A Clinical Trial Data Processing Method and System Based on IRT System

Abstract

This invention belongs to the field of data processing technology, specifically relating to a clinical trial data processing method and system based on an IRT (Initial Timeout) system. The method includes: after the coordinator's timeout failure, any participating system initiates a self-decision-making process: calculating the local transaction urgency related to business importance and waiting time; detecting and integrating the status of other participating systems; calculating the consensus confidence level; combining a preset asymmetric risk coefficient to assess the expected risk cost of either submit or rollback; and automatically executing the decision with the lower risk cost. This invention, through a decentralized autonomous decision-making mechanism, effectively avoids indefinite system waiting, achieves intelligent fault tolerance based on quantitative risk assessment, and improves the robustness and security of the clinical trial system.

Description

Technical Field

[0001] This invention relates to the field of data processing technology. More specifically, this invention relates to a clinical trial data processing method and system based on an IRT system. Background Technology

[0002] In the execution of modern clinical trials, multiple information systems typically need to work together, such as the IRT system for randomization of subjects and drug management, the EDC system for collecting clinical data, and the CTMS system for overall trial management; the data consistency between these systems is crucial to the compliance and scientific rigor of the trial.

[0003] To ensure the atomicity of cross-system operations, i.e., all operations either succeed or fail, existing technologies often use a two-phase commit protocol as the core technical means. This protocol introduces a centralized coordinator to uniformly schedule the commit or rollback of all transactions participating in the system.

[0004] However, this traditional two-phase commit protocol heavily relies on a centralized coordinator to uniformly schedule transactions. Once the coordinator experiences a single point of failure or becomes isolated from some participating systems, all participating systems that have completed preparation and locked resources will enter an uncertain state, unable to decide for themselves whether to commit or roll back, and can only passively wait indefinitely for the coordinator to recover. More importantly, in the special scenario of clinical trials, the business risks of different erroneous decisions are highly asymmetrical. For example, erroneously committing a drug allocation transaction that should fail may lead to a serious safety incident, the risk of which is far greater than the business process delay caused by erroneous rollback. Existing technologies do not fully consider this asymmetrical risk cost when dealing with coordinator failure issues, and their decision-making mechanisms lack quantitative assessment of business consequences, which may lead the system to make high-risk choices under uncertain circumstances. Summary of the Invention

[0005] To address the technical problem of existing technologies that heavily rely on a centralized coordinator in two-phase commit protocols, where all participating systems are trapped in an uncertain state of indefinite blocking when the coordinator fails, unable to make independent decisions and increasing system risk and business interruption time, this invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides a clinical trial data processing method based on an IRT system, comprising: initiating a distributed transaction involving multiple participating systems, and having a coordinator send a preparation request to each participating system; upon receiving the preparation request, each participating system enters a ready state and responds to the coordinator; if no second-stage instruction from the coordinator is received within a preset timeout threshold, the coordinator is deemed to have failed and the current participating system initiates a self-decision-making process as the decision initiator; the self-decision-making process includes: calculating the urgency of a local transaction, wherein the urgency of a local transaction is positively correlated with the business weight of the corresponding transaction and the waiting time since the ready state; obtaining other participating systems'... The system calculates the consensus confidence score based on the transaction status of all participating systems. The consensus confidence score is a weighted average of the state values ​​of all participating systems, and the weighted average is based on the system weight of each participating system. Based on the urgency of local transactions and a preset asymmetric risk coefficient, the system calculates the heuristic commit failure cost and the heuristic rollback failure cost. Based on the consensus confidence score, the system estimates the probability that the global transaction should eventually be committed and rolled back. Based on the heuristic commit failure cost and the heuristic rollback failure cost, as well as the probability that the global transaction should eventually be committed and rolled back, the system calculates the expected risk cost of taking a commit action and the expected risk cost of taking a rollback action, and executes the action with the lower expected risk cost as the final decision.

[0007] Compared to existing two-phase commit protocols that heavily rely on a centralized coordinator, this invention proposes a mechanism for participating systems to make autonomous decisions after the coordinator fails. By introducing a timeout-triggered self-decision-making process, any participating system can proactively resolve the transaction hanging problem after sensing the coordinator's failure. This process integrates the local transaction urgency characteristics representing business importance and waiting time, the consensus confidence characteristics representing collective wisdom, and the asymmetric risk coefficient in clinical trial scenarios. By calculating and comparing the expected risk costs of commit and rollback operations, the system ultimately executes the decision with the lowest risk. This solves the single point of failure and blocking problems of traditional solutions, enabling the system to make rational decisions aimed at minimizing business risks under uncertainty. This improves the robustness, availability, and security of distributed systems in high-risk, high-compliance scenarios such as clinical trials.

[0008] Preferably, the formula for calculating the urgency of the local transaction is: In the formula, The urgency of local affairs; This refers to the business weighting coefficient. This refers to the elapsed waiting time since the system entered the ready state. The preset timeout threshold for the coordinator instruction wait timer.

[0009] This invention combines static business weight coefficients with dynamic waiting times to quantify the urgency of business priorities and time passage into a single feature. This enables subsequent decision-making processes to prioritize more important and longer-waiting tasks, thereby achieving the perception and prioritization of business value. It ensures that core business processes with the greatest impact on clinical trials are given priority in resource competition and decision-making trade-offs, thus improving the intelligence and business relevance of decision-making.

[0010] Preferably, the consensus confidence level is calculated as follows: In the formula, Consensus confidence level; The total number of systems participating in this distributed transaction; For the first The system weights of each participating system; For from the first Each participating system acquires a status value.

[0011] This invention enables decision initiators to actively probe the states of other participating systems and perform a weighted average based on the system weights of each system, thereby forming a quantitative estimate of the global transaction state. This collective intelligence mechanism avoids decision-making misjudgments caused by insufficient local information, enhances the system's fault tolerance under complex network partitions or partial node failures, and makes the final decision more global and robust.

[0012] Preferably, the state value This is the numerical value mapped to the transaction status participating in the system: if its status is ready, then... If it has already been rolled back, then If no response is received or the system is in an initial state, then .

[0013] Preferably, the formulas for calculating the heuristic commit failure cost and the heuristic rollback failure cost are as follows: ; In the formula, Cost of heuristic commit failure; Cost of heuristic rollback failure; The urgency of local affairs; and For asymmetric risk coefficient, , All are positive real numbers, and Greater than .

[0014] When calculating heuristic failure costs, this invention explicitly points out that the coefficient representing the risk of submission failure is greater than the coefficient representing the risk of rollback failure. This solves the core pain point of existing technologies not fully considering the asymmetric risk in clinical trials. It models this business-level asymmetric risk, making the system more inclined to seek benefits and avoid risks when making decisions. In the case of uncertainty, it tends to choose the safer operation of rollback, thus integrating the principle of safety into the automated decision-making logic.

[0015] Preferably, the probability that the global transaction should eventually be committed and rolled back is calculated using the following formula: ; In the formula, The probability that a global transaction should eventually be committed; The probability that the global transaction should eventually be rolled back; The consensus confidence level.

[0016] Preferably, the expected risk cost of taking the commit action is equal to the product of the probability that the global transaction should eventually be rolled back and the cost of the heuristic commit failure, and the expected risk cost of taking the rollback action is equal to the product of the probability that the global transaction should eventually be committed and the cost of the heuristic rollback failure.

[0017] The basic principle of the decision theory in this invention is that expected risk equals the probability of an event occurring multiplied by the cost of that event. It calculates the expected risk cost of taking either submit or rollback actions, so that the final decision is no longer based on experience, but on a quantitative comparison of potential risks, ensuring that the system always chooses the action that minimizes the expected loss.

[0018] Preferably, the self-decision-making process further includes: recording the final decision and its basis in a local anti-tampering audit log, wherein the basis for the decision includes at least: local transaction urgency, consensus confidence, heuristic commit failure cost, heuristic rollback failure cost, expected risk cost of taking a commit action, and expected risk cost of taking a rollback action.

[0019] Preferably, the value range of the business weight is (0,1]; the value range of the system weight is (0,1]; and the value range of the timeout threshold is [5,30], in seconds.

[0020] Secondly, the present invention provides a clinical trial data processing system based on an IRT system, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned clinical trial data processing method based on an IRT system is implemented.

[0021] By adopting the above technical solution, a clinical trial data processing method based on the IRT system is generated into a computer program and stored in a memory for loading and execution by a processor. This allows for the creation of a terminal device based on the memory and processor, facilitating its use.

[0022] The beneficial effects of this invention are as follows:

[0023] Compared to existing two-phase commit protocols that heavily rely on a centralized coordinator, this invention proposes a mechanism for participating systems to make autonomous decisions after the coordinator fails. By introducing a timeout-triggered self-decision-making process, any participating system can proactively resolve the transaction hanging problem after sensing the coordinator's failure. This process integrates the local transaction urgency characteristics representing business importance and waiting time, the consensus confidence characteristics representing collective wisdom, and the asymmetric risk coefficient in clinical trial scenarios. By calculating and comparing the expected risk costs of commit and rollback operations, the system ultimately executes the decision with the lowest risk. This solves the single point of failure and blocking problems of traditional solutions, enabling the system to make rational decisions aimed at minimizing business risks under uncertainty. This improves the robustness, availability, and security of distributed systems in high-risk, high-compliance scenarios such as clinical trials. Attached Figure Description

[0024] Figure 1 This is a flowchart illustrating a clinical trial data processing method based on an IRT system according to the present invention;

[0025] Figure 2 This is a schematic diagram illustrating the local transaction urgency and consensus confidence of each participating system;

[0026] Figure 3 This is a flowchart illustrating step S4;

[0027] Figure 4 This is a schematic diagram illustrating the heuristic commit failure cost and heuristic rollback failure cost of each participating system;

[0028] Figure 5 This is a schematic diagram illustrating the expected risk cost of each participating system taking a submission action and the expected risk cost of taking a rollback action;

[0029] Figure 6 This is a schematic diagram illustrating the comparison between our side and the traditional 2PC protocol. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0032] This invention discloses a clinical trial data processing method based on an IRT system, referring to... Figure 1 This includes steps S1-S4:

[0033] S1: Initiate a distributed transaction, and each participating system monitors the timeout status of the coordinator's instructions.

[0034] It should be noted that this step is the prerequisite for triggering the entire self-decision-making mechanism. By setting up an independent local timeout monitor for each participating system, the original fault detection mode that relied entirely on the centralized coordinator is transformed into a distributed decentralized monitoring mode. This ensures that any participating system can promptly initiate the subsequent self-decision-making process when the coordinator is unresponsive for an extended period of time, preventing the entire transaction cluster from falling into an indefinite blocking state. This is the first line of defense to ensure system availability.

[0035] Specifically, when a cross-system business process is initiated, the coordinator node sends the first phase of a two-phase commit, namely a preparation request, to all systems participating in the transaction. After receiving the preparation request, each participating system executes its local transaction preparation, locks relevant data resources, and returns a ready vote response to the coordinator.

[0036] Furthermore, after sending a voting response, each participating system immediately starts a local coordinator instruction waiting timer, and the preset timeout threshold of this timer is... If within the preset timeout threshold If the coordinator fails to receive the second-stage instruction, namely global commit or global rollback, the coordinator is deemed to have failed, and the subsequent self-decision-making process is automatically triggered.

[0037] Among them, the timeout threshold This is the longest reasonable time that the participating system is willing to wait for the coordinator's second-phase instructions. Its value should be greater than the normal network round-trip time and the coordinator's processing time, but small enough to detect faults as quickly as possible. Therefore, its range can be set to [5, 30], in seconds. In this embodiment, […]. Set to 15 seconds.

[0038] It should be noted that by establishing a distributed timeout monitoring mechanism, the ability to detect single-point failures is extended to every participating system, ensuring that failures can be detected in a timely manner and providing a definite entry point for subsequent decision-making processes. This fundamentally solves the problem of infinite waiting when the coordinator fails in the traditional two-phase commit protocol.

[0039] S2: Calculate the local transaction urgency based on the transaction's business weight and waiting time.

[0040] It should be noted that in the abnormal situation of coordinator failure, system resources are locked, and not all suspended transactions have equal importance. Therefore, it is necessary to construct an indicator to quantify the business criticality of transactions, so that the subsequent decision-making process can have business awareness and transform abstract business priorities into calculable values. For example, randomization of subjects is more important than regular data synchronization, ensuring that high-priority business receives more attention in resource competition and decision trade-offs.

[0041] Specifically, when any participating system enters the self-decision-making process due to the coordinator's timeout, the local transaction urgency is calculated based on the transaction type and the waiting time. The specific calculation formula is as follows:

[0042]

[0043] In the formula, The urgency of local affairs; This refers to the business weighting coefficient. This refers to the elapsed waiting time since the system entered the ready state. The preset timeout threshold for the coordinator instruction wait timer.

[0044] in, This is a quantitative representation of the inherent importance of different business types, configured by the system administrator based on key nodes of the clinical trial protocol, with a value range of (0,1]. In this embodiment, for core businesses such as subject randomization, its business weight coefficient is... Set to 0.9; for regular data synchronization services, its business weight coefficient is... Set to 0.3.

[0045] It should be noted that the constructed local transaction urgency index combines business-level priorities with technical-level decision-making mechanisms. This index not only considers the static importance of the business but also dynamically incorporates the urgency brought about by the passage of time. This makes the entire self-decision-making system no longer blindly tolerate technical faults but can prioritize the core business processes that have the greatest impact on the clinical trial process, greatly improving the business value and intelligence level of the solution.

[0046] S3: The decision initiator probes the transaction status of other participating systems and calculates the consensus confidence based on the weighted average of the statuses.

[0047] It should be noted that in a distributed environment, no single node should make decisions that affect the whole system while remaining in an information silo. Instead, the participating systems that enter the decision-making process, i.e., the decision initiators, should actively investigate the status of other peers. By collecting this scattered information, a quantitative judgment on the current collective tendency of the entire transaction cluster can be formed. This consensus confidence level is an optimal estimate of the global state and provides key external environmental input for subsequent risk assessment.

[0048] Specifically, the decision initiator that enters the timeout state broadcasts a state probe request to all other participating systems in this distributed transaction; upon receiving the request, the other participating systems return their current transaction state.

[0049] Furthermore, the decision initiator calculates the consensus confidence level based on the collected data on the participating system states. The specific calculation formula is as follows:

[0050]

[0051] In the formula, Consensus confidence level; The total number of systems participating in this distributed transaction; For the first The system weights of each participating system; For from the first Each participating system acquires a status value; status value This is the numerical value mapped to the transaction status participating in the system: if its status is ready, then... If it has already been rolled back, then If no response is received or the system is in an initial state, then .

[0052] It should be noted that, It sums the results of all participating systems in this distributed transaction except for the decision initiator.

[0053] Among them, system weights This reflects the core position of different systems in the business process: the EDC system, as the data source, should have a higher system weight than the downstream CTMS system, and the system weight ranges from (0,1]. Therefore, in this embodiment, the EDC system weight is set to 1.0, the IRT system weight is set to 0.8, and the CTMS system weight is set to 0.5.

[0054] It should be noted that by introducing peer-to-peer detection and weighted consensus mechanisms, the decision-making basis is shifted from relying on a single and potentially ineffective coordinator to relying on the collective wisdom of the entire participating system cluster. This not only enhances the system's fault tolerance under complex network partitions and partial node failures, but also makes the final decision more robust and has a global perspective, effectively avoiding misjudgments caused by insufficient local information.

[0055] For example, for the five participating systems, they are represented by transaction IDs: TX1001 to TX1005, where TX1001 is a heuristic commit and TX1002 to TX1005 are heuristic rollbacks. When the coordinator times out, the local transaction urgency is calculated based on the transaction type and the waiting time, and the consensus confidence is calculated based on the collected participating system states. A diagram illustrating the local transaction urgency and consensus confidence of each participating system is shown below. Figure 2 As shown.

[0056] S4: Based on the urgency of local transactions and the preset asymmetric risk coefficient, calculate the heuristic commit failure cost and heuristic rollback failure cost. Based on the consensus confidence, estimate the probability that the global transaction should eventually be committed and rolled back. Then calculate the expected risk cost of taking the commit action and the expected risk cost of taking the rollback action, and execute the action with the lower expected risk cost as the final decision.

[0057] It should be noted that by quantifying the drastically different business consequences of two different failure modes—error submission and error rollback—and calculating and comparing the expected risk costs of taking different actions, the system can make more prudent and responsible choices, which is crucial for the field of clinical trials where there is zero tolerance for risk.

[0058] The flowchart for step S4 is shown below. Figure 3 The process includes steps S401 to S403, specifically as follows:

[0059] S401: Calculate the heuristic commit failure cost and heuristic rollback failure cost based on the local transaction urgency and the preset asymmetric risk coefficient.

[0060] It should be noted that linking the cost of errors to the importance or urgency of the transaction itself means that the more important the transaction, the higher the cost of a wrong decision. At the same time, by introducing two unequal risk coefficients, the model directly reflects the business's differentiated perception of the two types of errors, thereby quantifying the consequences of different decision-making errors.

[0061] Specifically, heuristic submission failure costs are defined for the two potential erroneous decisions. and the cost of heuristic rollback failure The specific calculation formula is as follows:

[0062]

[0063]

[0064] In the formula, Cost of heuristic commit failure; Cost of heuristic rollback failure; The urgency of local affairs; and This refers to the asymmetric risk coefficient. and These are the amplification factors for the relative business risk of erroneous commits and rollbacks. These two coefficients are preset by business experts based on the transaction type, and their values ​​are both positive real numbers. Furthermore, for high-risk transactions... It should be much larger than Therefore, in this embodiment, for the high-risk task of drug distribution, Set to 10. Setting it to 1.5 explicitly indicates that the cost of an erroneous commit is several times the cost of an erroneous rollback.

[0065] It should be noted that by establishing a cost function related to business urgency and risk coefficient, a data foundation is provided for subsequent expected risk cost analysis, making the entire decision-making model highly business interpretable.

[0066] For example, for the five participating systems, the heuristic commit failure cost and heuristic rollback failure cost are calculated respectively. A schematic diagram of the heuristic commit failure cost and heuristic rollback failure cost for each participating system is shown below. Figure 4 As shown, more pressing matters are accompanied by higher potential costs associated with any wrong decision-making; for example, the bar pairs of TX1005 are higher than those of TX1004.

[0067] S402: Based on consensus confidence, estimate the probability that a global transaction should eventually be committed and rolled back.

[0068] It should be noted that the probability of obtaining different outcomes is a prerequisite for calculating the expected risk cost. Since the coordinator has failed, the final global state is unknown. Therefore, this step utilizes the consensus confidence level, which represents collective wisdom. As the best estimate of the global state under the current information, it is smoothly converted into a probability value, enabling rational reasoning under uncertainty.

[0069] Specifically, based on consensus confidence level Estimate the probability that the global transaction should eventually be committed. And the probability of rollback The specific calculation formula is as follows:

[0070]

[0071]

[0072] In the formula, The probability that a global transaction should eventually be committed; The probability that the global transaction should eventually be rolled back; The consensus confidence level.

[0073] It should be noted that this step transforms the confidence index describing collective tendency into a probabilistic input necessary for calculating expected risk through a linear mapping, thus building a bridge from state perception to probabilistic reasoning in the decision model, making subsequent cost-minimizing decisions possible.

[0074] S403: Calculate the expected risk cost of taking the submit action and the expected risk cost of taking the rollback action, and execute the action with the lower expected risk cost as the final decision.

[0075] It should be noted that under uncertainty, a rational decision-maker should choose the action plan that minimizes the expected loss or cost. Therefore, by combining the cost function and outcome probability calculated in the previous steps, this step calculates the expected risk score for both the submit and rollback options, so that the final decision is no longer an ambiguous judgment.

[0076] Specifically, the decision initiators each calculate the expected risk cost of taking the submission action. And the expected risk cost of taking a rollback action The specific calculation formula is as follows:

[0077]

[0078]

[0079] In the formula, The expected risk cost of taking the submission action; The expected risk cost of taking a rollback action; The probability that the global transaction should eventually be rolled back; Cost of heuristic commit failure; The probability that a global transaction should eventually be committed; Cost of heuristic rollback failure.

[0080] For example, for each of the five participating systems, the expected risk cost of taking a submit action and the expected risk cost of taking a rollback action are calculated. A schematic diagram illustrating the expected risk costs of each participating system taking a submit action and taking a rollback action is shown below. Figure 5 As shown, the arrows and labels (i.e., choose to commit or choose to roll back) indicate the action chosen by the algorithm, always pointing to the bar with the lower height.

[0081] Furthermore, the decision initiator executes the final decision based on the principle of minimizing expected risk costs: if If the condition is met, a heuristic commit of the local transaction is executed; otherwise, a heuristic rollback of the local transaction is executed.

[0082] Furthermore, upon completion, the decision and its complete decision-making chain will be used as the basis, including , , , , , The value is recorded in detail in the local anti-tampering audit log.

[0083] It should be noted that by ultimately making decisions by minimizing expected risk costs, the system's autonomous behavior ensures that even in the extreme anomaly of coordinator failure, the primary goal is to reduce potential business risks. Compared to traditional decision-making mechanisms based solely on technical status, this risk-based decision-making approach enhances the system's security and reliability in critical business scenarios and guarantees the auditability of all decision-making processes.

[0084] For example, a schematic diagram comparing this solution with the traditional 2PC protocol is shown below. Figure 6 As shown, in the traditional 2PC protocol, the high and unavoidable costs of system blocking and indefinite waiting when the coordinator fails; while this solution successfully resolves transactions while minimizing the expected risk cost. The huge difference between the two shows that this solution transforms the high-cost failure mode that leads to system stagnation into a low-cost and manageable solution. It not only avoids blocking but also actively manages by selecting the lowest-risk path, thereby greatly reducing operational risk costs.

[0085] This invention also discloses a clinical trial data processing system based on an IRT system, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement a clinical trial data processing method based on an IRT system according to the present invention.

[0086] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

Claims

1. An IRT system-based clinical trial data processing method, characterized in that, The method comprises the following steps: starting a distributed transaction comprising a plurality of participating systems, and sending a prepare request to each participating system by a coordinator; after receiving the prepare request, each participating system enters a prepare-ready state and responds to the coordinator; when a second-stage instruction from the coordinator is not received within a preset timeout threshold, it is determined that the coordinator fails, and a self-decision process is started by the current participating system as a decision initiator; the self-decision process comprises the following steps: calculating a local transaction urgency, which is positively correlated with a business weight of a corresponding transaction and a waiting time length since the prepare-ready state; obtaining transaction states of other participating systems and calculating a consensus confidence, which is a weighted average of state values of all participating systems, the weighted average being based on system weights of the participating systems; calculating a heuristic commit failure cost and a heuristic rollback failure cost according to the local transaction urgency and a preset asymmetric risk coefficient; estimating a probability that a global transaction should finally be committed and rolled back based on the consensus confidence; calculating an expected risk cost of taking a commit action and an expected risk cost of taking a rollback action based on the heuristic commit failure cost and the heuristic rollback failure cost, and the probability that the global transaction should finally be committed and rolled back, and performing an action with a lower expected risk cost as a final decision; the asymmetric risk coefficient is a magnification factor of relative business risks of false commitment and false rollback, which is preset by a business expert according to a transaction type, and is a positive real number.

2. The method of processing clinical trial data based on IRT system according to claim 1, characterized in that, the calculation formula of the local transaction urgency is: ； wherein is a local transaction urgency; is a business weight coefficient; is a waiting duration that has elapsed since the participation system entered a ready state; is a preset timeout threshold of the coordinator instruction waiting timer.

3. The method of claim 1, wherein the IRT system is a system for processing clinical trial data. the calculation formula of the consensus confidence is: ； In the formula, is the consensus confidence; is the total number of systems participating in the current distributed transaction; is the system weight of the participating system; is the state value obtained from the participating system.

4. The method of processing clinical trial data based on IRT system according to claim 3, characterized in that, The state value is a value mapped from the state of the transaction participating in the system: if its state is ready, then ; If rolled back, then ; If no response is received or if in the initial state, then .

5. The method of processing clinical trial data based on IRT system according to claim 1, characterized in that, the calculation formula of the heuristic commit failure cost and the heuristic rollback failure cost is: ； ； wherein is the heuristic commit failure cost; is the heuristic rollback failure cost; is the local transaction urgency; and is the asymmetric risk factor, , are positive real numbers, and is greater than .

6. The method of processing clinical trial data based on IRT system according to claim 1, characterized in that, the calculation formula of the probability that the global transaction should finally be committed and rolled back is: ； ； In the formula, is the probability that the global transaction will eventually commit; is the probability that the global transaction will eventually roll back; is the consensus confidence.

7. The method of processing clinical trial data based on IRT system according to claim 1, characterized in that, the expected risk cost of taking the commit action is equal to the probability that the global transaction should finally be rolled back multiplied by the heuristic commit failure cost, and the expected risk cost of taking the rollback action is equal to the probability that the global transaction should finally be committed multiplied by the heuristic rollback failure cost.

8. The method of processing clinical trial data based on IRT system according to claim 1, characterized in that, the self-decision process further comprises the following steps: recording the final decision and its decision basis in a local tamper-proof audit log, the decision basis at least comprising the local transaction urgency, the consensus confidence, the heuristic commit failure cost, the heuristic rollback failure cost, the expected risk cost of taking the commit action, and the expected risk cost of taking the rollback action.

9. The method of processing clinical trial data based on IRT system according to claim 1, characterized in that, the value range of the business weight is (0, 1], the value range of the system weight is (0, 1], and the value range of the timeout threshold is [5, 30] seconds.

10. An IRT system-based clinical trial data processing system, characterized by, The method comprises the following steps: a processor and a memory, the memory storing computer program instructions which, when executed by the processor, implement a clinical trial data processing method based on an IRT system according to any one of claims 1-9.

Images