Clinical test patient screening and dynamic follow-up visit management method based on artificial intelligence

Abstract

The invention relates to the technical field of medical information, in particular to a clinical test patient screening and dynamic follow-up visit management method based on artificial intelligence, and the method comprises the following steps: collecting and standardizing admission, examination, medication and image data, mapping test items to form associated health data, comparing the group entering standard, screening qualified patients, and performing dynamic follow-up visit management. And extracting distribution mark difference crowds of medical units and patients, analyzing a physiological data fluctuation trend, and adjusting a follow-up visit sequence to generate a dynamic management plan. According to the method, by collecting structured health information and mapping test items, accurate screening and synchronous rhythm control are achieved, distribution of patients and medical units is analyzed, differential groups are recognized, resource allocation is optimized, a time sequence is constructed through physiological data, the difference value is calculated, and the health fluctuation trend is captured. The follow-up visit node is combined to adjust time arrangement, the execution efficiency is improved, and the patient management response speed and the data processing integrity are enhanced through multi-link cooperation.

Description

Technical Field

[0001] The present invention relates to the field of medical information technology, and particularly to a method for screening clinical trial patients and dynamically follow-up management based on artificial intelligence. Background Art

[0002] The field of medical information technology includes the collection, storage, processing, analysis, and management of medical data. The core contents involved in this field include the digitization of patient health data, the interconnection of medical systems, the management of electronic medical records, the support system for disease diagnosis and treatment, the data management of clinical trials, and the health monitoring of patients. With the continuous development of information technology, the application of medical information technology in clinical decision-making, disease prediction and diagnosis, the formulation of personalized treatment plans, and health management has become increasingly extensive, promoting the digital and intelligent transformation of the medical industry. The technology in this field is constantly improving to achieve the improvement of the efficiency and quality of medical services.

[0003] Among them, the method for screening clinical trial patients and dynamically follow-up management based on artificial intelligence refers to optimizing the screening and dynamic management of patients in clinical trials through artificial intelligence technology. The technical matters targeted by this patent theme include comprehensively analyzing the health data, historical cases, genetic information, and other relevant data of patients, automatically identifying the patient groups meeting the conditions of clinical trials, and conducting dynamic health management based on the individual's health status and follow-up requirements. Specifically, by constructing a patient screening model and a follow-up management model, and combining artificial intelligence algorithms to monitor patient information in real time, conduct data analysis, risk assessment, and follow-up reminders, to ensure that the health status of patients during clinical trials is managed and tracked in a timely manner.

[0004] Existing medical information technology mainly relies on static data processing rules, lacks flexibility when processing clinical trial data, and is difficult to adapt to rapidly changing trial conditions and patient status, resulting in the inability to identify eligible patients in a timely manner. The dispersion of data sources and the lack of a unified data format also hinder effective data integration, restricting the scope and accuracy of patient screening. The intermittency of health management and the lack of real-time monitoring result in insufficient tracking of patient conditions, the inability to detect and handle health fluctuations in a timely manner, and affect the timeliness of medical intervention. In the formulation and implementation of follow-up plans, overlapping time arrangements and unreasonable resource allocation increase the waste of medical resources, and may also lead to interruptions in data collection, affecting the data continuity and accuracy of clinical trials. For example, due to the lack of an effective data integration mechanism, the integration of cross-institutional medical records is not timely enough, which may lead to the omission of important medical information, thereby affecting the reliability of the results of clinical trials. Summary of the Invention

[0005] The purpose of the present invention is to solve the deficiencies existing in the prior art, and to propose a method for screening clinical trial patients and dynamically follow-up management based on artificial intelligence.

[0006] To achieve the above object, the present invention adopts the following technical solutions: An artificial intelligence-based method for screening clinical trial patients and dynamic follow-up management, comprising the following steps: S1: Deploy a data collection channel in the medical data network, collect admission information, examination records, medication and imaging data, extract key information to generate a data table in a unified format, map the trial items, and obtain trial-related health data; S2: Based on the trial-related health data, compare with the inclusion criteria, screen out the records that match the criteria, calculate the average time interval between adjacent records, adjust the data update rhythm, and obtain a group of patients initially screened as qualified by artificial intelligence; S3: Based on the group of patients initially screened as qualified by artificial intelligence, extract the associated medical units and the number of patients, calculate and classify the record differences, and mark the population with distribution differences to obtain a dynamic follow-up management monitoring list; S4: Based on the dynamic follow-up management monitoring list, extract the physiological data of the patients' continuous records, construct a sequence in chronological order, calculate the adjacent differences, identify the continuously changing segments, and establish analysis tags to obtain the health fluctuation trend of the dynamically monitored population; S5: Based on the health fluctuation trend of the dynamically monitored population, find the corresponding time nodes of the follow-up arrangement, identify time overlaps and misalignments and adjust the order, organize them into update items, and generate an artificial intelligence-driven patient dynamic management plan.

[0007] The associated health data specifically includes admission information, examination records, medication data, and imaging data. The group of patients initially screened as qualified by artificial intelligence includes records that match the screening conditions, time interval data, and data update frequency. The dynamic follow-up management monitoring list includes medical unit identifiers, the number of patients, and distribution difference marks. The health fluctuation trend of the dynamically monitored population specifically refers to time series, health differences, and fluctuation tags. The artificial intelligence-driven patient dynamic management plan includes follow-up time nodes, order adjustment details, and update items.

[0008] As a further solution of the present invention, the specific steps of S1 are: S101: Deploy a data collection channel in the medical data network, collect medical records, tests, imaging and admission records, examination items, imaging descriptions, and drug information, unify the field format, perform field mapping and standardize the data type to obtain a unified field table; S102: Based on the field items in the unified field table, perform structure analysis and content recognition, screen the field records, calculate the field integrity, and eliminate the records that do not meet the requirements through a set threshold to obtain the standard records of the structural fields; S103: Based on the above-mentioned structural field standard record, match the record fields with the requirements of the experimental design, compare the fields that meet the requirements, establish a mapping relationship, and combine with the evaluation of the dynamic follow-up management node to screen the records that meet the experimental requirements, so as to obtain the health data related to the experiment.

[0009] As a further solution of the present invention, the specific steps of S2 are as follows: S201: Based on the above-mentioned health data related to the experiment, compare each field item by item according to the preset inclusion criteria, record the field characteristics that are consistent with the standard content, and generate a field matching statistical result; S202: Based on the above-mentioned field matching statistical result, sort the records that meet the inclusion criteria according to the receiving time, calculate the receiving time interval between adjacent records, obtain the time difference between every two records, and generate an average record time interval; S203: Based on the above-mentioned average record time interval, adjust the synchronization rhythm of data batch update, calculate the data update frequency, and perform batch allocation according to the update frequency. Incorporate the patient records that meet the inclusion criteria into the update queue to generate a group of patients initially screened as qualified by artificial intelligence.

[0010] As a further solution of the present invention, the specific formula for calculating the data update frequency is as follows: ; Wherein, represents the data update frequency, represents the receiving time of the i-th record, represents the difference between the i-th record and the previous record, represents the total number of records, represents the time difference between adjacent records, represents the absolute value of the difference from the previous record.

[0011] As a further solution of the present invention, the specific steps of S3 are as follows: S301: Based on the above-mentioned group of patients initially screened as qualified by artificial intelligence, extract the name of the medical unit associated with each record and the corresponding number of patients, group the records by medical unit, count the number of patients in each unit, and calculate the difference in the number of patients between units to generate a statistical table of the difference in the number of patients between units; S302: Based on the above-mentioned statistical table of the difference in the number of patients between units, partition the range of the difference in the number of patients between units, classify the units with deviated differences into high-priority monitoring areas, mark the units as areas with intensive data update frequency, and extract the corresponding patient records to generate a classification and marking table of unit differences; S303: Based on the unit difference classification and marking table, incorporate the patient records in the high-priority area into the scope of attention for dynamic follow-up management, generate a schedule for dynamic follow-up management tasks, and synchronously update the monitoring list to obtain the dynamic follow-up management monitoring list.

[0012] As a further solution of the present invention, the specific calculation formula for the quantity difference between units is: ; where represents the quantity difference between units, represents the number of patients in the th medical unit, represents the number of patients in the th medical unit, represents the adjustment value of unit , represents the adjustment value of unit .

[0013] As a further solution of the present invention, the specific steps of S4 are: S401: Based on the dynamic follow-up management monitoring list, extract the body temperature, blood sugar, and blood pressure data of each patient, sort the data in chronological order, and generate a sorted data sequence; S402: Based on the sorted data sequence, calculate the numerical difference between adjacent records, obtain the daily change amount, and mark the change direction of the daily data according to the change value, identify the records with the same trend for multiple days, and generate the change trend direction data; S403: Based on the change trend direction data, screen the records with the same change trend for multiple days, mark the records with obvious trend characteristics as abnormal fluctuations and summarize them to obtain the health fluctuation trend of the dynamically monitored population.

[0014] As a further solution of the present invention, the specific calculation formula for the numerical difference between adjacent records is: ; where represents the numerical difference between adjacent records, represents the data record value on the th day, represents the data record value of the previous day, represents the weight value on the th day, represents the data weight value on the th day.

[0015] As a further solution of the present invention, the specific steps of S5 are: S501: Based on the health fluctuation trend of the dynamically monitored population, extract the patient numbers whose follow-up plans need to be adjusted, call the scheduling records of individual patients, search for and match the associated follow-up nodes, obtain the time intervals of each node, and generate the time intervals of patient follow-up nodes. S502: Based on the time intervals of patient follow-up nodes, compare the time arrangements of the nodes with the preset time for consistency, identify the node contents with time overlaps and inconsistencies, record and mark the nodes that need to be adjusted, and generate a time node adjustment mark table. S503: Based on the time node adjustment mark table, adjust the trigger order of the nodes with overlaps and inconsistencies, set the identification of newly added task nodes, and synchronize all updated information to the individual patient schedule to generate an artificial intelligence-driven dynamic patient management plan.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In the present invention, by collecting structured health information and mapping test items, precise screening and synchronous rhythm control are achieved, the distribution of patients and medical units is analyzed, different groups are identified, resource allocation is optimized, physiological data is used to construct a time series and calculate the difference to capture the health fluctuation trend. Combining with the adjustment of follow-up nodes for time arrangements, the execution efficiency is improved, and multi-link collaboration enhances the response speed of patient management and the integrity of data processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a schematic flow chart of the steps of the present invention; DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not used to limit the present invention.

[0019] In the description of the present invention, it should be understood that the orientation or positional relationships indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. are based on the orientation or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention. In addition, in the description of the present invention, "a plurality of" means two or more unless otherwise specifically defined.

[0020] Please refer to Figure 1 , the method for screening and dynamically following up clinical trial patients based on artificial intelligence includes the following steps: S1: Establish a data acquisition channel in the medical data network to connect medical records, examinations, images, and medications. Collect the patient's admission information, examination records, medication lists, and image descriptions. Extract the patient identification, recording time, test values, drug names, and image keywords. Normalize and write the data into the task registration form according to a unified format. Establish the corresponding relationship between clinical trials and data fields. Identify screening conditions and evaluate dynamic follow-up management nodes to obtain trial-related health data; S2: Based on the trial-related health data, match and mark each field with the inclusion criteria item by item. Include the records with the number of matching fields reaching a fixed threshold in the passing list. Arrange the data reception times of the passing records in order, calculate the average time interval between adjacent records, and adjust the data batch update rhythm to obtain a group of patients initially screened as qualified by artificial intelligence; S3: Based on the group of patients initially screened as qualified by artificial intelligence, extract the names of the medical units associated with each record and the corresponding number of patients. Calculate the quantity difference between the units. Classify and mark according to the difference range. Include the patients corresponding to the units with intensive record update frequencies after classification in the scheduling attention range of dynamic follow-up management to generate a dynamic follow-up management monitoring list; S4: Based on the dynamic follow-up management monitoring list, extract the recent body temperature, blood sugar, and blood pressure records of each patient on the list. Construct a continuous data sequence in chronological order, calculate the difference between adjacent values, and mark the daily change direction. Identify the data segments showing a consistent change trend in multiple consecutive days. Include the records with trend characteristics in the recognition results to obtain the health fluctuation trend of the dynamically monitored population; S5: Based on the health fluctuation trend of the dynamically monitored population, extract the patient numbers that need to adjust the follow-up plan. Call the scheduling record to find the associated follow-up nodes. Compare the time interval between the trend data and the current node time. Identify the node content with time overlap or incorrect time sequence. Adjust the trigger order and set a new node identifier. Synchronize all updated information to the patient individual schedule to generate an artificial intelligence-driven patient dynamic management plan.

[0021] The trial-related health data specifically refers to admission information, examination records, medication data, and image data. The group of patients initially screened as qualified by artificial intelligence includes screening condition matching records, time interval data, and data update frequency. The dynamic follow-up management monitoring list includes medical unit identifiers, patient numbers, and distribution difference marks. The health fluctuation trend of the dynamically monitored population specifically refers to time series, health differences, and fluctuation labels. The artificial intelligence-driven patient dynamic management plan includes follow-up time nodes, order adjustment details, and update items.

[0022] The specific steps of S1 are as follows: S101: Deploy a data collection channel in the medical data network to collect medical records, test, imaging and admission records, examination items, imaging descriptions and drug information, unify the field format, perform field mapping and normalize data types to obtain a unified field table; First, deploy a data collection channel in the medical data network, use the interface to connect with the medical information system, and obtain field data such as patient admission records, test data, imaging examination results, and drug information. After data collection, format it by establishing a standardized data field structure so that the information provided by different data sources can meet the unified format standard. By comparing the data structures in different medical systems, a data table in a unified format is generated to ensure the standardized management and effective integration of subsequent data. For example, the patient's admission information includes the patient's name, admission date, ward number, etc., and the test items include fields such as blood routine and urine routine. Through data docking and mapping, these fields are unified into a format that meets the requirements to ensure that the data can be processed and stored by the system. In addition, the image description provided by the imaging system may contain complex text information and needs to be standardized through preset parsing rules. For example, "chest X-ray shows lung shadows" can be unified as "chest image shadows", and drug information such as drug name, dosage, and usage methods can be compatible through unified field mapping. Next, the data type of the collected fields is standardized, such as the date field is unified into the YYYY-MM-DD format, and the numeric field is unified into the floating number format. Finally, all the normalized data are organized into a unified field table and used as the basis for subsequent screening and analysis. Through this process, a unified field table is obtained.

[0023] S102: Performing structural analysis and content recognition based on field items in the unified field table, screening field records, calculating field completeness, and eliminating records that do not meet the requirements by setting a threshold to obtain standard records of structural fields; First, based on the unified field table, extract the identity identifier, test data, recording time, drug information, and imaging description of each patient. For example, the identity identifier of patient A is "12345", the test data includes blood glucose of "5.2 mmol / L", the drug information is "amoxicillin", the imaging description is "right lung shadow", and the recording time is "2025-04-08 09:00:00". Through these fields, the basic information of the patient and various test and medical data can be directly obtained. Next, screen these fields, and the screening conditions are the integrity and validity of the inspection data. For example, if the test data is empty or the date field format is non-compliant, it is regarded as an invalid record. Then, calculate the field integrity of each record to check whether there is missing or incorrect data. The field integrity is calculated according to a predetermined rule (such as "none of the 5 fields are empty"). Taking patient A as an example for the calculation of field integrity: if the "test data" and "recording time" of this record meet the requirements, the integrity of this record is 100%. If a certain field is missing (such as the test data is empty), the integrity of this record will be reduced and it will ultimately fail to pass the screening. On this basis, set the field integrity threshold. For example, it is stipulated that records with a field integrity lower than 80% need to be excluded, and such records will be marked and removed. Finally, all the records that pass the screening will be sorted out and recorded in a new table to form a standard record of structural fields.

[0024] S103: Based on the standard record of structural fields, match the record fields with the requirements of the trial design, compare the compliant fields and establish a mapping relationship, and combine the evaluation of dynamic follow-up management nodes to screen the records that meet the trial requirements to obtain the trial-related health data; First, extract the field values in each record. For example, the record time field is extracted as "April 8, 2025 10:30:00", the test field value is extracted as "Blood glucose = 6.1 mmol / L", the imaging field value is "Lung shadow", and the drug field is "Amoxicillin". Subsequently, compare each field value with the corresponding indicators in the test design requirements one by one. It is required that the blood glucose is between 4.4 - 6.4 mmol / L and not empty, the record time should be after the test start date "April 1, 2025", the drug field should not contain prohibited drug keywords such as "Aspirin", and the imaging field should identify and contain preset keywords such as "shadow" or "infiltration". The field comparison operation is completed using conditional judgment. That is, set the upper and lower limits of the test field range as [4.4, 6.4]. If the field value falls within this range, it is judged to be satisfied. For the record time field, use the timestamp comparison method to perform a difference operation with the start date converted to a timestamp to determine whether it is positive. For the imaging keywords, use the set of matching keywords to determine the string inclusion relationship with the field description. If there is a keyword match, it is considered satisfied. For the drug field, call the list of excluded key drugs to screen the intersection of the set of matching drug names and the field value to determine whether there is a prohibited drug item. If there is, it is judged not to be satisfied. After completing the comparison of each field, accumulate the conforming field items by scoring. Set the field matching threshold to 80% of the total number of fields. For example, when the total number of fields is 10, at least 8 field matching requirements need to be met. If a record meets more than 8 field items, it enters the next step of processing, establishing the mapping relationship between the field values and the test standards. Through the field label name conversion table, the local field names are uniformly mapped to the field names used in the test standards. For example, "Blood glucose value" is uniformly mapped to "GLU", and then fill in the mapped data structure in the field order. After establishing the mapping, perform node evaluation processing. Extract the record time field in each record and determine which follow-up node evaluation interval it falls into. The intervals are set as 7-day, 14-day, and 30-day nodes according to the test plan. Set the node evaluation rule as that if the time difference between the record time and the nearest follow-up node is less than 3 days, the record count of this node is included. If the difference is negative or greater than 3 days, it is counted as a deviated record. Set the threshold for the proportion of deviated nodes higher than 30% as the record items that require node evaluation attention. For example, if 2 out of 5 patient records deviate from the node, that is, the deviation rate reaches 40%, which is higher than the set threshold. Finally, screen out the records that meet the field comparison requirements and pass the node evaluation range as the records that meet the test requirements to obtain the test-related health data.

[0025] The specific steps of S2 are as follows: S201: Based on the test-related health data, compare each field item by item according to the preset inclusion criteria, and record the field characteristics that are consistent with the standard content to generate the field matching statistical results; First, based on the trial-related health data, extract the key information from each record, including the patient's age, gender, diagnosis name, and prohibited drug information. For example, the record of patient A contains "Age: 60", "Gender: Male", "Diagnosis name: Hypertension", and "Prohibited drug: Aspirin". These fields will be compared item by item according to the preset inclusion criteria. Suppose the inclusion criteria require the age to be between 40 and 70 years old, the gender is not restricted, the diagnosis name is hypertension or diabetes, and the prohibited drugs are aspirin or warfarin. Then, the record of patient A will be considered to meet the criteria during the comparison because its age, gender, and diagnosis name all meet the criteria, and the prohibited drug field is also valid. Next, record the number of fields that meet the requirements for each record. For example, the record of patient A meets the requirements of 3 fields (age, gender, diagnosis name). After the field comparison is completed, the system will record the number of fields that meet the criteria as the basis for further screening. This process ensures that each record is strictly compared according to the inclusion criteria and provides data support for subsequent screening and processing.

[0026] Then, the system will screen out all records that meet the criteria. All records with complete and compliant fields will be included in the subsequent analysis, and those records with any non-compliant field will be excluded. For example, if the age in the record of patient B is 80 years old, which exceeds the age range of the inclusion criteria, this record will be excluded. Finally, all records that meet the criteria will be organized into a set of data to generate the field matching statistical results.

[0027] S202: Based on the field matching statistical results, sort the records that meet the inclusion criteria by the reception time, calculate the reception time interval between adjacent records, obtain the time difference between every two records, and generate the average record time interval; First, based on the field matching statistics results, we sort all the records that meet the inclusion criteria by the receiving time. For example, if there are three records recording the receiving times of patients A, B, and C as "2025-04-01 10:00:00", "2025-04-01 14:30:00", and "2025-04-02 09:00:00" respectively, the system will sort them from the earliest to the latest according to these receiving times. The execution of this step ensures that all data can be reasonably arranged in time for subsequent operations. Then, the system calculates the time interval between adjacent records. Based on the above example, the time difference between record A and record B is 4 hours and 30 minutes, and the time difference between record B and record C is 18 hours and 30 minutes. In this step, the time difference between every two records is calculated. For each adjacent record pair, the system subtracts the two receiving times to obtain the corresponding time difference. Next, the system calculates the mean value of the time intervals based on these time differences. For example, if the time difference between record A and B is 4 hours and 30 minutes, and the time difference between record B and C is 18 hours and 30 minutes, then the average value of the two time differences is (4.5 + 18.5) / 2 = 11 hours. This mean value reflects the average receiving time interval between all records and is used for subsequent optimization of the time arrangement. Through these steps, the mean value of the record time intervals is finally generated.

[0028] S203: Based on the mean value of the record time intervals, adjust the synchronization rhythm of data batch updates, calculate the data update frequency, and perform batch allocation according to the update frequency. Incorporate the patient records that meet the inclusion criteria into the update queue to generate a group of patients initially screened as qualified by artificial intelligence. The specific calculation formula for the data update frequency is: ; Where, represents the data update frequency, represents the receiving time of the i-th record, represents the difference between the i-th record and the previous record, represents the total number of records, represents the time difference between adjacent records, represents the absolute value of the difference from the previous record; Calculation process: Suppose there are 5 records with time intervals of 15 minutes, 20 minutes, 10 minutes, and 15 minutes respectively, and differences of 2mg, 4mg, 1mg, and 3mg respectively. Calculate according to the formula :

[0029] Calculate the time intervals: minutes, minutes, minute minutes Calculate the difference: mg mg mg mg Calculate the adjustment: , , , ; Calculate the update frequency: ; ; ; This result represents the data update frequency is 14.1, which means that during the data synchronization process of the entire record set, the update frequency per unit time is 14.1. This value helps to optimize the data processing strategy. Specifically, this update frequency value can be used to set the update cycle of data batches or adjust the time points of data synchronization to ensure that the system can balance the update load and effectively synchronize between different records. As the patient data changes, the value of this update frequency will also change accordingly, thereby adjusting the synchronization strategy to keep the data processing efficient and flexible.

[0030] The specific steps of S3 are as follows: S301: Based on the group of initially screened eligible patients by artificial intelligence, extract the names of medical units associated with each record and the corresponding number of patients, group the records by medical unit, count the number of patients in each unit, and calculate the difference in the number between units to generate a statistical table of the difference in the number of patients per unit; The specific calculation formula for the difference in the number between units is: ; Among them, represents the difference in the number between units, represents the number of patients in the th medical unit, represents the number of patients in the th medical unit, represents the adjustment value of unit , represents the adjustment value of unit ; Calculation steps: Assume that the number of patients in unit and unit are respectively: ; ; Meanwhile, assume the unit adjustment values are respectively: ; ; Calculate the difference in the number of patients: ; ; ; ; ; Calculate unit and unit The standardized difference between: ; This result shows that Indicates that the unit and unit After standardization, the difference in the number of patients between is 0.189. The relative difference in the number of patients between the two units is small. A value close to 0 indicates a small difference in the number of patients between the two units, while a value close to 1 indicates a significant difference. Introducing the adjustment values and Helps to handle the scale difference in the number of patients of different units, making the result more reasonable and comparable.

[0031] S302: Based on the statistical table of the difference in the number of patients per unit, partition the range of the difference in the number of patients between units, classify the units with deviated differences into high-priority monitoring areas, mark the units as areas with a high frequency of data update, and extract the corresponding patient records to generate a classification mark table of unit differences; First, according to the statistical table of the difference in the number of patients per unit, we need to distinguish the differences in the number of patients in each unit. Taking the actual data as an example, assume there are three medical units A, B, and C, with the number of patients being 100, 150, and 200 respectively. The differences in the number of patients between adjacent units A and B, and B and C are 50 and 50 respectively, and the difference in the number of patients between unit A and C is 100. The system partitions according to these difference ranges. For example, the difference intervals are set as 0 - 50, 51 - 100, 101 - 150, etc., and different units are classified into different intervals according to their differences in the number of patients. Next, the system will further classify the differences in the number of units through the set difference threshold. In the above example, if the difference is greater than or equal to 50, the unit is classified as a "high - priority monitoring area". The differences between unit A and B, and B and C are both less than 50, so they are classified as regular monitoring areas, while the difference between unit A and C exceeds 100, so it is classified as a high - priority monitoring area. In this process, the specific setting of the difference threshold is closely related to the difference in the number of patients in the unit and the actual situation, so the threshold can be adjusted according to the distribution of the actual data. Then, based on the partitioning results of these difference ranges, the high - priority monitoring areas of the units are marked as areas with a high frequency of data update, which means that these units require higher - frequency monitoring and updating. For example, the difference in the number of patients between unit A and C is large, so it will be marked as a unit that needs to update data frequently. This step is of great significance for the subsequent real - time update and adjustment tasks of the data, and can ensure the timeliness and accuracy of data update. Finally, the system extracts the corresponding patient records according to the above steps, further processes them and generates a classification mark table for unit differences.

[0032] S303: Based on the classification mark table for unit differences, incorporate the patient records in the high - priority area into the scope of attention for dynamic follow - up management, generate a schedule for dynamic follow - up management tasks, and synchronously update the monitoring list to obtain a dynamic follow - up management monitoring list; First, extract the patient records classified as high-priority monitoring areas from the unit difference classification marker table. The records of these patients are from those units with significant differences (such as a large difference in the number of patients between Unit A and Unit B). Due to their higher monitoring requirements, the system will mark and automatically include these records in the scope of dynamic follow-up management attention. For example, if the difference between Unit A and Unit B exceeds the threshold, the system will automatically select the patient data related to these units and incorporate it into the next follow-up management task arrangement. Next, based on the selected patient records in the high-priority area, the system generates a dynamic follow-up management task schedule according to conditions such as record time and patient diagnosis. The system will schedule the follow-up tasks for each patient according to preset rules, such as the severity of the patient's condition and the follow-up time limit. For example, Patient A will be scheduled for multiple follow-ups in the short term due to his high-risk assessment result, while the task of Patient B will be scheduled at longer time intervals. This process ensures that patients with different priorities can be appropriately followed up. Subsequently, the system combines the generated task schedule information with the existing patient data and synchronously updates the monitoring list. This operation will update the schedule and follow-up tasks for all patient follow-ups and ensure that the tasks of high-priority patients are followed up in a timely manner. For example, if the follow-up time of Patient A changes, the system will automatically update the patient's monitoring list and notify the relevant medical staff through the system. Finally, all patient data and follow-up tasks will be integrated in the system to generate a dynamic follow-up management monitoring list.

[0033] The specific steps of S4 are as follows: S401: Based on the dynamic follow-up management monitoring list, extract the body temperature, blood sugar, and blood pressure data of each patient, sort the data in chronological order, and generate a sorted data sequence; First, obtain the basic health data of each patient from the dynamic follow-up management monitoring list, including body temperature, blood sugar, and blood pressure. These data are used to continuously track the patient's health status. Specifically, the system extracts the latest health index data such as body temperature, blood sugar, and blood pressure according to each patient's records. For example, assuming that the records of patient A show a body temperature of 37.2°C, a blood sugar level of 6.1 mmol / L, and a blood pressure of 120 / 80 mmHg, these data will be extracted as basic health information. Next, the system sorts these data in chronological order. The purpose of this sorting is to ensure that the time series of the data is continuous, so as to accurately reflect the trend of the patient's health changes during subsequent analysis. For patient A, for example, the system sorts the data according to the time stamps of the data (such as "08:00 on April 8, 2025"), ensuring that each data record corresponds to its corresponding date and time one by one, and avoiding incorrect analysis caused by disordered data time. After sorting, all the data will form a sorted data sequence. This sequence contains the changes in body temperature, blood sugar, and blood pressure of each patient at different time points. Assume the records of patient A at different time points are as follows: 08:00 on April 8, 2025, body temperature 37.2°C, blood sugar 6.1 mmol / L, blood pressure 120 / 80 mmHg; 12:00 on April 8, 2025, body temperature 37.4°C, blood sugar 6.4 mmol / L, blood pressure 122 / 82 mmHg. The system will arrange these data in chronological order and assign appropriate serial numbers to the data at each time node for subsequent trend analysis or early warning judgment. Through this process, a continuous health data sequence is established for each patient.

[0034] S402: Based on the sorted data sequence, calculate the numerical difference between adjacent records, obtain the daily change amount, and mark the change direction of the daily data according to the change value. Identify the records with the same trend over multiple days and generate the change trend direction data; The specific calculation formula for the numerical difference between adjacent records is: ; Where, represents the numerical difference between adjacent records, represents the data record value on the th day, represents the data record value of the previous day, represents the weight value on the th day, represents the data weight value on the th day.

[0035] Assume today's body temperature is (unit: °F), yesterday's body temperature is , the data is automatically collected by the body temperature detection device at hourly granularity and then averaged daily, with weights set , ; Calculate the absolute value of the change amplitude: ; Calculate the square root of the product of adjacent values: ; Calculate the weighted mean part: ; Substitute the above three parts into the formula to get: ; This result indicates that there is a standardized change amplitude of 0.9 between the current record and the previous record. This value represents the intensity of the impact of temperature changes on the stability of an individual's physical signs in dynamic trend recognition and can be used subsequently to judge continuous upward trends and reach the warning set threshold.

[0036] S403: Based on the change trend direction data, screen the records with consistent change trends in multiple days, mark the records with obvious trend characteristics as abnormal fluctuations and summarize them to obtain the health fluctuation trend of the dynamically monitored population; First, the system will analyze the change directions of indicators such as body temperature, blood sugar, and blood pressure in the multi-day data according to the change trend direction data and screen out the records with consistent change trends in multiple days. Specifically, the system will check the fluctuation trends of data such as body temperature, blood sugar, and blood pressure of each patient in consecutive days. For example, if a patient's body temperature shows an upward trend in 3 days, or the blood sugar continues to rise in three consecutive days, it is determined that the record has a consistent fluctuation trend. The key to this process is to identify the directionality of data changes and compare the data points in chronological order to ensure that only those data that continuously change and have the same change direction in multiple days are selected. For example, the body temperature data of patient A is: 37.1°C on April 8, 2025, 37.3°C on April 9, 2025, and 37.5°C on April 10, 2025. This continuous record shows an upward trend in body temperature, so the data for these three days will be identified as records with a consistent change trend. Next, the system marks the data that meets the consistent change trend as abnormal fluctuations to attract the attention of doctors. For example, if patient A's blood sugar rises continuously for three days to near the hyperglycemic range (e.g., from 5.8 mmol / L to 7.2 mmol / L), then this data point will be marked as abnormal fluctuation data. The system will judge whether it is an abnormal trend by comparing historical data or setting thresholds when marking the data. Finally, the system will summarize all the records with obvious trend characteristics to generate the health fluctuation trend of the dynamically monitored population.

[0037] The specific steps of S5 are as follows: S501: Based on the health fluctuation trend of the dynamically monitored population, extract the patient numbers whose follow-up plans need to be adjusted, call the scheduling records of individual patients, search for and match the associated follow-up nodes, obtain the time intervals of each node, and generate the time intervals of patient follow-up nodes; First, continuously monitor the dynamic health data. Divide the monitoring time range into multiple 24-hour cycles. Record the maximum value, minimum value, average value, and standard deviation of the fluctuation curves of patient physical signs parameters such as heart rate, blood pressure, blood oxygen, and body temperature within each cycle. By setting interval judgment criteria, for example, if the daily fluctuation range of the heart rate exceeds 30 beats per minute, the difference between the daily maximum and minimum values of blood pressure is greater than 40 mmHg, the daily average value of blood oxygen is lower than 92%, or the body temperature exceeds 37.5°C for three consecutive days, it is marked as abnormal fluctuation. According to the above judgment criteria, extract the individual numbers of all individuals with abnormal values continuously appearing in any index within the monitoring cycle as the key objects whose follow-up rhythms need to be adjusted. Output the list of numbers as the basic data for subsequent processing. For each individual number in the number list, call its corresponding follow-up scheduling record file, extract the follow-up nodes corresponding to each time point. The extraction process includes index positioning of the data associated with each number, reading the node fields, screening the node items with the follow-up mark field in the "activated" state, and excluding the node items marked in the "invalid" or "cancelled" state. By reading the planned time field of each valid node and combining the node historical execution log, calculate the time interval of the node. If the actual completion time of the node is more than 3 days earlier than the planned time, the time interval is counted as "ahead of schedule". If the actual completion time is no more than 7 days later than the planned time, it is counted as "normal". Otherwise, it is marked as "delayed". In this way, standardize the node execution record status. During the matching process, match all the follow-up node records of each number and construct a time interval classification label within the record. Finally, generate the time interval information of the follow-up nodes of each patient and output the result table in a structured manner. The fields include patient number, node number, planned time, actual time, time interval classification, etc.

[0038] S502: Based on the time intervals of patient follow-up nodes, compare the time arrangements of the nodes with the preset time for consistency, identify the node contents with time overlap and inconsistency, record and mark the nodes that need to be adjusted, and generate a time node adjustment mark table; First, obtain the time information of each follow-up node of the patient and record the start and end times of each node. For example, assume that the follow-up node time of patient A is from April 1, 2025 to April 5, 2025, and the time stamps of the body temperature data collected by the system are from April 2, 2025 to April 4, 2025. The system will compare this part of the data with the time interval of the follow-up node. First, through the comparison operation, the system will find the overlapping part of the time stamps of the trend data and the time interval of the follow-up node to ensure the time alignment between the data and the node and avoid omission or conflict. Next, the system will judge whether there is a situation of time coincidence or time mismatch based on these data. For example, if the time stamp of a certain data occurs outside the range of the follow-up node, such as the body temperature data is on April 6, 2025, then the system will mark this part of the data as inconsistent data that needs to be further adjusted or the follow-up node needs to be rearranged. Suppose there are multiple data with time misalignment, the system will collect and mark these inconsistent data and record the node information that needs to be adjusted. For the nodes that need to be adjusted, the system will mark them and update the corresponding patient follow-up plan to ensure that it conforms to the new time interval. Suppose for patient B, the system finds that there are time mismatches in the data of multiple follow-up nodes. The system will rearrange and generate a new follow-up plan based on the existing time data. These adjusted nodes will be added to an adjustment mark table for further follow-up and analysis. Finally, a time node adjustment mark table is generated.

[0039] S503: Based on the time node adjustment mark table, adjust the trigger order of the overlapping and inconsistent nodes, set the identification of the newly added task nodes, synchronize all the updated information to the patient individual schedule, and generate an artificial intelligence-driven patient dynamic management plan; First, the system adjusts the marker table according to time nodes, obtains all overlapping or inconsistent follow-up nodes, and compares them with the nodes in the original plan. For example, assume there are two overlapping time nodes in the follow-up plan of patient A. One originally scheduled follow-up node is from May 10, 2025 to May 12, 2025, and the other is from May 10, 2025 to May 13, 2025. The system will identify the time overlap of these two nodes and determine them as inconsistent. The system will automatically adjust the trigger order of the nodes. By rearranging the task order, it avoids time conflicts. During this process, the system compares the time intervals of each node to find out if there are inconsistent time nodes, and determines whether to adjust the trigger order by judging whether the time intervals overlap. For example, if a certain node occurs in a misaligned time period, the system will reorder and mark the follow-up nodes based on the latest health data and the patient's medical history. If there are still problems with the trigger order of the nodes after reordering, the system will continue to adjust until all time nodes comply with the rules, ensuring that all follow-up tasks can be executed smoothly. Next, the system sets new task node identifiers to clearly identify the adjusted nodes. For example, if the follow-up node of patient B needs to be postponed due to non-adjustability, the system will generate a new task identifier for this node and mark it as a new node. The system will ensure that the newly added task nodes do not conflict with other nodes and keep the patient's health data consistent with the follow-up plan. Finally, the system synchronizes all adjusted task information to the patient's individual schedule, ensuring that the follow-up tasks of each patient are updated in a timely manner and reflected in the personal health management plan. All update and adjustment information will be archived in the form of an update record generated by the system for future reference. Ultimately, the system will generate an AI-driven dynamic patient management plan based on the adjusted task arrangement.

[0040] The above are only the preferred embodiments of the present invention and do not limit the present invention in other forms. Any person skilled in the art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. An artificial intelligence-based method for screening clinical trial patients and dynamically following up and managing them, characterized in that, The following steps are involved: S1: Deploy data collection channels in the medical data network to collect admission information, examination records, medication and imaging data, extract key information to generate a unified format data table, map the test items, and obtain the test-related health data; S2: Based on the test-related health data, the data is compared with the inclusion criteria, the records matching the criteria are screened out, the mean time interval between adjacent records is calculated, the data update rhythm is adjusted, and the AI ​​initial screening qualified patient group is obtained; S3: Based on the AI ​​preliminary screening of qualified patient groups, extract the number of related medical institutions and patients, calculate and classify the record differences, mark the population with distribution differences, and obtain a dynamic follow-up management monitoring list; S4: Based on the dynamic follow-up management monitoring list, extract the continuously recorded physiological data of the patient, construct a sequence by time order, calculate adjacent differences, identify continuous change segments, establish analysis labels, and obtain the health fluctuation trend of the dynamically monitored population; S5: Based on the health fluctuation trends of the dynamically monitored population, find the corresponding time nodes for follow-up arrangements, identify time overlaps and misalignments and adjust the order, organize them into update items, and generate an AI-driven dynamic patient management plan.

2. The method for screening and dynamically following up clinical trial patients based on artificial intelligence according to claim 1, wherein The associated health data specifically include admission information, examination records, medication data, and imaging data. The group of patients qualified by the artificial intelligence initial screening includes screening condition matching records, time interval data, and data update frequency. The dynamic follow-up management monitoring list includes medical unit identification, number of patients, and distribution difference marks. The health fluctuation trend of the dynamically monitored population specifically refers to time series, health difference, and fluctuation labels. The artificial intelligence-driven dynamic patient management plan includes follow-up time nodes, sequence adjustment details, and update items.

3. The method for screening and dynamically following up and managing clinical trial patients based on artificial intelligence according to claim 1, wherein The specific steps of S1 are: S101: Deploy a data collection channel in the medical data network to collect medical records, test, imaging and admission records, examination items, imaging descriptions and drug information, unify the field format, perform field mapping and normalize data types to obtain a unified field table; S102: Performing structural analysis and content recognition based on the field items in the unified field table, screening field records, calculating field completeness, and eliminating records that do not meet the requirements by setting a threshold to obtain standard records of structural fields; S103: Based on the standard record of the structure field, the record fields are matched with the test design requirements, the fields that meet the requirements are compared and a mapping relationship is established, and combined with the dynamic follow-up management node evaluation, the records that meet the test requirements are screened to obtain the test-related health data.

4. The method for screening and dynamically following up and managing clinical trial patients based on artificial intelligence according to claim 1, wherein, The specific steps of S2 are: S201: Based on the test-related health data, each field is compared item by item according to the preset inclusion criteria, and field features consistent with the standard content are recorded to generate field matching statistical results; S202: Based on the field matching statistical results, the records that meet the inclusion criteria are sorted by reception time, the reception time intervals of adjacent records are calculated, the time difference between every two records is obtained, and the mean value of the record time interval is generated; S203: Based on the mean of the recorded time intervals, adjust the synchronization rhythm of data batch updates, calculate the data update frequency, and perform batch allocation according to the update frequency. Incorporate patient records that meet the inclusion criteria into the update queue to generate a group of patients initially screened as qualified by artificial intelligence.

5. The method for screening and dynamically following up clinical trial patients based on artificial intelligence according to claim 4, wherein The specific formula for calculating the data update frequency is as follows: ； Among them, represents the data update frequency, represents the reception time of the i-th record, represents the difference between the i-th record and the previous record, represents the total number of records, represents the time difference between adjacent records, represents the absolute value of the difference from the previous record.

6. The method for screening and dynamic follow-up management of clinical trial patients based on artificial intelligence according to claim 1, characterized in that, The specific steps of S3 are as follows: S301: Based on the group of patients initially screened as qualified by artificial intelligence, extract the name of the medical unit associated with each record and the corresponding number of patients. Group the records by medical unit, count the number of patients in each unit, and calculate the difference in the number of patients between units to generate a statistical table of the difference in the number of patients between units. S302: Based on the statistical table of the difference in the number of patients between units, partition the range of the difference in the number of patients between units. Classify the units with a deviated difference as high-priority monitoring areas, mark the units as data update frequency intensive areas, and extract the corresponding patient records to generate a classification and marking table of the difference between units. S303: Based on the classification and marking table of the difference between units, incorporate the patient records in the high-priority areas into the scope of attention of dynamic follow-up management, generate a schedule for dynamic follow-up management tasks, and synchronously update the monitoring list to obtain a dynamic follow-up management monitoring list.

7. The method for screening and dynamically following up and managing clinical trial patients based on artificial intelligence according to claim 1, characterized in that, The specific formula for calculating the difference in the number of patients between units is as follows: ； Among them, represents the quantity difference between units, represents the number of patients in the th medical unit, represents the number of patients in the th medical unit, represents the adjustment value of unit , represents the adjustment value of unit .

8. The method for screening and dynamically following up and managing clinical trial patients based on artificial intelligence according to claim 1, wherein The specific steps of S4 are as follows: S401: Based on the dynamic follow-up management monitoring list, extract the body temperature, blood glucose, and blood pressure data of each patient, sort the data in chronological order to generate a sorted data sequence. S402: Based on the sorted data sequence, calculate the numerical difference between adjacent records, obtain the daily change amount, and mark the change direction of the daily data according to the change value. Identify records with a consistent trend over multiple days to generate data on the change trend direction. S403: Based on the data on the change trend direction, filter out records with a consistent change trend over multiple days, mark records with obvious trend characteristics as abnormally fluctuating and summarize them to obtain the health fluctuation trend of the dynamically monitored population.

9. The method for screening and dynamically following up and managing clinical trial patients based on artificial intelligence according to claim 1, wherein The specific formula for calculating the numerical difference between adjacent records is as follows: ； Among them, represents the numerical difference between adjacent records, represents the data record value on the th day, represents the data record value of the previous day, represents the weight value on the th day, represents the data weight value on the th day.

10. The method for screening and dynamically following up and managing clinical trial patients based on artificial intelligence according to claim 1, wherein, The specific steps of S5 are as follows: S501: Based on the health fluctuation trend of the dynamically monitored population, extract the patient numbers of those who need to adjust the follow-up plan, call the scheduling records of individual patients, search for and match the associated follow-up nodes, and obtain the time interval of each node to generate the time interval of the patient follow-up nodes. S502: Based on the time interval of the patient follow-up nodes, compare the time arrangements of the nodes with the preset time to identify node contents with time overlap or inconsistency, record and mark the nodes that need to be adjusted to generate a marking table for time node adjustment. S503: Based on the marking table for time node adjustment, adjust the trigger order of the nodes with overlap or inconsistency, set the identification of new task nodes, and synchronize all update information to the individual patient schedule to generate an artificial intelligence-driven dynamic management plan for patients.

Images