A therapeutic effect evaluation system for HIV patients with hyperlipidemia based on PCSK9 intervention

By designing a therapeutic evaluation system for patients with HIV combined with hyperlipidemia based on PCSK9 intervention, using the knowledge base and artificial convolutional neural network model, the problem of incomplete and accurate evaluation in the existing technology was solved, and efficient and accurate efficacy evaluation was achieved. Improves the accuracy of data processing and analysis.

CN118782265BActive Publication Date: 2024-11-12核工业四一六医院
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202411274998.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-12
Publication Date
2024-11-12
Estimated Expiration
2044-09-12

AI Technical Summary

Technical Problem

The prior art is difficult to effectively evaluate the efficacy of HIV-combined hyperlipidemia patients on PCSK9 intervention therapy, and the lack of intelligent data collection and analysis methods, resulting in incomplete and accurate assessment.

Method used

Design a therapeutic evaluation system for patients with HIV combined with hyperlipidemia based on PCSK9 intervention. By building a knowledge base module, data acquisition and processing module and data analysis module, the artificial convolutional neural network model is used to realize the comprehensive collection, processing and analysis of patient data. , generate accurate results for efficacy evaluation.

Benefits of technology

It improves the comprehensiveness and accuracy of patient data processing and analysis, realizes structured processing and feature extraction of unstructured data, lays the foundation for subsequent intelligent chemotherapy efficacy evaluation, and improves the efficiency and accuracy of the evaluation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118782265B_ABST
    Figure CN118782265B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of smart medical technology, and specifically discloses a therapeutic effect evaluation system for HIV patients with hyperlipidemia based on PCSK9 intervention. The system builds a knowledge base module, a data acquisition and processing module, a data analysis module, and a visual output module, respectively establishes inclusion criteria, exclusion criteria, and elimination criteria knowledge graphs, collects patients' diverse medical data for processing and analysis, obtains patients' diverse medical data feature vectors, and uses patients' medical data feature vectors and artificial convolutional neural network models to achieve comprehensive evaluation of patient efficacy, thereby improving the comprehensiveness and accuracy of the evaluation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of smart medical technology, and specifically relates to a PCSK9 intervention-based efficacy evaluation system for HIV-related hyperlipidemia patients. Background Art

[0002] HIV-infected people have a higher incidence of atherosclerotic cardiovascular disease (ASCVD) than non-HIV-infected people, and the age of onset is younger. At present, evidence from epidemiological, genetic and clinical intervention studies has fully confirmed that low-density lipoprotein cholesterol is a pathogenic risk factor for ASCVD. The latest American Heart Association / American College of Cardiology blood cholesterol management guidelines identify HIV infection as a risk-increasing factor for dyslipidemia. Therefore, strengthening blood lipid management in HIV patients and rationally selecting lipid-lowering drugs are key links in reducing the risk of ASCVD in this population. PCSK9 is a serine protease mainly synthesized and secreted by the liver. It mainly reduces the clearance rate of low-density lipoprotein cholesterol (LDL-C) by binding to the degradation of low-density lipoprotein receptors (LDL-R). Many studies have found that serum PCSK9 levels are closely related to the degree of atherosclerotic plaque burden, and serum PCSK9 levels are linearly correlated with the degree and amount of necrotic core tissue in coronary atherosclerosis. Studies have found that PCSK9 inhibitors can reduce the risk of myocardial infarction by 20%, suggesting that PCSK9 inhibitors may be an emerging therapy to reduce the risk of ASCVD and major adverse cardiovascular events. It is very necessary to clarify the impact of PCSK9 on cardiovascular risk in HIV patients and its safety, tolerability and efficacy.

[0003] With the development of artificial intelligence and big data technology, more intelligent requirements have been put forward for patient efficacy evaluation. Combining the technical advantages of artificial intelligence and big data, optimization of data collection, storage, processing and analysis is achieved. In the context of smart medical care, this application aims to provide a PCSK9 intervention-based HIV combined with hyperlipidemia patient efficacy evaluation system, so that the collected data is more comprehensive, and the implicit correlation between a large amount of data is excavated, making the patient efficacy evaluation more efficient, comprehensive and accurate. Summary of the invention

[0004] The purpose of the present invention is to overcome the above-mentioned problems existing in the prior art, and provide a therapeutic effect evaluation system for HIV patients with hyperlipidemia based on PCSK9 intervention. The system establishes inclusion criteria, exclusion criteria and rejection criteria knowledge graphs by building a knowledge base module, a data acquisition and processing module, a data analysis module and a visualization output module, collects patients' diverse medical data for processing and analysis, and uses an artificial convolutional neural network model to achieve comprehensive evaluation of patient efficacy.

[0005] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions:

[0006] A PCSK9 intervention-based efficacy evaluation system for HIV patients with hyperlipidemia, including the following functional modules:

[0007] A knowledge base module, including a patient inclusion criteria knowledge base unit, a patient exclusion criteria knowledge base unit, and an elimination criteria knowledge base unit;

[0008] The patient inclusion standard knowledge base unit is used to input the patient inclusion standard condition document, perform knowledge extraction on the inclusion standard condition document, and obtain the inclusion standard condition entity set EI and the relationship attribute set RI between the entities; the entity set EI includes 4 types of primary entities: age entity, HIV infection level entity, fasting cholesterol entity and fasting triglyceride entity; the relationship attribute set RI includes the numerical range corresponding to the attribute of each entity in the entity set EI; construct the patient inclusion standard condition knowledge graph A according to the EI and RI;

[0009] The patient exclusion standard knowledge base unit is used to input the patient exclusion standard condition document, extract knowledge from the exclusion standard condition document, and obtain the exclusion standard condition entity set EO and the relationship attribute set RO between the entities; the entity set EO includes 7 types of first-level entities: drug use entities, other clinical trial acceptance entities, organ function entities, basic disease entities, bad living habits entities, compliance degree entities and special physical entity entities; among which the basic disease entity includes 5 types of second-level entities: diabetes entities, hypertension entities, heart disease entities, cerebral hemorrhage entities, and malignant tumor entities; the relationship attribute set RO includes the numerical range and degree level corresponding to the attributes of each entity in the entity set EO; the patient exclusion standard condition knowledge graph B is constructed according to the EO and RO;

[0010] The patient exclusion standard knowledge base unit is used to input patient exclusion standard condition documents and historical patient exclusion event data, perform knowledge extraction on the exclusion standard condition documents and historical patient exclusion event data, and obtain the exclusion standard condition entity set EA and the relationship attribute set RA between entities; the entity set EA includes 4 types of first-level entities: allergy entities, patient willingness entities, researcher willingness entities and exclusion event entities; the relationship attribute set RA includes the level range and attribute events corresponding to the attributes of each entity in the entity set EA; construct a patient exclusion standard condition knowledge graph C based on the EA and RA;

[0011] Data collection and processing module: including patient inclusion-related data collection and processing unit, patient medication data collection and processing unit, and patient physiological index data collection and processing unit before and after medication;

[0012] The patient inclusion-related data collection and processing unit: collects basic information data documents and medical record data documents of each HIV combined with hyperlipidemia patient; pre-establishes a medical dictionary, the medical dictionary includes medical indicator keywords and corresponding attribute keywords; uses the medical dictionary to perform structured processing on the basic information data documents and medical record data documents to obtain structured document data of each HIV combined with hyperlipidemia patient, the structured document data is in key-value form, key represents the identified medical indicator, and value represents the attribute value corresponding to the medical indicator; uses the pre-established knowledge graph A to match and screen each structured document data to obtain structured document data of patients that meet the inclusion criteria; uses the pre-established knowledge graph B to match and filter the structured documents of patients that meet the inclusion criteria, filters out the structured documents of patients that meet the exclusion criteria, obtains the final structured document data of patients that meet the criteria and sequentially codes the patients therein; evenly groups the patients, with patients with odd numbers being classified as the statin group, and patients with even numbers being classified as the PCSK9 inhibitor group, and formulates a theoretical medication plan for each patient;

[0013] The patient medication data collection and processing unit: collects the patient's actual medication data, including medication time, medication dosage, number of medications, whether there is an allergic reaction after medication, specific symptoms and allergy time; summarizes and processes the patient's actual medication data every time period T1 to obtain the patient's actual medication index data set and allergic reaction data set; analyzes each patient's theoretical medication regimen to obtain its theoretical medication index data set in the current time period; calculates the difference between the actual medication index data set and the theoretical medication index data set to obtain the patient's compliance assessment level; uses the allergic reaction data set to obtain the patient's allergy assessment level and persistent allergy condition; uses the knowledge graph C, the patient's compliance assessment level, the allergy assessment level and the persistent allergy condition to match and screen the patients after medication every time period T1, and eliminates patients who meet the elimination criteria;

[0014] The patient's physiological index data collection and processing unit before and after medication: collects relevant data of the patient before medication to form a patient basic data set, a medical history data set and a patient's physiological index data set before medication; the patient's basic data set includes gender, age, height and weight; the medical history data set includes HIV history, hypertension history, diabetes history, coronary heart disease, cerebral infarction, smoking history, and drinking history; the patient's physiological index data set before medication includes blood lipid data, blood sugar data, blood pressure data, liver function, kidney function, creatine kinase, CD4, HIV virus load, blood routine data, electrocardiogram, neck vascular ultrasound data, heart ultrasound data, and limb arteriosclerosis index data;

[0015] The patient's pre-medication physiological index data set is processed, and the indicators involved are divided into two categories: normal and abnormal, taking the smallest subdivision index corresponding to each indicator data as the unit, and the<key,value,p> The format of the triplet is stored in the normal index data set before medication and the abnormal index data set before medication respectively; the key represents the subdivision index, the value represents the numerical value corresponding to the subdivision index, and p is a flag used to indicate whether the index is normal or abnormal;

[0016] Extracting features from the patient's basic data set, medical history data set, normal index data set before medication, and abnormal index data set before medication, respectively, to obtain a patient's basic data feature vector, a patient's medical history feature vector, and a patient's physiological index feature vector before medication;

[0017] The calculation method of the patient's basic data feature vector is as follows: a basic data level mapping table is pre-established, which includes the age range and the corresponding level, which is divided into youth, middle-aged and elderly; including BMI and the corresponding level, which is divided into underweight, normal, overweight and obese; using the basic data level mapping table, the patient's basic data set is processed to obtain each patient's basic portrait data triple <gender, age level, obesity level>, and the patient's basic portrait data is assigned and mapped to obtain the feature values ​​of the three data of gender, age level and obesity level, each feature value range is [0,1], and the three feature values ​​are spliced ​​together to form the patient's basic data feature vector;

[0018] The calculation method of the medical history feature vector is as follows: the data in the medical history data set is divided into four categories of features, namely HIV infection, hypertension history and diabetes history, heart disease history and bad living habits, the eigenvalues ​​corresponding to the four categories of features are calculated, and the four eigenvalues ​​are concatenated to form the patient's medical history feature vector; the eigenvalues ​​corresponding to the HIV infection status of all patients are assigned to 1, and the formula is used Calculate the last three eigenvalues, where fa represents the eigenvalue, n represents the number of indicators corresponding to a certain type of feature, and y i Indicates the specific value of the indicator. When there is a history of a certain disease, the value of the indicator is 1, otherwise it is 0;

[0019] The calculation method of the characteristic vector of physiological indicators before medication is as follows: the number of indicators corresponding to the normal indicator data set before medication and the abnormal indicator data set before medication are respectively counted to obtain the number of normal indicators p1 and the number of abnormal indicators p2, and the percentage of the number of normal indicators to the total number of indicators p1 / (p1+p2) is calculated as the characteristic value corresponding to the normal indicator, and the percentage of the number of abnormal indicators to the total number of indicators p2 / (p1+p2) is calculated as the characteristic value corresponding to the abnormal indicator, and the two characteristic values ​​are spliced ​​together to form the characteristic vector of the patient's physiological indicators before medication;

[0020] At every time period T2, relevant data of the patient after taking the medicine is collected to form a physiological index data set and a post-medication adverse event data set of the patient; the post-medication physiological index data set includes blood lipid data, blood sugar data, blood pressure data, liver function, kidney function, creatine kinase, CD4, HIV viral load, blood routine data, electrocardiogram, neck vascular ultrasound data, cardiac ultrasound data, limb arteriosclerosis index, and coagulation function; the post-medication adverse event data set includes adverse reaction symptoms, adverse reaction time, and adverse reaction frequency of the patient after taking the medicine;

[0021] Extracting features from the data in the post-medication physiological index dataset and the patient adverse event dataset respectively to obtain a feature vector of the patient's post-medication physiological index and a feature vector of the adverse event;

[0022] The patient's physiological index data set after medication is processed, and the indicators involved are divided into two categories: normal and abnormal, taking the smallest subdivision indicator corresponding to each indicator data as the unit, and<key,value,p> The format of the triplet is stored in the normal index data set before medication and the abnormal index data set before medication respectively; the key represents the subdivision index, the value represents the numerical value corresponding to the subdivision index, and p is a flag used to indicate whether the index is normal or abnormal;

[0023] The calculation method of the characteristic vector of the physiological index of the patient after taking the medicine is as follows: the number of corresponding indicators in the normal index data set after taking the medicine and the abnormal index data set after taking the medicine are respectively counted to obtain the number of normal indicators p1 and the number of abnormal indicators p2, and the percentage of the number of normal indicators to the total number of indicators p1 / (p1+p2) is calculated as the characteristic value corresponding to the normal indicator, and the percentage of the number of abnormal indicators to the total number of indicators p2 / (p1+p2) is calculated as the characteristic value corresponding to the abnormal indicator, and the two characteristic values ​​are spliced ​​together to form the characteristic vector of the physiological index of the patient after taking the medicine;

[0024] The calculation method of adverse event feature vector is as follows: count the number of adverse reaction types k and the number of occurrences of each type of adverse reaction in the patient adverse event data set (n1, n2, ... n k ), where n1~n k Respectively represent the number of occurrences of the 1st to kth adverse reactions; the proportion of the number of occurrences of each adverse reaction to the total number of occurrences is taken as the eigenvalue corresponding to the adverse reaction, and the k eigenvalues ​​are concatenated to form the patient's adverse event feature vector;

[0025] The data analysis module includes a manual efficacy evaluation unit and a neural network efficacy evaluation unit;

[0026] The artificial efficacy evaluation unit: respectively selects patient samples accounting for 80% of the total number of samples in the statin group and the PCSK9 inhibitor group for artificial efficacy evaluation, and by comparing and analyzing the patient basic data set, the previous medical history data set, the patient physiological index data set before medication, and the patient physiological index data set after medication, the researcher analyzes the patient's indicators before and after medication based on clinical experience and medical common sense, and obtains the patient's corresponding efficacy evaluation grade label, the label includes excellent, good and poor, and uses the efficacy evaluation grade label to label the 80% patient samples to obtain the labeled statin group training sample set and PCSK9 inhibitor group training sample set;

[0027] The neural network efficacy evaluation unit: using the labeled statin group training sample set and PCSK9 inhibitor group training sample set, determine the basic data feature vector of each patient in the two training sample sets, the patient's previous medical history feature vector, the patient's physiological index feature vector before medication, the patient's physiological index feature vector after medication, and the feature vector of adverse events after medication, respectively splice them into a comprehensive feature vector of the statin group patients and a comprehensive feature vector of the PCSK9 inhibitor group patients, respectively input them into the pre-constructed artificial convolutional neural network models A and B for training iterations, and obtain converged statin group efficacy evaluation models and PCSK9 inhibitor group efficacy evaluation models; select the remaining patient sample data accounting for 20% of the total sample number in the statin group and the PCSK9 inhibitor group, respectively input them into the statin group efficacy evaluation model and the PCSK9 inhibitor group efficacy evaluation model, and obtain the statin group efficacy evaluation results and the PCSK9 inhibitor group efficacy evaluation results;

[0028] Visual output module: outputs the evaluation results in the form of charts.

[0029] By adopting the above technology, compared with the prior art, the beneficial effects of the present invention are:

[0030] This application uses big data and knowledge graph technology to collect, process and analyze patient data, improving the comprehensiveness and accuracy of patient data processing and analysis.

[0031] This application uses a variety of methods to process the collected unstructured patient data, realizes the extraction of structured data features, and lays the foundation for the subsequent successful use of artificial intelligence technology to realize data analysis.

[0032] This application first conducts efficacy analysis on some patient samples based on manual experience, and labels some patient samples based on the analysis results. It then uses the labeled data to train an artificial neural network model, and performs intelligent efficacy evaluation on the remaining patient data, thereby improving the data processing capacity of the efficacy evaluation and laying a theoretical foundation for subsequent medical research. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0034] Figure 1 This is a functional module diagram of a PCSK9 intervention-based efficacy evaluation system for HIV-related hyperlipidemia patients in this application;

[0035] Figure 2 This is the functional flow chart of the knowledge base module of this application;

[0036] Figure 3 A functional flow chart of the relevant data collection and processing unit included in this application;

[0037] Figure 4 This is a functional flow chart of the patient medication data collection and processing unit of this application;

[0038] Figure 5 This is a functional flow chart of the data collection and processing unit for the physiological indicators of patients before and after medication in this application;

[0039] Figure 6 This is the functional flow chart of the data analysis module of this application. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. The following description of at least one exemplary embodiment is actually only illustrative and is by no means intended to limit the present invention and its application or use. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] Example 1

[0042] like Figure 1 As shown, this embodiment provides a therapeutic effect evaluation system for HIV combined with hyperlipidemia patients based on PCSK9 intervention, including the following functional modules:

[0043] A knowledge base module, including a patient inclusion criteria knowledge base unit, a patient exclusion criteria knowledge base unit, and an elimination criteria knowledge base unit;

[0044] The patient inclusion standard knowledge base unit is used to input the patient inclusion standard condition document and construct the patient inclusion standard condition knowledge graph A;

[0045] The patient exclusion criteria knowledge base unit is used to input patient exclusion criteria condition documents and construct a patient exclusion criteria condition knowledge graph B;

[0046] The patient rejection standard knowledge base unit is used to input patient rejection standard condition documents and historical patient rejection event data to construct a patient rejection standard condition knowledge graph C;

[0047] Data collection and processing module: including patient inclusion-related data collection and processing unit, patient medication data collection and processing unit, and patient physiological index data collection and processing unit before and after medication;

[0048] The patients are included in the relevant data collection and processing unit, and the basic information data files and medical record data files of each HIV combined with hyperlipidemia patient are collected, and the structured document data of each HIV combined with hyperlipidemia patient is obtained through processing; then the structured document data of the patients who finally meet the conditions are obtained and the patients therein are sequentially coded, and a theoretical medication plan is formulated for each patient;

[0049] The patient medication data collection and processing unit collects the patient's actual medication data, processes it to obtain the patient's actual medication index data set and allergic reaction data set; processes it to obtain the patient's compliance assessment level, the patient's allergy assessment level and the persistence process status;

[0050] The patient's physiological index data collection and processing unit before and after medication collects relevant data of the patient before medication to form a patient basic data set, a previous medical history data set and a patient's physiological index data set before medication;

[0051] The patient's pre-medication physiological index data set is processed, and the indicators involved are divided into two categories: normal and abnormal, taking the smallest subdivision index corresponding to each indicator data as the unit, and the<key,value,p> The format of the triplet is stored in the normal index data set before medication and the abnormal index data set before medication respectively; the key represents the subdivision index, the value represents the numerical value corresponding to the subdivision index, and p is a flag used to indicate whether the index is normal or abnormal;

[0052] Extracting features from the patient's basic data set, medical history data set, normal index data set before medication, and abnormal index data set before medication, respectively, to obtain a patient's basic data feature vector, a patient's medical history feature vector, and a patient's physiological index feature vector before medication;

[0053] At every time period T2, relevant data of the patient after taking the medicine is collected to form a physiological index data set and a post-medication adverse event data set of the patient; feature extraction is performed on the data in the post-medication physiological index data set and the patient adverse event data set to obtain a feature vector of the patient's physiological index after taking the medicine and a feature vector of the adverse event;

[0054] Data analysis module: including manual efficacy evaluation unit and neural network efficacy evaluation unit;

[0055] The artificial efficacy evaluation unit obtains a labeled statin group training sample set and a PCSK9 inhibitor group training sample set;

[0056] The neural network efficacy evaluation unit obtains the efficacy evaluation results of the statin group and the PCSK9 inhibitor group;

[0057] Visual output module: outputs the evaluation results in the form of charts.

[0058] Example 2

[0059] like Figure 2 As shown, this embodiment provides a therapeutic effect evaluation system for HIV combined with hyperlipidemia patients based on PCSK9 intervention, wherein the knowledge base module specifically includes:

[0060] The patients are included in the standard knowledge base unit, and the entity set EI includes 4 types of first-level entities: age entity, HIV infection level entity, fasting cholesterol entity and fasting triglyceride entity; the relationship attribute set RI includes the numerical range corresponding to the attribute of each entity in the entity set EI.

[0061] The patient exclusion criteria knowledge base unit, the entity set EO includes 7 types of first-level entities: drug use entities, other clinical trial acceptance entities, organ function entities, basic disease entities, bad living habits entities, compliance degree entities and special physical entity entities; the basic disease entity includes 5 types of second-level entities: diabetes entities, hypertension entities, heart disease entities, cerebral hemorrhage entities, and malignant tumor entities; the relationship attribute set RO includes the numerical range and degree level corresponding to the attributes of each entity in the entity set EO;

[0062] During patient screening, the research subjects were determined to be patients with HIV infection and dyslipidemia.

[0063] The content of the patient inclusion criteria document is as follows:

[0064] ①Age: 18 to 65 years old.

[0065] ② Documented HIV infection, HIV-1 RNA level below 200 copies / mL for at least 12 weeks, CD4 + T cells ≥ 200 cells / mm 3 .

[0066] ③Fasting LDL-C ≥ 100 mg / dL or non-HDL-C ≥ 130 mg / dL.

[0067] ④Fasting triglycerides ≤ 600 mg / dL (6.8 mmol / L).

[0068] The patient exclusion criteria document is as follows:

[0069] ① Use of lipid-lowering drugs such as statins, PCSK9 inhibitors, ezetimibe, and incretin in the past half month

[0070] patient;

[0071] ② Those who have participated in other clinical trials within 1 month before screening and received treatment with investigational drugs or devices;

[0072] ③ Patients with severe lung, liver, and kidney dysfunction;

[0073] ④ Diabetes with unstable blood sugar control (fasting blood sugar ≥ 8.6mmol / L) or with diabetes mellitus

[0074] Those with complications (diabetic nephropathy, diabetic neuropathy) and those receiving daily insulin therapy;

[0075] ⑤ Subjects with poorly controlled hypertension (blood pressure ≥160 / 95mmhg);

[0076] ⑥ Subjects who have experienced congestive heart failure, New York Heart Association functional grade III, IV, or left ventricular ejection fraction <35% in the past 6 months;

[0077] ⑦ History of acute myocardial infarction, unstable angina, cerebral hemorrhage or stroke in the past year;

[0078] ⑧ Those with a history of substance abuse, drug addiction, and alcoholism (tolerance, withdrawal, and impaired control over drinking behavior) in the past year;

[0079] ⑨ Patients with a history of malignant tumors in the past 3 years;

[0080] ⑩ Those with poor compliance and unable to cooperate with the study requirements;

[0081] ⑪ Women who are pregnant or breastfeeding.

[0082] The content of the patient exclusion criteria document is as follows:

[0083] ① Those with severe allergic reactions;

[0084] ② Those with persistent allergic reactions;

[0085] ③The research participant requests to withdraw from the clinical research;

[0086] ④ Patients who the researcher considers unsuitable to continue participating in this clinical study.

[0087] The historical patient exclusion event data includes patient exclusion events that occurred in the historical trial research process of our hospital, including the patient's basic information and the reason for the patient's exclusion.

[0088] The patient exclusion standard knowledge base unit, the entity set EA includes 4 types of first-level entities: allergy entities, patient willingness entities, researcher willingness entities and exclusion event entities; the relationship attribute set RA includes the level range and attribute events corresponding to the attributes of each entity in the entity set EA.

[0089] Example 3

[0090] like Figure 3 As shown, this embodiment provides an evaluation of the efficacy of HIV combined with hyperlipidemia patients based on PCSK9 intervention, wherein the patient is included in the relevant data collection and processing unit specifically including:

[0091] The patient is included in the relevant data collection and processing unit: basic information data documents and medical record data documents of each HIV combined with hyperlipidemia patient are collected; a medical dictionary is pre-established, and the medical dictionary includes medical indicator keywords and corresponding attribute keywords; using the medical dictionary, the basic information data documents and medical record data documents are structured to obtain structured document data of each HIV combined with hyperlipidemia patient, and the structured document data is in key-value format, where key represents the identified medical indicator and value represents the attribute value corresponding to the medical indicator;

[0092] Using the pre-established knowledge graph A, each of the structured document data is matched and screened to obtain the structured document data of patients that meet the inclusion criteria;

[0093] Using the pre-established knowledge graph B, match and filter the structured documents of patients that meet the inclusion criteria, filter out the structured documents of patients that meet the exclusion criteria, obtain the final structured document data of patients that meet the criteria and sequentially encode the patients therein;

[0094] The patients were evenly divided into groups, with those with odd numbers assigned to the statin group and those with even numbers assigned to the PCSK9 inhibitor group, and a theoretical medication plan was developed for each patient.

[0095] This study was conducted in the form of a controlled trial. HIV-infected patients with dyslipidemia who visited the infectious department of our hospital and met the inclusion criteria were numbered, with odd-numbered patients assigned to the statin group (Group A) and even-numbered patients assigned to the PCSK9 inhibitor group (Group B).

[0096] Statin group (Group A): The theoretical medication regimen is to give antiviral treatment and add atorvastatin (oral 20 mg / night) as basic lipid-lowering treatment.

[0097] PCSK9 inhibitor group (Group B): The theoretical medication regimen is to add PCSK9 inhibitors (subcutaneous injection, 150 mg, once every 2 weeks) to the antiviral basic group for lipid-lowering treatment.

[0098] Example 4

[0099] like Figure 4 As shown, this embodiment provides a therapeutic effect evaluation system for HIV patients with hyperlipidemia based on PCSK9 intervention, wherein the patient medication data collection and processing unit specifically includes:

[0100] Collect the patient's actual medication data, including medication time, medication dosage, medication frequency, whether there is an allergic reaction after medication, specific symptoms and allergic time;

[0101] Every time period T1, the actual medication data of the patient is aggregated and processed to obtain the patient's actual medication index data set and allergic reaction data set;

[0102] Analyze the theoretical medication regimen of each patient to obtain the theoretical medication index data set in the current time period;

[0103] Calculating the difference between the actual medication index data set and the theoretical medication index data set to obtain a patient compliance assessment grade;

[0104] Using the allergic reaction data set, obtaining the patient's allergy assessment level and persistent allergy condition;

[0105] At each time period T1, the knowledge graph C, the patient compliance assessment level, the allergy assessment level and the persistent allergy condition are used to perform matching screening on the patients after medication, and the patients who meet the exclusion criteria are eliminated.

[0106] The patient compliance assessment level reflects the patient's cooperation and accuracy in the implementation of the theoretical medication plan. Both the actual medication indicator data set and the theoretical medication indicator data set include medication time, medication frequency and medication dosage. By comparing the medication time, medication frequency and medication dosage within the statistical period, the number of times that do not meet the theoretical medication indicators is obtained. If the number reaches more than 60% of the total number, the patient's compliance assessment level is considered to be poor. When the number is between 20% and 60% of the total number, the patient's compliance assessment level is considered to be good. When the number is less than 20% of the total number, the patient's compliance assessment level is considered to be excellent.

[0107] The patient's allergy assessment level reflects the abnormal reaction of the patient's immune system to certain substances. These substances are harmless to most people, but for people with allergies, the immune system will identify them as harmful substances and produce immune responses; the symptoms of allergic reactions can range from mild skin symptoms (such as rash and itching) to severe systemic reactions (such as anaphylactic shock). The allergic reaction data set includes the medication time, the duration and repetition of allergies, the number of allergies and allergic symptoms; the number of allergic symptoms after medication is calculated. If the number is greater than 50% of the total number of medications, the patient's allergy assessment level is considered high. If the number is between 20% and 50% of the total number of times, the patient's allergy assessment level is considered medium. If the number is less than 20% of the total number of times, the patient's allergy assessment level is considered low; if the allergic symptoms last for half an hour after medication, or if the allergic symptoms recur more than 2 times before the next medication, the patient is considered to have persistent allergies.

[0108] Example 5

[0109] like Figure 5 As shown, this embodiment provides a system for evaluating the efficacy of HIV combined with hyperlipidemia based on PCSK9 intervention, wherein the patient's physiological index data collection and processing unit before and after medication specifically includes:

[0110] Collect relevant data of patients before medication to form a basic data set of patients, a medical history data set and a physiological index data set of patients before medication; the basic data set of patients includes gender, age, height and weight; the medical history data set includes HIV history, hypertension history, diabetes history, coronary heart disease, cerebral infarction, smoking history, and drinking history; the physiological index data set of patients before medication includes blood lipid data, blood sugar data, blood pressure data, liver function, kidney function, creatine kinase, CD4, HIV viral load, blood routine data, electrocardiogram, neck vascular ultrasound data, cardiac ultrasound data, and limb arteriosclerosis index data;

[0111] The patient's pre-medication physiological index data set is processed, and the indicators involved are divided into two categories: normal and abnormal, taking the smallest subdivision index corresponding to each indicator data as the unit, and the<key,value,p> The format of the triplet is stored in the normal index data set before medication and the abnormal index data set before medication respectively; the key represents the subdivision index, the value represents the numerical value corresponding to the subdivision index, and p is a flag used to indicate whether the index is normal or abnormal;

[0112] Since some indicators in the physiological indicator data set are composite indicators, such as blood pressure data, which is evaluated by measuring systolic pressure and diastolic pressure, the subdivided indicators corresponding to the blood pressure data are systolic pressure and diastolic pressure; for example, liver function is completed by detecting the specific values ​​of several subdivided indicators. The specific subdivided indicators are subdivided according to the indicators recognized in the medical field, and this application will not go into details.

[0113] Extracting features from the patient's basic data set, medical history data set, normal index data set before medication, and abnormal index data set before medication, respectively, to obtain a patient's basic data feature vector, a patient's medical history feature vector, and a patient's physiological index feature vector before medication;

[0114] The calculation method of the patient's basic data feature vector is as follows: a basic data level mapping table is pre-established, which includes age ranges and corresponding levels, divided into youth, middle-aged and elderly; including BMI and corresponding levels, divided into underweight, normal, overweight and obese; using the basic data level mapping table, the patient's basic data set is processed to obtain each patient's basic portrait data triplet <gender, age level, obesity level>, and the patient's basic portrait data is assigned and mapped to obtain the feature values ​​of the three data of gender, age level and obesity level, each feature value range is [0,1], and the three feature values ​​are spliced ​​together to form the patient's basic data feature vector.

[0115] Researchers set specific assignment rules based on actual conditions, and this application does not limit this;

[0116] The calculation method of the medical history feature vector is as follows: the data in the medical history data set is divided into four categories of features, namely HIV infection, hypertension history and diabetes history, heart disease history and bad living habits, the eigenvalues ​​corresponding to the four categories of features are calculated, and the four eigenvalues ​​are concatenated to form the patient's medical history feature vector; the eigenvalues ​​corresponding to the HIV infection status of all patients are assigned to 1, and the formula is used Calculate the last three eigenvalues, where fa represents the eigenvalue, n represents the number of indicators corresponding to a certain type of feature, and y i Indicates the specific value of the indicator. When there is a history of a certain disease, the value of the indicator is 1, otherwise it is 0;

[0117] Since all the subjects were HIV-infected patients, the characteristic values ​​corresponding to the HIV infection status of all patients were assigned a value of 1. Calculate the last three eigenvalues, where fa represents the eigenvalue, n represents the number of indicators corresponding to a certain type of feature, and y iIt represents the specific value of the indicator. When there is a certain medical history, the indicator takes the value of 1, otherwise it is 0. For example, for the history of hypertension and diabetes, the specific indicators are divided into the history of hypertension and the history of diabetes. At this time, n=2, y1 represents the history of hypertension, and y2 represents the history of diabetes. If the patient has a history of hypertension but no history of diabetes, the characteristic value fa calculated by the formula is 0.5. Similarly, the history of heart disease and bad living habits also calculate the corresponding characteristic values ​​according to the above formula, which will not be repeated here.

[0118] The calculation method of the characteristic vector of physiological indicators before medication is as follows: the number of indicators corresponding to the normal indicator data set before medication and the abnormal indicator data set before medication are respectively counted to obtain the number of normal indicators p1 and the number of abnormal indicators p2, and the percentage of the number of normal indicators to the total number of indicators p1 / (p1+p2) is calculated as the characteristic value corresponding to the normal indicator, and the percentage of the number of abnormal indicators to the total number of indicators p2 / (p1+p2) is calculated as the characteristic value corresponding to the abnormal indicator, and the two characteristic values ​​are spliced ​​together to form the characteristic vector of the patient's physiological indicators before medication;

[0119] At every time period T2, relevant data of the patient after taking the medicine is collected to form a physiological index data set and a post-medication adverse event data set of the patient; the post-medication physiological index data set includes blood lipid data, blood sugar data, blood pressure data, liver function, kidney function, creatine kinase, CD4, HIV viral load, blood routine data, electrocardiogram, neck vascular ultrasound data, cardiac ultrasound data, limb arteriosclerosis index, and coagulation function; the post-medication adverse event data set includes adverse reaction symptoms, adverse reaction time, and adverse reaction frequency of the patient after taking the medicine;

[0120] Extracting features from the data in the post-medication physiological index dataset and the patient adverse event dataset respectively to obtain a feature vector of the patient's post-medication physiological index and a feature vector of the adverse event;

[0121] The patient's physiological index data set after medication is processed, and the indicators involved are divided into two categories: normal and abnormal, taking the smallest subdivision indicator corresponding to each indicator data as the unit, and<key,value,p> The format of the triplet is stored in the normal index data set before medication and the abnormal index data set before medication respectively; the key represents the subdivision index, the value represents the numerical value corresponding to the subdivision index, and p is a flag used to indicate whether the index is normal or abnormal;

[0122] The processing method is the same as that of the patient's physiological indicator data set before medication, which will not be repeated here.

[0123] The calculation method of the characteristic vector of the physiological index of the patient after taking the medicine is as follows: the number of corresponding indicators in the normal index data set after taking the medicine and the abnormal index data set after taking the medicine are respectively counted to obtain the number of normal indicators p1 and the number of abnormal indicators p2, and the percentage of the number of normal indicators to the total number of indicators p1 / (p1+p2) is calculated as the characteristic value corresponding to the normal indicator, and the percentage of the number of abnormal indicators to the total number of indicators p2 / (p1+p2) is calculated as the characteristic value corresponding to the abnormal indicator, and the two characteristic values ​​are spliced ​​together to form the characteristic vector of the physiological index of the patient after taking the medicine;

[0124] The calculation method of adverse event feature vector is as follows: count the number of adverse reaction types k and the number of occurrences of each type of adverse reaction in the patient adverse event data set (n1, n2, ... n k ), where n1~n k They represent the number of occurrences of the 1st to kth adverse reactions respectively; the proportion of the number of occurrences of each adverse reaction to the total number of occurrences is taken as the eigenvalue corresponding to the adverse reaction, and the k eigenvalues ​​are concatenated to form the patient's adverse event feature vector.

[0125] In this application, adverse reactions represent the side effects after taking the medicine. Side effects will cause certain damage to the patient. Therefore, taking the side effects into consideration as an indicator for evaluating the efficacy of the drug can make the efficacy evaluation data more comprehensive.

[0126] Example 6

[0127] like Figure 6 As shown, this embodiment provides a therapeutic effect evaluation system for HIV combined with hyperlipidemia patients based on PCSK9 intervention, wherein the data analysis module specifically includes:

[0128] Manual efficacy evaluation unit and neural network efficacy evaluation unit;

[0129] The artificial efficacy evaluation unit: respectively selects patient samples accounting for 80% of the total number of samples in the statin group and the PCSK9 inhibitor group for artificial efficacy evaluation, and by comparing and analyzing the patient basic data set, the previous medical history data set, the patient physiological index data set before medication, and the patient physiological index data set after medication, the researcher analyzes the patient's indicators before and after medication based on clinical experience and medical common sense, and obtains the patient's corresponding efficacy evaluation grade label, the label includes excellent, good and poor, and uses the efficacy evaluation grade label to label the 80% patient samples to obtain the labeled statin group training sample set and PCSK9 inhibitor group training sample set;

[0130] The neural network efficacy evaluation unit: uses the labeled statin group training sample set and PCSK9 inhibitor group training sample set to determine the basic data feature vector of each patient in the two training sample sets, the patient's previous medical history feature vector, the patient's physiological index feature vector before medication, the patient's physiological index feature vector after medication, and the feature vector of adverse events after medication, and splices them into a comprehensive feature vector of the statin group patients and a comprehensive feature vector of the PCSK9 inhibitor group patients, respectively, and inputs them into the pre-constructed artificial convolutional neural network models A and B for training iterations, respectively, to obtain a converged statin group efficacy evaluation model and a PCSK9 inhibitor group efficacy evaluation model; selects the remaining patient sample data accounting for 20% of the total sample number in the statin group and the PCSK9 inhibitor group, and inputs them into the statin group efficacy evaluation model and the PCSK9 inhibitor group efficacy evaluation model, respectively, to obtain the statin group efficacy evaluation results and the PCSK9 inhibitor group efficacy evaluation results.

[0131] The artificial convolutional neural network model is trained using a common model in the prior art, which is not limited in this application.

[0132] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, computer devices, or computer-readable storage media. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware.

[0133] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0134] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. According to the technical essence of the present invention, within the spirit and principles of the present invention, any simple modification, equivalent replacement and improvement made to the above embodiment still falls within the protection scope of the technical solution of the present invention.

Claims

1. A PCSK9 intervention-based therapeutic effect evaluation system for HIV patients with hyperlipidemia, characterized in that: Includes the following functional modules: A knowledge base module, including a patient inclusion criteria knowledge base unit, a patient exclusion criteria knowledge base unit and a patient rejection criteria knowledge base unit; The patient inclusion standard knowledge base unit is used to input the patient inclusion standard condition document and construct the patient inclusion standard condition knowledge graph A; The patient exclusion criteria knowledge base unit is used to input patient exclusion criteria condition documents and construct a patient exclusion criteria condition knowledge graph B; The patient rejection standard knowledge base unit is used to input patient rejection standard condition documents and historical patient rejection event data to construct a patient rejection standard condition knowledge graph C; Data collection and processing module: including patient inclusion-related data collection and processing unit, patient medication data collection and processing unit, and patient physiological index data collection and processing unit before and after medication; The patients are included in the relevant data collection and processing unit, and the basic information data files and medical record data files of each HIV combined with hyperlipidemia patient are collected, and the structured document data of each HIV combined with hyperlipidemia patient is obtained through processing; then the structured document data of the patients who finally meet the conditions are obtained and the patients therein are sequentially coded, and a theoretical medication plan is formulated for each patient; The patient medication data collection and processing unit collects the patient's actual medication data and processes it to obtain the patient's actual medication index data set and allergic reaction data set; The patient compliance assessment level and patient allergy assessment level were obtained; The patient's physiological index data collection and processing unit before and after medication collects relevant data of the patient before medication to form a patient basic data set, a previous medical history data set and a patient's physiological index data set before medication; The patient's pre-medication physiological index data set is processed, and the indicators involved are divided into two categories: normal and abnormal, taking the smallest subdivision index corresponding to each indicator data as the unit, and the<key,value,p> The format of the triplet is stored in the normal index data set before medication and the abnormal index data set before medication respectively; the key represents the subdivision index, the value represents the numerical value corresponding to the subdivision index, and p is a flag used to indicate whether the subdivision index is normal or abnormal; Extracting features from the patient's basic data set, medical history data set, normal index data set before medication, and abnormal index data set before medication, respectively, to obtain a patient's basic data feature vector, a patient's medical history feature vector, and a patient's physiological index feature vector before medication; At every time period T2, relevant data of the patient after taking the medicine is collected to form a physiological index data set and a post-medication adverse event data set of the patient; feature extraction is performed on the data in the post-medication physiological index data set and the patient adverse event data set to obtain a feature vector of the patient's physiological index after taking the medicine and a feature vector of the adverse event; Data analysis module: including manual efficacy evaluation unit and neural network efficacy evaluation unit; The artificial efficacy evaluation unit obtains a labeled statin group training sample set and a PCSK9 inhibitor group training sample set; The neural network efficacy evaluation unit obtains the efficacy evaluation results of the statin group and the PCSK9 inhibitor group; Visual output module: outputs the evaluation results in the form of charts; The knowledge base module specifically includes: Knowledge extraction is performed on the inclusion criteria condition document to obtain an inclusion criteria condition entity set EI and a relationship attribute set RI between entities; the entity set EI includes four types of primary entities: age entity, HIV infection level entity, fasting cholesterol entity, and fasting triglyceride entity; the relationship attribute set RI includes the numerical range corresponding to the attribute of each entity in the entity set EI; a patient inclusion criteria condition knowledge graph A is constructed based on the EI and RI; The knowledge of the exclusion criteria condition document is extracted to obtain the exclusion criteria condition entity set EO and the relationship attribute set RO between the entities; the entity set EO includes 7 types of first-level entities: drug use entities, other clinical trial acceptance entities, organ function entities, basic disease entities, bad living habits entities, compliance degree entities and special physical entity entities; among which the basic disease entities include 5 types of second-level entities: diabetes entities, hypertension entities, heart disease entities, cerebral hemorrhage entities, and malignant tumor entities; the relationship attribute set RO includes the numerical range and degree level corresponding to the attributes of each entity in the entity set EO; the patient exclusion criteria condition knowledge graph B is constructed according to the EO and RO; Knowledge extraction is performed on the exclusion standard condition document and the exclusion event data of historical patients to obtain the exclusion standard condition entity set EA and the relationship attribute set RA between entities; the entity set EA includes 4 types of first-level entities: allergy entities, patient willingness entities, researcher willingness entities and exclusion event entities; the relationship attribute set RA includes the level range and attribute events corresponding to the attributes of each entity in the entity set EA; based on the EA and RA, a patient exclusion standard condition knowledge graph C is constructed; The patient inclusion related data acquisition and processing unit specifically includes: Pre-establish a medical dictionary, the medical dictionary including medical indicator keywords and corresponding attribute keywords; use the medical dictionary to perform structured processing on the basic information data document and the medical record data document to obtain structured document data of each HIV-combined hyperlipidemia patient, the structured document data is in key-value format, the key represents the identified medical indicator, and the value represents the attribute value corresponding to the medical indicator; Using the pre-established knowledge graph A, each of the structured document data is matched and screened to obtain the structured document data of patients that meet the inclusion criteria; Using the pre-established knowledge graph B, match and filter the structured documents of patients that meet the inclusion criteria, filter out the structured documents of patients that meet the exclusion criteria, obtain the final structured document data of patients that meet the criteria and sequentially encode the patients therein; The patients are evenly grouped, with patients with odd numbers being grouped into a statin group and patients with even numbers being grouped into a PCSK9 inhibitor group, and a theoretical medication plan is formulated for each patient; the patient's physiological index data collection and processing unit before and after medication specifically includes: The medical history data set includes HIV history, hypertension history, diabetes history, coronary heart disease, cerebral infarction, smoking history, and drinking history; the patient's physiological index data set before medication includes blood lipid data, blood sugar data, blood pressure data, liver function, kidney function, creatine kinase, CD4, HIV viral load, blood routine data, electrocardiogram, neck vascular ultrasound data, heart ultrasound data, and limb arteriosclerosis index data; The calculation method of the medical history feature vector is as follows: the data in the medical history data set is divided into four categories of features, namely HIV infection, hypertension history and diabetes history, heart disease history and bad living habits, the eigenvalues ​​corresponding to the four categories of features are calculated, and the four eigenvalues ​​are concatenated to form the patient's medical history feature vector; the eigenvalues ​​corresponding to the HIV infection status of all patients are assigned to 1, and the formula is used Calculate the last three eigenvalues, where fa represents the eigenvalue, n represents the number of indicators corresponding to a certain type of feature, and y i Indicates the specific value of the indicator. When there is a history of a certain disease, the value of the indicator is 1, otherwise it is 0; The calculation method of the characteristic vector of physiological indicators before medication is as follows: the number of indicators corresponding to the normal indicator data set before medication and the abnormal indicator data set before medication are respectively counted to obtain the number of normal indicators p1 and the number of abnormal indicators p2, and the percentage of the number of normal indicators to the total number of indicators p1 / (p1+p2) is calculated as the characteristic value corresponding to the normal indicator, and the percentage of the number of abnormal indicators to the total number of indicators p2 / (p1+p2) is calculated as the characteristic value corresponding to the abnormal indicator, and the two characteristic values ​​are spliced ​​together to form the characteristic vector of the patient's physiological indicators before medication; The patient's physiological index data collection and processing unit before and after medication specifically includes: The physiological index data set after medication includes blood lipid data, blood sugar data, blood pressure data, liver function, kidney function, creatine kinase, CD4, HIV viral load, blood routine data, electrocardiogram, neck vascular ultrasound data, heart ultrasound data, limb arteriosclerosis index, and coagulation function; The patient's physiological index data set after medication is processed, and the indicators involved are divided into two categories: normal and abnormal, taking the smallest subdivision indicator corresponding to each indicator data as the unit, and<key,value,p> The format of the triplet is stored in the normal index data set before medication and the abnormal index data set before medication respectively; the key represents the subdivision index, the value represents the numerical value corresponding to the subdivision index, and p is a flag used to indicate whether the index is normal or abnormal; The calculation method of the characteristic vector of the physiological index of the patient after taking the medicine is as follows: the number of corresponding indicators in the normal index data set after taking the medicine and the abnormal index data set after taking the medicine are respectively counted to obtain the number of normal indicators p1 and the number of abnormal indicators p2, and the percentage of the number of normal indicators to the total number of indicators p1 / (p1+p2) is calculated as the characteristic value corresponding to the normal indicator, and the percentage of the number of abnormal indicators to the total number of indicators p2 / (p1+p2) is calculated as the characteristic value corresponding to the abnormal indicator, and the two characteristic values ​​are spliced ​​together to form the characteristic vector of the physiological index of the patient after taking the medicine; The data analysis module specifically includes: The artificial efficacy evaluation unit: respectively selects patient samples accounting for 80% of the total number of samples in the statin group and the PCSK9 inhibitor group for artificial efficacy evaluation, and by comparing and analyzing the patient basic data set, the previous medical history data set, the patient physiological index data set before medication, and the patient physiological index data set after medication, the researcher analyzes the patient's indicators before and after medication based on clinical experience and medical common sense, and obtains the patient's corresponding efficacy evaluation grade label, the label includes excellent, good and poor, and uses the efficacy evaluation grade label to label the 80% patient samples to obtain the labeled statin group training sample set and PCSK9 inhibitor group training sample set; The neural network efficacy evaluation unit: uses the labeled statin group training sample set and PCSK9 inhibitor group training sample set to determine the basic data feature vector of each patient in the two training sample sets, the patient's previous medical history feature vector, the patient's physiological index feature vector before medication, the patient's physiological index feature vector after medication, and the feature vector of adverse events after medication, and splices them into a comprehensive feature vector of the statin group patients and a comprehensive feature vector of the PCSK9 inhibitor group patients, respectively, and inputs them into the pre-constructed artificial convolutional neural network models A and B for training iterations, respectively, to obtain a converged statin group efficacy evaluation model and a PCSK9 inhibitor group efficacy evaluation model; selects the remaining patient sample data accounting for 20% of the total sample number in the statin group and the PCSK9 inhibitor group, and inputs them into the statin group efficacy evaluation model and the PCSK9 inhibitor group efficacy evaluation model, respectively, to obtain the statin group efficacy evaluation results and the PCSK9 inhibitor group efficacy evaluation results.

2. The therapeutic effect evaluation system for HIV combined with hyperlipidemia patients based on PCSK9 intervention as claimed in claim 1, characterized in that: The patient medication data collection and processing unit specifically includes: Collect the actual medication data of the patients, including medication time, medication dosage, number of medications, whether there is an allergic reaction after medication, specific symptoms and allergy time; summarize and process the actual medication data of the patients every time period T1 to obtain the actual medication index data set and allergic reaction data set of the patients; analyze the theoretical medication regimen of each patient to obtain the theoretical medication index data set in the current time period; calculate the difference between the actual medication index data set and the theoretical medication index data set to obtain the patient's compliance assessment level; use the allergic reaction data set to obtain the patient's allergy assessment level and persistent allergy condition; every time period T1, use the knowledge graph C, the patient's compliance assessment level, allergy assessment level and persistent allergy condition to match and screen the patients after medication, and eliminate the patients who meet the elimination criteria.

3. The therapeutic effect evaluation system for HIV combined with hyperlipidemia patients based on PCSK9 intervention as claimed in claim 1, characterized in that: The patient's physiological index data collection and processing unit before and after medication specifically includes: The patient basic data set includes gender, age, height and weight; the calculation method of the patient basic data feature vector is as follows: a basic data level mapping table is pre-established, which includes age ranges and corresponding levels, divided into youth, middle-aged and elderly; including BMI and corresponding levels, divided into underweight, normal, overweight and obese; using the basic data level mapping table, the patient basic data set is processed to obtain each patient's basic portrait data triple <gender, age level, obesity level>, and the patient's basic portrait data is assigned and mapped to obtain the feature values ​​of the three data of gender, age level and obesity level, each feature value range is [0,1], and the three feature values ​​are spliced ​​together to form the patient's basic data feature vector.

4. The therapeutic effect evaluation system for HIV combined with hyperlipidemia patients based on PCSK9 intervention as claimed in claim 1, characterized in that: The patient's physiological index data collection and processing unit before and after medication specifically includes: The adverse event data set after medication includes adverse reaction symptoms, adverse reaction time, and adverse reaction frequency after medication. The adverse event feature vector is calculated as follows: Count the number of adverse reaction types k and the number of occurrences of each type of adverse reaction in the patient's adverse event data set (n1, n2, ... n k ), where n1~n k They represent the number of occurrences of the 1st to kth adverse reactions respectively; the proportion of the number of occurrences of each adverse reaction to the total number of occurrences is taken as the eigenvalue corresponding to the adverse reaction, and the k eigenvalues ​​are concatenated to form the patient's adverse event feature vector.

Citation Information

Patent Citations

  • Curative effect and quantification system for treating cervical spondylotic radiculopathy through three-stage syndrome differentiation traction method

    CN111710392A

  • Test method for preventing and treating cognitive impairment after acute stroke by spleen strengthening, kidney tonifying and turbid dissolving method

    CN114283940A

  • DRG high-compilation high-suit behavior identification method and device, electronic equipment and medium

    CN117786123A

  • Equipment system efficiency evaluation method and device based on evaluation knowledge representation

    CN117973683A

  • Multi-drug treatment response prediction method based on blood viscosity change

    CN117995430A