An intelligent adverse event reporting method for clinical trials of a new drug for endocrine nephropathy

Abstract

The present application relates to the technical field of medical informatics, in particular to a kind of endocrine nephropathy new drug clinical test adverse event intelligent reporting method.Method includes: obtaining the different index data of endocrine nephropathy subject before and after new drug test;Based on the deviation of each kind of index data and benchmark value at each collection time, abnormal index is filtered and abnormal period is determined;Intersection accumulation processing is carried out to abnormal index in abnormal period to obtain intersection period;According to the time sequence waveform similarity of each kind of index data in intersection period, obtain physiological state consistent factor, and determine core period in abnormal period;According to the abnormal index and physiological state consistent factor in core period, the degree of drug-related risk is evaluated;Global risk factor is obtained by comprehensively considering drug-related risk degree and the length of abnormal period, and then corresponding hierarchical safety control and adverse event reporting operation is executed.The present application improves the accuracy of adverse event reporting of endocrine nephropathy new drug clinical test.

Description

Technical Field

[0001] This invention relates to the field of medical informatics technology, specifically to an intelligent reporting method for adverse events in clinical trials of a new drug for endocrine nephropathy. Background Technology

[0002] With the rapid development of precision medicine, artificial intelligence, and innovative drug development technologies, the research and development of new drugs in the field of endocrine nephrology has entered a stage of rapid development. Endocrine nephropathy refers to diseases that damage the structure and function of the kidneys due to abnormalities in endocrine metabolism. Among them, diabetic nephropathy, metabolic syndrome-related nephropathy, obesity-related nephropathy, and hyperuricemic nephropathy have become important causes of end-stage renal disease. In recent years, new drugs targeting mechanisms of action such as renal inflammation, fibrosis, metabolic reprogramming, and immune regulation have continuously entered the clinical trial stage. In the process of new drug clinical trials, adverse event reporting is a core part of the drug safety evaluation system. Researchers need to promptly identify, accurately determine, and standardize the reporting of various adverse events that occur during the trial.

[0003] Existing clinical trial safety monitoring systems mostly use single indicator thresholds to determine risk, which cannot distinguish between the simultaneous, unidirectional, slow deterioration of multiple indicators caused by drug accumulation and the random fluctuations caused by the subject's underlying kidney disease and blood sugar disorders. Furthermore, interference from high-salt diets and the combined use of potassium-sparing drugs can generate a large number of false abnormalities and a massive amount of invalid warnings. Summary of the Invention

[0004] To address the problem of numerous false alarms in existing methods for reporting adverse events in clinical trials of new drugs for endocrine nephropathy, the present invention aims to provide an intelligent reporting method for adverse events in clinical trials of new drugs for endocrine nephropathy. The specific technical solution adopted is as follows: This invention provides an intelligent reporting method for adverse events in clinical trials of new drugs for endocrine nephropathy, the method comprising the following steps: To obtain different indicator data of subjects with endocrine nephropathy before and after the new drug trial; Based on the deviation of each indicator data from the baseline value corresponding to the subject at each collection time, abnormal indicators at each collection time are screened and abnormal time periods are determined; the abnormal indicators at each collection time within the abnormal time period are subjected to intersection accumulation processing to divide each intersection time period; based on the temporal waveform similarity of each indicator data within the intersection time period, the physiological state consistency factor within the intersection time period is obtained, and the core time period within the abnormal time period is determined. Based on the distribution of abnormal indicators within the core time period and the physiological state consistency factor, the drug-related risk level of the abnormal time period is evaluated; by combining the drug-related risk level of all historical abnormal time periods and the duration of the abnormal time periods, a global risk factor is obtained. Based on the global risk factors, execute the corresponding hierarchical security control and adverse event reporting operations.

[0005] Preferably, the step of filtering abnormal indicators at each collection time based on the deviation of each indicator data from the corresponding baseline value of the subject includes: For any data collection time after a new drug trial: The time period of a preset duration that is prior to and adjacent to any of the acquisition times is taken as the reference time period for any of the acquisition times; Based on the difference between each indicator data and the corresponding benchmark value at each collection time within the reference period, the deviation value of each indicator data at each collection time is obtained. A linear fit is performed on the deviation values ​​of each indicator data at all collection times within the reference period. Based on the slope of the fitted line and the maximum value of the deviation values ​​of each indicator data at all collection times within the reference period, the degree of deviation corresponding to each indicator data at any collection time is obtained. Data points with deviations greater than or equal to a preset deviation threshold are recorded as abnormal indicators.

[0006] Preferably, determining the abnormal time period includes: The degree of abnormality at each collection time is obtained by the ratio of the number of abnormal indicators to the total number of all indicator data types, and the minimum deviation of all abnormal indicators. The collection time when the degree of abnormality is greater than or equal to the preset abnormality threshold is judged as an abnormal time; Consecutive adjacent abnormal moments constitute an abnormal time period.

[0007] Preferably, the step of performing intersection and accumulation processing on the abnormal indicators at each collection time within the abnormal period and dividing each intersection period includes: taking the period consisting of consecutive collection times where the abnormal indicators have an intersection within each abnormal period as the intersection period.

[0008] Preferably, obtaining the physiological state consistency factor within the intersection period based on the temporal waveform similarity of each indicator data within the intersection period includes: For any type of indicator data: calculate the Pearson correlation coefficient between the indicator data and the corresponding time index sequence within the candidate intersection period; determine the similarity score of the indicator data within the candidate intersection period based on the Pearson correlation coefficient; The index data with similarity scores greater than or equal to the preset similarity threshold are recorded as the target dimension; The physiological state consistency factor within the candidate intersection period is calculated based on the proportion of the target dimensions and the minimum similarity score of the target dimensions within the candidate intersection period. The candidate intersection time period is any intersection time period.

[0009] Preferably, the determination of the core time period within the abnormal time period includes: For any abnormal period, the intersection period corresponding to the largest physiological state consistency factor within that abnormal period is taken as the core period of that abnormal period.

[0010] Preferably, the step of evaluating the degree of drug-related risk during abnormal periods based on the distribution of abnormal indicators within the core time period and the physiological state consistency factor includes: Based on the physiological state consistency factor of the core period to be analyzed, the proportion of abnormal indicators in the core period to be analyzed, and the average of the degree of abnormality at all collection times in the core period to be analyzed, the degree of drug-related risk of the abnormal period in which the core period to be analyzed is located is obtained. The core time period to be analyzed can be any core time period.

[0011] Preferably, the global risk factor, obtained by comprehensively considering the drug-related risk levels and durations of all historical abnormal periods, includes: The global risk factor is obtained by weighting the drug-related risk levels of all historical abnormal periods by using the duration of each abnormal period as the weight. Among them, the historical abnormal period refers to all abnormal periods before the current moment.

[0012] Preferably, the step of performing corresponding hierarchical security control and adverse event reporting operations based on the global risk factor includes: When the global risk factor is less than or equal to the first risk threshold, the subject is determined to be in a level one risk state, and the frequency of indicator data collection remains unchanged. When the global risk factor is greater than the first risk threshold and less than the second risk threshold, the subject is determined to be in a level 2 risk state, marked as an observation subject, and additional special testing items are added. When the global risk factor is greater than or equal to the second risk threshold, the subject is determined to be in a level three risk state, an expedited re-examination work order is generated and reported, and the frequency of indicator data collection is increased. The first risk threshold is less than the second risk threshold.

[0013] Preferably, the acquisition of the benchmark value includes: taking the median of each indicator data of the subject before the new drug trial as the benchmark value of each indicator data of the subject.

[0014] The present invention has at least the following beneficial effects: This invention acquires multi-indicator data from subjects before and after new drug trials, and filters out abnormal indicators based on the deviations of the data from individualized baseline values ​​at each collection time. This effectively eliminates inherent interference from subjects' underlying diseases and daily physiological fluctuations. Furthermore, by accumulating and dividing the intersection periods of abnormal indicators within abnormal timeframes, it can accurately pinpoint abnormal intervals where multiple indicators co-occur continuously, avoiding misjudgments caused by occasional deviations of a single indicator and reducing the false positive rate of adverse event reporting. By analyzing the temporal waveform similarity of data from each indicator within the intersection period, a physiological state consistency factor is obtained, thereby determining the core timeframe within the abnormal period. This allows for the precise extraction of truly drug-related safety signals from a temporal evolution perspective, ensuring that subsequent risk assessments focus on sub-intervals with clearly defined cumulative drug effects, improving the accuracy and scientific basis of drug-related risk assessment. Combined with the abnormal data within the core timeframe... The distribution of normal indicators and the consistency factor with physiological state are used to evaluate the degree of drug-related risk in each abnormal period. By combining the risk degree and duration of all abnormal periods in history, a global risk factor is generated. This global risk factor reflects both the severity of the current abnormal event and the cumulative effect and evolution trend of historical abnormal events. It can comprehensively and dynamically depict the overall safety risk status of subjects throughout the trial period, avoiding decision-making bias caused by single or short-term abnormal signals. Based on the global risk factor, corresponding stratified safety management and adverse event reporting operations are implemented, improving the standardization of adverse event management and the accuracy of reporting in clinical trials of new drugs for endocrine nephropathy. Attached Figure Description

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart of an intelligent reporting method for adverse events in a clinical trial of a new drug for endocrine nephropathy, provided as an embodiment of the present invention. Detailed Implementation

[0017] The following describes in detail, with reference to the accompanying drawings, the specific scheme of the intelligent reporting method for adverse events in clinical trials of a new drug for endocrine nephropathy provided by the present invention.

[0018] Example of an intelligent reporting method for adverse events in a clinical trial of a new drug for endocrine nephropathy: This embodiment proposes an intelligent reporting method for adverse events in clinical trials of new drugs for endocrine nephropathy, such as... Figure 1As shown in this embodiment, an intelligent reporting method for adverse events in a clinical trial of a new drug for endocrine nephropathy includes the following steps: Step S1: Obtain different indicator data of subjects with endocrine nephropathy before and after the new drug trial.

[0019] Timely identification and reporting of adverse events during clinical trials are crucial for ensuring patient safety and improving drug development. For example, diabetic nephropathy, a type of endocrine kidney disease, has increasingly diverse treatment options. However, during new drug clinical trials, patients may experience various adverse reactions, including but not limited to metabolic disorders, renal failure, and other systemic reactions. Therefore, the proper and effective reporting of adverse events is essential for timely risk assessment and prompt intervention, ensuring the continuation of research and patient safety.

[0020] Subjects with endocrine nephropathy already have diabetes and chronic kidney disease, and their indicators naturally fluctuate. If baseline data is collected after medication is started, physiological changes will be mixed with drug effects, making it impossible to distinguish between spontaneous fluctuations in the underlying disease and adverse drug reactions. Therefore, it is necessary to record the natural physiological state of subjects during drug-free periods to provide a dedicated control benchmark for subsequent assessment of indicator deviations, further reducing the false alarm rate of early warnings.

[0021] Before conducting clinical trials of new drugs for endocrine nephropathy, after the subjects sign informed consent forms, a baseline data collection period is set. The baseline data collection period can be 4 weeks before drug administration. During the baseline data collection period, the subjects do not take any investigational drugs to truly reflect the subjects' natural physiological state without intervention from the investigational drugs. During this period, different types of indicator data are continuously collected from the subjects. The types of indicator data may include weight, blood glucose, blood pressure, blood oxygen, 24-hour total urine output, daily activity level, etc. Blood glucose, blood pressure, blood oxygen, and daily activity level can be collected using wearable monitoring devices, such as smartwatches. In this embodiment, the indicator data collection frequency is once a day.

[0022] After collecting data on each indicator from the subjects before the new drug trial, the median of the collected data for that indicator before the new drug trial was used as the baseline value for that indicator.

[0023] After the subjects begin the new drug trial, different types of indicator data are continuously collected from the subjects, and the collected data are saved.

[0024] Thus, through the above methods, we obtained different indicator data of subjects with endocrine nephropathy at each collection time before and after the new drug trial.

[0025] Step S2: Based on the deviation of each indicator data from the baseline value corresponding to the subject at each collection time, abnormal indicators at each collection time are screened and abnormal time periods are determined; the abnormal indicators at each collection time within the abnormal time period are subjected to intersection accumulation processing to divide each intersection time period; based on the temporal waveform similarity of each indicator data within the intersection time period, the physiological state consistency factor within the intersection time period is obtained, and the core time period within the abnormal time period is determined.

[0026] The physiological state of subjects with endocrine nephropathy is greatly affected by occasional factors such as daily diet, activity, and emotions. Single-day abnormalities are often false signals. Only coordinated abnormalities over several consecutive days can reflect the true trend of drug-related adverse events and reduce the false positive rate of adverse event reporting.

[0027] Subjects with chronic kidney disease and diabetes may experience small, periodic fluctuations in weight, urine output, and blood pressure. These fluctuations are inherent to the underlying disease. Therefore, it is necessary to compare the data from subsequent medication phases with the individual's specific baseline values, rather than using universal reference values ​​for healthy individuals, to avoid directly diagnosing them as abnormal and generating false alarms.

[0028] The following example uses a data collection point after a new drug trial as an example. The method provided in this example can be used for other data collection points.

[0029] Specifically, for any data collection point after a new drug trial: the time period preceding and adjacent to that data collection point, with a preset duration, is used as the reference time period for that data collection point; the preset duration can be one week. It should be noted that if the time period between the data collection point and the first day of drug administration is less than the preset duration, the historical data from the baseline data collection period most recent one week prior to the current time is combined with the post-drug data to form the reference time period.

[0030] For any type of indicator data, the absolute value of the difference between the indicator data and the baseline value of that indicator data at each collection time within the reference period is calculated. This absolute value is taken as the difference between the indicator data and the corresponding baseline value at each collection time within the reference period. This difference is then normalized, and the normalized result is taken as the deviation value of the indicator data at each collection time. When normalizing the difference, a maximum-minimum normalization method can be used, where the maximum and minimum values ​​are the maximum and minimum values ​​of the natural fluctuation difference of the indicator data for the same type of subjects within the baseline period, respectively. Specifically, if the difference is less than the minimum value of the natural fluctuation difference, the normalized result is set to 0; if the difference is greater than the maximum value of the natural fluctuation difference, the normalized result is set to 1. The least squares method is used to fit a straight line to the deviation value of the indicator data at all collection times within the reference period, obtaining a fitted straight line. The horizontal axis of the fitted straight line represents the collection time, and the vertical axis represents the deviation value. The least squares method for straight line fitting is an existing technique and will not be elaborated further here. Obtain the slope of the fitted line. Based on the slope of the fitted line and the maximum deviation of the index data at all collection times within the reference period, obtain the degree of deviation of the index data at that collection time.

[0031] As a specific implementation method, the deviation degree of the index data at the acquisition time can be obtained in the following way: obtain the maximum value between the slope of the fitted straight line and the constant 0, and multiply the maximum value of the deviation value of the index data at all acquisition times within the reference period as the deviation degree of the index data at the acquisition time.

[0032] By using the methods described above, we can obtain the degree of deviation corresponding to each indicator data.

[0033] Among the multidimensional indicators of subjects with endocrine nephropathy, occasional deviations of a single indicator are relatively common. However, deviations of a single indicator do not necessarily indicate an abnormality. In this embodiment, the proportion (breadth) of abnormal indicators is nonlinearly coupled with the degree of deviation (depth) to accurately characterize the comprehensive abnormal state of the subjects and avoid misjudging occasional fluctuations in a single dimension as drug-related abnormal signals.

[0034] Specifically, data points deviating from a preset deviation threshold are recorded as abnormal indicators. The preset deviation threshold is used to filter out occasional minor fluctuations in physiological indicators caused by the subject's underlying disease and daily dietary activities. This value needs to be calibrated in conjunction with the natural fluctuation range of similar endocrine-related kidney diseases within the baseline period to ensure that the selected indicator deviations have clinical pharmacological significance. For example, its value can be 0.1.

[0035] Then, based on the ratio of the number of abnormal indicators at each collection time to the total number of all indicator data types, and the minimum deviation of all abnormal indicators, the degree of abnormality at each collection time is obtained.

[0036] As a specific implementation method, the degree of anomaly can be determined as follows: For any data acquisition time, the ratio of the number of abnormal indicators at that acquisition time to the total number of all indicator data types is taken as the ratio of the number of abnormal indicators to the total number of indicator data types at that acquisition time; the product of this ratio and the minimum deviation value corresponding to all abnormal indicators at that acquisition time is taken as the degree of anomaly at that acquisition time. If there are no abnormal indicators at that acquisition time, the degree of anomaly at that acquisition time is set to 0.

[0037] The higher the proportion of abnormal indicators and the more severe the abnormal state, the more it indicates that multiple indicators are simultaneously and significantly deviating from the baseline value. Furthermore, the wider the affected area and the greater the deviation of individual indicators, the more consistent this is with the characteristics of drug-induced water and sodium retention, and the more attention is required. Therefore, collection times with an abnormality level greater than or equal to a preset abnormality threshold are defined as abnormal times. The preset abnormality threshold is used to determine the overall severity of simultaneous abnormalities in multiple indicators. Since adverse events related to endocrine nephropathy drugs typically manifest as synergistic abnormalities in multiple physiological systems, this threshold setting needs to balance the breadth of the abnormality and the depth of deterioration of individual indicators to exclude one-dimensional false alarms; its value can be 0.15.

[0038] Using the above methods, multiple abnormal moments can be screened from the collection time after the new drug trial, and consecutive adjacent abnormal moments constitute an abnormal time period.

[0039] Considering that the abnormal periods identified by the method can only reflect deviations from the baseline values ​​of indicators over multiple consecutive days, it cannot distinguish between persistent drug toxicity and random disturbances caused by diet or underlying kidney disease. Because adverse events related to endocrine nephropathy drugs have a clear pharmacodynamic evolution pattern, under the cumulative effect of drug efficacy, the daily abnormal high-frequency time-series waveforms (heart rate, blood pressure, etc.) exhibit a synchronous, unidirectional, and slowly deteriorating evolution characteristic over multiple consecutive days, with highly similar waveforms and involving coordinated abnormalities of multiple physiological systems. In contrast, abnormalities caused by the subject's own underlying kidney disease, blood glucose disorders, etc., manifest as random oscillating noise, with indicators deviating from the baseline values ​​over multiple consecutive days, and the daily high-frequency time-series waveform trends lack consistency, resulting in extremely low waveform similarity scores.

[0040] Based on the above characteristics, it is possible to further pinpoint the sub-intervals with the most stable co-occurrence of multiple indicators, the strongest consistency of physiological state, and the most significant cumulative drug effect characteristics from abnormal periods. These sub-intervals can then be used as core periods to ensure that subsequent risk level assessments and adverse event reporting decisions focus on genuine drug-related safety signals, rather than random fluctuations in underlying diseases.

[0041] Specifically, for any abnormal time period, the time period consisting of consecutive collection times where the abnormal indicators within that abnormal time period intersect is taken as the intersection time period. That is, the set of abnormal indicators corresponding to each collection time within the abnormal time period is obtained. Starting from the first collection time of that abnormal time period, the intersection of the abnormal indicator sets of multiple consecutive collection times is taken sequentially until the intersection is empty. The multiple consecutive collection times before the current intersection becomes empty are then divided into an intersection time period. The starting calculation position is shifted forward by one collection time, and the intersection is taken again, until all collection times within the abnormal time period are traversed, resulting in several intersection time periods. For example: [The text abruptly ends here, likely due to an incomplete sentence or missing information.] The abnormal indicators at the first data collection time within a segment are data of the first, second, third, and sixth types; the abnormal indicators at the second data collection time are data of the second, third, and sixth types; the abnormal indicators at the third data collection time are data of the second and third types; and the abnormal indicator at the fourth data collection time is data of the fifth type. Therefore, there is an overlap between the abnormal indicators at the first three data collection times, but no overlap between the abnormal indicators at the fourth data collection time and those at the first three data collection times. Thus, the first three data collection times constitute an overlapping time period. Using this method, each abnormal time period can be divided into multiple overlapping time periods.

[0042] The following embodiment will use any intersection time period as an example for explanation. The method provided in this embodiment can be used to process other intersection time periods.

[0043] Any intersection period is designated as a candidate intersection period. For any type of index data, calculate the Pearson correlation coefficient between that index data and its corresponding time index sequence within the candidate intersection period. The time index sequence for each intersection period starts from 1 and increments by 1 step. For example, if a candidate intersection period contains 3 data collection times, the corresponding time index sequence is (1, 2, 3); if a candidate intersection period contains 4 data collection times, the corresponding time index sequence is (1, 2, 3, 4).

[0044] Furthermore, the maximum value between the Pearson correlation coefficient and 0 between the index data and the corresponding time index sequence within the candidate intersection period is taken as the similarity score of the index data within the candidate intersection period.

[0045] The higher the Pearson correlation coefficient between the index data and the corresponding time index sequence within the candidate intersection period, the higher the similarity of the index data within the candidate intersection period, i.e., the higher the similarity score of the index data within the candidate intersection period.

[0046] Specifically, if the number of consecutive acquisition times contained within the candidate intersection time period is 1 or 2, then its corresponding similarity score is directly recorded as 0. If the number of consecutive acquisition times contained within the candidate intersection time period is greater than 2, then the similarity score is calculated using the method described above.

[0047] After obtaining the similarity score of each indicator data within the candidate intersection period using the above method, the indicator data with a similarity score greater than or equal to the preset similarity threshold are recorded as the target dimension. Based on the proportion of target dimensions and the minimum similarity score of the target dimension within the candidate intersection period, the physiological state consistency factor within the candidate intersection period is calculated. The preset similarity threshold is used to assess the stability of continuous deterioration of the indicator, to distinguish between drug toxicity accumulation (strong linear unidirectional deterioration) and random oscillations caused by underlying diseases; its value can be set to 0.7.

[0048] As a specific implementation method, the physiological state consistency factor can be determined as follows: the ratio between the number of target dimensions and the number of all indicator data types is taken as the proportion of the target dimension; the product of the proportion of the target dimension and the minimum similarity score of all target dimensions within the candidate intersection period is taken as the physiological state consistency factor within the candidate intersection period. Specifically, if the number of target dimensions is 0, then the physiological state consistency factor within the candidate intersection period is set to 0.

[0049] Significant differences exist in the stability and temporal consistency of multi-indicator co-occurrence across different intersection periods within the same abnormal time period. To ensure that subsequent risk level assessments and adverse event reporting decisions focus on the core components most relevant to the drug, it is necessary to select the intersection period with the highest physiological state consistency within each abnormal time period. This allows for the precise identification of the sub-interval within the abnormal time period that exhibits the most stable multi-indicator co-occurrence, the strongest physiological state consistency, and best reflects the cumulative characteristics of drug effects. Therefore, for any abnormal time period, the intersection period corresponding to the largest physiological state consistency factor within that period is designated as the core period for that abnormal time period. Specifically, if multiple intersection periods have the same maximum physiological state consistency factor, the intersection period with the earliest starting time is selected as the core period.

[0050] Using the above methods, the core time period of each abnormal period can be obtained.

[0051] Step S3: Evaluate the degree of drug-related risk during abnormal periods based on the distribution of abnormal indicators within the core time period and the physiological state consistency factor; obtain a global risk factor by combining the degree of drug-related risk and the duration of all historical abnormal periods.

[0052] Risk assessment of adverse events from endocrine nephropathy medications requires comprehensive consideration of multiple dimensions. Furthermore, the risk at the current moment depends not only on the intensity of the most recent adverse event, but also on the cumulative effect and evolution trend of multiple adverse events in the past.

[0053] The following explanation uses any one core time period as an example. The method provided in this embodiment can be used to process other core time periods.

[0054] Specifically, any core time period is designated as the core time period to be analyzed. Based on the physiological state consistency factor of the core time period to be analyzed, the proportion of abnormal indicators within the core time period to be analyzed, and the average of the degree of abnormality at all collection times within the core time period to be analyzed, the drug-related risk level of the abnormal time period in which the core time period to be analyzed is obtained.

[0055] As a specific implementation method, the degree of drug-related risk can be determined in the following way: the ratio of the number of abnormal indicators in the core period to be analyzed to the number of all types of indicator data is taken as the proportion of the number of abnormal indicators in the core period to be analyzed; the product of the physiological state consistency factor of the core period to be analyzed, the proportion of the number of abnormal indicators in the core period to be analyzed, and the mean of the degree of abnormality at all collection times in the core period to be analyzed is taken as the degree of drug-related risk in the abnormal period in which the core period to be analyzed is located.

[0056] Using the above methods, the degree of drug-related risk for each abnormal period can be obtained.

[0057] All abnormal periods up to and including the current time are recorded as historical abnormal periods. A weighted average of the drug-related risk levels across all historical abnormal periods is calculated, using the duration of each historical abnormal period as a weight, to obtain the global risk factor.

[0058] As a specific implementation method, a detailed formula for calculating the global risk factor is provided. The global risk factor can be expressed as: Where F represents the global risk factor, and S represents the number of historical abnormal periods. This indicates the degree of drug-related risk during the s-th historical anomaly period. This represents the duration of the s-th historical anomaly period.

[0059] Adverse events of endocrine nephropathy drugs usually have a cumulative effect over time. The longer the abnormal period lasts, the more stable the risk signal is, the higher the probability that it is induced by the investigational drug, and the greater the threat to the overall safety of the subjects. Therefore, it is given a higher weight in the overall assessment. On the other hand, short abnormalities are more likely to be random fluctuations of the underlying disease, and their weight is naturally reduced.

[0060] Specifically, when the number of historical abnormal periods is 0, the drug-related risk level of the current abnormal period is directly used as the global risk factor.

[0061] Using the above methods, the global risk factors were determined.

[0062] Step S4: Based on the global risk factor, perform the corresponding hierarchical security control and adverse event reporting operations.

[0063] In this embodiment, the global risk factor is determined in step S3. Next, the risk level of the subject at the current moment will be judged by comparing the relationship between the global risk factor and the threshold.

[0064] Specifically, when the global risk factor is less than or equal to the first risk threshold, it indicates that the subject is currently in a low-risk state. That is, in all historical monitoring data before the current moment, the subject has not yet formed a continuous, stable, and repeatable abnormal physiological pattern, and the fluctuations of its various monitoring indicators are within the normal range allowed by the individualized baseline. There are no drug-related safety signals that researchers need to pay attention to. Therefore, the subject is determined to be in a first-level risk state at this time, and the frequency of data collection for the subject's indicators remains unchanged.

[0065] When the global risk factor is greater than the first risk threshold but less than the second risk threshold, it indicates that the subject is currently in a medium-risk state. At this time, the subject is determined to be in a level 2 risk state, and is marked as an observation subject. In addition to the original in-hospital follow-up plan, specific tests for serum potassium and urine microalbumin / creatinine ratio are added. The first risk threshold is less than the second risk threshold. The first risk threshold is used to sensitively capture early suspicious signals, and the first risk threshold can be 0.4. The second risk threshold is used to confirm the exact drug toxicity and enforce an emergency response (such as expedited reporting) to prevent irreversible organ damage to the subject, and the second risk threshold can be 0.7.

[0066] When the global risk factor is greater than or equal to the second risk threshold, it indicates that the subject is currently in a high-risk state. At this time, the subject is judged to be in a level 3 risk state, and an expedited retest work order is immediately generated and reported to the trial investigator. The subject is required to complete a full set of venous blood tests for renal function and electrolytes within 72 hours. At the same time, the frequency of data collection is increased. For example, data can be collected once every 3 hours, that is, the frequency of home multi-dimensional vital signs collection is continuously increased to track the time-series changes of various indicators without interruption.

[0067] Thus, by using the method provided in this embodiment, intelligent reporting of adverse events in clinical trials of new drugs for endocrine nephropathy has been achieved.

[0068] This embodiment acquires multi-indicator data of subjects before and after the new drug trial, and filters abnormal indicators based on the deviation of the indicator data from individualized baseline values ​​at each collection time. This effectively eliminates inherent interference from the subjects' underlying diseases and daily physiological fluctuations. Furthermore, by performing intersection and accumulation processing on abnormal indicators within abnormal periods and dividing the intersection periods, it is possible to accurately pinpoint abnormal intervals where multiple indicators co-occur continuously, avoiding misjudgments caused by occasional deviations of a single indicator and reducing the false positive rate of adverse event reporting. By analyzing the temporal waveform similarity of each indicator data within the intersection periods, a physiological state consistency factor is obtained, thereby determining the core period within the abnormal period. This allows for the precise extraction of truly drug-related safety signals from a temporal evolution perspective, ensuring that subsequent risk assessments focus on sub-intervals with clearly defined cumulative drug effects, improving the accuracy and scientific basis of drug-related risk assessment. Combined with the abnormal indicators within the core period... The distribution of normal indicators and the consistency factor with physiological state are used to evaluate the degree of drug-related risk in each abnormal period. By combining the risk degree and duration of all abnormal periods in history, a global risk factor is generated. This global risk factor reflects both the severity of the current abnormal event and the cumulative effect and evolution trend of historical abnormal events. It can comprehensively and dynamically depict the overall safety risk status of subjects throughout the trial period, avoiding decision-making bias caused by single or short-term abnormal signals. Based on the global risk factor, corresponding stratified safety management and adverse event reporting operations are implemented, improving the standardization of adverse event management and the accuracy of reporting in clinical trials of new drugs for endocrine nephropathy.

Claims

1. A method for intelligent reporting of adverse events in clinical trials of a new drug for endocrine nephropathy, characterized in that, The method includes the following steps: To obtain different indicator data of subjects with endocrine nephropathy before and after the new drug trial; Based on the deviation of each indicator data from the baseline value corresponding to the subject at each collection time, abnormal indicators at each collection time are screened and abnormal time periods are determined; the abnormal indicators at each collection time within the abnormal time period are subjected to intersection accumulation processing to divide each intersection time period; based on the temporal waveform similarity of each indicator data within the intersection time period, the physiological state consistency factor within the intersection time period is obtained, and the core time period within the abnormal time period is determined. Based on the distribution of abnormal indicators within the core time period and the physiological state consistency factor, the drug-related risk level of the abnormal time period is evaluated; by combining the drug-related risk level of all historical abnormal time periods and the duration of the abnormal time periods, a global risk factor is obtained. Based on the global risk factors, execute the corresponding hierarchical security control and adverse event reporting operations.

2. The intelligent reporting method for adverse events in a clinical trial of a new drug for endocrine nephropathy according to claim 1, characterized in that, The method of filtering abnormal indicators at each collection time based on the deviation of each indicator data from the corresponding baseline value of the subject includes: For any data collection time after a new drug trial: The time period of a preset duration that is prior to and adjacent to any of the acquisition times is taken as the reference time period for any of the acquisition times; Based on the difference between each indicator data and the corresponding benchmark value at each collection time within the reference period, the deviation value of each indicator data at each collection time is obtained. A linear fit is performed on the deviation values ​​of each indicator data at all collection times within the reference period. Based on the slope of the fitted line and the maximum value of the deviation values ​​of each indicator data at all collection times within the reference period, the degree of deviation corresponding to each indicator data at any collection time is obtained. Data points with deviations greater than or equal to a preset deviation threshold are recorded as abnormal indicators.

3. The intelligent reporting method for adverse events in a clinical trial of a new drug for endocrine nephropathy according to claim 2, characterized in that, The determination of abnormal time periods includes: The degree of abnormality at each collection time is obtained by the ratio of the number of abnormal indicators to the total number of all indicator data types, and the minimum deviation of all abnormal indicators. The collection time when the degree of abnormality is greater than or equal to the preset abnormality threshold is judged as an abnormal time; Consecutive adjacent abnormal moments constitute an abnormal time period.

4. The intelligent reporting method for adverse events in a clinical trial of a new drug for endocrine nephropathy according to claim 1, characterized in that, The step of performing intersection and accumulation processing on the abnormal indicators at each collection time within the abnormal period and dividing each intersection period includes: taking the period consisting of consecutive collection times where the abnormal indicators have an intersection within each abnormal period as the intersection period.

5. The intelligent reporting method for adverse events in a clinical trial of a new drug for endocrine nephropathy according to claim 1, characterized in that, The step of obtaining the physiological state consistency factor within the intersection period based on the temporal waveform similarity of each indicator data within the intersection period includes: For any type of indicator data: calculate the Pearson correlation coefficient between the indicator data and the corresponding time index sequence within the candidate intersection period; determine the similarity score of the indicator data within the candidate intersection period based on the Pearson correlation coefficient; The index data with similarity scores greater than or equal to the preset similarity threshold are recorded as the target dimension; The physiological state consistency factor within the candidate intersection period is calculated based on the proportion of the target dimensions and the minimum similarity score of the target dimensions within the candidate intersection period. The candidate intersection time period is any intersection time period.

6. The intelligent reporting method for adverse events in a clinical trial of a new drug for endocrine nephropathy according to claim 1, characterized in that, The core time period within the identified abnormal time period includes: For any abnormal period, the intersection period corresponding to the largest physiological state consistency factor within that abnormal period is taken as the core period of that abnormal period.

7. The intelligent reporting method for adverse events in a clinical trial of a new drug for endocrine nephropathy according to claim 2, characterized in that, The evaluation of the drug-related risk level during abnormal periods based on the distribution of abnormal indicators within the core time period and the physiological state consistency factor includes: Based on the physiological state consistency factor of the core period to be analyzed, the proportion of abnormal indicators in the core period to be analyzed, and the average of the degree of abnormality of all collection times in the core period to be analyzed, the degree of drug-related risk of the abnormal period in which the core period to be analyzed is located is obtained. The core time period to be analyzed can be any core time period.

8. The intelligent reporting method for adverse events in a clinical trial of a new drug for endocrine nephropathy according to claim 1, characterized in that, The global risk factor is derived by comprehensively analyzing the drug-related risk levels and durations of all abnormal periods in history, including: The global risk factor is obtained by weighting the drug-related risk levels of all historical abnormal periods by using the duration of each abnormal period as the weight. Among them, the historical abnormal period refers to all abnormal periods before the current moment.

9. The intelligent reporting method for adverse events in a clinical trial of a new drug for endocrine nephropathy according to claim 1, characterized in that, The step of performing corresponding hierarchical security control and adverse event reporting operations based on the global risk factor includes: When the global risk factor is less than or equal to the first risk threshold, the subject is determined to be in a level one risk state, and the frequency of indicator data collection remains unchanged. When the global risk factor is greater than the first risk threshold and less than the second risk threshold, the subject is determined to be in a level 2 risk state, marked as an observation subject, and additional special testing items are added. When the global risk factor is greater than or equal to the second risk threshold, the subject is determined to be in a level three risk state, an expedited re-examination work order is generated and reported, and the frequency of indicator data collection is increased. The first risk threshold is less than the second risk threshold.

10. The intelligent reporting method for adverse events in a clinical trial of a new drug for endocrine nephropathy according to claim 1, characterized in that, The acquisition of the baseline values ​​includes: taking the median of each indicator data of the subject before the new drug trial as the baseline value of each indicator data of the subject.

Images