Management system applied to medical clinical test

Abstract

The invention discloses a management system applied to a medical clinical test, and relates to the technical field of data processing, and the system comprises an authentication management layer which carries out the identity verification of an accessor, generates a dynamic authority token exclusive to each accessor, converts the identity of the accessor, the current test stage and the subject information into the authority level of the accessor, and carries out the verification of the authority level of the accessor; the authority of the accessor is dynamically adjusted based on the change of the test stage; and the projection layer and operation mapping module is used as a middle buffer layer between the authentication management layer and the data storage layer, sets a high-concurrency threshold value, starts hot data layer projection and establishes an operation mapping table when the request frequency in unit time reaches the high-concurrency threshold value, and performs pre-aggregation and optimization scheduling on repeated requests and concurrent modification operations of the same data group. According to the method, in a high-concurrency scene, the hot data layer can be projected according to the dynamic permission token of the accessor and the test stage.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a management system applied to medical clinical trials. Background Art

[0002] A data management system is a set of technical solutions and software platforms for collecting, storing, processing, integrating, analyzing, and distributing data. It supports internal information development and is also a key enabler for cross-institutional and cross-platform data collaboration.

[0003] In the cross-institutional data collaboration scenario, the existing system faces the following technical bottlenecks:

[0004] Data heterogeneity leads to inefficient transmission: Different organizations often use proprietary databases, storage structures, or data formats (e.g., relational databases, NoSQL databases, XML / JSON documents, etc.). When a system needs to integrate data from various organizations, standardizing data formats and matching semantics becomes the first technical barrier. Data heterogeneity often requires extensive conversion, mapping, and even cleansing operations, which prolongs transmission and processing time.

[0005] The lack of dynamic permission control leads to data leakage risks: When collaborating across institutions, the original permission management model of each system or institution is based on static roles and preset permissions. However, in complex and changing scenarios, static permissions cannot meet the needs of adjusting access rights according to real-time conditions (such as user behavior, access environment, and data importance). For example: When a data interface of an institution is accessed at an unusual time, its permission requirements are different from usual times, but because the permission settings are static, it is difficult to take effect immediately and adapt to abnormal situations;

[0006] System response delays under high concurrency: When multiple organizations and users access the system simultaneously, resources such as the database, application servers, and network bandwidth can easily compete, resulting in some requests not being responded to in a timely manner. For example, when multiple organizations retrieve and update data at the same time, the database's locking mechanism can cause request blocking under high concurrency, resulting in response delays and data synchronization issues due to frequent requests.

[0007] Announcement No. CN118153097B discloses an anti-theft management system and method for key data in drug clinical trials, including an intelligent audit module, a chip module, an autonomous data control module, and an abnormal behavior detection module. It can monitor and analyze test data in real time to ensure that there are no omissions or errors in the data, and generate detailed audit reports so that people can trust the accuracy of the data. It introduces a dedicated chip as hardware-level security protection, making the data storage and processing process more secure and reliable, preventing data from being stolen or illegally accessed.

[0008] Although existing technologies (such as patent publication number CN118153097B, which discloses a method and method for preventing theft of key data from pharmaceutical clinical trials) incorporate specialized chips for hardware-level security protection to prevent data theft and illegal access, the risk of complete data leakage still exists when retrieving data internally. Furthermore, in high-concurrency access scenarios, the database's locking mechanism can cause request blocking and response delays.

[0009] To this end, there is an urgent need for a collaborative management system that supports efficient data synchronization, dynamic permission management, and reduces high-concurrency access latency to address the above technical deficiencies. Summary of the Invention

[0010] One of the purposes of the present invention is to provide a management system for medical clinical trials, which protects entity data and protects the security of complete data through a partially open approach combined with dynamic permission configuration. During high-concurrency access, structured data is projected through a projection layer to process high-concurrency access and reduce latency problems.

[0011] To achieve the above objectives, the present invention is implemented through the following technical solutions: A management system for medical clinical trials, comprising:

[0012] The data collection and integration layer collects experimental data and subject sample data, and processes them uniformly through ETL, performing standardization, deduplication, and error detection.

[0013] The data storage layer stores structured data and adopts a hot and cold data tiering strategy. The hot data layer stores real-time updated data and is deployed in memory cache or high-speed databases. The cold data layer stores underlying data and is deployed in disk storage or archive databases. The data between the cold and hot data layers is adjusted during the trial phase.

[0014] The authentication management layer authenticates the accessor and generates a dynamic permission token for each accessor. It converts the accessor's identity, the current trial stage, and the subject information into the accessor's permission level and dynamically adjusts the accessor's permissions based on changes in the trial stage.

[0015] The projection layer and operation mapping module serve as an intermediate buffer layer between the authentication management layer and the data storage layer. They set a high concurrency threshold. When the request frequency reaches the high concurrency threshold within a unit time, they start the hot data layer projection and establish an operation mapping table. They retrieve the accessor's dynamic permission token, and determine whether the access request of the accessor is submitted to the hot data layer or the cold data layer based on the dynamic permission token. They pre-aggregate and optimize the scheduling of repeated requests and concurrent modification operations for the same data group.

[0016] In one or more embodiments of the present invention, the management system further includes:

[0017] Audit log recording module records the login, data access and modification operation logs of all access users, and performs auditing, anomaly detection and security monitoring;

[0018] The system monitoring and early warning module monitors the operating status, access volume, response delay, and resource load of the management system, and monitors the high concurrency threshold based on access requests. When the access request frequency reaches the high concurrency threshold, an early warning is triggered.

[0019] The data query and analysis module queries, counts and analyzes the data in the data storage layer, and generates test progress reports and data visualization charts;

[0020] The data management layer manages the test data, subject sample data and missing data.

[0021] In one or more embodiments of the present invention, the data management layer includes:

[0022] The trial management module records the entire process from trial establishment to registration, generates a unique trial number and basic information file, collects and integrates subject information at the data collection layer, and screens subjects based on the trial setting criteria. The screening is divided into two types: random grouping strategy and preset grouping strategy;

[0023] The recruitment management module sets screening conditions based on missing data in the trial, generates screening results and synchronizes recruitment progress, monitors recruitment progress and missing data in real time, sets scheduling, reminders, and early warning mechanisms, and synchronously feeds recruitment progress back to the trial management module to optimize missing data and thus change screening criteria.

[0024] In one or more embodiments of the present invention, the weight of each accessor in the experiment is determined based on the accessor's contribution to the experiment data and behavioral performance during the experiment:

[0025] Calculate the weight W of accessor i during the experiment i :

[0026] W i =a·F i +b·Q i +c·R i +d·S i ;

[0027] Among them, F is the activity index, Q is the data quality index, R is the experiment contribution index, S is the safety index, and F i represents the activity score of access user i, Q i represents the data quality score of access point i, R i represents the experimental contribution score of accessor i, S irepresents the security score of accessor i, a, b, c, and d represent activity, data quality, experiment contribution, and security weight coefficient, respectively.

[0028] In one or more embodiments of the present invention, the dynamic permission token is updated as follows:

[0029] The monitoring module embedded in the authorization authentication layer records the user's operational behavior, data quality, security record, and test contribution. The operational behavior, data quality, security record, and test contribution are summarized at fixed time intervals to calculate the user's latest weight.

[0030] The dynamic permission token carries the latest weight value and permission level of the accessor, and embeds the permission level field in the dynamic permission token. When the accessor makes an interaction request, the projection layer performs permission verification and filtering;

[0031] Record the accessor's abnormal operations, data errors, and security issues, adjust the security index S, and reduce the accessor's weight.

[0032] In one or more embodiments of the present invention, when access requests do not reach a high concurrency threshold, access requests from access users are routed based on dynamic permission tokens:

[0033] For high-weight access users and medium-weight access users for pre-approval operations, access requests are directly routed to the hot data layer, achieving high responsiveness and real-time data interaction.

[0034] For medium-weight and low-weight access users who have not obtained pre-approval operations, their access requests are routed to the cold data layer, and information is obtained from stable and periodically updated data.

[0035] In one or more embodiments of the present invention, when the request frequency reaches a high concurrency threshold, the projection layer implements hot data projection based on the hot data layer:

[0036] Obtain the accessor's dynamic permission token to determine the corresponding experiment. The projection layer extracts the experimental data group from the hot data layer and pre-builds the data and materialized views.

[0037] Based on the experimental phases, a phase relationship network is established to associate the experimental data and subject data corresponding to each phase. The survival time and expiration rules corresponding to each phase are set. In combination with write-through or write-back strategies, the data in the projection layer is ensured to be consistent with the data in the hot data layer.

[0038] Monitor access requests and, in combination with dynamic permission tokens, determine whether the requested data belongs to the test data group. If the test data group is hit, it will be directly returned by the data storage layer. If the test data group is not hit, it will be forwarded to the cold data layer.

[0039] In one or more embodiments of the present invention, the relationship network is constructed during the trial phase:

[0040] Divide the pilot process into several key phases, clarify the business scope and data entry requirements for each phase, and create a data baseline to determine the business scope and data entry baseline range for each phase;

[0041] Use the unified identification of the test process in the data storage layer to establish the relationship between each stage;

[0042] Construct a phase relationship network and describe the dependency and transfer relationship of experimental data between phases through a graphical structure;

[0043] Dynamically update test data based on changes in the test process.

[0044] In one or more embodiments of the present invention, the effective time and expiration rules are set:

[0045] Based on the frequency of changes and real-time requirements of stage test data, set the effective time corresponding to different stages, establish a dynamic effective time mechanism, and dynamically adjust the effective time according to the access frequency and update frequency of real-time monitoring data;

[0046] Set the time expiration, data change expiration and system load expiration rules for the test data of each stage. According to the baseline range of the data baseline of each stage, the corresponding data in the projection layer is marked as invalid after the stage data baseline, and the next stage test data group is extracted and updated.

[0047] In one or more embodiments of the present invention, pre-aggregation and optimized scheduling are performed by operating a mapping table:

[0048] Based on the experiment identifier and data group identifier in the token, the projection layer determines the hot data group corresponding to the request;

[0049] Operation mapping table to obtain records:

[0050] If there are operation entries with the same data group in the operation record, the new request will be merged into the entry and pre-aggregated according to the time window;

[0051] If the operation record does not exist, a new record is created to record the initial request information;

[0052] The projection layer judges the parsed dynamic permission token information:

[0053] With modification permission: request submission back to the projection layer;

[0054] With read-only permissions only: Requests are submitted to the cold data layer, and query results are returned using pre-built materialized views and aggregated data;

[0055] For repeated requests or concurrent modification operations within the same data group and within the same time window, the operation mapping table is pre-aggregated:

[0056] For the same query request, pre-aggregated results are returned to reduce multiple accesses;

[0057] For concurrent modification operations, conflict detection and batch updates are performed, and a serialized scheduling strategy is adopted to ensure data consistency.

[0058] Through the above technical solution, the present invention has the following beneficial effects:

[0059] 1. In high-concurrency scenarios, the present invention can project the hot data layer according to the accessor's dynamic permission token and the test phase, and establish an operation mapping table in the projection layer to pre-aggregate and optimize the scheduling of repeated requests and concurrent modification operations for the same data group.

[0060] 2. By implementing hot data projection at the projection layer, the system can significantly improve response speed in high-concurrency scenarios while ensuring the stability and consistency of the core database.

[0061] 3. Establish a phase relationship network to divide the entire trial process requirements into several phases, and associate and map the data and subject information corresponding to each phase to form a logically interconnected network. When transitioning to different phases, the data corresponding to the previous phase loses its timeliness and is updated in real time according to changes in the trial phase.

[0062] 4. By establishing an operation mapping table, repeated access requests and concurrent modification operations within the same data group can be pre-aggregated and optimized for scheduling; by parsing the accessor's dynamic permission token, it is possible to determine in real time whether the request belongs to the hot data layer or the cold data layer for processing.

[0063] Other features and advantages of the present invention will be described in the following description, and part of them will become obvious from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 Schematic diagram of the management system of the present invention. DETAILED DESCRIPTION

[0065] The following drawings illustrate various embodiments of the present invention. For clarity, many practical details are included in the following description. However, it should be understood that these practical details are not intended to limit the present invention. In other words, in some embodiments of the present invention, these practical details are optional. Furthermore, features from different embodiments may be interchangeably applicable, where practically possible.

[0066] Unless otherwise defined, all words used herein (including technical and scientific terms) have their ordinary meanings as understood by those skilled in the art. Furthermore, the definitions of the above-mentioned words in commonly used dictionaries should be interpreted in the context of this specification as having the same meanings as those in the relevant field of the present invention. Unless otherwise explicitly defined, these words should not be interpreted as having idealized or overly formal meanings.

[0067] See Figure 1 The present invention provides a management system for medical clinical trials, which manages structured data, improves the security of structured data, and reduces system response delay under high concurrent access.

[0068] The management system includes:

[0069] The data collection and integration layer collects experimental data and subject sample data, and processes them uniformly through ETL, performing standardization, deduplication, and error detection.

[0070] The data storage layer stores structured data and adopts a hot and cold data tiering strategy. The hot data layer stores real-time updated data and is deployed in memory cache or high-speed databases. The cold data layer stores underlying data and is deployed in disk storage or archive databases. The data between the cold and hot data layers is adjusted during the trial phase.

[0071] The authentication management layer authenticates the accessor and generates a dynamic permission token for each accessor. It converts the accessor's identity, the current trial stage, and the subject information into the accessor's permission level and dynamically adjusts the accessor's permissions based on changes in the trial stage.

[0072] The projection layer and operation mapping module serve as an intermediate buffer layer between the authentication management layer and the data storage layer. They set a high concurrency threshold. When the request frequency reaches the high concurrency threshold within a unit time, they start the hot data layer projection and establish an operation mapping table. They retrieve the accessor's dynamic permission token, and determine whether the access request of the accessor is submitted to the hot data layer or the cold data layer based on the dynamic permission token. They pre-aggregate and optimize the scheduling of repeated requests and concurrent modification operations for the same data group.

[0073] In one feasible method, experimental data and subject sample data are divided into different data groups according to their types. Each data group contains complete experimental data and subject sample data of the corresponding type. Frequent access requests to the same data group and concurrent modification operations will cause the data group to be in a high-concurrency state. However, since not all requests are consistent during high-concurrency access, there are different states such as query and modification. The response state is adjusted according to the different states, which greatly reduces the response speed.

[0074] By projecting the hot data layer through the projection layer, requests can be diverted, further improving the response speed of the data storage layer to requests, adjusting the accessor's authority level, and distinguishing accessors. Due to the different permissions of accessors, there is a difference between being able to only query or modify data. Therefore, after determining the accessor's authority level, different responses can be made.

[0075] The access rights of the access person also correspond to the data scope of the subject that can be viewed, further solving the problem of leakage of the subject's complete data.

[0076] In one embodiment, the management system further comprises:

[0077] Audit log recording module records the login, data access and modification operation logs of all access users, and performs auditing, anomaly detection and security monitoring;

[0078] The system monitoring and early warning module monitors the operating status, access volume, response delay, and resource load of the management system, and monitors the high concurrency threshold based on access requests. When the access request frequency reaches the high concurrency threshold, an early warning is triggered.

[0079] The data query and analysis module queries, counts and analyzes the data in the data storage layer, and generates test progress reports and data visualization charts;

[0080] The data management layer manages the test data, subject sample data and missing data.

[0081] In one feasible method, the system is monitored and warned during the movement of the management system, and the movement status of the management system can be determined. The system monitoring and warning module and the projection layer both monitor the frequency of access requests, which can doubly ensure stable projection of hot data under high concurrency conditions.

[0082] Optionally, the high concurrency threshold sets two response states. First, when the access request for the same data group reaches a certain frequency, in this state, the access request reaches a certain number of times, and the projection layer projects the hot data layer to reduce the system load and improve the response speed; second, when the access request for multiple data groups of the management system reaches a certain number of times, the projection layer projects the hot data layer to reduce the access request frequency of multiple data groups.

[0083] In one embodiment, the data management layer includes:

[0084] The trial management module records the entire process from trial establishment to registration, generates a unique trial number and basic information file, collects and integrates subject information at the data collection layer, and screens subjects based on the trial setting criteria. The screening is divided into two types: random grouping strategy and preset grouping strategy;

[0085] The recruitment management module sets screening conditions based on missing data in the trial, generates screening results and synchronizes recruitment progress, monitors recruitment progress and missing data in real time, sets scheduling, reminders, and early warning mechanisms, and synchronously feeds recruitment progress back to the trial management module to optimize missing data and thus change screening criteria.

[0086] In one feasible method, the subject sample data collected by the data collection integration layer cannot fully meet the test data in the test process. It is necessary to further expand the test sample based on the missing data of the test, determine the missing data of the test, and determine the corresponding recruitment screening conditions. Since there is a time interval between recruitment and the completion of missing data, the recruitment progress is fed back to the test management module, which can pre-supplement some missing data and adjust the screening criteria for the missing data that cannot be supplemented again, thereby completing the missing data.

[0087] Exemplary: for randomization strategy;

[0088] Simple randomization: The most commonly used method, based on the principle of complete randomness. For example, groups are determined using a random number table, a computer-generated random sequence, or coin tossing.

[0089] Stratified randomization: After stratification based on certain important characteristics (such as age group and gender), random assignment is performed within each stratum to ensure that these key characteristics are roughly evenly distributed among the groups.

[0090] Block randomization: The subjects are divided into several blocks and allocated to the experimental group and the control group in equal proportions within each block to ensure the balance of the groups throughout the experiment.

[0091] Dynamic randomization: Dynamically adjust grouping based on real-time data from the experiment to minimize the impact of known confounding factors.

[0092] Forecast grouping strategy;

[0093] Disease-specific grouping: Participants are divided into different groups based on the severity of their disease, disease duration, or other baseline characteristics (such as genotype).

[0094] Grouping by medication history: Different groups are set up based on the patients' previous treatment regimens or drug use in order to study their impact on the trial intervention.

[0095] Grouping by trial objectives: Based on the trial objectives, specific populations are pre-divided, such as "high-risk population" and "general population".

[0096] During the experiment, two strategies are combined according to the experimental goals:

[0097] Preset groups first, then randomize them: After subjects with specific characteristics are grouped according to preset groups, the treatment methods (such as medication and placebo) are randomly assigned within each group.

[0098] Randomization and stratification: Combining random grouping strategy with stratification technology allows the trial to fully focus on the distribution of baseline characteristics of special populations while controlling bias.

[0099] In one embodiment, the weight of each participant in the experiment is determined based on the participant's contribution to the experiment data and behavioral performance during the experiment:

[0100] Calculate the weight W of accessor i during the experiment i :

[0101] W i =a·F i +b·Q i +c·R i +d·S i ;

[0102] Among them, F is the activity index, Q is the data quality index, R is the experiment contribution index, S is the safety index, and F i represents the activity score of access user i, Q i represents the data quality score of access point i, R i represents the experimental contribution score of accessor i, S i represents the security score of accessor i, a, b, c, and d represent activity, data quality, experiment contribution, and security weight coefficient, respectively.

[0103] In one feasible approach, the test data and subject sample data are divided into basic data, sensitive data, and core data. The degree of openness of the accessor's dynamic permission token is determined by the accessor's weight in the test:

[0104] High-authority access: Fully open, including access to core data, sensitive data, and basic data in the hot data layer, data modification, and data review permissions, allowing for real-time data updates and decision-making operations;

[0105] Medium-weight access users: Open sensitive data and basic data, query sensitive data and basic data, and modify basic data. Sensitive data only has read-only permissions and partial pre-approval modification.

[0106] Low-weight access users: only have read-only access to basic data.

[0107] For example, high-weight access W i≥90, high-weight access users are usually experienced, have made significant contributions to experiments, and strictly abide by safety regulations. They can perform real-time data updates and key decision-making operations;

[0108] Medium-weight access users 60≤W i <90: Medium-weighted access users are suitable for daily data processing and data verification work. Preventing medium-weighted access users from accessing core data can avoid risks caused by improper operations;

[0109] Low-weight access, W i <60, used for learning access users in the supervision and training stages or access users who need to be temporarily restricted after multiple violations.

[0110] By assigning weights to accessors, basic data, sensitive data and core data can be protected separately. When accessors with different weights log in to the management system to query data, different data openness levels correspond to different weight levels.

[0111] In one embodiment, the dynamic permission token is updated as follows:

[0112] The monitoring module embedded in the authorization authentication layer records the user's operational behavior, data quality, security record, and test contribution. The operational behavior, data quality, security record, and test contribution are summarized at fixed time intervals to calculate the user's latest weight.

[0113] The dynamic permission token carries the latest weight value and permission level of the accessor, and embeds the permission level field in the dynamic permission token. When the accessor makes an interaction request, the projection layer performs permission verification and filtering;

[0114] Record the accessor's abnormal operations, data errors, and security issues, adjust the security index S, and reduce the accessor's weight.

[0115] In one feasible method, the dynamic permission token of the access user is unique. The dynamic permission token of each access user is combined with the permission level. When the access user logs into the management system, different data is opened in a targeted manner to avoid the problem of repeated permission level confirmation.

[0116] In one embodiment, when access requests do not reach a high concurrency threshold, access requests from access users are routed based on dynamic permission tokens:

[0117] For high-weight access users and medium-weight access users for pre-approval operations, access requests are directly routed to the hot data layer, achieving high responsiveness and real-time data interaction.

[0118] For medium-weight and low-weight access users who have not obtained pre-approval operations, their access requests are routed to the cold data layer, and information is obtained from stable and periodically updated data.

[0119] In one feasible approach, accessors with low, medium, and high weights are routed to the hot data layer and the cold data layer respectively. Due to the different response speeds of the hot and cold data layers, the data contained in the cold data layer is more stable, while the hot data layer requires frequent real-time modifications. Therefore, the faster response speed of the hot data layer can meet the timeliness of real-time data.

[0120] It not only improves data security and prevents improper operations from affecting test data, but also provides a flexible and scientific authority management mechanism for the management system that can adapt to changing test requirements.

[0121] In one embodiment, when the request frequency reaches a high concurrency threshold, the projection layer implements hot data projection based on the hot data layer:

[0122] Obtain the accessor's dynamic permission token to determine the corresponding experiment. The projection layer extracts the experimental data group from the hot data layer and pre-builds the data and materialized views.

[0123] Based on the experimental phases, a phase relationship network is established to associate the experimental data and subject data corresponding to each phase. The survival time and expiration rules corresponding to each phase are set. In combination with write-through or write-back strategies, the data in the projection layer is ensured to be consistent with the data in the hot data layer.

[0124] Monitor access requests and, in combination with dynamic permission tokens, determine whether the requested data belongs to the test data group. If the test data group is hit, it will be directly returned by the data storage layer. If the test data group is not hit, it will be forwarded to the cold data layer.

[0125] In one possible implementation, a write-through strategy or a write-back strategy is adopted based on the access request time:

[0126] When a user modifies hot data, the system uses a write-through mechanism. This means that the modification is written directly to the hot data layer's storage while also updating the corresponding cache and materialized views in the projection layer, ensuring real-time consistency. This mechanism is currently primarily suitable for operations requiring high data real-time performance.

[0127] For certain non-real-time sensitive modifications, the projection layer can temporarily store the modifications and then "write them back" to the hot data layer after a certain period of time. This can reduce the pressure of real-time synchronization while maintaining final data consistency.

[0128] At the same time, the invalidation mechanism and version control ensure the consistency of the hot data layer and the projection layer. For example, timestamp or version number comparison is used to ensure that no data conflicts occur during concurrent modifications.

[0129] If the requested data is not within the experimental data group (for example, the request does not belong to the current hotspot or pre-built experimental data), the request is forwarded to the cold data layer for processing. The cold data layer typically uses standard query methods to respond to requests with lower real-time requirements. Intelligent routing also promptly updates the hotspot data group's concepts and projection layer cache content based on real-time access volume and dynamic permission adjustment strategies, ensuring that the system can quickly adapt to load changes under long-term high concurrency.

[0130] In one embodiment, during the trial phase, the relationship network is constructed:

[0131] Divide the pilot process into several key phases, clarify the business scope and data entry requirements for each phase, and create a data baseline to determine the business scope and data entry baseline range for each phase;

[0132] Use the unified identification of the test process in the data storage layer to establish the relationship between each stage;

[0133] Construct a phase relationship network and describe the dependency and transfer relationship of experimental data between phases through a graphical structure;

[0134] Dynamically update test data based on changes in the test process.

[0135] In one feasible approach, a phase relationship network is established to divide the entire trial process requirements into several phases, and the data and subject information corresponding to each phase are associated and mapped to form a logically interconnected network. When transitioning to different phases, the data corresponding to the previous phase loses its timeliness and is updated in real time according to changes in the trial phase.

[0136] In one embodiment, the effective time and expiration rules are set:

[0137] Based on the frequency of changes and real-time requirements of stage test data, set the effective time corresponding to different stages, establish a dynamic effective time mechanism, and dynamically adjust the effective time according to the access frequency and update frequency of real-time monitoring data;

[0138] Set the time expiration, data change expiration and system load expiration rules for the test data of each stage. According to the baseline range of the data baseline of each stage, the corresponding data in the projection layer is marked as invalid after the stage data baseline, and the next stage test data group is extracted and updated.

[0139] In one implementation, the failure strategy is combined with a write-through or write-back strategy:

[0140] Write-through: The modification operation updates the backend data and the projection layer cache simultaneously. In this case, the invalidation rules are mainly triggered by time and events.

[0141] Write-back: The cache first saves the modification operation and synchronizes the update in batches. The invalidation rule needs to additionally consider the synchronization invalidation problem after the batch update.

[0142] For example, in the screening stage, data updates are more frequent (such as the first entry of subject information, changes in preliminary screening results), and the effective time is set to be shorter (a few minutes to more than ten minutes).

[0143] Random grouping stage: The data baseline is relatively stable, and there is a need for re-grouping and adjustment after grouping. The effective time is moderate (30 minutes or one hour).

[0144] Treatment phase: Data changes rapidly (such as medication records and adverse reactions), requiring a highly real-time effectiveness time (1-5 minutes).

[0145] Follow-up stage: Data update frequency is low, and the effective time can be set longer (several hours to one day).

[0146] By establishing a trial phase relationship network, clinical trial data can be finely divided according to business processes. Appropriate validity periods and expiration rules can be set for each phase to ensure the validity of pre-built data and materialized views in the projection layer. This mechanism enables rapid response to user requests in high-concurrency scenarios while ensuring data consistency with the hot data layer.

[0147] In one embodiment, pre-aggregation and optimized scheduling are performed by operating a mapping table:

[0148] Based on the experiment identifier and data group identifier in the token, the projection layer determines the hot data group corresponding to the request;

[0149] Operation mapping table to obtain records:

[0150] If there are operation entries with the same data group in the operation record, the new request will be merged into the entry and pre-aggregated according to the time window;

[0151] If the operation record does not exist, a new record is created to record the initial request information;

[0152] The projection layer judges the parsed dynamic permission token information:

[0153] With modification permission: request submission back to the projection layer;

[0154] With read-only permissions only: Requests are submitted to the cold data layer, and query results are returned using pre-built materialized views and aggregated data;

[0155] For repeated requests or concurrent modification operations within the same data group and within the same time window, the operation mapping table is pre-aggregated:

[0156] For the same query request, only pre-aggregated results are returned to reduce multiple accesses;

[0157] For concurrent modification operations, conflict detection and batch updates are performed, and a serialized scheduling strategy is adopted to ensure data consistency.

[0158] One feasible approach involves establishing an operation mapping table to pre-aggregate and optimize scheduling for repeated access requests and concurrent modification operations within the same data group. By parsing the accessor's dynamic permission token, it can determine in real time whether the request belongs to the hot or cold data tier. If the request hits the hot data tier, it can be quickly responded to using pre-built materialized views and optimized cache. If it misses, it is forwarded to the cold data tier for processing. Furthermore, batch update strategies, such as write-through or write-back, ensure consistency between the projection layer and the underlying data during data updates.

[0159] In summary, the technical solutions disclosed in the above embodiments of the present invention have at least the following advantages:

[0160] Although the present invention is disclosed in conjunction with the above embodiments, it is not intended to limit the present invention. Anyone familiar with this technology can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention should be based on the definition of the attached claims.

Claims

1. A management system for medical clinical trials, characterized in that: include: The data collection and integration layer collects experimental data and subject sample data, and processes them uniformly through ETL, performing standardization, deduplication, and error detection. The data storage layer stores structured data and adopts a hot and cold data tiering strategy. The hot data layer stores real-time updated data and is deployed in memory cache or high-speed databases. The cold data layer stores underlying data and is deployed in disk storage or archive databases. The data between the cold and hot data layers is adjusted during the trial phase. The authentication management layer authenticates the accessor and generates a dynamic permission token for each accessor. It converts the accessor's identity, the current trial stage, and the subject information into the accessor's permission level and dynamically adjusts the accessor's permissions based on changes in the trial stage. The projection layer and operation mapping module serve as an intermediate buffer layer between the authentication management layer and the data storage layer. They set a high concurrency threshold. When the request frequency reaches the high concurrency threshold within a unit time, they start the hot data layer projection and establish an operation mapping table. They retrieve the accessor's dynamic permission token, and determine whether the access request of the accessor is submitted to the hot data layer or the cold data layer based on the dynamic permission token. They pre-aggregate and optimize the scheduling of repeated requests and concurrent modification operations for the same data group.

2. A management system for medical clinical trials according to claim 1, characterized in that: The management system also includes: Audit log recording module records the login, data access and modification operation logs of all access users, and performs auditing, anomaly detection and security monitoring; The system monitoring and early warning module monitors the operating status, access volume, response delay, and resource load of the management system, and monitors the high concurrency threshold based on access requests. When the access request frequency reaches the high concurrency threshold, an early warning is triggered. The data query and analysis module queries, counts and analyzes the data in the data storage layer, and generates test progress reports and data visualization charts; The data management layer manages the test data, subject sample data and missing data.

3. A management system for medical clinical trials according to claim 2, characterized in that: The data management layer includes: The trial management module records the entire process from trial establishment to registration, generates a unique trial number and basic information file, collects and integrates subject information at the data collection layer, and screens subjects based on the trial setting criteria. The screening is divided into two types: random grouping strategy and preset grouping strategy; The recruitment management module sets screening conditions based on missing data in the trial, generates screening results and synchronizes recruitment progress, monitors recruitment progress and missing data in real time, sets scheduling, reminders, and early warning mechanisms, and synchronously feeds recruitment progress back to the trial management module to optimize missing data and thus change screening criteria.

4. A management system for medical clinical trials according to claim 3, characterized in that: Based on the participant's contribution to the test data and their performance during the test, the weight of each participant in the test is determined: Calculate the weight W of accessor i during the experiment i : W i =a·F i +b·Q i +c·R i +d·S i ; Among them, F is the activity index, Q is the data quality index, R is the experiment contribution index, S is the safety index, and F i represents the activity score of access user i, Q i represents the data quality score of access point i, R i represents the experimental contribution score of accessor i, S i represents the security score of accessor i, a, b, c, and d represent activity, data quality, experiment contribution, and security weight coefficient, respectively.

5. A management system for medical clinical trials according to claim 4, characterized in that: The dynamic permission token is updated as follows: The monitoring module embedded in the authorization authentication layer records the user's operational behavior, data quality, security record, and test contribution. The operational behavior, data quality, security record, and test contribution are summarized at fixed time intervals to calculate the user's latest weight. The dynamic permission token carries the latest weight value and permission level of the accessor, and embeds the permission level field in the dynamic permission token. When the accessor makes an interaction request, the projection layer performs permission verification and filtering; Record the accessor's abnormal operations, data errors, and security issues, adjust the security index S, and reduce the accessor's weight.

6. A management system for medical clinical trials according to claim 5, characterized in that: When access requests do not reach the high concurrency threshold, the access requests of the accessor are routed based on the dynamic permission token: For high-weight access users and medium-weight access users for pre-approval operations, access requests are directly routed to the hot data layer, achieving high responsiveness and real-time data interaction. For medium-weight and low-weight access users who have not obtained pre-approval operations, their access requests are routed to the cold data layer, and information is obtained from stable and periodically updated data.

7. A management system for medical clinical trials according to claim 6, characterized in that: When the request frequency reaches the high concurrency threshold, the projection layer implements hot data projection based on the hot data layer: Obtain the accessor's dynamic permission token to determine the corresponding experiment. The projection layer extracts the experimental data group from the hot data layer and pre-builds the data and materialized views. Based on the experimental phases, a phase relationship network is established to associate the experimental data and subject data corresponding to each phase. The survival time and expiration rules corresponding to each phase are set. In combination with write-through or write-back strategies, the data in the projection layer is ensured to be consistent with the data in the hot data layer. Monitor access requests and, in combination with dynamic permission tokens, determine whether the requested data belongs to the test data group. If the test data group is hit, it will be directly returned by the data storage layer. If the test data group is not hit, it will be forwarded to the cold data layer.

8. A management system for medical clinical trials according to claim 7, characterized in that: Network building during the pilot phase: Divide the pilot process into several key phases, clarify the business scope and data entry requirements for each phase, and create a data baseline to determine the business scope and data entry baseline range for each phase; Use the unified identification of the test process in the data storage layer to establish the relationship between each stage; Construct a phase relationship network and describe the dependency and transfer relationship of experimental data between phases through a graphical structure; Dynamically update test data based on changes in the test process.

9. A management system for medical clinical trials according to claim 8, characterized in that: Set the effective time and expiration rules: Based on the frequency of changes and real-time requirements of stage test data, set the effective time corresponding to different stages, establish a dynamic effective time mechanism, and dynamically adjust the effective time according to the access frequency and update frequency of real-time monitoring data; Set the time expiration, data change expiration and system load expiration rules for the test data of each stage. According to the baseline range of the data baseline of each stage, the corresponding data in the projection layer is marked as invalid after the stage data baseline, and the next stage test data group is extracted and updated.

10. A management system for medical clinical trials according to claim 9, characterized in that: Pre-aggregation and optimized scheduling are performed through the operation mapping table: Based on the experiment identifier and data group identifier in the token, the projection layer determines the hot data group corresponding to the request; Operation mapping table to obtain records: If there are operation entries with the same data group in the operation record, the new request will be merged into the entry and pre-aggregated according to the time window; If the operation record does not exist, a new record is created to record the initial request information; The projection layer judges the parsed dynamic permission token information: With modification permission: request submission back to the projection layer; With read-only permissions only: Requests are submitted to the cold data layer, and query results are returned using pre-built materialized views and aggregated data; For repeated requests or concurrent modification operations within the same data group and within the same time window, the operation mapping table is pre-aggregated: For the same query request, pre-aggregated results are returned to reduce multiple accesses; For concurrent modification operations, conflict detection and batch updates are performed, and a serialized scheduling strategy is adopted to ensure data consistency.

Images