Clinical test data quality supervision method and system based on risk characteristics

Abstract

The invention discloses a clinical test data quality supervision method and system based on risk characteristics, and relates to the field of quality supervision, and the method comprises the steps: determining a clinical test characteristic type and a test control parameter based on a clinical test design scheme; the method comprises the following steps: acquiring initial feature information through historical body data and medicine taking data of a subject; obtaining a type correlation coefficient; setting a risk feature recognition threshold value in combination with the analysis type correlation coefficient and the initial feature information; acquiring real-time body information of a subject, and acquiring clinical test data according to a clinical test design scheme; analyzing the clinical test data based on the risk feature identification threshold, and obtaining an abnormal coefficient; comparing the abnormal coefficient with the body information of the subject, and judging whether a false result phenomenon exists or not; analyzing test control parameters and clinical test data according to a false result phenomenon judgment condition, and setting an abnormal information threshold value; and monitoring the clinical test data according to the abnormal information threshold, and evaluating the quality of the clinical test data.

Description

Technical Field

[0001] The present invention relates to the field of quality supervision, and specifically to a method and system for clinical trial data quality supervision based on risk characteristics. Background Art

[0002] Clinical trials are an important method of medical research used to evaluate the safety, effectiveness, and feasibility of new treatment methods, drugs, medical devices, or interventions. Through strictly designed and controlled experiments, clinical trials provide scientific evidence for the medical field. Risk characteristics are a concept that comprehensively describes and characterizes various attributes and features of risks, providing a basis for risk assessment and management by presenting the essence and characteristics of risks.

[0003] Existing clinical trial data quality supervision methods adopt a unified inspection and audit strategy, making it difficult to prioritize the risk characteristics of different data, resulting in excessive resources being invested in low-risk data while high-risk data is ignored. Different patient groups have significant differences in treatment responses and side effects, and existing quality supervision methods often struggle to distinguish the characteristics between different patients, leading to potential side effect data not being discovered in a timely manner, making clinical trial data imperfect and thus affecting the reliability of research results. Summary of the Invention

[0004] To solve the above technical problems, a method and system for clinical trial data quality supervision based on risk characteristics are provided. This technical solution solves the problems in the above background art, namely, excessive resources are invested in low-risk data while high-risk data is ignored, it is difficult to distinguish the characteristics between different patients, leading to potential side effect data not being discovered in a timely manner, making clinical trial data imperfect and thus affecting the reliability of research results.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A method for clinical trial data quality supervision based on risk characteristics, comprising:

[0007] Obtaining and analyzing a clinical trial design plan to determine the types of clinical trial characteristics and trial control parameters;

[0008] Collecting and organizing the historical physical data and historical medication data of subjects to obtain initial characteristic information;

[0009] Obtaining a type correlation coefficient based on the initial characteristic information;

[0010] Setting a risk characteristic recognition threshold based on the type correlation coefficient and the initial characteristic information;

[0011] Based on a portable medical device, obtain the real-time physical information of the subject, provide treatment to the subject according to the clinical trial design plan, and obtain clinical trial data;

[0012] Based on the risk characteristic identification threshold, analyze the clinical trial data to obtain an anomaly coefficient;

[0013] According to the anomaly coefficient, analyze the real-time physical information of the subject to determine whether there is a false result phenomenon;

[0014] According to the judgment of the false result phenomenon, analyze the test control parameters and the clinical trial data, and set the anomaly information threshold;

[0015] According to the anomaly information threshold, supervise the clinical trial data, determine whether the clinical trial data is true, and evaluate the quality of the clinical trial data.

[0016] Preferably, the obtaining of the type correlation coefficient specifically includes:

[0017] Sort out the historical physical data of the subject to generate physical type data, and the physical type data includes heart rate, blood pressure, blood sugar, blood lipid, lung function index, liver function index, gastrointestinal function index and respiratory rate;

[0018] Analyze the physical type data of each subject, use the heart rate as the reference data, and obtain the corresponding relationship of physiological data;

[0019] Classify the initial feature information according to the clinical trial feature type to generate feature performance information;

[0020] Analyze the feature performance information according to the physical type data to determine the individual difference degree;

[0021] Analyze the historical medication data of the subject to determine the mechanism of action of the drug and the type of influencing data;

[0022] Classify the subjects according to the mechanism of action of the drug to generate the same drug mechanism group;

[0023] Based on the historical medication records of the subject, analyze the historical physical data of each subject in the same drug mechanism group to generate the historical physical data of the subject corresponding to the type of drug influencing data, and obtain the drug effect absorption degree of the subject;

[0024] Classify the subjects according to the drug effect absorption degree of the subject, group the subjects with the same drug effect absorption degree into one group to generate the same physical constitution type group, and obtain the number of subjects in each type group;

[0025] Randomly select two subjects from the same physical constitution type group, analyze the corresponding relationship of physiological data, calculate the average physiological fluctuation of heart rate of the two subjects, and in the group with the same physical constitution type, repeat this calculation step, and the number of repetitions is equal to the number of subjects in this physical constitution type group to obtain multiple groups of average physiological fluctuations of heart rate;

[0026] Perform weighted averaging on multiple groups of average physiological fluctuations of heart rate according to the individual difference degree to obtain the type correlation coefficient.

[0027] Preferably, the setting of the risk feature recognition threshold specifically includes:

[0028] According to the clinical trial design plan, obtain the corresponding numerical relationship between the clinical trial feature type and the initial feature information;

[0029] Analyze the initial feature information according to the individual difference degree, and respectively obtain the body type data thresholds of each subject;

[0030] Analyze the body type data thresholds of each subject according to the corresponding numerical relationship between the clinical trial feature type and the initial feature information to obtain the abnormal feature manifestations of each subject;

[0031] Preprocess the abnormal feature manifestations of each subject and set the first risk feature recognition threshold;

[0032] Analyze the body type data thresholds of each subject according to the type correlation coefficient to obtain the heart beat cycle thresholds of each subject, and set the second risk feature recognition threshold.

[0033] Preferably, the obtaining of the abnormal coefficient specifically includes:

[0034] Analyze the clinical trial feature type according to the initial feature information to obtain the data acquisition feature trend;

[0035] Based on the data acquisition feature trend, analyze the clinical trial data to obtain the recording parameters;

[0036] Compare the recording parameters with the first risk feature recognition threshold and the second risk feature recognition threshold respectively to obtain the first deviation parameter and the second deviation parameter;

[0037] Analyze the first deviation parameter according to the first risk feature recognition threshold to obtain the parameter deviation frequency;

[0038] Analyze the second deviation parameter according to the second risk feature recognition threshold to obtain the parameter deviation amplitude;

[0039] Analyze the parameter deviation frequency and the parameter deviation amplitude according to the individual difference degree to obtain the abnormal coefficient.

[0040] Preferably, the determination of whether there is a false result specifically includes:

[0041] Determine the standard physical fitness threshold according to the initial characteristic information;

[0042] Obtain the individualized control parameters through the test control parameters and the anomaly coefficient;

[0043] Calculate the standard deviation of the individualized control parameters, determine the standard control parameters, and obtain the test control distribution trend;

[0044] Perform weighted processing on the standard physical fitness threshold according to the individual differences of each subject to obtain the subject difference threshold;

[0045] Compare the clinical trial data with the subject difference threshold one by one to obtain the data anomaly situation recognition result;

[0046] Compare the data anomaly situation recognition result with the test control distribution trend to determine whether the two are consistent. If so, there is no false result phenomenon. If not, there is a false result phenomenon, and generate the false result phenomenon judgment situation.

[0047] Preferably, the setting of the anomaly information threshold specifically includes:

[0048] Sort out the false result phenomenon judgment situation to obtain the false result subject information, classify the false result subject information in sequence according to age and health status, and respectively obtain the distribution of false results in each age and each health status;

[0049] Analyze the distribution of false results in each age and each health status respectively to obtain the average value of the age distribution data and the average value of the health status distribution;

[0050] According to the standard control parameters, analyze the average value of the age distribution data and the average value of the health status distribution respectively to obtain the standard deviation of the age distribution data and the standard deviation of the health status distribution data;

[0051] Sort out the age and health status of the subjects, and summarize the body type data of the subjects whose age is within one standard deviation range of the average value of the age distribution data and whose health status is within one standard deviation range of the average value of the health status to obtain the average peak value of the body type data and the average valley value of the body type data;

[0052] Perform a difference processing on the average peak value of the body type data and the average valley value of the body type data as the anomaly information threshold.

[0053] Furthermore, a clinical trial data quality supervision system based on risk characteristics is proposed for implementing the above-mentioned supervision method, which is characterized in that it specifically includes:

[0054] A data acquisition module, which is used to obtain a clinical trial design plan, collect the historical ecological data and historical medication data of the subjects, and transmit the collected data to the feature recognition module;

[0055] A feature recognition module, which is used to analyze the received data and extract features, obtain initial feature information, type correlation coefficients and abnormal feature manifestations, generate feature performance information, and transmit the data to the data analysis module, the risk assessment module and the data monitoring module;

[0056] A data analysis module, which is used to analyze, sort out and compare the received data, obtain the individual difference degree, determine the mechanism of action of the drug, the type of data affected by the drug, the drug effect absorption degree of the subject, the parameter deviation frequency and the parameter deviation amplitude, and transmit the data to the risk assessment module and the data monitoring module;

[0057] A risk assessment module, which is used to sort out and summarize the received data, set the risk feature recognition threshold and the abnormal information threshold, and transmit the set thresholds to the data monitoring module;

[0058] A data monitoring module, which is used to compare the received data, analyze the real-time physical information of the subject through a constant coefficient to judge whether there is a false result phenomenon, and compare the clinical trial data through the abnormal information threshold to judge whether the clinical trial data is true, and transmit the judgment result to the quality control module;

[0059] A quality control module, which is used to analyze the received judgment result and evaluate the quality of the clinical trial data.

[0060] Preferably, the feature recognition module specifically includes:

[0061] A first recognition unit, which is used to recognize the historical physical data and historical medication data of the subject to obtain initial feature information;

[0062] A second recognition unit, which is used to classify the initial feature information according to the clinical trial feature type to generate feature performance information;

[0063] A third recognition unit, which is used to perform feature recognition on the initial feature information to obtain type correlation coefficients.

[0064] Preferably, the data analysis module specifically includes:

[0065] The first analysis unit is configured to analyze the characteristic performance information based on the body type data to determine the degree of individual difference;

[0066] The second analysis unit is configured to analyze the historical medication data of the subject to determine the mechanism of action of the drug and the type of impact data;

[0067] Preferably, the risk assessment module specifically includes:

[0068] The first assessment unit is configured to analyze the body type data threshold of each subject through the corresponding numerical relationship between the clinical trial characteristic type and the initial characteristic information, obtain the abnormal characteristic performance of each subject, and set the first risk characteristic identification threshold according to the abnormal characteristic performance of each subject;

[0069] The second assessment unit is configured to analyze the body type data threshold of each subject through the type correlation coefficient, obtain the heart rate cycle threshold of each subject, and set the second risk characteristic identification threshold;

[0070] The third assessment unit is configured to obtain the distribution of false results in each age and each health condition according to the judgment of false result phenomena, and set the abnormal information threshold by analyzing the distribution of false results in each age and each health condition.

[0071] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0072] The present invention proposes a clinical trial data quality supervision scheme based on risk characteristics. Based on the clinical trial design scheme, the types of clinical trial characteristics and trial control parameters are determined. By collecting and organizing the historical physical data and historical medication data of the subjects, initial characteristic information is obtained. According to the initial characteristic information, type correlation coefficients are obtained. By combining the analysis of the type correlation coefficients and the initial characteristic information, a risk characteristic identification threshold is set. Based on a portable medical device, the real-time physical information of the subjects is obtained, and the subjects are treated according to the clinical trial design scheme to obtain clinical trial data. Based on the risk characteristic identification threshold, the clinical trial data is analyzed to obtain an anomaly coefficient. According to the anomaly coefficient, the real-time physical information of the subjects is analyzed to determine whether there is a false result phenomenon, and the judgment situation of the false result phenomenon is obtained. According to the judgment situation of the false result phenomenon, the trial control parameters and the clinical trial data are analyzed to set an abnormal information threshold. Based on the abnormal information threshold, the clinical trial data is supervised to determine whether the clinical trial data is true, and the quality of the clinical trial data is evaluated. In this way, the risk characteristics of different data are sorted, the allocation efficiency of resources between low-risk data and high-risk data is improved, a detailed differential evaluation is carried out on different patient groups, and the integrity, accuracy, and reliability of the data on the treatment response and side effects of the subjects are ensured, providing a solid data foundation for the subsequent evaluation of drugs or treatment methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 It is a flowchart of the method for supervising the quality of clinical trial data based on risk characteristics proposed by the present invention;

[0074] Figure 2 It is a flowchart of the method for obtaining type correlation coefficients in the present invention;

[0075] Figure 3 It is a flowchart of the method for setting a risk characteristic identification threshold in the present invention;

[0076] Figure 4 It is a flowchart of the method for obtaining an anomaly coefficient in the present invention;

[0077] Figure 5 It is a flowchart of the method for determining whether there is a false result in the present invention;

[0078] Figure 6 It is a flowchart of the method for setting an abnormal information threshold in the present invention;

[0079] Figure 7 It is a structural diagram of the system for supervising the quality of clinical trial data based on risk characteristics proposed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0080] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments in the following description are only examples, and other obvious variations can be conceived by those skilled in the art.

[0081] Referring to Figure 1 as shown, a method for quality supervision of clinical trial data based on risk characteristics includes:

[0082] Obtain and analyze the clinical trial design plan to determine the clinical trial characteristic type and test control parameters;

[0083] Collect and sort out the historical physical data and historical medication data of the subjects to obtain the initial characteristic information;

[0084] Obtain the type correlation coefficient according to the initial characteristic information;

[0085] Set the risk characteristic recognition threshold according to the type correlation coefficient and the initial characteristic information;

[0086] Based on a portable medical device, obtain the real-time physical information of the subject, treat the subject according to the clinical trial design plan, and obtain the clinical trial data;

[0087] Analyze the clinical trial data based on the risk characteristic recognition threshold to obtain the anomaly coefficient;

[0088] Analyze the real-time physical information of the subject according to the anomaly coefficient to determine whether there is a false result phenomenon;

[0089] Analyze the test control parameters and the clinical trial data according to the judgment situation of the false result phenomenon, and set the anomaly information threshold;

[0090] Supervise the clinical trial data according to the anomaly information threshold, judge whether the clinical trial data is true, and evaluate the quality of the clinical trial data.

[0091] The present invention proposes a clinical trial data quality supervision scheme based on risk characteristics. Based on the clinical trial design scheme, the clinical trial characteristic types and trial control parameters are determined. By collecting and sorting the historical physical data and historical medication data of the subjects, initial characteristic information is obtained. According to the initial characteristic information, type correlation coefficients are obtained. By combining and analyzing the type correlation coefficients and the initial characteristic information, a risk characteristic identification threshold is set. Based on a portable medical device, the real-time physical information of the subjects is obtained, and treatment is carried out for the subjects according to the clinical trial design scheme to obtain clinical trial data. Based on the risk characteristic identification threshold, the clinical trial data is analyzed to obtain an anomaly coefficient. According to the anomaly coefficient, the real-time physical information of the subjects is analyzed to determine whether there is a false result phenomenon, and the judgment situation of the false result phenomenon is obtained. According to the judgment situation of the false result phenomenon, the trial control parameters and the clinical trial data are analyzed to set an anomaly information threshold. Taking the anomaly information threshold as a benchmark, the clinical trial data is supervised to determine whether the clinical trial data is true and evaluate the quality of the clinical trial data.

[0092] Referring to Figure 2 as shown, obtaining the type correlation coefficients specifically includes:

[0093] Sort out the historical physical data of the subjects to generate body type data, where the body type data includes heart rate, blood pressure, blood sugar, blood lipids, lung function indicators, liver function indicators, gastrointestinal function indicators, and respiratory rate;

[0094] Analyze the body type data of each subject, taking the heart rate as the reference data, to obtain the corresponding relationship of physiological data;

[0095] Classify the initial characteristic information according to the clinical trial characteristic types to generate characteristic performance information;

[0096] Analyze the characteristic performance information according to the body type data to determine the individual difference degree;

[0097] Analyze the historical medication data of the subjects to determine the mechanism of action of the drug and the type of data affected;

[0098] Classify the subjects according to the mechanism of action of the drug to generate the same drug mechanism group;

[0099] Based on the historical medication records of the subjects, analyze the historical physical data of each subject in the same drug mechanism group to generate the historical physical data of the subjects corresponding to the type of data affected by the drug, obtain the drug effect absorption degree of the subjects. The mechanism of action of the drug and the type of data affected can determine whether each organ of the human body is affected and whether it has an impact on the body type data. Compare the affected body type data before and after the subjects take the drug to obtain the difference in body type data, and take the ratio of the difference in body type data to the trial control parameters as the drug effect absorption degree of the subjects;

[0100] Classify the subjects according to the drug effect absorption degree of the subjects. Group the subjects with the same drug effect absorption degree into one group to generate the same physical constitution type group, and obtain the number of subjects in each type group.

[0101] Randomly select two subjects from the same physical constitution type group, analyze the corresponding relationship of physiological data, and calculate the average value of the physiological fluctuations of the heart rates of the two subjects. In the group with the same physical constitution type, repeat this calculation step, and the number of repetitions is equal to the number of subjects in this physical constitution type group to obtain multiple groups of average values of physiological fluctuations of the heart rates.

[0102] Perform a weighted average on multiple groups of average values of physiological fluctuations of the heart rates according to the degree of individual differences to obtain the type correlation coefficient.

[0103] It is understandable that body type data reflects various functions of the human body. Heart rate reflects the heart function and the balance state of the autonomic nervous system. A normal heart rate is a basic sign of heart health, which shows the number of times the heart beats per minute. In a quiet state, the normal heart rate of adults is generally between 60 and 100 beats per minute. An accelerated heart rate is caused by various factors such as emotional excitement, hyperthyroidism, and heart diseases. A bradycardia is caused by side effects of certain drugs or problems with the cardiac conduction system. By monitoring the heart rate, the body's response degree to various internal and external stimuli can be reflected. Blood pressure reflects the pressure of the blood in the blood vessels on the vessel walls and is a key indicator for measuring the health of the cardiovascular system. Blood pressure includes systolic blood pressure and diastolic blood pressure. Systolic blood pressure is the highest pressure of the blood on the vessel walls when the heart contracts, and diastolic blood pressure is the lowest pressure of the blood on the vessel walls when the heart relaxes. Normal blood pressure is generally lower than 120 / 80 mmHg. High blood pressure is manifested as a continuous value higher than 140 / 90 mmHg and is an important risk factor for cardiovascular diseases. Long-term high blood pressure can cause damage to organs such as the heart, brain, kidneys, and eyes. Low blood pressure is manifested as a blood pressure lower than 90 / 60 mmHg, which can cause symptoms such as dizziness, fatigue, and blackouts in front of the eyes and is caused by insufficient blood volume, weakened heart function, and blood vessel dilation. Blood sugar mainly reflects the body's sugar metabolism. The blood sugar level is regulated by hormones such as insulin, and factors such as food intake, exercise, and emotions can also affect it. Normal fasting blood sugar is generally between 3.9 and 6.1 mmol / L. High blood sugar is usually the main feature of diabetes. Long-term high blood sugar can damage blood vessels, nerves, and various organs, leading to complications such as diabetic nephropathy, diabetic retinopathy, and diabetic foot. Hypoglycemia is manifested as a blood sugar lower than 2.8 mmol / L can cause symptoms such as sweating, palpitations, tremors, and hunger. In severe cases, coma may occur. Blood lipids include indicators such as cholesterol, triglycerides, low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C), which reflect the content and metabolism of lipids in the blood. Cholesterol and triglycerides are important components of blood lipids. Excessively high concentrations of them in the blood increase the risk of atherosclerosis. LDL-C is called "bad cholesterol" as it deposits within the blood vessel wall and promotes plaque formation, while HDL-C is called "good cholesterol" as it helps transport cholesterol from the blood vessel wall back to the liver for metabolism. Abnormal blood lipid levels are closely related to cardiovascular diseases such as coronary heart disease and cerebrovascular diseases. Lung function indicators reflect the ventilation function of the lungs. If the lung function indicators of a person are low, it indicates obstructive ventilation dysfunction in the airway of the person. Liver function indicators reflect the metabolic, synthetic, detoxifying, and excretory functions of the liver. When liver cells are damaged, the blood concentration of the person will increase. Gastrointestinal function indicators represent the digestive ability of the gastrointestinal tract to food. Low gastrointestinal function indicators will result in symptoms such as indigestion and loss of appetite. Respiratory rate reflects the working efficiency of the respiratory system and the body's metabolic needs. The normal respiratory rate of an average adult at rest is generally 12 - 20 times per minute. An increased respiratory rate is caused by exercise, fever, lung diseases, and heart diseases. A too slow respiratory rate is usually due to the influence of certain drug poisonings, nervous system diseases, etc. on the regulatory function of the respiratory center. Respiratory rate reflects the working efficiency of the respiratory system and the body's metabolic needs. By processing the historical body data of different subjects, blood pressure, blood sugar, blood lipids, lung function indicators, liver function indicators, gastrointestinal function indicators, and respiratory rate corresponding to different heart rate data of different subjects can be generated. According to the body type data, it reflects various functions of the human body. The characteristic manifestations are associated with various body type data. By separately calculating the average values of the various body data of different subjects when characteristic manifestations occur, standard ecological data of characteristic manifestations are generated. The body data of different subjects when characteristic manifestations occur are subtracted from the standard ecological data of characteristic manifestations and then divided by the standard ecological data of characteristic manifestations to obtain the individual difference degree.

[0104] Refer to Figure 3 As shown, set the risk characteristic recognition threshold, specifically including:

[0105] According to the clinical trial design plan, obtain the corresponding numerical relationship between the clinical trial characteristic type and the initial characteristic information;

[0106] According to the individual difference degree, analyze the initial characteristic information and separately obtain the body type data threshold of each subject;

[0107] According to the corresponding numerical relationship between the clinical trial characteristic type and the initial characteristic information, analyze the body type data threshold of each subject to obtain the abnormal characteristic manifestations of each subject;

[0108] Preprocess the abnormal feature manifestations of each subject and set the first risk feature recognition threshold;

[0109] Analyze the body type data thresholds of each subject according to the type correlation coefficient, obtain the heart beat cycle thresholds of each subject, and set the second risk feature recognition threshold.

[0110] It can be understood that in clinical trials, the physical conditions of subjects vary widely, and these differences cover multiple dimensions. From the basic physiological characteristics, the age span can range from teenagers to the elderly, and the physical functions at different ages, such as metabolic rate and organ function activity, are very different. In terms of gender, men and women are naturally different in hormone levels and body fat distribution, which will affect the absorption, metabolism and distribution of drugs. In terms of health status, some subjects have chronic diseases, such as diabetes and hypertension, and their in-vivo biochemical environment is completely different from that of healthy people. Some have a history of allergies and a more sensitive immune system. At the same time, lifestyle cannot be ignored. The lung and liver functions of subjects who smoke and drink alcohol for a long time are more damaged than those of subjects who maintain a healthy lifestyle. Therefore, the data collected in clinical trials need to fully consider the impact of individual factors on the results to ensure the scientificity and reliability of the research conclusions. According to the individual difference degree, perform weighted processing on the corresponding numerical relationship between the clinical trial feature type and the initial feature information, and different body type data thresholds of subjects with feature manifestations can be obtained. According to the body type data thresholds of different subjects with feature manifestations, divide the corresponding subject body type data, and set the divided data value of the subject as the first risk feature recognition threshold. The heart beat cycle value represents the cycle value of the normal heart beating. The cycle value of the normal heart beating can be obtained by calculating the heart rate. The heart rate can intuitively reflect whether the heart beats regularly. The difference change of the heart beat cycle value reflects the regulatory function of the nervous system on the heart. If the difference change of the heart beat cycle value is relatively high, it means that the body is in a relatively relaxed state, while if the difference change of the heart beat cycle value is relatively low, it means that the body may be under stress or there may be potential pathology. Using the heart beat cycle as the second risk feature recognition threshold, it is possible to judge whether there is a potential risk.

[0111] Refer to Figure 4 As shown, obtain the abnormal coefficient, specifically including:

[0112] Analyze the clinical trial feature type according to the initial feature information to obtain the data collection feature trend;

[0113] Based on the data collection feature trend, analyze the clinical trial data to obtain the recording parameters;

[0114] Compare the recorded parameters with the first risk feature recognition threshold and the second risk feature recognition threshold respectively to obtain a first deviation parameter and a second deviation parameter;

[0115] Analyze the first deviation parameter according to the first risk feature recognition threshold to obtain the parameter deviation frequency;

[0116] Analyze the second deviation parameter according to the second risk feature recognition threshold to obtain the parameter deviation amplitude;

[0117] Analyze the parameter deviation frequency and the parameter deviation amplitude according to the individual difference degree to obtain the anomaly coefficient.

[0118] It can be understood that the physical conditions of the subjects are different, and their sensitivities to drugs also vary. The clinical trial feature type reflects the possible side effect feature manifestations of the subjects' bodies in the clinical trial. By comparing the initial feature information and the subject feature manifestation information, the feature manifestation information brought by the clinical trial to the subjects can be determined. By collecting the time distribution of the clinical trial data corresponding to the feature manifestation information brought by the clinical trial to the subjects, and arranging and sorting the time distribution, the data collection feature trend can be determined. Different drug doses, administration methods, and treatment cycles will result in different data collection methods. Different drug doses may lead to different degrees of efficacy and adverse reactions. Therefore, during the data collection process, various observation indicators of each dose group should be recorded in detail to observe the trend of the data as the dose changes. The administration method is also crucial. For example, oral administration, intravenous injection, etc. The drug absorption speed and effect are different for different methods. During data collection, attention should be paid to the collection and analysis of the drug onset time and blood drug concentration change data under different administration methods. The recorded parameter represents the specific value of the subject's body type data collected. The parameter deviation frequency represents the ratio of the number of times the parameter deviates from the first risk feature recognition threshold to the total number of observations during the trial. The parameter deviation amplitude refers to the degree of difference between the parameter value and the second risk feature recognition threshold during the trial. According to the individual difference degree, the parameter deviation frequency and the parameter deviation amplitude of the subjects are weighted successively to generate the anomaly coefficient.

[0119] Refer to Figure 5 shown to determine whether there are false results, specifically including:

[0120] Determine the standard physical fitness threshold according to the initial feature information;

[0121] Obtain the individual control parameter through the test control parameter and the anomaly coefficient;

[0122] Calculate the standard deviation of the individual control parameter to determine the standard control parameter and obtain the test control distribution trend;

[0123] The standard physical fitness threshold is weighted according to the individual difference degree of each subject to obtain the subject difference threshold;

[0124] The clinical trial data is compared with the subject difference threshold one by one to obtain the recognition result of data abnormality;

[0125] The recognition result of data abnormality is compared with the test control distribution trend to judge whether the two are consistent. If so, there is no false result phenomenon. If not, there is a false result phenomenon, and the judgment situation of the false result phenomenon is generated.

[0126] It can be understood that in clinical trials, due to various factors, such as data errors, experimental design defects, and operator biases, incorrect or inaccurate conclusions may be produced. This situation is usually called false results. False results in clinical trials may lead to ineffective or harmful drugs or therapies entering the market, delaying the treatment of patients, and seriously endangering the life and health of patients. Therefore, it is necessary to judge and process the false results of the collected subject data. By accumulating and averaging the historical body data of all subjects, the average body type data of the subjects can be obtained as the standard physical fitness threshold. Multiplying the test control parameter by the abnormality coefficient can obtain the individualized control parameter. The individualized control parameter represents the independent clinical trial control parameter of each subject. The test control distribution trend refers to the distribution trend of the data where the subject body type data is regulated during the clinical trial process. By subtracting the collected subject body type data from the individualized control parameter, using the difference between the body type data of different subjects and the individualized control parameter as the reference item, the data is sorted and a line chart, the test control distribution trend, is generated.

[0127] Refer to Figure 6 As shown, set the abnormal information threshold, specifically including:

[0128] The judgment situation of the false result phenomenon is sorted to obtain the information of the subjects with false results, and the information of the subjects with false results is classified according to age and health status in turn to obtain the distribution of false results in each age and each health status;

[0129] The distribution of false results in each age and each health status is analyzed respectively to obtain the mean value of age distribution data and the mean value of health status distribution;

[0130] According to the standard control parameter, the mean value of age distribution data and the mean value of health status distribution are analyzed respectively to obtain the standard deviation of age distribution data and the standard deviation of health status distribution data;

[0131] Sort out the age and health status of the subjects, summarize the body type data of the subjects whose age is within one standard deviation of the mean of the age distribution data and whose health status is within one standard deviation of the mean of the health status, and obtain the average peak value and average trough value of the body type data;

[0132] Perform a difference operation on the average peak value and average trough value of the body type data as the abnormal information threshold.

[0133] It can be understood that the abnormal information threshold refers to the numerical boundary used to identify and distinguish normal and abnormal situations during data supervision. When the data is within the set threshold, it is regarded as abnormal information. The abnormal threshold is usually set according to historical data, experience or statistical models to detect abnormal fluctuations, deviations or errors in the data. Based on the abnormal information threshold, compare the clinical trial data to determine whether the clinical trial data is within the abnormal information threshold. If so, determine that the clinical trial data is abnormal and mark it. If not, it means that the clinical trial data is accurate and record it. Calculate the ratio of the accurate clinical trial data in the total obtained clinical trial data as the quality of the clinical trial data.

[0134] Further, referring to Figure 7 shown, a clinical trial data quality supervision system based on risk characteristics is proposed to implement the supervision method as described above, which is characterized in that it specifically includes:

[0135] A data collection module, which is used to obtain the clinical trial design plan, collect the historical ecological data and historical medication data of the subjects, and transmit the collected data to the feature recognition module;

[0136] A feature recognition module, which is used to analyze the received data and extract features, obtain initial feature information, type correlation coefficients and abnormal feature manifestations, generate feature manifestation information, and transmit the data to the data analysis module, risk assessment module and data monitoring module;

[0137] A data analysis module, which is used to analyze, sort out and compare the received data, obtain the individual difference degree, determine the mechanism of action of the drug, the type of drug impact data, the drug effect absorption degree of the subject, the parameter deviation frequency and the parameter deviation amplitude, and transmit the data to the risk assessment module and the data monitoring module;

[0138] A risk assessment module, which is used to sort out and summarize the received data, set the risk feature recognition threshold and the abnormal information threshold, and transmit the set thresholds to the data monitoring module;

[0139] A data monitoring module, which is used to compare the received data, analyze the real-time physical information of the subjects through constant coefficients, determine whether there is a false result phenomenon, compare the clinical trial data through an abnormal information threshold, determine whether the clinical trial data is true, and transmit the judgment result to the quality control module;

[0140] A quality control module, which is used to analyze the received judgment result and evaluate the quality of clinical trial data.

[0141] Further, a feature recognition module, specifically including:

[0142] A first recognition unit, which is used to recognize the historical physical data and historical medication data of the subjects to obtain initial feature information;

[0143] A second recognition unit, which is used to classify the initial feature information according to the clinical trial feature type to generate feature performance information;

[0144] A third recognition unit, which is used to perform feature recognition on the initial feature information to obtain a type correlation coefficient.

[0145] Further, a data analysis module, specifically including:

[0146] A first analysis unit, which is used to analyze the feature performance information through body type data to determine the individual difference degree;

[0147] A second analysis unit, which is used to analyze the historical medication data of the subjects to determine the mechanism of action of the drug and the type of influencing data;

[0148] Further, a risk assessment module, specifically including:

[0149] A first assessment unit, which is used to analyze the body type data threshold of each subject through the corresponding numerical relationship between the clinical trial feature type and the initial feature information, obtain the abnormal feature performance of each subject, and set a first risk feature recognition threshold according to the abnormal feature performance of each subject;

[0150] A second assessment unit, which is used to analyze the body type data threshold of each subject through the type correlation coefficient, obtain the heart rate cycle threshold of each subject, and set a second risk feature recognition threshold;

[0151] The third evaluation unit, which is used to obtain the distribution of false results in different ages and health conditions according to the judgment of false result phenomena, and set the abnormal information threshold by analyzing the distribution of false results in different ages and health conditions.

[0152] In summary, the advantages of the present invention are as follows: it realizes the ranking of risk characteristics of different data, improves the allocation efficiency of resources between low-risk data and high-risk data, conducts a detailed differential evaluation on different patient groups, ensures the integrity, accuracy and reliability of the data on the treatment response and side effects of the subjects, and provides a solid data basis for the subsequent evaluation of drugs or treatment methods.

[0153] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and the principles described in the above embodiments and the specification are only the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for quality supervision of clinical trial data based on risk characteristics, characterized in that, Including: Obtain and analyze the clinical trial design plan, and determine the clinical trial feature type and trial control parameters; Collect and collate the historical physical data and historical medication data of the subjects to obtain initial feature information; Obtain the type correlation coefficient according to the initial feature information; Set the risk feature recognition threshold according to the type correlation coefficient and the initial feature information; Based on a portable medical device, obtain the real-time physical information of the subject, treat the subject according to the clinical trial design plan, and obtain clinical trial data; Analyze the clinical trial data based on the risk feature recognition threshold to obtain the anomaly coefficient; Analyze the real-time physical information of the subject according to the anomaly coefficient to determine whether there is a false result phenomenon; Analyze the trial control parameters and clinical trial data according to the judgment of the false result phenomenon, and set the anomaly information threshold; Supervise the clinical trial data according to the anomaly information threshold, judge whether the clinical trial data is true, and evaluate the quality of the clinical trial data.

2. The method for quality supervision of clinical trial data based on risk characteristics according to claim 1, wherein, The obtaining of the type correlation coefficient specifically includes: Sort out the historical physical data of the subject to generate physical type data, and the physical type data includes heart rate, blood pressure, blood sugar, blood lipid, lung function index, liver function index, gastrointestinal function index and respiratory rate; Analyze the physical type data of each subject, use the heart rate as the reference data, and obtain the corresponding relationship of physiological data; Classify the initial feature information according to the clinical trial feature type to generate feature performance information; Analyze the feature performance information according to the physical type data to determine the individual difference degree; Analyze the historical medication data of the subject to determine the mechanism of action of the drug and the type of data affected; Classify the subjects according to the mechanism of action of the drug to generate the same drug mechanism group; Based on the historical medication records of the subjects, analyze the historical physical data of each subject in the same drug mechanism group to generate the historical physical data of the subject corresponding to the type of data affected by the drug, and obtain the drug effect absorption degree of the subject; Classify the subjects according to the drug effect absorption degree of the subject, divide the subjects with the same drug effect absorption degree into one group to generate the same physical type group, and obtain the number of subjects in each type group; Randomly select two subjects in the same physical type group, analyze the corresponding relationship of physiological data, calculate the average value of the heart rate physiological fluctuations of the two subjects, and repeat this calculation step in the group with the same physical type. The number of repetitions is equal to the number of subjects in this physical type group to obtain multiple groups of average values of heart rate physiological fluctuations; Perform weighted averaging on multiple groups of average values of heart rate physiological fluctuations according to the individual difference degree to obtain the type correlation coefficient.

3. The method for quality supervision of clinical trial data based on risk characteristics according to claim 2, wherein The setting of the risk feature recognition threshold specifically includes: Obtain the corresponding numerical relationship between the clinical trial feature type and the initial feature information according to the clinical trial design plan; Analyze the initial feature information according to the individual difference degree, and respectively obtain the physical type data threshold of each subject; Analyze the physical type data threshold of each subject according to the corresponding numerical relationship between the clinical trial feature type and the initial feature information to obtain the abnormal feature performance of each subject; Preprocess the abnormal feature manifestations of each subject and set the first risk feature recognition threshold; Analyze the body type data threshold of each subject according to the type correlation coefficient, obtain the heart beat cycle threshold of each subject, and set the second risk feature recognition threshold.

4. A method for quality supervision of clinical trial data based on risk characteristics according to claim 3, characterized in that The obtaining of the abnormal coefficient specifically includes: Analyze the clinical trial feature type according to the initial feature information to obtain the data acquisition feature trend; Analyze the clinical trial data based on the data acquisition feature trend to obtain the recording parameters; Compare the recording parameters with the first risk feature recognition threshold and the second risk feature recognition threshold respectively to obtain the first deviation parameter and the second deviation parameter; Analyze the first deviation parameter according to the first risk feature recognition threshold to obtain the parameter deviation frequency; Analyze the second deviation parameter according to the second risk feature recognition threshold to obtain the parameter deviation amplitude; Analyze the parameter deviation frequency and the parameter deviation amplitude according to the individual difference degree to obtain the abnormal coefficient.

5. The method for quality supervision of clinical trial data based on risk characteristics according to claim 4, characterized in that, The judgment of whether there is a false result specifically includes: Determine the standard physical fitness threshold according to the initial feature information; Obtain the individual control parameter through the test control parameter and the abnormal coefficient; Calculate the standard deviation of the individual control parameter to determine the standard control parameter and obtain the test control distribution trend; Weight the standard physical fitness threshold according to the individual difference degree of each subject to obtain the subject difference threshold; Compare the clinical trial data with the subject difference threshold one by one to obtain the data anomaly situation recognition result; Compare the data anomaly situation recognition result with the test control distribution trend to judge whether the two are consistent. If so, there is no false result phenomenon. If not, there is a false result phenomenon, and generate the false result phenomenon judgment situation.

6. The method for quality supervision of clinical trial data based on risk characteristics according to claim 5, wherein, The setting of the abnormal information threshold specifically includes: Sort out the false result phenomenon judgment situation to obtain the false result subject information, classify the false result subject information according to age and health status in turn, and respectively obtain the distribution of false results in each age and each health status; Analyze the distribution of false results in each age and each health status respectively to obtain the mean value of age distribution data and the mean value of health status distribution; Analyze the mean value of age distribution data and the mean value of health status distribution according to the standard control parameter respectively to obtain the standard deviation of age distribution data and the standard deviation of health status distribution data; Sort out the age and health status of the subjects, and summarize the body type data of the subjects whose age is within one standard deviation of the mean value of age distribution data and whose health status is within one standard deviation of the mean value of health status to obtain the average peak value of body type data and the average valley value of body type data; Perform a difference processing on the average peak value of body type data and the average valley value of body type data as the abnormal information threshold.

7. A clinical trial data quality supervision system based on risk characteristics, applicable to the supervision method described in any one of claims 1-6, characterized in that, Specifically include: A data acquisition module, which is used to obtain the clinical trial design scheme, collect the historical ecological data and historical medication data of the subjects, and transmit the collected data to the feature recognition module; A feature recognition module, which is used to analyze the received data, extract features, obtain initial feature information, type correlation coefficients, and abnormal feature manifestations, generate feature manifestation information, and transmit the data to the data analysis module, risk assessment module, and data monitoring module; A data analysis module, which is used to analyze, sort out, and compare the received data, obtain individual difference degrees, determine the mechanism of action of drugs, the types of drug impact data, the drug effect absorption degrees of subjects, parameter deviation frequencies, and parameter deviation amplitudes, and transmit the data to the risk assessment module and data monitoring module; A risk assessment module, which is used to sort out and summarize the received data, set risk feature recognition thresholds and abnormal information thresholds, and transmit the set thresholds to the data monitoring module; A data monitoring module, which is used to compare the received data, analyze the real-time physical information of subjects through constant coefficients to determine whether there is a false result phenomenon, compare the clinical trial data through the abnormal information threshold to determine whether the clinical trial data is true, and transmit the judgment result to the quality control module; A quality control module, which is used to analyze the received judgment results and evaluate the quality of clinical trial data.

8. A quality supervision system for clinical trial data based on risk characteristics according to claim 7, characterized in that, The feature recognition module specifically includes: A first recognition unit, which is used to recognize the historical physical data and historical medication data of subjects to obtain initial feature information; A second recognition unit, which is used to classify the initial feature information according to the clinical trial feature types to generate feature manifestation information; A third recognition unit, which is used to perform feature recognition on the initial feature information to obtain type correlation coefficients.

9. A quality supervision system for clinical trial data based on risk characteristics according to claim 7, characterized in that The data analysis module specifically includes: A first analysis unit, which is used to analyze the feature manifestation information through body type data to determine the individual difference degree; A second analysis unit, which is used to analyze the historical medication data of subjects to determine the mechanism of action of drugs and the types of impact data.

10. A quality supervision system for clinical trial data based on risk characteristics according to claim 7, characterized in that, The risk assessment module specifically includes: A first assessment unit, which is used to analyze the body type data thresholds of each subject through the corresponding numerical relationship between the clinical trial feature types and the initial feature information, obtain the abnormal feature manifestations of each subject, and set the first risk feature recognition threshold according to the abnormal feature manifestations of each subject; A second assessment unit, which is used to analyze the body type data thresholds of each subject through the type correlation coefficient, obtain the heart rate cycle thresholds of each subject, and set the second risk feature recognition threshold; A third assessment unit, which is used to obtain the distribution of false results in each age and each health condition according to the judgment of the false result phenomenon, and set the abnormal information threshold by analyzing the distribution of false results in each age and each health condition.

Images