A clinical trial data intelligent analysis method and system

Abstract

The present invention relates to the field of data processing technology, and more specifically, to a method and system for intelligent analysis of clinical trial data. The method comprises: obtaining a data sequence of each physiological indicator of a patient, calculating the degree of deviation by quantifying the difference between a target data point and the normal range of the corresponding physiological indicator, calculating an abnormality index based on the degree of deviation and the distribution characteristics of the corresponding data in the data sequence of the corresponding physiological indicator, and obtaining a fitting curve for each physiological indicator by fitting the abnormality index. The method further comprises screening time periods in which all fitting curves show an upward trend, selecting physiological indicator pairs with correlations greater than a preset value within the time period that meet the screening conditions as target indicator pairs, and constructing a causal graph based on the target indicator pairs using a PC algorithm to identify the causal relationships between the physiological indicators. The present invention can reduce the complexity of constructing the causal graph using the PC algorithm while retaining the important causal relationships between the physiological indicators.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and more specifically, to a method and system for intelligent analysis of clinical trial data. Background Art

[0002] With the rapid development of modern clinical trials, data collection technologies are becoming increasingly diverse. Data sources such as electronic medical records, gene sequencing, wearable sensors, and medical imaging generate massive amounts of heterogeneous data. In remote patient monitoring scenarios, wearable devices collect multi-dimensional physiological indicators such as heart rate, blood pressure, blood sugar levels, and exercise intensity in real time, generating high-density, high-dimensional time-series data streams. The scale and complexity of this data pose significant challenges to traditional causal inference methods.

[0003] In related technologies, constraint-based causal discovery algorithms, such as the Peter-Clark Algorithm (PC) algorithm, can identify causal relationships between different health indicators and are often used to construct causal graphs between health indicators and predict intervention effects. The core steps of the PC algorithm include: (1) constructing a completely undirected graph; (2) gradually eliminating irrelevant edges through zero-order and higher-order conditional independence tests; and (3) determining the direction of causal relationships based on directional rules.

[0004] However, in high-dimensional, large-scale data scenarios, high-order conditional independence tests require traversing an exponentially growing number of conditional variable combinations, resulting in a sharp increase in computational complexity. For example, for a dataset containing dozens of physiological indicators such as heart rate, blood pressure, and blood sugar levels, the algorithm must perform millions of independence tests. This is not only time-consuming but also prone to misjudgments (such as false causal relationships or missing true associations) due to multiple hypothesis testing. This can affect the prediction of the effectiveness of intervention measures, lead to biased clinical decision-making, and reduce system credibility. Summary of the Invention

[0005] In order to solve the problem that the PC algorithm has high computational complexity in identifying the causal relationship between various physiological indicators due to the huge amount of clinical trial data, which in turn affects the accuracy of the constructed causal graph, the present invention provides a method and system for intelligent analysis of clinical trial data.

[0006] According to a first aspect of the present invention, a method for intelligent analysis of clinical trial data is provided, comprising:

[0007] Acquire and preprocess the patient's clinical trial data to obtain time-aligned data series of various physiological indicators, including heart rate, blood pressure, and blood sugar levels;

[0008] Selecting a target data point from the data sequence of the target physiological indicator, and determining the degree of deviation of the target data point by quantifying the difference between the target data point and the normal range of the target physiological indicator at the corresponding time;

[0009] Based on the degree of deviation, and the frequency and average time interval of data points with the same value as the target data point in the data sequence of the target physiological indicator, an abnormality index of the target data point is calculated. The abnormality index is positively correlated with the frequency and the degree of deviation, and negatively correlated with the average time interval.

[0010] By fitting the abnormal index of the data points in the data series of each physiological indicator, the fitting curve of each physiological indicator is obtained, and the time period in which all the fitting curves show an upward trend is screened. The physiological indicator pairs with a correlation greater than the preset value in the time period that meets the screening conditions are used as target indicator pairs. The PC algorithm is used to construct a causal graph based on the target indicator pairs to identify the causal relationship of each physiological indicator.

[0011] The present invention comprehensively considers the abnormal index and time persistence, screens target indicator pairs, and can accurately identify physiological indicators that have a significant impact on the patient's health. The causal graph constructed based on the screened target indicators can intuitively display the causal relationship between the patient's various physiological indicators, thereby reducing the amount of data that needs to be processed when the PC algorithm constructs the causal graph while ensuring the accuracy of the constructed causal graph.

[0012] Preferably, the method for obtaining the deviation degree of the target data point includes:

[0013] If the value of the target data point is greater than the upper limit of the normal range, the difference between the value of the target data point and the upper limit, and the ratio of the upper limit, is taken as the deviation degree of the target data point;

[0014] If the value of the target data point is less than the lower limit of the normal range, the difference between the lower limit and the value of the data point, and the ratio of the lower limit, is taken as the deviation degree of the target data point;

[0015] If the value of the target data point is within the normal range, the deviation degree of the target data point is set to zero.

[0016] The present invention can standardize the degree of deviation to a relative ratio, which enables the deviation degrees of different indicators to be compared on the same scale, avoiding the difficulty of comparison caused by differences in indicator dimensions or numerical ranges.

[0017] Preferably, when the healthy range of the target physiological indicator is affected by the patient's exercise volume, the method for obtaining the normal range includes:

[0018] The difference between the patient's exercise volume at the time corresponding to the target data point and the preset exercise volume threshold is multiplied by the preset weight as the variation to obtain the static health range of the target physiological indicator. Based on the variation, the lower and upper bounds of the static health range are updated through a summation operation;

[0019] Get the maximum value of the lower bound values ​​before and after the update, and the minimum value of the upper bound values ​​before and after the update, and use the range consisting of the maximum and minimum values ​​as the normal range.

[0020] By dynamically adjusting the normal range of each physiological indicator, the present invention can avoid misjudging normal changes in physiological indicators affected by the patient's exercise volume as abnormalities, thereby ensuring the accuracy of the deviation degree of each data point.

[0021] Preferably, when the healthy range of the target physiological indicator is a fixed range, the static healthy range of the target physiological indicator is used as the normal range of the target physiological indicator at each moment.

[0022] Preferably, the method for obtaining the abnormality index of the target data point includes:

[0023] The opposite number of the average time interval is used as the power of the exponential function, and the power operation is performed to obtain the severity index;

[0024] The occurrence frequency is normalized, the normalized value obtained is added to the severity index, and the product of the sum and the severity is normalized to obtain the anomaly index of the target data point.

[0025] The present invention can comprehensively consider the frequency and time interval of anomalies, and obtain a quantitative anomaly index through weighted accumulation and normalization processing. It is comparable, so that the anomaly indices of different data points can be compared on the same scale, thereby providing a quantitative basis for subsequent screening operations.

[0026] Preferably, obtaining a fitting curve of each physiological indicator by fitting the abnormal index of the data point in the data sequence of each physiological indicator includes:

[0027] The least square method was used to perform curve fitting on the abnormal index of all data points in the data series of each physiological index to obtain the fitting curve of each physiological index.

[0028] Preferably, the time period in which all the fitting curves show an upward trend is selected, including:

[0029] Obtain the slope value of each data point in each fitted curve. If the slope values ​​of all data points on the fitted curve in any time period are positive, retain that time period to obtain the time period that meets the screening conditions.

[0030] Preferably, the method for obtaining the target indicator pair includes:

[0031] For any time period that meets the screening conditions, the absolute value of the Spearman correlation coefficient of the data series of any two physiological indicators in any time period is calculated to obtain the correlation of the corresponding physiological indicator pairs, and the physiological indicator pairs with a correlation greater than the preset correlation threshold are taken as the target indicator pairs.

[0032] The present invention utilizes correlation to screen target indicator pairs, which can ensure that the screened target indicator pairs have a relatively important causal relationship, thereby providing a data basis for the subsequent construction of a causal graph.

[0033] Preferably, when constructing a causal diagram based on the target indicator pair using the PC algorithm, the conditional variable of any target indicator pair is a physiological indicator whose average correlation with any target indicator pair in the same time period is greater than a set value.

[0034] According to a second aspect of the present invention, a method for intelligent analysis of clinical trial data is provided. The system includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the first aspect of the present invention.

[0035] The present invention has the following effects:

[0036] The present invention determines the abnormality index of each data point by integrating multiple indicators, and can accurately evaluate the possibility of abnormality of each data point. It can also facilitate the determination of the changing trend of each physiological indicator through fitting. Therefore, based on the abnormal duration of each physiological indicator, by evaluating the correlation between any two physiological indicator pairs, physiological indicator pairs with important relationships, i.e., target indicator pairs, can be quickly and accurately screened out. When constructing a causal graph based on the screened target indicator pairs, the causal relationship between the physiological indicators can be more accurately identified, thereby effectively reducing the computational complexity of the PC algorithm when processing large-scale data while ensuring the accuracy of the constructed causal graph. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the following detailed description with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an illustrative and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0038] Figure 1 This is a schematic diagram of the steps of a method for intelligent analysis of clinical trial data according to an embodiment of the present invention. DETAILED DESCRIPTION

[0039] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work shall fall within the scope of protection of the present invention.

[0040] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0041] Reference Figure 1 A method for intelligent analysis of clinical trial data includes steps S1 to S4, specifically as follows:

[0042] S1: Obtain the patient's clinical trial data and preprocess it to obtain a time-aligned data sequence of various physiological indicators, including heart rate, blood pressure, and blood sugar level.

[0043] Specifically, the patient's clinical trial data (such as heart rate, blood pressure, blood sugar level, etc.) in the past day can be continuously collected from the patient's wearable devices, sensors or electronic medical record systems at a fixed sampling frequency, such as 1 time / minute, to obtain a heart rate data sequence, a blood pressure data sequence, and a blood sugar level data sequence, etc.

[0044] Optionally, the total number of samples can be recorded as , then the heart rate data series can be Recorded as ; Blood pressure data series Recorded as ; Transform blood glucose level data series Recorded as Where, 、 as well as Respectively Heart rate, blood pressure and blood sugar level values ​​collected; is the sampling order; is the total number of sampling times.

[0045] It should be noted that in order to eliminate the influence of different physiological index dimensions and align the data series of each physiological index in time, it is necessary to perform normalization processing on the data series of each physiological index separately, such as maximum and minimum value normalization processing.

[0046] Furthermore, when different types of sensors have different sampling frequencies, resampling techniques (such as upsampling or downsampling) may be required to temporally align the data series of each physiological indicator.

[0047] Optionally, when any physiological indicator at any sampling moment has a missing value, linear interpolation is used to replace the missing value to ensure the data integrity of the physiological indicator at each sampling moment. Among them, normalization processing, resampling technology and linear interpolation are all existing technologies and are not described in detail in this embodiment.

[0048] S2: Select a target data point from the data sequence of the target physiological indicator, and determine the degree of deviation of the target data point by quantifying the difference between the target data point and the normal range of the target physiological indicator at the corresponding moment.

[0049] Among them, the target physiological indicator refers to a randomly selected physiological indicator; the target data point refers to a data point randomly selected from the data sequence of the target physiological indicator; and the degree of deviation refers to the deviation between the value of the target data point and the normal level of the target physiological indicator.

[0050] In an exemplary embodiment of the present invention, when the healthy range of the target physiological indicator is affected by the patient's exercise amount, the normal range of the target physiological indicator at each moment can be determined by the following steps:

[0051] The difference between the patient's exercise volume at the time corresponding to the target data point and the preset exercise volume threshold, multiplied by the preset weight, is used as the variation to obtain the static health range of the target physiological indicator, and based on the variation, the lower and upper bounds of the static health range are updated through a summation operation; the maximum value of the lower bound values ​​before and after the update, and the minimum value of the upper bound values ​​before and after the update are obtained, and the range consisting of the maximum and minimum values ​​is taken as the normal range.

[0052] It should be noted that changes in physiological indicators (such as heart rate and blood sugar levels) during exercise are normal physiological reactions and not health issues. Using the healthy range at rest (static healthy range) to evaluate data during exercise can lead to misjudgments. Therefore, this embodiment proposes a method for dynamically adjusting the healthy range. For physiological indicators affected by the patient's exercise level, the normal range is adjusted in real time based on the patient's exercise status, thereby avoiding misjudgments and improving assessment accuracy.

[0053] Specifically, the normal range of heart rate at any time satisfies the following relationship:

[0054] ;

[0055] Where, For heart rate Normal range at all times; 60 and 100 are the lower and upper limits of the static healthy range of heart rate, respectively. The static healthy range of heart rate refers to the healthy range of heart rate monitored when the patient's exercise volume is lower than the set exercise volume threshold, usually 60-100 beats / minute; 、 They are functions that return maximum and minimum values ​​respectively; For patients in The amount of exercise at any given moment; To preset the exercise threshold, The value is the doctor's assessment of the patient's actual exercise capacity, which can effectively distinguish the patient's physiological state; 0.01 is the preset weight, which is used to control the adjustment range of the heart rate boundary value.

[0056] Normal range of blood sugar levels at any given time , satisfying the following relationship:

[0057] ;

[0058] Where, For blood sugar levels Normal range at all times; 3.9 and 6.1 are the lower and upper limits of the static healthy range of blood sugar levels, respectively. The static healthy range of blood sugar levels refers to the healthy range of blood sugar levels monitored when the patient's exercise volume is lower than the set exercise volume threshold, usually 3.9-6.1mmol / L; 0.02 is the preset weight of the blood sugar level, which is used to control the adjustment range of the blood sugar level boundary value.

[0059] In an exemplary embodiment of the present invention, when the healthy range of the target physiological indicator is a fixed range, the static healthy range of the target physiological indicator is used as the normal range of the target physiological indicator at each moment.

[0060] For example, for physiological indicators whose healthy range is not affected by the patient's exercise level, such as blood pressure, the static healthy range of blood pressure can be directly used as the normal range of blood pressure at each moment. The static healthy range of blood pressure refers to the healthy range of blood pressure monitored when the patient's exercise level is below the set exercise threshold, usually 70-100 mmHg.

[0061] Furthermore, after the normal range of each physiological indicator at each moment is determined, the degree of deviation of each data point in the data sequence of each physiological indicator can be calculated.

[0062] In an exemplary embodiment of the present invention, the following steps may be used to determine the degree of deviation of each data point in the data sequence of each physiological indicator:

[0063] If the value of the target data point is greater than the upper limit of the normal range, the difference between the value of the target data point and the upper limit, and the ratio of the upper limit, is taken as the deviation degree of the target data point;

[0064] If the value of the target data point is less than the lower limit of the normal range, the difference between the lower limit and the value of the data point, and the ratio of the lower limit, is taken as the deviation degree of the target data point;

[0065] If the value of the target data point is within the normal range, the deviation degree of the target data point is set to zero.

[0066] Optionally, if the target data point is the first data points, the heart rate can be The lower limit of the normal range at a moment is recorded as , the heart rate The upper limit of the normal range at a moment is recorded as , then the degree of deviation of the target data point satisfies the relationship:

[0067] ;

[0068] Where, is the first The degree of deviation of the data points; is the first data points; 、 Heart rate in the The lower and upper limits of the normal range at each moment; represent The value of the heart rate is within the normal range at this moment.

[0069] It should be noted that the method for determining the deviation degree of each data point in the data series of other physiological indicators is the same as the method for determining the deviation degree of each data point in the heart rate data series, and is not described in detail in this embodiment.

[0070] S3: Based on the degree of deviation, as well as the frequency and average time interval of data points with the same value as the target data point in the data sequence of the target physiological indicator, calculate the abnormality index of the target data point. The abnormality index is positively correlated with the frequency and the degree of deviation, and negatively correlated with the average time interval.

[0071] In an exemplary embodiment of the present invention, the abnormality index of each data point in the data sequence of each physiological indicator can be determined by the following steps:

[0072] The opposite of the average time interval is used as the power of the exponential function, and the power operation is performed to obtain the severity index; the occurrence frequency is normalized, and the cumulative sum of the normalized value and the severity index is calculated, and the product of the cumulative sum and the severity is normalized to obtain the anomaly index of the target data point.

[0073] For example, when the target data point is the first When there are data points, the abnormal index of the target data point satisfies the following relationship:

[0074] ;

[0075] Where, is the first Anomaly index of data points; is the first The degree of deviation of the data points; The value of the heart rate data sequence is The number of data points that are identical; is the total number of data points in the heart rate data series; The value of the heart rate data sequence is The average time interval between data points with the same data points; The natural constant An exponential function with base ; is the normalization function.

[0076] in, Reflects the difference between the value in the heart rate data sequence and the The frequency of occurrence of data points with the same data points; Reflects the first The severity index of the data point. The larger the value, the more serious the heart rate data sequence is. The more densely the data points are distributed, the more likely it is that When it is larger, it means that the probability of abnormal heart rate at that moment is greater, and the corresponding abnormal index is relatively large.

[0077] Optionally, other normalization methods, such as a sigmoid function, may be used for normalization. This embodiment does not specifically limit the selected normalization method.

[0078] It should be noted that the method for obtaining the abnormality index of each data point in the data series of other physiological indicators is the same as the method for determining the abnormality index of each data point in the heart rate data series, and is not described in detail in this embodiment.

[0079] In another embodiment, the relationship: , calculate the abnormal index of each data point in the heart rate data series.

[0080] S4: By fitting the abnormal index of the data points in the data sequence of each physiological indicator, the fitting curve of the abnormal index of each physiological indicator is obtained, and the time period in which all the fitting curves show an upward trend is screened. The physiological indicator pairs with a correlation greater than the preset value in the time period that meets the screening conditions are used as target indicator pairs. The PC algorithm is used to construct a causal graph based on the target indicator pairs to identify the causal relationship of each physiological indicator.

[0081] It should be noted that when the present invention uses the PC algorithm to construct a causal diagram, it can reduce the number of physiological indicators that need to be processed by screening and retaining more important variables, thereby effectively reducing the computational complexity while ensuring the accuracy of the causal relationship in the constructed causal diagram.

[0082] Among them, the present invention only adds the data screening process when constructing the causal graph, and does not improve other contents of the PC algorithm, such as the construction method of the undirected graph, the test process of conditional independence, and the determination process of directed edges.

[0083] In an exemplary embodiment of the present invention, the determination of the fitting curve of each physiological indicator can be achieved by the following steps:

[0084] The least square method was used to perform curve fitting on the abnormal index of all data points in the data series of each physiological index to obtain the fitting curve of each physiological index.

[0085] For example, for a heart rate data sequence, the least squares method can be used to perform curve fitting on the anomaly index of all data points in the heart rate data sequence, thereby obtaining a fitted curve for the heart rate. Similarly, the fitting curves for blood glucose level and blood pressure can be determined using the same method for determining the fitting curve for the heart rate. It should be noted that the process of curve fitting using the least squares method is conventional and will not be described in detail in this embodiment.

[0086] Optionally, other fitting methods, such as spline interpolation, may be used to perform curve fitting. This embodiment does not impose any particular limitation on the selected fitting method.

[0087] Furthermore, after obtaining the fitted curves for each physiological indicator, the time intervals in which all the fitted curves show a synchronous upward trend can be screened out, thereby determining one or more qualifying time periods. It should be noted that the present invention aims to measure the duration of increases in the abnormality index of all physiological indicators by screening the time intervals in which all the fitted curves show a synchronous upward trend. These screened time periods are precisely the periods in which all physiological indicators are most likely to be abnormal, thus providing an important basis for the subsequent determination of target indicator pairs.

[0088] In an exemplary embodiment of the present invention, the time period that satisfies the screening condition may be determined by the following steps:

[0089] Obtain the slope value of each data point in each fitted curve. If the slope values ​​of all data points on the fitted curve in any time period are positive, retain that time period to obtain the time period that meets the screening conditions.

[0090] It should be noted that in the mathematical representation of a fitted curve, the first-order derivative (i.e., slope) of each data point essentially reflects the rate of change of the physiological indicator around that time point. Its sign and magnitude directly quantify the direction and intensity of the trend. Based on this characteristic, the present invention accurately assesses the changing trend of each physiological indicator by calculating the slope value of each data point in each fitted curve, thereby screening one or more time periods that meet the criteria.

[0091] In an exemplary embodiment of the present invention, the determination of the target indicator pair can be achieved by the following steps:

[0092] For any time period that meets the screening conditions, the absolute value of the Spearman correlation coefficient of the data series of any two physiological indicators in any time period is calculated to obtain the correlation of the corresponding physiological indicator pairs, and the physiological indicator pairs with a correlation greater than the preset correlation threshold are taken as the target indicator pairs.

[0093] For example, the correlation threshold can be set to 0.5, and any time period that meets the screening conditions is recorded as [ ], the Spearman correlation coefficient between any two data series in the heart rate data series, blood pressure data series, and blood sugar level data series of the period can be calculated, denoted as ,in, as well as are any two different physiological indicators. When 、 These two physiological indicators are used to obtain a target indicator pair within the time period, and then all target indicator pairs within the time period that meet the screening conditions can be obtained. This embodiment does not specifically limit the size of the correlation threshold.

[0094] In another embodiment, the absolute value of the Pearson correlation coefficient of any two data sequences can also be used as the correlation between the corresponding two data sequences. It should be noted that the determination methods of the Spearman correlation coefficient and the Pearson correlation coefficient are both existing technologies and will not be described in detail in this embodiment.

[0095] Furthermore, after all target indicator pairs are determined, the PC algorithm can be used to construct corresponding causal graphs for all target indicator pairs through the steps of initializing a completely undirected graph, gradually testing conditional independence, and determining edge directions. It should be noted that the process of constructing a causal graph using the PC algorithm given known input data is conventional technology and will not be described in detail in this embodiment.

[0096] In an exemplary embodiment of the present invention, when constructing a causal graph based on target indicator pairs using a PC algorithm, the conditional variable of any target indicator pair is a physiological indicator whose average correlation with the target indicator pair in the same time period is greater than a set value.

[0097] For example, when determining in [ During this period, 、 When the target index is the right one, we can calculate 、 The data series of physiological indicators other than 、 The correlation of the data series in the time period is calculated and averaged. If the resulting average value is greater than a set value, such as 0.6, the corresponding physiological indicator can be used as the conditional variable for the target indicator pair. Based on the determined conditional variable, a stepwise conditional independence test step can be performed on the target indicator pair, thereby improving the test efficiency when performing the stepwise conditional independence test step on each target indicator pair. The method for determining the correlation in this embodiment is the same as the method for calculating the correlation when screening the target indicator pair.

[0098] Furthermore, once a patient's causal graph is determined, the causal relationships between the patient's various physiological indicators can be identified by comparing the patient's causal graph with that of the general population. For example, if a patient's causal graph shows that blood sugar levels have a significant impact on heart rate, special attention should be paid to the patient's blood sugar control, thereby providing doctors with more targeted reference for diagnosis and treatment.

[0099] The present invention also provides a clinical trial data intelligent analysis system, which includes a memory and a processor, and a computer program is stored in the memory. The computer program integrates the functions of a clinical trial data intelligent analysis method. When the computer program is executed, the clinical trial data intelligent analysis method can reduce the complexity of the PC algorithm in constructing a causal graph while retaining the important causal relationship between various physiological indicators.

[0100] In the description of this specification, "multiple" and "several" mean at least two, such as two, three or more, unless otherwise clearly defined.

[0101] While several embodiments of the present invention have been shown and described herein, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Numerous modifications, variations, and alternatives will occur to those skilled in the art without departing from the concept and spirit of the present invention. It should be understood that various alternatives to the embodiments of the present invention described herein may be employed in practicing the present invention.

Claims

1. A method for intelligent analysis of clinical trial data, characterized in that: include: Acquiring and preprocessing patient clinical trial data to obtain time-aligned data sequences of various physiological indicators, including heart rate, blood pressure, and blood glucose levels; Selecting a target data point from the data sequence of the target physiological indicator, and determining the degree of deviation of the target data point by quantifying the difference between the target data point and the normal range of the target physiological indicator at the corresponding time; When the healthy range of the target physiological indicator is affected by the patient's exercise volume, the method for obtaining the normal range includes: The difference between the patient's exercise volume at the time corresponding to the target data point and a preset exercise volume threshold is multiplied by a preset weight as a variation to obtain a static healthy range of the target physiological indicator, and based on the variation, the lower and upper bounds of the static healthy range are updated through a summation operation; Obtaining the maximum value of the lower bound values ​​before and after the update, and the minimum value of the upper bound values ​​before and after the update, and taking the range formed by the maximum value and the minimum value as the normal range; Based on the degree of deviation, and the frequency and average time interval of data points with the same value as the target data point in the data sequence of the target physiological indicator, an abnormality index of the target data point is calculated. The abnormality index is positively correlated with the frequency and the degree of deviation, and negatively correlated with the average time interval. By fitting the abnormal index of the data points in the data series of each physiological indicator, the fitting curve of each physiological indicator is obtained, and the time period in which all the fitting curves show an upward trend is screened. The physiological indicator pairs with a correlation greater than a preset value in the time period that meets the screening conditions are used as target indicator pairs. Based on the target indicator pairs, a causal graph is constructed using the PC algorithm to identify the causal relationship of each physiological indicator; When it is determined in [ During this period, 、 When the target index is the same, the calculation is divided by 、 The data series of physiological indicators other than 、 The correlation of the data series in this time period is calculated, and the sum is taken and the average is taken. If the average value obtained is greater than the set value, the corresponding physiological indicator is used as the conditional variable of the target indicator pair, so as to perform a stepwise conditional independence test step on the target indicator pair based on the determined conditional variable to improve the test efficiency when performing the stepwise conditional independence test step on each target indicator pair.

2. A clinical trial data intelligent analysis method according to claim 1, characterized in that: The method for obtaining the deviation degree of the target data point includes: If the value of the target data point is greater than the upper limit of the normal range, the difference between the value of the target data point and the upper limit, and the ratio of the upper limit, is used as the deviation degree of the target data point; If the value of the target data point is less than the lower limit of the normal range, the difference between the lower limit and the value of the data point, and the ratio of the lower limit, is used as the deviation degree of the target data point; If the value of the target data point is within the normal range, the deviation degree of the target data point is set to zero.

3. A clinical trial data intelligent analysis method according to claim 2, characterized in that: When the healthy range of the target physiological indicator is a fixed range, the static healthy range of the target physiological indicator is used as the normal range of the target physiological indicator at each moment.

4. A clinical trial data intelligent analysis method according to claim 1, characterized in that: The method for obtaining the abnormality index of the target data point includes: Taking the opposite number of the average time interval as the power of the exponential function, performing a power operation to obtain a severity index; The occurrence frequency is normalized, the normalized value obtained is added to the severity index, and the product of the added value and the severity is normalized to obtain an abnormality index of the target data point.

5. The method for intelligent analysis of clinical trial data according to claim 1, characterized in that: The step of obtaining a fitting curve for each physiological indicator by fitting the abnormality index of the data points in the data sequence of each physiological indicator includes: The least square method was used to perform curve fitting on the abnormal index of all data points in the data series of each physiological index to obtain the fitting curve of each physiological index.

6. A clinical trial data intelligent analysis method according to claim 5, characterized in that: The time period in which all the fitting curves are screened and show an upward trend includes: Obtain the slope value of each data point in each of the fitting curves. If the slope values ​​of all data points on the fitting curves in any time period are positive, retain the time period to obtain the time period that meets the screening condition.

7. The method for intelligent analysis of clinical trial data according to claim 1, characterized in that: The method for obtaining the target indicator pair includes: For any time period that meets the screening conditions, the absolute value of the Spearman correlation coefficient of the data series of any two physiological indicators in any time period is calculated to obtain the correlation of the corresponding physiological indicator pairs, and the physiological indicator pairs with a correlation greater than the preset correlation threshold are taken as the target indicator pairs.

8. The method for intelligent analysis of clinical trial data according to claim 1, characterized in that: When constructing a causal diagram based on the target indicator pair using the PC algorithm, the conditional variable of any target indicator pair is a physiological indicator whose average correlation with any target indicator pair in the same period is greater than a set value.

9. A clinical trial data intelligent analysis system, characterized in that: The clinical trial data intelligent analysis system includes a memory and a processor, the memory stores a computer program, and the processor executes the computer program to implement the steps of the clinical trial data intelligent analysis method as described in any one of claims 1-8.

Images