is less than 10 mmHg, the module will issue a signal of poor efficacy. Finally, if the efficacy monitoring module indicates that further optimization of treatment is needed, the medication plan adjustment module will use the Bayesian network to update the patient's efficacy probability distribution and consider new treatment options. For example, if the current plan is to use drug X, based on the new data obtained, it is inferred that drug Y may bring better effects. By calculating the posterior probabilities of the two plans, if the efficacy probability of drug Y is significantly higher, it is recommended to adjust the treatment plan to drug Y.

[0048] Step S105, construct a knowledge graph of interactions between multiple drugs, collect drug-drug, drug-food, and drug-disease interaction data, and provide a reference for the selection of combinations of multiple drugs.

[0049] Medical literature and drug instructions are analyzed using text mining technology to extract information about the properties and mechanisms of action of drugs. Information in the food ingredient database is used in combination with relevant knowledge of nutrition to determine the impact that specific food ingredients may have on drug efficacy. In-depth analysis of disease-related data is performed to identify the correlation between drug mechanisms of action and disease characteristics. By identifying these correlations, it is possible to know which drugs may have therapeutic or side effect risks for specific diseases. Drug metabolism analysis tools can be used to better understand the metabolic process of drugs in the body, especially the interaction between metabolites and other drugs or nutrients. By analyzing medical guidelines and treatment manuals, medical professionals can be provided with recommendations for the combined use of multiple drugs to help develop the most reasonable combination treatment plan. By tracking drug concentration monitoring data, the impact of combined medication on drug plasma concentration can be observed in real time, and the trend of drug concentration changes under various conditions can be analyzed. In analyzing side effect data, disease characteristics will be considered to determine the risk of side effects that different patient groups may encounter when using specific drugs. Based on the above collected data and previous research results, a drug efficacy impact assessment model is constructed, which aims to predict the interaction between different drug combinations and evaluate how they affect the treatment effect. Integrate the data and evaluation results obtained in all the above steps, and use pattern recognition and machine learning algorithms to deploy a dynamic knowledge graph update mechanism. This mechanism can automatically incorporate new research results and real-time data, thereby continuously improving the knowledge graph and providing medical workers with scientific decision-making support on drug combination therapy.

[0050] For example, when using text mining technology to analyze medical literature and drug instructions, natural language processing (NLP) algorithms such as entity recognition (NER) can be used to identify key information such as drug name, dosage and mechanism of action. For example, a model based on conditional random fields (CRF) can be used to extract information about "aspirin", "antiplatelet" and "reducing myocardial infarction" from a paper about aspirin. Next, combined with food ingredient information in a database, such as the food ingredient database of the United States Department of Agriculture (USDA), it can be determined that specific ingredients such as furanocoumarins in grapefruit juice inhibit the drug CYP3A4 enzyme, which may affect the metabolic rate of the drug. Through the synergistic analysis of nutritional components and drug effects, a logistic regression model can be used to predict the probability of the food component interacting with a specific drug, and then evaluate its degree of influence. In the correlation analysis of disease characteristics and drug mechanisms of action, genomic association analysis (GWAS) can be used to identify the therapeutic efficiency of drugs in specific disease states based on gene expression profile data and disease biomarkers. For example, for statins, inhibitors of HMG-CoA reductase, a key enzyme in the cholesterol synthesis process, GWAS analysis can find that patients with certain genetic variants may respond better to statins. In terms of in-depth understanding of drug metabolism, pharmacokinetic (PK) models, such as nonlinear mixed effects models (NONMEM), can be used to simulate and predict the absorption, distribution, metabolism and excretion (ADME) processes of drugs in different individuals. For example, this model can be used to predict that when drugs A and B coexist in the same patient, the half-life of drug A increases from 8 hours to 10 hours. When analyzing drug side effect data, the information in the Adverse Event Reporting System (FAERS) database is used to predict and classify the severity of patients' side effects to a certain drug using the back propagation neural network (BPNN) model. For example, it may be found that there is a 25% probability of hypoglycemia in elderly diabetic patients receiving a blood pressure drug. The construction of the drug effect evaluation model can combine multiple models and data, including the aforementioned food-drug, disease-drug, drug metabolism model and drug side effect model, and use ensemble learning methods such as random forest models to predict the therapeutic effect and side effect risk of different drug combinations. When drug A and drug B are used together, the model calculates that the risk of side effects of combined use increases by 15%, while the therapeutic effect increases by 10%. Finally, the construction and update of the knowledge graph can be achieved using the graph database Neo4j, combining the above analysis results with the built-in Cypher query language of the graph database to optimize the connection between drugs. For example, when updating the graph, drug interactions discovered in new studies will be automatically added, and the therapeutic effect improvement index when drug A and drug B are used together will be calculated, and this will be used as a basis to provide decision support for medical professionals.The knowledge graph will integrate various models and database information, as well as real-time clinical data, to form a self-updating and dynamic decision-making support system to improve the safety and effectiveness of drug combinations. Through this series of comprehensive data analysis and automated models, a more sophisticated and self-learning medical information technology framework can be built.

[0051] Step S106, collect the patient's compliance and lifestyle information during multi-drug treatment, establish a model to evaluate the correlation between these factors and medication efficacy and adverse reactions, and provide health management suggestions.

[0052] Collect detailed information about patients' medication treatment, including the name of the drug, dosage, frequency of medication, and specific time period of medication, to establish a basic database for subsequent analysis; collect data on patients' compliance with medical treatment, including whether patients take medications on time according to doctor's orders, monitor and mark incidents of missed or wrong medications, which provides direct evidence for evaluating the actual implementation of medication treatment; conduct a comprehensive survey of patients' lifestyles, including evaluation and recording of diet, exercise, sleep, and ways of coping with stress, to form a lifestyle data set related to health status; combine patients' basic health information, such as medical history and family history, to build a comprehensive database that comprehensively records the overall health status of individuals and provides the necessary background information for further analysis; use data mining technology to conduct a comprehensive survey of medication treatment An in-depth analysis of compliance and its relationship with patient health is aimed at identifying key factors that may affect the effectiveness of drug treatment; statistical methods are used to study the relationship between lifestyle and drug efficacy and adverse reactions, and to explore the potential connection between lifestyle habits and drug safety; based on the above analysis, a prediction model is established, which can evaluate the possibility of different lifestyles affecting patients' drug responses, and then the model is calibrated and verified to ensure its accuracy; based on the results predicted by the model, personalized health management suggestions are provided to patients to help them make adjustments in drug use and encourage them to improve their lifestyle habits; dynamic monitoring of patients' drug treatment compliance and lifestyle changes is achieved, and these data are reviewed regularly to continuously optimize the model's prediction accuracy and adjust health management measures accordingly.

[0053] Exemplary, in the process of establishing a basic database, for example, an electronic health record system may be used to automatically capture and record the patient's medication information, including the use of a certain antihypertensive drug, such as lovastatin 20mg, once a day in the evening. Subsequently, data on compliance behavior can be collected using smart medicine boxes, which can record the time when the drug is taken out and compare it with the scheduled medication time, and automatically mark missed medication events, such as one day in five consecutive days when the patient did not take the medicine at 8 pm. Further, through patient self-reporting and wearable devices to monitor their lifestyle, it can be learned that the patient's average number of steps per week is 6,000 steps, and he insists on 7 hours of sleep and a high-fat diet every night. Combined with disease history, such as a 10-year history of hypertension and a family history of heart disease, a comprehensive database containing comprehensive health information of individuals is obtained. Through data mining tools and algorithms, such as logistic regression analysis, the relationship between drug compliance and patient health indicators can be analyzed. For example, the model may show that for every 1% decrease in medication compliance, systolic blood pressure increases by an average of 0.5 mmHg. Using correlation analysis in statistics, we can study the relationship between lifestyle and drug efficacy. For example, we found that the increase in exercise is positively correlated with the antihypertensive effect of drugs, with a correlation coefficient of 0.3. Next, based on the collected data and analysis results, a machine learning model, such as random forest, is established to predict the impact of different lifestyles on drug response. Through cross-validation, the performance of the model on different data sets is checked to adjust the parameters and improve the accuracy and generalization ability of the model. For example, the model accuracy reaches 80% and the F1 score is 0.75. Based on the model, personalized health management suggestions are provided. For example, for patients who do not exercise enough, it is recommended to gradually increase to 8,000 steps per day and monitor their drug efficacy changes. Finally, big data technology is used to monitor patients' long-term drug treatment compliance and lifestyle changes in real time, and the prediction model is continuously optimized through deep learning methods. The model is recalibrated every quarter to adapt to new patient data and treatment response patterns. For example, the model weight is adjusted to increase the impact of medication compliance by 5% to reflect its significance in treatment effect.

[0054] Step S107, design an intelligent voice assistant for medication reminder and monitoring, conduct disease tracking, consulting services, etc., to help patients use multiple drugs correctly.

[0055] The patient's health and medication records are collected through the disease tracking system to monitor and analyze the patient's disease progression and drug usage patterns. Using the health data stored in the patient's medical file and comparing the drug characteristics in the drug information database, the system automatically checks the potential interactions of current medications to prevent unsafe drug combinations. Based on the analysis results of drug interactions, the compound pharmacology algorithm is used to calculate safe and effective drug dosages, and a personalized medication reminder plan is adjusted for the patient to ensure the accuracy of medication timing and dosage. The intelligent voice interaction function allows patients to verbally query their medication records and medical advice through voice recognition technology to achieve quick self-management. When patients have questions about medications or diseases, they are automatically connected to relevant medical professionals through the consulting service interface to provide instant professional consulting services. Appropriate information is extracted from the medication guidance knowledge base to educate patients on the correct use of drugs and enhance their ability to follow medical advice. Using the disease management statistics function, disease progress reports containing patient disease progression information are regularly generated, and patient treatment plans are adjusted based on these reports to help doctors evaluate treatment effectiveness. Real-time synchronization of patient feedback and health data ensures that all relevant medical information is kept up to date, allowing medical staff to conduct effective monitoring and timely intervention. Continuously monitor system performance, assess potential risks or existing problems, and regularly optimize system functions based on collected data and user feedback to continuously improve service accuracy and user experience.

[0056] For example, patient A weighs 60 kg, is 7 meters tall, and is 35 years old. Patient A is currently taking a drug called "aspirin" at a dose of 100 mg per day. Doctor B has told patient A that this drug may cause some side effects, including gastrointestinal discomfort and bleeding. The following health data is collected: weight is 60 kg, height is 1.7 meters, age is 35 years old, drug name is aspirin, and drug dose is 100 mg per day. Use the drug characteristics in the drug information database to check the potential interactions of current medications. The drug information database shows that aspirin may cause the following side effects: gastrointestinal discomfort includes nausea, vomiting, stomach pain, constipation, etc., and bleeding includes bleeding tendency, hematuria, etc. Calculate the potential interaction risk of current medications to help patients prevent possible side effects. According to the calculation results, it is recommended that patient A add gastrointestinal protection measures when taking aspirin, such as daily use of proton pump inhibitors to relieve gastrointestinal discomfort. According to the analysis results of drug interactions, calculate safe and effective drug doses. For example, patient A needs to take 80 mg of aspirin per day to achieve the therapeutic effect. Check whether the current dosage is safe and effective, and automatically adjust the dosage of patient A based on the calculation results. Finally, generate a course report containing information about the patient's condition progress based on the patient's condition progress, and adjust the patient's treatment plan based on these reports. If patient A's condition does not have any serious side effects during the medication period, doctor B may consider increasing patient A's dosage to achieve better treatment results.

[0057] Step S108, re-test the patient's hemostasis parameters, evaluate the changes in blood viscosity before and after multi-drug treatment, establish a database to collect effective cases, and provide support for new drug research and development.

[0058] In order to accurately monitor the changes in patients' hemostatic parameters, an automated blood analysis system is first deployed to track patients' hemostatic parameters to ensure that detailed data before and after treatment are collected. Using the collected hemostatic parameters as the basis for research, the blood state before and after multi-drug treatment is analyzed at the cortical level through hemodynamic simulation software to grasp the specific changes in blood viscosity. In order to improve data management efficiency, an effective database architecture is constructed to integrate hemostatic parameters and blood viscosity change data to ensure the integrity and consistency of these key information. Inspired by existing data, data mining algorithms are applied for optimization to provide real-time and effective decision support for ongoing new drug research and development. Define and improve efficacy evaluation criteria, and use them to analyze and verify the actual effects of treatment plans recorded in the database. Using the defined efficacy evaluation criteria combined with statistical methods, further analyze the correlation between hemostatic parameters and blood viscosity before and after treatment to identify significant differences. According to the data trends obtained from the analysis, necessary adjustments are made to the existing multi-drug treatment plans to improve the accuracy and pertinence of the treatment plans. Implement a dynamic update mechanism to ensure that the latest research findings, treatment results and related parameters can be reflected in the database in real time. Establish a regular reporting system to automatically generate comprehensive treatment effectiveness reports containing treatment adjustment recommendations and changes in patient hemostasis parameters.

[0059] Exemplarily, in order to accurately monitor the hemostasis parameters of patients, an automated blood analysis system is used, such as a coagulation function monitor used by patients treated with anticoagulant drugs, such as DDAVP, which can continuously track and record the patient's coagulation time (CT) and fibrinogen time (FT), so that detailed data can be obtained by comparing the CT and FT values ​​before and after treatment. For example, patient A had a CT of 120 seconds and a FT of 600 seconds before treatment, while the CT dropped to 60 seconds and the FT dropped to 400 seconds after treatment, thus proving the effectiveness of the treatment. After collecting these coagulation parameters, the blood state of the patient before and after treatment was simulated and analyzed by hemodynamic simulation software, such as Hemodyn. By inputting the patient's blood parameters and rheological properties, such as viscosity and density, the software can display the changes in blood flow state in a computer simulation. By analyzing the blood flow velocity diagram simulated by the software before and after treatment, specific changes in viscosity can be detected, such as a drop from 4.5 cP to 2.8 cP. In order to efficiently manage these data, a database architecture is constructed to integrate all coagulation parameters and blood viscosity data. SQL statements were used to establish associations, such as the "JOIN" command to associate various parameters, ensuring the integrity and consistency of the data. For example, a "hemodynamics" table was created, which contained fields such as patient ID, treatment time point, CT, FT, and viscosity, and data association was maintained through patient ID. Based on these data, data mining algorithms, such as random forest classifiers, were used for optimization. On this basis, a real-time decision support system was developed that could provide feedback to the drug development team based on real-time data. For example, the system might indicate that when viscosity drops below a certain threshold, reducing the dose of anticoagulants can achieve the same therapeutic effect. New efficacy evaluation criteria were defined, such as setting a viscosity drop of more than 5% as a significant therapeutic effect, and then analyzing the actual effects of the treatment plans recorded in the database. In this way, the accuracy and applicability of the data collected in the database in the treatment strategy can be verified. Statistical methods, such as Pearson correlation analysis, were used to further study the relationship between hemostatic parameters and blood viscosity. The correlation coefficients between patient hemostatic parameters and blood viscosity were analyzed, and it was found that their correlation was above 0.85, indicating that the two had a significant positive correlation. According to the trend of data analysis, the existing treatment plan was adjusted. For example, if improvements in coagulation parameters are observed but viscosity decreases are not significant, the dose of anticoagulants may be reduced while increasing drugs that promote blood circulation. Finally, a dynamic update mechanism is implemented, such as by writing automated scripts to periodically extract data from experimental instruments and update the database with new research findings, treatment results, and parameter changes in real time.

[0060] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for predicting multi-drug treatment response based on blood viscosity changes, characterized in that: The steps include: The blood rheology parameters are dynamically detected by the fully automatic blood viscosity monitor to evaluate the changes in blood viscosity and provide a basis for formulating multi-drug treatment plans; include: Use an automatic blood viscosity analyzer to test patient blood samples to collect data and evaluate blood fluidity; analyze blood rheology parameters, compare to determine differences between samples and set standard ranges for abnormal values; if blood viscosity is detected to deviate from the standard range, use a support vector machine algorithm to predict possible clinical diseases that cause the abnormality; based on the disease prediction results, access the medical drug database to obtain information on targeted therapeutic drugs; if the prediction results show that there are multiple treatment options, use a decision tree algorithm to analyze the effectiveness and interactions between various drugs; use an algorithm to analyze the various drug compatibility options that have been distinguished, calculate their effects on blood viscosity, and derive the optimal combination; the optimal drug combination will be compared and analyzed with historical case data, and the final personalized treatment plan will be determined with the assistance of an artificial intelligence system; the screened treatment plan will be entered into the drug management system, and the system will automatically generate a personalized drug use plan and a corresponding early warning mechanism; Construct a large sample database of patients of different ages, genders, and underlying diseases, collect clinical data on their treatment responses and adverse reactions during multi-drug treatment, use deep learning technology to analyze the characteristics and outcomes of these patients, and establish an accurate prognosis prediction model; including: Through the construction of a professional database, the patient's age, gender and existing underlying diseases are recorded as an important basis for primary screening of patients; the collected clinical data are processed using data cleaning technology, the key is to extract information such as the patient's treatment drugs, reactions to these drugs, and possible adverse reactions; the cleaned data are classified, and based on the accepted treatment plan, multiple sub-databases can be created to analyze the reactions caused by different drug treatments; through text mining technology, the key narratives describing the treatment effects and detailed information about complications are deeply mined and sorted out in the clinical data; the feature selection algorithm is used, combined with an in-depth analysis of patient characteristics, to screen out feature factors that are closely related to the treatment effect and have a great influence; these screened patient characteristics and treatment effect data are input into the deep learning model, which is used as the basis for the initial training of the model and the prediction of the patient's prognosis; Design an intelligent multi-drug compatibility system, use machine learning algorithms to predict the efficacy and risks of different medication modes based on patient individual parameters and disease severity, and provide decision support for multi-drug treatment plans; including: Collect individual parameters of the patient's medical history and vital signs, and integrate this information to ensure the accuracy of each data; use convolutional neural network algorithms to analyze historical medical data, including treatment effects for similar cases; build a primary drug compatibility model based on these data, and define an initial weight score for each drug based on the patient's individual parameters and relevant medical literature and research. This score reflects the possible effectiveness and risk level of the drug. By further adjusting these initial weights, the selection and compatibility of drugs can be optimized until a combination suitable for a specific patient is found; conduct risk assessment based on the severity of the patient's current condition to help determine whether medication is urgently needed and prevent possible side effects in advance; apply convolutional neural network algorithms to predict and analyze the potential efficacy under different medication modes and propose multiple treatment plans; evaluate the long-term and short-term risks of these plans, conduct in-depth analysis and comparison to identify the risks that patients may face; use the drug interaction database to check the drug compatibility in these plans to ensure treatment safety; based on the feedback of real-time medical monitoring data, continuously analyze the changes in signs during treatment and adjust the drug plan and dosage accordingly; use expert systems as decision support tools to combine expert knowledge with algorithm-generated plans to provide a tailored multi-drug treatment plan; Wearable sensors are used to monitor the patient's vital signs in real time during multi-drug treatment, evaluate the safety of different dosing regimens, and adjust the medication plan as needed in conjunction with the above-mentioned blood viscosity monitoring results; including: Wearable sensor technology is used to monitor the patient's vital signs and collect real-time data. The data is processed by an intelligent data analysis system to obtain detailed physiological parameters and viscosity monitoring results; the collected data is transmitted in real time to the data storage and management module, which is responsible for recording the patient's health history information and using this information as an auxiliary basis for the clinical decision support system; the personalized medical strategy module analyzes the monitoring data according to the patient's specific situation and determines whether the current medication regimen needs to be adjusted to ensure the best treatment effect; when the monitoring data shows that the patient has potential symptoms or risks of maladaptation, the risk warning system can promptly issue an alarm to the doctor or patient; the safety assessment module continuously evaluates the patient's real-time vital signs data to determine whether there are adverse reactions caused by drugs; if the assessment results show potential risks, the medication plan adjustment module is activated to re-formulate the treatment plan, and the multi-drug treatment analysis module is used to evaluate the interaction between different drugs to optimize the treatment strategy; the treatment effect monitoring module continuously tracks the new treatment plan to evaluate the efficacy of the plan and ensure that the treatment goals are achieved; Construct a knowledge graph of drug-drug interactions, collect drug-drug, drug-food, and drug-disease interaction data, and provide a reference for the selection of drug combinations for multi-drug combination therapy; Collect information on patients' compliance and lifestyle during multi-drug treatment, establish a model to evaluate the correlation between these factors and medication efficacy and adverse reactions, and provide health management advice; including: Collect detailed information about patients' medication treatment, including the name of the drug, dosage, frequency of medication, and specific time period of medication, to establish a basic database for subsequent analysis; collect data on patients' compliance with medical treatment, including whether patients take medication on time according to doctor's orders, monitor and mark incidents of missed or wrong medications; conduct a comprehensive survey of patients' lifestyles, including evaluation and recording of diet, exercise, sleep, and ways of coping with stress, to form a lifestyle data set related to health status; combine patients' basic health information to build a comprehensive database that comprehensively records the overall health status of individuals; use data mining technology to conduct in-depth analysis of medication compliance and its relationship with patient health, aiming to identify Identify key factors that may affect the effectiveness of drug treatment; use statistical methods to study the relationship between lifestyle and drug efficacy and adverse reactions, and explore the potential connection between lifestyle habits and drug safety; based on the above analysis, establish a prediction model that can evaluate the possibility of different lifestyles affecting patients' drug responses, and then perform calibration and verification operations on the model to ensure its accuracy; provide patients with personalized health management suggestions based on the results predicted by the model, help patients make adjustments in drug use, and encourage them to improve their lifestyle habits; realize dynamic monitoring of patients' drug treatment compliance and lifestyle changes, review these data regularly to continuously optimize the prediction accuracy of the model, and adjust health management measures accordingly; Design an intelligent voice assistant for medication reminders and monitoring, provide condition tracking and consulting services, and help patients use multiple medications correctly; including: The patient's health and drug use records are collected through the disease tracking system to monitor and analyze the patient's disease progression and drug use patterns. The system automatically checks the potential interactions of current medications by comparing the health data stored in the patient's medical records with the drug characteristics in the drug information database to prevent unsafe drug combinations. Based on the analysis results of drug interactions, the compound pharmacology algorithm is used to calculate safe and effective drug dosages, and a personalized medication reminder plan is adjusted for the patient to ensure the accuracy of medication timing and dosage. The intelligent voice interaction function allows patients to verbally query their medication records and medical advice through voice recognition technology to achieve quick self-management. When patients have questions about drugs or diseases, they are automatically connected to relevant medical professionals through the consulting service interface to provide instant professional consulting services. Appropriate information is extracted from the medication guidance knowledge base to educate patients on the correct use of drugs and enhance their ability to follow medical advice. The disease management statistics function is used to regularly generate disease progress reports containing patient disease progression information, and adjust the patient's treatment plan based on these reports to help doctors evaluate the effectiveness of treatment. Re-test the patient's hemostasis parameters, evaluate the changes in blood viscosity before and after multi-drug treatment, establish a database to collect effective cases, and provide support for new drug research and development; including: First, an automated blood analysis system is deployed to track the patient's hemostatic parameters in order to accurately monitor changes in the patient's hemostatic parameters and ensure that detailed data before and after treatment are collected; using the collected hemostatic parameters as the basis for research, the blood status before and after multi-drug treatment is analyzed at the cortical level through hemodynamic simulation software in order to understand the specific changes in blood viscosity; to improve data management efficiency, an effective database architecture is constructed to integrate hemostatic parameters and blood viscosity change data to ensure the integrity and consistency of these key information; inspired by existing data, data mining algorithms are applied for optimization to provide real-time and effective decision support for ongoing new drug research and development.

2. The method according to claim 1, characterized in that The method of dynamically detecting blood rheological parameters by a fully automatic blood viscosity monitor and evaluating changes in blood viscosity to provide a basis for formulating a multi-drug treatment plan also includes: After the drug management program was implemented, the patient's blood viscosity was monitored again using an automated blood viscosity analyzer to verify the continuity and stability of the treatment effect.

3. The method according to claim 1, characterized in that: The method constructs a large sample database of patients of different ages, genders, and underlying diseases, collects clinical data on their treatment responses and adverse reactions during multi-drug treatment, uses deep learning technology to analyze the characteristics and outcomes of these patients, and establishes an accurate prognosis prediction model: The cross-validation method is used to evaluate the stability and accuracy of the prognosis prediction model, which can ensure that the model meets the statistical requirements of the study; after the model evaluation meets expectations, it is further optimized and adjusted using incremental learning technology to improve the model's prediction accuracy while continuously introducing new clinical data; the optimized prognosis prediction model is deployed in the clinical decision support system to help formulate individualized treatment plans for individual patients based on the accurate prognostic information provided by the model.

4. The method according to claim 1, characterized in that: The intelligent multi-drug combination system is designed to use machine learning algorithms to predict the efficacy and risks of different medication modes based on factors such as individual patient parameters and disease severity, and provide decision support for multi-drug treatment plans. It also includes: designing a feedback mechanism to continuously track the patient's treatment effects, optimizing the convolutional neural network algorithm through actual results, and improving the accuracy and safety of the treatment plan.

5. The method according to claim 1, characterized in that The wearable sensor is used to monitor the changes in the patient's vital signs during multi-drug treatment in real time, evaluate the safety of different dosing regimens, and adjust the medication plan as needed in conjunction with the above-mentioned blood viscosity monitoring results. If the efficacy monitoring module feedback shows that the efficacy is not good, the medication plan adjustment module will be started again and continuously optimized until the most suitable treatment plan for the patient is found.

6. The method according to claim 1, characterized in that The construction of the multi-drug interaction knowledge graph collects drug-drug, drug-food, and drug-disease interaction data to provide a reference for the selection of multi-drug combination medications, including: Use text mining technology to analyze medical literature and drug instructions to extract information on drug properties and mechanisms of action; use information from food ingredient databases, combined with relevant knowledge of nutrition, to determine the impact that specific food ingredients may have on drug efficacy; conduct in-depth analysis of disease-related data to identify the correlation between drug mechanisms of action and disease characteristics; by identifying these correlations, we can know which drugs may have therapeutic or side effect risks for specific diseases; through drug metabolism analysis tools, we can better understand the metabolic process of drugs in the body, especially the interaction between metabolites and other drugs or nutrients; by analyzing medical guidelines and treatment manuals, we provide medical professionals with recommendations for the combined use of multiple drugs to help formulate The most reasonable combination therapy plan; by tracking drug concentration monitoring data, the effect of combined medication on drug plasma concentration can be observed in real time, and the changing trend of drug concentration under various conditions can be analyzed; in terms of analyzing side effect data, disease characteristics will be considered to determine the risk of side effects that different patient groups may encounter when using specific drugs; based on the data collected above and previous research results, a drug efficacy impact assessment model is constructed; integrating the data and evaluation results obtained in all the above steps, and using pattern recognition and machine learning algorithms to deploy a dynamic knowledge graph update mechanism; this mechanism can automatically incorporate new research results and real-time data, thereby continuously improving the knowledge graph and providing medical workers with scientific decision-making support on drug combination therapy.

7. The method according to claim 1, characterized in that The intelligent voice assistant designed for medication reminder and monitoring, disease tracking and consulting services, and helping patients use multiple drugs correctly also includes: Synchronize patient feedback and health data in real time to ensure that all relevant medical information remains up to date, allowing medical staff to conduct effective monitoring and timely intervention; continuously monitor system performance, assess potential risks or existing problems, and regularly optimize system functions based on collected data and user feedback to continuously improve service accuracy and user experience.

8. The method according to claim 1, characterized in that The retesting of the patient's hemostasis parameters, evaluating the changes in blood viscosity before and after multi-drug treatment, establishing a database to collect effective cases, and providing support for new drug development also includes: Define and improve efficacy evaluation criteria, and use them to analyze and verify the actual effects of treatment plans recorded in the database; use the defined efficacy evaluation criteria combined with statistical methods to further analyze the correlation between hemostasis parameters and blood viscosity before and after treatment to identify significant differences; make necessary adjustments to existing multi-drug treatment plans based on the data trends obtained from the analysis to improve the accuracy and pertinence of the treatment plans; Implement a dynamic update mechanism to ensure that the latest research findings, treatment results and related parameters can be reflected in the database in real time; establish a regular reporting system to automatically generate a comprehensive treatment effect report that includes treatment adjustment recommendations and changes in patient hemostasis parameters.

Images