A method and system for automatic optimization of clinical trial plans

Abstract

The present application relates to the technical field of scheme optimization, and discloses a method and system for automatic optimization of clinical trial schemes, comprising: constructing a scheme map based on pre-acquired clinical trial data by analyzing the correlation between test steps and corresponding parameters; reordering the structure of the scheme map by analyzing the node priorities of the scheme map to obtain an updated scheme map; performing a path search in the updated scheme map in response to clinical trial requirements input by a user to obtain a scheme generation path and a scheme verification path; generating and optimizing a clinical trial scheme based on the scheme generation path and the scheme verification path through a preset dual-channel scheme optimization model to obtain an optimized clinical trial scheme; the present application can comprehensively consider the correlation between test steps, quickly search for core steps that meet user needs, generate and optimize corresponding clinical trial schemes, and improve the effectiveness of clinical trial schemes.

Description

Technical Field

[0001] The present invention relates to the technical field of protocol optimization, and in particular to a method and system for automatically optimizing a clinical trial protocol. Background Art

[0002] Clinical trials are a critical step in transforming medical innovations, such as new drug development and medical device validation. The quality of clinical trial protocols directly determines their success or failure. Traditional clinical trial protocol design relies on the personal experience and knowledge of medical experts. This leads to inefficient protocol development and the risk of subjective judgment bias, making it unable to meet the current demands for accurate and timely protocols in clinical trials.

[0003] The existing technology has the following problems: it uses a single experimental knowledge retrieval and matching method, which cannot combine the correlation between data, and the constructed experimental plan has poor relevance and feasibility; it is impossible to quickly search for the corresponding relevant experimental steps when formulating the experimental plan, resulting in low efficiency in plan formulation; when verifying the plan, a single verification standard is used for inspection and verification, which is inefficient and prone to subjective judgment deviations, and cannot guarantee the compliance and effectiveness of the experimental plan; in order to solve at least one of the above problems, the present application proposes a method and system for automatic optimization of clinical trial plans. Summary of the Invention

[0004] In response to the shortcomings of the prior art, the main purpose of the present invention is to provide a method and system for automatically optimizing clinical trial protocols, which can effectively solve the problems in the background technology. The specific technical solutions of the present invention are as follows:

[0005] A method for automatically optimizing a clinical trial protocol, comprising:

[0006] Based on pre-acquired clinical trial data, a protocol map is constructed by analyzing the correlation between the trial steps and the corresponding parameters;

[0007] By analyzing the node priorities of the solution map, the structure of the solution map is reordered to obtain an updated solution map;

[0008] In response to the clinical trial requirements input by the user, a path search is performed in the updated protocol map to obtain a protocol generation path and a protocol verification path;

[0009] Based on the protocol generation path and the protocol verification path, a clinical trial protocol is generated and optimized through a preset dual-channel protocol optimization model to obtain an optimized clinical trial protocol, so as to automatically optimize the clinical trial protocol. In the dual-channel protocol optimization model, the first channel generates a corresponding initial clinical trial protocol according to the protocol generation path, and the second channel verifies and optimizes the initial clinical trial protocol according to the protocol verification path.

[0010] Specifically, the scheme map is constructed by analyzing the correlation between the test steps and the corresponding parameters based on the pre-acquired clinical trial data, including:

[0011] Identify trial step nodes and corresponding parameter nodes from pre-acquired clinical trial data;

[0012] Analyze the association relationship between each test step node and the corresponding parameter node to obtain the corresponding node association relationship;

[0013] According to the node association relationship, connections are established between corresponding test step nodes and parameter nodes to construct a solution map.

[0014] Specifically, by analyzing the node priorities of the solution map, the structure of the solution map is reordered to obtain an updated solution map, including:

[0015] Analyze the influence relationship between each node in the scheme map and the effect of the experimental scheme to obtain the node priority of each node, wherein the nodes include experimental step nodes and parameter nodes;

[0016] The structure of the solution map is reordered according to the node priorities to obtain an updated solution map.

[0017] Specifically, the structure of the solution map is reordered according to the node priorities to obtain an updated solution map, including:

[0018] Sort the test step nodes and parameter nodes according to the node priority from high to low, and obtain the test step node priority order and the corresponding parameter node priority order;

[0019] According to the priority order of the test step nodes, the test step nodes whose order is before the preset first order threshold are used as core test step nodes;

[0020] According to the parameter node priority order, the parameter node whose order is before the preset second order threshold is used as the core parameter node of the corresponding test step node;

[0021] Arrange the core test step nodes from the center of the graph outward according to the priority order of the corresponding test step nodes from high to low to obtain a core step area;

[0022] Arrange the core parameter nodes of each test step node from the center of the test step node outward according to the corresponding parameter node priority order from high to low to obtain the core parameter area;

[0023] In combination with the core step area and the core parameter area, the structure of the solution map is updated to obtain an updated solution map.

[0024] Specifically, in response to the clinical trial requirements input by the user, a path search is performed in the updated protocol map to obtain a protocol generation path and a protocol verification path, including:

[0025] According to the clinical trial requirements input by the user, the corresponding required test step node set is extracted;

[0026] According to the required test step node set, a corresponding solution generation path is searched in the update solution map;

[0027] Combined with the solution generation path, the corresponding solution verification path is matched in the update solution map.

[0028] Specifically, according to the required test step node set, a corresponding solution generation path is searched in the update solution map, including:

[0029] Each node in the requirement test step node set is used as the starting node;

[0030] According to the starting node, in the core step area of ​​the update solution map, according to the connection relationship between the nodes, a first search path set is searched;

[0031] According to the first search path set, searching in a non-core step area of ​​the update solution map to obtain a second search path set;

[0032] By analyzing the similarities between the paths, the paths in the first search path set and the second search path set are merged to obtain a first candidate path set and a second candidate path set;

[0033] Combining the paths in the first candidate path set with the paths in the second candidate path set to obtain a candidate solution generation path set;

[0034] By calculating the correlation between each path in the candidate solution generation path set and the experimental solution effect, the path with the highest correlation is selected as the solution generation path.

[0035] Specifically, in combination with the solution generation path, the corresponding solution verification path is matched in the update solution map, including:

[0036] According to each path node in the solution generation path, by expanding the associated nodes, the corresponding experimental knowledge associated node set is matched in the updated solution graph;

[0037] By analyzing the correlation between each node in the test knowledge associated node set and the corresponding path node, the nodes having the correlation greater than a preset third threshold are screened out as the candidate test knowledge associated node set;

[0038] Connecting the nodes in the candidate test knowledge associated node set according to the connection relationship between the nodes in the update solution graph to obtain a candidate solution verification path set;

[0039] By analyzing the correlation between each path in the candidate solution verification path set and the effect of the experimental solution, the path with the highest correlation is selected as the solution verification path.

[0040] Specifically, based on the protocol generation path and the protocol verification path, a clinical trial protocol is generated and optimized through a preset dual-channel protocol optimization model to obtain an optimized clinical trial protocol, thereby automatically optimizing the clinical trial protocol, including:

[0041] According to the protocol generation path, the initial clinical trial protocol is generated through the first channel model in the preset dual-channel protocol optimization model;

[0042] According to the protocol verification path, the initial clinical trial protocol is verified and optimized through the second channel model in the preset dual-channel protocol optimization model to obtain an optimized clinical trial protocol, so as to automatically optimize the clinical trial protocol.

[0043] Specifically, according to the protocol verification path, the initial clinical trial protocol is verified and optimized by the second channel model in the preset dual-channel protocol optimization model to obtain an optimized clinical trial protocol, so as to automatically optimize the clinical trial protocol, including:

[0044] According to the protocol verification path, the initial clinical trial protocol is verified through the second channel model in the preset dual-channel protocol optimization model to obtain the protocol verification results;

[0045] According to the solution verification result, searching for path nodes that need to be optimized in the solution verification path to obtain a set of path nodes to be optimized;

[0046] Based on the update solution graph, searching for the optimization node with the highest correlation with each node in the set of path nodes to be optimized;

[0047] The initial clinical trial plan is optimized according to the optimization nodes to obtain an optimized clinical trial plan, so as to automatically optimize the clinical trial plan.

[0048] A clinical trial protocol automatic optimization system, used to implement the clinical trial protocol automatic optimization method, comprising:

[0049] The protocol map construction module constructs a protocol map based on pre-acquired clinical trial data by analyzing the correlation between trial steps and corresponding parameters;

[0050] A solution map updating module reorders the structure of the solution map by analyzing the node priorities of the solution map to obtain an updated solution map;

[0051] A path search module, in response to a clinical trial requirement input by a user, performs a path search in the updated protocol map to obtain a protocol generation path and a protocol verification path;

[0052] The clinical trial protocol optimization module generates and optimizes the clinical trial protocol based on the protocol generation path and the protocol verification path through a preset dual-channel protocol optimization model to obtain an optimized clinical trial protocol, so as to automatically optimize the clinical trial protocol. In the dual-channel protocol optimization model, the first channel generates a corresponding initial clinical trial protocol according to the protocol generation path, and the second channel verifies and optimizes the initial clinical trial protocol according to the protocol verification path.

[0053] Compared with the prior art, this application has the following beneficial effects:

[0054] This application constructs a scheme map by analyzing the correlation between experimental steps and parameters, analyzes the node priority based on the impact of each node on the effect of the experimental scheme, and reorders the map structure to obtain an updated scheme map. In the updated scheme map, a scheme generation path and a scheme verification path that meet user needs are searched, and based on the scheme generation path and the scheme verification path, a clinical trial scheme is generated and optimized to automatically generate and optimize the clinical trial scheme. The correlation between the experimental steps can be comprehensively considered, and the core steps that meet user needs can be quickly searched, and the corresponding clinical trial scheme is generated and optimized to improve the effect of the clinical trial scheme. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 This is a workflow diagram of a method for automatically optimizing a clinical trial protocol in Example 1 of the present invention;

[0056] Figure 2 Schematic diagram of the scheme map in Example 1 of the present invention;

[0057] Figure 3 Schematic diagram of the update scheme map in Example 1 of the present invention;

[0058] Figure 4 A schematic diagram of a path search process for generating a solution in Example 1 of the present invention;

[0059] Figure 5This is a schematic diagram of the solution verification path matching process in Example 1 of the present invention;

[0060] Figure 6 Schematic diagram of the structure of a clinical trial plan automatic optimization system in Example 2 of the present invention. DETAILED DESCRIPTION

[0061] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0062] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0063] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0064] Example 1

[0065] This embodiment provides a method for automatically optimizing a clinical trial protocol. Figure 1 As shown, the method for automatic optimization of a clinical trial protocol comprises:

[0066] S101. Construct a protocol map by analyzing the relationship between the trial steps and corresponding parameters based on pre-acquired clinical trial data;

[0067] S102: Reorder the structure of the solution map by analyzing the node priorities of the solution map to obtain an updated solution map;

[0068] S103. In response to the clinical trial requirements input by the user, a path search is performed in the updated protocol map to obtain a protocol generation path and a protocol verification path;

[0069] S104. Based on the protocol generation path and the protocol verification path, a clinical trial protocol is generated and optimized through a preset dual-channel protocol optimization model to obtain an optimized clinical trial protocol, so as to automatically optimize the clinical trial protocol, wherein the first channel in the dual-channel protocol optimization model generates a corresponding initial clinical trial protocol according to the protocol generation path, and the second channel verifies and optimizes the initial clinical trial protocol according to the protocol verification path.

[0070] The current method for formulating clinical trial plans is based on experimental experience and manual design. This embodiment analyzes the correlation between clinical trial steps and constructs a plan map. Based on the plan map, the plan generation path and the plan verification path are searched, and the corresponding clinical trial plan is generated and optimized. Compared with the traditional method of formulating clinical trial plans, the embodiment of the present application combines map construction, map update, plan formulation and plan update, which can comprehensively consider all aspects of clinical trials. The generated plan is more feasible and effective, which improves the success rate of clinical trials and the reliability of research results.

[0071] In this embodiment, first, the pre-acquired clinical trial data is analyzed to obtain the association relationship between the test steps and the corresponding parameters, and connections are established between the corresponding nodes based on the association relationship to construct a scheme map; the clinical trial contains multiple test steps, each test step has corresponding parameters, and the clinical trial data is obtained from historical clinical trial reports, research documents, and databases. The clinical trial data includes a detailed description of the test steps, parameters of each step, the value range of the parameters, the sequence of each step, and other information; each test step and parameter in the clinical trial scheme is extracted as a test step node and a parameter node, respectively, the association relationship between the test steps and the parameters is analyzed, and connections are made between the nodes with the association relationship to construct a scheme map; by analyzing the relationship between the test steps and the parameters, the corresponding test scheme can be quickly formulated according to the scheme map, thereby improving the efficiency and scientificity of the scheme formulation.

[0072] For example, for clinical trials of new antihypertensive drugs, the trial steps include "determining the trial period", "screening subjects", "drug administration", "monitoring blood pressure changes" and other steps; related parameters include "trial period duration", "subject age range", "daily dose of drug", "blood pressure measurement frequency" and other parameters; analysis shows that "determining the trial period" will affect the frequency of "monitoring blood pressure changes"; "screening subjects" depends on parameters such as "subject age range"; these steps and parameters are used as nodes, and the corresponding association relationships are used as edges to construct a solution map.

[0073] Specifically, after constructing the solution map, the node priorities of the solution map are analyzed, and the structure of the solution map is reordered according to the node priorities to obtain an updated solution map; through priority analysis, taking into account the different importance and impact of different nodes on clinical trials, the criticality of each node in the trial process, the sensitivity of parameter nodes, the execution frequency of nodes and other influencing factors are analyzed to determine the node priority, and according to the calculated node priority values, the nodes in the solution map are reordered, and the nodes with high priority are adjusted to a position close to the center of the map, and the edges between the adjusted nodes are reconnected to obtain an updated solution map; by calculating the node priority to structurally update the solution map, key nodes and key paths can be quickly searched when formulating clinical trial plans, so as to quickly find solution paths that meet user needs and improve the efficiency of solution formulation and optimization.

[0074] After the user inputs the clinical trial requirements, the clinical trial requirements are analyzed to obtain the user's test requirements in terms of test objectives, conditions, restrictions, etc. Based on the analyzed test requirements, a path search is performed in the updated plan map, and paths that can meet the requirements are searched in the map, including plan generation paths and plan verification paths; among them, the plan generation path is used to generate clinical trial plans, and the plan verification path verifies and optimizes the generated clinical trial plans; combining user needs and the updated plan map, it is possible to quickly and accurately search for paths that meet user needs in the map according to user needs, provide a clear direction for generating and optimizing clinical trial plans, and improve the efficiency and quality of clinical trial plan formulation.

[0075] Specifically, according to the searched solution generation path and solution verification path, the clinical trial solution is generated and optimized through the preset dual-channel solution optimization model to obtain the optimized clinical trial solution; the dual-channel solution optimization model includes two channels. The first channel combines the experimental steps and parameters represented by the nodes and edges on the path according to the solution generation path according to the corresponding solution generation rules to obtain the initial clinical trial solution; the second channel verifies the rationality and feasibility of the clinical trial solution according to the solution verification path, and optimizes the solution according to the verification results to finally obtain the optimized clinical trial solution; the clinical trial solution and solution optimization are generated respectively through the dual-channel model to ensure that the generated clinical trial solution can meet user needs and improve the feasibility and effectiveness of the solution; at the same time, the automatic optimization of the clinical trial solution is realized to improve the efficiency of the clinical trial solution optimization.

[0076] This application constructs a scheme map by analyzing the correlation between experimental steps and parameters, analyzes the node priority based on the impact of each node on the effect of the experimental scheme, and reorders the map structure to obtain an updated scheme map. In the updated scheme map, a scheme generation path and a scheme verification path that meet user needs are searched, and based on the scheme generation path and the scheme verification path, a clinical trial scheme is generated and optimized to automatically generate and optimize the clinical trial scheme. The correlation between the experimental steps can be comprehensively considered, and the core steps that meet user needs can be quickly searched, and the corresponding clinical trial scheme is generated and optimized to improve the effect of the clinical trial scheme.

[0077] Furthermore, the method of constructing a protocol map based on pre-acquired clinical trial data by analyzing the correlation between the trial steps and corresponding parameters includes:

[0078] S201. Identify test step nodes and corresponding parameter nodes from pre-acquired clinical trial data;

[0079] S202, analyzing the association relationship between each test step node and the corresponding parameter node to obtain the corresponding node association relationship;

[0080] S203: Establish connections between corresponding test step nodes and parameter nodes according to the node association relationship to construct a solution map.

[0081] In this embodiment, first, the pre-acquired clinical trial data is cleaned and pre-processed by removing duplicate, erroneous or irrelevant data records. Based on the pre-processed clinical trial data, the operation steps in the implementation of the clinical trial and the key parameters of each operation step are identified respectively, and the operation steps and key parameters are used as test step nodes and corresponding parameter nodes; the text content in the data is analyzed by natural language processing technology, and operationally meaningful statements are extracted from the test process as test step nodes; the parameters of each test step node are quantified to obtain the corresponding parameter nodes; by identifying the test step nodes and parameter nodes, basic elements are provided for constructing the scheme map, thereby ensuring the accuracy and effectiveness of the map construction.

[0082] Secondly, after identifying the test step nodes and the corresponding parameter nodes, the association relationship between each test step node and the corresponding parameter node is analyzed to obtain the corresponding node association relationship; by analyzing the association relationship between the nodes, taking into account the mutual influence and mutual restriction relationship between the test step node and the corresponding parameter node, the association relationship between each test step node and the parameter node is analyzed according to the preset association relationship analysis rules. When the parameter directly determines the specific operation method or condition of the test step, there is a determination relationship between the parameter node and the test step node; when the execution of the test step causes the value or state of the parameter to change, there is an influence relationship; by determining the association relationship between the nodes, an accurate association basis is provided for the node connection in the scheme map, ensuring that the map can truly reflect the relationship between the various elements of the clinical trial, and improving the practicality and reliability of the map.

[0083] like Figure 2 As shown, combined with the identified test step nodes, parameter nodes and the association relationship between the nodes, connections are established between the corresponding nodes, connections are made between the test step nodes and parameter nodes with association relationships, and at the same time, connections are made between the test step nodes in the order of the test steps to construct a scheme map; by constructing the scheme map, the association between the steps and parameters can be combined when formulating the clinical trial plan to formulate a highly feasible test plan, and the corresponding associated nodes can be quickly searched according to the scheme map to formulate the corresponding experimental plan, thereby improving the efficiency of clinical trial plan formulation.

[0084] Furthermore, by analyzing the node priorities of the solution map, the structure of the solution map is reordered to obtain an updated solution map, including:

[0085] S301, analyzing the influence relationship between each node in the scheme map and the experimental scheme effect to obtain the node priority of each node, wherein the nodes include experimental step nodes and parameter nodes;

[0086] S302: Reorder the structure of the solution map according to the node priorities to obtain an updated solution map.

[0087] After constructing the protocol map, this embodiment reorders the map structure by analyzing the node priorities and updates the protocol map, which can make the key information in the map more prominent. By combining key nodes and key paths to formulate clinical trial plans, the quality and reliability of clinical trial plans can be improved and the probability of trial success can be increased.

[0088] In this embodiment, the influence relationship between each node in the solution map and the effect of the test solution is analyzed. By analyzing the importance of each node in achieving the test goal, ensuring the test quality and reliability, the priority of each node is determined. By analyzing the node priority, when formulating the test solution, the solution can be formulated around the nodes with high priority, which helps users screen key factors and improve the pertinence and effectiveness of the solution formulation.

[0089] For example, based on the type and objectives of the clinical trial, multiple evaluation indicators are formulated, including the impact on the accuracy of the trial results, the impact on the trial safety, the impact on the trial cost, the impact on the trial cycle, etc. For each evaluation indicator, a corresponding quantitative evaluation standard is formulated, and each indicator of each node is scored according to the corresponding quantitative evaluation standard. The scoring results of each indicator are added together to obtain the priority value of each node.

[0090] Specifically, based on the node priorities obtained from the analysis, the structure of the scheme map is reordered, and nodes with high priority are placed in more core and prominent positions of the map, and nodes with low priority are placed at the edge. After adjusting the node positions, the connection relationship between the nodes is adjusted. By updating the structure of the scheme map, the structure of the scheme map can highlight the core nodes more. When formulating clinical trial plans, it can quickly locate the key nodes and paths that have a great impact on the effectiveness of the trial plans, thereby improving the efficiency of plan formulation and optimization.

[0091] Furthermore, the structure of the solution map is reordered according to the node priorities to obtain an updated solution map, including:

[0092] S401, sorting the test step nodes and parameter nodes from high to low according to the node priority, to obtain the test step node priority sequence and the corresponding parameter node priority sequence;

[0093] S402: According to the priority order of the test step nodes, the test step nodes whose order is before a preset first order threshold are used as core test step nodes;

[0094] S403: according to the priority order of the parameter nodes, taking the parameter nodes whose order is before the preset second order threshold as the core parameter nodes of the corresponding test step nodes;

[0095] S404, arranging the core test step nodes from the center of the graph outward according to the corresponding test step node priority order from high to low to obtain a core step area;

[0096] S405, arranging the core parameter nodes of each test step node from the center of the test step node outward according to the corresponding parameter node priority order from high to low to obtain a core parameter area;

[0097] S406 , combining the core step area and the core parameter area, updating the structure of the solution map to obtain an updated solution map.

[0098] In this embodiment, based on the calculated node priority values, the test step nodes are sorted from high to low priority, and for the parameter nodes corresponding to each test step node, the parameter nodes are sorted from high to low priority, and the test step node priority order and the parameter node priority order corresponding to each test step node priority order are obtained respectively; by sorting the test step nodes and the parameter nodes respectively, considering that the test step nodes and the parameter nodes play different roles in clinical trials, after obtaining the node order, the key nodes and secondary nodes in different types of nodes can be distinguished, which provides a basis for screening core nodes and adjusting the graph structure, reduces the complexity of the node screening process, and improves the efficiency and accuracy of the graph structure reordering.

[0099] Specifically, according to the priority order of the test step nodes, the test step nodes that are before the preset first order threshold are regarded as core test step nodes. According to the type and complexity of the clinical trial, the corresponding first order threshold is set. For example, for simple clinical trials, the first order threshold is set to the top 30%; for complex multi-center clinical trials, the first order threshold is set to the top 20%. According to the set first order threshold, the corresponding test step nodes are screened out to obtain the core test step nodes. By screening the core test step nodes, a basis is provided for updating the atlas structure, so that the updated atlas can quickly locate the key test steps in the clinical trial, improving the efficiency of plan formulation.

[0100] At the same time, for the parameter nodes of each test step node, according to the parameter node priority order, the parameter nodes before the preset second order threshold are used as the core parameter nodes of the corresponding test step node; the second order threshold can be set according to the data accuracy requirements of the clinical trial. For example, in clinical trials with high parameter accuracy requirements, the second order threshold is set to the top 15%; in general clinical trials, it is set to the top 30%. The parameter nodes are screened according to the set second order threshold to obtain the core parameter nodes; by screening the core parameter nodes corresponding to each test step node, the core parameters closely related to each test step can be clearly identified. During the design and implementation of the test plan, the focus can be on the setting, monitoring and control of the core parameters, reducing the poor test results and error risks caused by unreasonable parameters.

[0101] Specifically, the screened core test step nodes are arranged from high to low according to the priority order of the corresponding test step nodes from the center of the map to the outside, so as to obtain the core step area; the core test step node with the highest priority is placed at the center of the map, and the nodes with the second highest priority are distributed outward around the center node, and so on, the positions of the test step nodes are updated. By arranging the test step nodes from the center position to the outside, it is convenient to establish an association between the core test step node at the center position and the surrounding nodes, thereby strengthening the important position of the core test step node in the map. When formulating a plan, the corresponding plan generation path node is searched for in the core step area first, and then the corresponding plan generation path node is searched for in the non-core step area according to the connection relationship between the test step nodes between the core step area and the non-core step area. By searching in the core step area first, the key test steps can be quickly located, and the efficiency of grasping the core steps in the test process can be improved. Secondly, searching in the non-core step area can ensure the integrity of the plan generation path.

[0102] At the same time, according to the core parameter nodes of each test step node, the corresponding parameter nodes are arranged from the center of the test step node outward in order of priority from high to low to obtain the core parameter area; the parameter nodes are distributed outward around the center of the test step node according to the priority from high to low, and the core parameter node with the highest priority is placed at a position closer to the center of the test step node, and the nodes with lower priority are expanded outward in turn; by updating the position of the core parameter nodes, the close relationship between the core test steps and the core parameters can be paid attention to, and path search and parameter matching are given priority in the core step area when formulating the plan. By giving priority to setting and controlling the core parameters of each test step, the effect of the test plan is improved, and the search efficiency of the plan generation path is improved.

[0103] like Figure 3 , according to the core step area and core parameter area after rearrangement, the structure of the solution map is updated to obtain an updated solution map; for the nodes after the updated position, the connection edges between the nodes are adjusted according to the association relationship between the corresponding nodes, and the updated solution map is comprehensively checked to ensure that the relationship between the nodes and the corresponding connection edges is accurate; by updating the solution map, the updated solution map can highlight the core test steps and core parameters, and when formulating the test plan, the key links and parameters that need to be improved can be quickly identified, thereby improving the pertinence and efficiency of the solution optimization.

[0104] Furthermore, in response to the clinical trial requirements input by the user, a path search is performed in the updated protocol map to obtain a protocol generation path and a protocol verification path, including:

[0105] S501. Extracting a corresponding set of required test step nodes based on the clinical trial requirements input by the user;

[0106] S502: Searching for a corresponding solution generation path in the update solution map according to the required test step node set;

[0107] S503: Combine the solution generation path and match the corresponding solution verification path in the update solution map.

[0108] In this embodiment, based on the clinical trial requirements input by the user, the requirements for the trial objectives, processes, etc. in the clinical trial requirements are analyzed, and the test steps involved in the requirements are extracted to form a requirement test step node set. By extracting the requirement test steps, the user's vague requirements are converted into a clear test step node set. Accurate path search results can be obtained during path search and node matching, thereby improving the accuracy and pertinence of the path search and ensuring that the search results can meet the user's actual needs.

[0109] For example, the user input requirement is "to conduct a drug safety trial for patients with hypertension, which is required to be completed within 6 months, and to screen patients with mild to moderate hypertension, and to evaluate drug safety by regularly measuring blood pressure and testing liver and kidney function"; after parsing the requirement, keywords such as "patient screening", "drug administration", "regular blood pressure measurement", "liver and kidney function testing", and "safety assessment" are extracted, and the keywords are matched with the test step nodes to obtain the required test step node set: patient screening, drug administration, regular blood pressure measurement, liver and kidney function testing, and safety assessment.

[0110] like Figure 4 According to the extracted required test step node set, the test step node that matches each node in the required test step node set is searched in the update solution map, and the corresponding path is searched according to the connection relationship between the nodes to obtain the solution generation path composed of test steps and related parameters that can meet user needs; by matching and searching the solution generation path in the update solution map, the optimal solution generation path that can meet user needs can be quickly searched, thereby improving the efficiency of solution formulation.

[0111] like Figure 5After searching for the solution generation path, the solution generation path is combined with the solution generation path to search for relevant nodes and edges related to the solution generation path in the update solution graph for verifying the solution, forming a solution verification path; the feasibility, effectiveness, and compliance of the clinical trial solution can be verified through the solution verification path; by matching the corresponding solution verification path, the clinical trial solution can be verified, and problems and risks in the solution can be optimized in advance to avoid trial failure or adverse consequences due to solution defects, thereby improving the success rate and reliability of the trial.

[0112] Furthermore, according to the required test step node set, a corresponding solution generation path is searched in the update solution map, including:

[0113] S601, taking each node in the requirement test step node set as a starting node;

[0114] S602: Based on the starting node, in the core step area of ​​the update solution graph, search and obtain a first search path set according to the connection relationship between nodes;

[0115] S603: Searching for a second search path set in a non-core step region of the update solution graph based on the first search path set;

[0116] S604: merging the paths in the first search path set and the second search path set respectively by analyzing the similarity between the paths to obtain a first candidate path set and a second candidate path set;

[0117] S605: Combine the paths in the first candidate path set with the paths in the second candidate path set to obtain a candidate solution generation path set;

[0118] S606: Calculate the correlation between each path in the candidate solution generation path set and the experimental solution effect, and select the path with the highest correlation as the solution generation path.

[0119] In this embodiment, each node in the required test step node set is used as the starting node, and the corresponding path is searched from each starting node. This can fully explore the paths related to these key steps in the update solution map, avoid missing the solution generation path, and search for all test step combinations that meet user needs from multiple angles, providing a reference for screening the optimal solution generation path.

[0120] Specifically, according to the starting node, taking each starting node as the starting point, searching is performed in the core step area according to the connection relationship between nodes to search for a first search path set related to the core step, and according to the starting node, searching is performed in the core step area for core step nodes related to the starting node to search for all related core step nodes, and combining the connection relationship between the corresponding core step nodes to obtain the corresponding first search path, and combining the first search paths to obtain the first search path set; by searching in the core step area first, the core test step nodes and corresponding important paths that are highly correlated with the effect of the test plan can be quickly located, thereby reducing invalid searches in non-critical areas and improving search efficiency; at the same time, the path constructed based on the core step can form a higher quality clinical trial plan when generating the corresponding test plan.

[0121] After obtaining the first search path set by searching in the core step area, the search is continued in the non-core step area based on each path node in the first search path set to supplement the test steps and obtain the second search path set; by continuing to search in the non-core step area, the path information of the entire test plan process can be supplemented, including the auxiliary links, preparatory work and subsequent processing of the test, etc., and the paths in the non-core step area are combined with the paths in the core step area to obtain more comprehensive and complete solution generation path candidates, thereby ensuring the integrity and feasibility of the solution.

[0122] Specifically, based on the similarity between the paths, the paths in the first search path set and the second search path set are merged respectively to obtain the first candidate path set and the second candidate path set; similar paths are merged to reduce redundant paths; the merged candidate path set retains representative paths; whether the paths are similar is judged based on factors such as the proportion of the number of identical nodes in the paths, the order consistency of the nodes, and whether the test step nodes are the same; for example, when the proportion of the number of identical nodes in the two paths accounts for more than 70% of the total number of path nodes, and the test step nodes are the same, the two paths are considered similar and the similar paths are merged; similar paths are searched for in the first search path set and the second search path set respectively and merged to obtain the first candidate path set and the second candidate path set; by merging similar paths, redundant information in the path set can be reduced, representative paths can be retained, and it is helpful to more accurately screen out solution generation paths that meet user needs.

[0123] After obtaining the first candidate path set and the second candidate path set, the paths in the first candidate path set and the paths in the second candidate path set are combined to obtain paths of all combinations. The paths in the first candidate path set are obtained by searching in the core step area based on the starting node, reflecting the corresponding core steps; the paths in the second candidate path set are obtained by searching in the non-core step area, reflecting the corresponding non-core steps; by combining the first candidate path and the second candidate path, a complete path of all combination results can be obtained, and a candidate solution generation path set is obtained, which provides a complete solution selection for the screening of the solution generation path, and the optimal solution generation path is screened out.

[0124] Specifically, for each path in the candidate solution generation path set, the correlation between each path and the effect of the test solution is calculated separately, and the performance of each path in achieving the test objectives, ensuring the test quality, and controlling the test costs is analyzed, and the path with the highest correlation is selected as the solution generation path; according to the goals and characteristics of the clinical trial, evaluation indicators for the correlation between the path and the effect of the test solution are formulated, including the integrity of the key test steps in the path, the rationality of the parameter setting, the impact on the accuracy of the test results, etc., and corresponding scoring standards are formulated for each indicator. The score of each path in each indicator is obtained according to the scoring standard, and the indicator scores of each path are added together to obtain the correlation score of each path, and the path with the highest correlation score is selected as the solution generation path; by screening out the path that has the greatest impact on the test results, user needs can be met to the greatest extent, the quality and feasibility of the clinical trial plan can be improved, and the probability of trial success can be increased.

[0125] Furthermore, in combination with the solution generation path, a corresponding solution verification path is matched in the update solution map, including:

[0126] S701, according to each path node in the solution generation path, by expanding the associated nodes, respectively matching the corresponding experimental knowledge associated node set in the updated solution map;

[0127] S702: Analyze the correlation between each node in the test knowledge associated node set and the corresponding path node, and select nodes whose correlation is greater than a preset third threshold as candidate test knowledge associated node sets;

[0128] S703: Connect the nodes in the candidate test knowledge associated node set according to the connection relationship between the nodes in the update solution graph to obtain a candidate solution verification path set;

[0129] S704: Analyze the correlation between each path in the candidate solution verification path set and the experimental solution effect, and select the path with the highest correlation as the solution verification path.

[0130] In this embodiment, according to each path node in the solution generation path, matching is performed through extended association in the updated solution map, and all potential knowledge nodes associated with the path node are screened out to form a set of experimental knowledge associated nodes; for each path node, in the updated solution map, according to the connection relationship between the nodes, the surrounding nodes connected to it are searched, and all the explored nodes are recorded to form a set of experimental knowledge associated nodes corresponding to the path node. By searching for experimental knowledge nodes related to the solution generation path node, the clinical trial solution generated according to the solution generation path can be verified, and potential problems and risks in the solution can be identified.

[0131] Specifically, the correlation between each node in the experimental knowledge association node set and the corresponding path node is analyzed, and the nodes with a correlation greater than a preset third threshold are screened out as the candidate experimental knowledge association node set; according to the characteristics and objectives of the clinical trial, indicators for evaluating node correlation are formulated, including the necessity of the knowledge represented by the node for the path node operation, the degree of influence on the accuracy of the test results, etc., and corresponding scoring standards are formulated for each indicator. According to the scoring standards, the score of each indicator in the corresponding indicator is calculated respectively, and the score of each indicator is added to obtain the correlation score between each node in the experimental knowledge association node set and the corresponding path node; according to the type and stability requirements of the clinical trial, a third threshold is set, and the node combination with a correlation score greater than the third threshold is taken as the candidate experimental knowledge association node set; by screening out nodes that are of great value to the scheme verification and removing redundant information, the scheme verification process is made more efficient, ensuring that the nodes in the candidate experimental knowledge association node set are closely related to the scheme generation path, and improving the effectiveness of the scheme verification.

[0132] Specifically, according to the connection relationship between the nodes in the update scheme graph, the nodes with associated relationships in the candidate test knowledge associated node set are connected to obtain a set of candidate scheme verification paths; by connecting the nodes in the candidate test knowledge associated node set, multiple scheme verification paths are quickly generated, providing options for selecting the optimal verification path, ensuring the comprehensiveness and flexibility of scheme verification.

[0133] Similarly, the correlation between each path in the candidate solution verification path set and the effect of the experimental solution is analyzed, and the path with the highest correlation is selected as the solution verification path; the performance of each path in achieving the experimental objectives, ensuring the experimental quality, and controlling the experimental costs is analyzed, and the path with the highest correlation is selected as the solution generation path; according to the objectives and characteristics of the clinical trial, evaluation indicators for the correlation between the path and the effect of the experimental solution are formulated, including the integrity of the key experimental steps in the path, the rationality of the parameter setting, the impact on the accuracy of the experimental results, etc., and corresponding scoring standards are formulated for each indicator. The score of each path in each indicator is obtained according to the scoring standard, and the indicator scores of each path are added together to obtain the correlation score of each path, and the path with the highest correlation score is selected as the solution verification path; by screening the solution verification paths, it can be ensured that the selected solution verification path can ensure the quality and reliability of the clinical trial plan and increase the probability of trial success.

[0134] Furthermore, based on the protocol generation path and the protocol verification path, a clinical trial protocol is generated and optimized through a preset dual-channel protocol optimization model to obtain an optimized clinical trial protocol, so as to automatically optimize the clinical trial protocol, including:

[0135] S801. Generate an initial clinical trial plan based on the plan generation path by using the first channel model in the preset dual-channel plan optimization model;

[0136] S802. According to the protocol verification path, the initial clinical trial protocol is verified and optimized through the second channel model in the preset dual-channel protocol optimization model to obtain an optimized clinical trial protocol, so as to automatically optimize the clinical trial protocol.

[0137] In this embodiment, according to the protocol generation path, the experimental steps and parameters in the protocol generation path are combined through the first channel model in the preset dual-channel protocol optimization model to construct an initial clinical trial protocol. The first channel model in this embodiment includes but is not limited to a sequence-to-sequence model. The sequence-to-sequence model is trained using a large number of existing clinical trial protocols and their corresponding protocol generation path data to obtain a pre-trained sequence-to-sequence model. The protocol generation path is input into the pre-trained sequence-to-sequence model, and the model will output an initial clinical trial protocol including a description of the trial objectives, subject recruitment criteria, drug administration regimen, testing arrangements, and efficacy evaluation methods. The first channel model can quickly convert the protocol generation path into a specific clinical trial protocol, and generate a corresponding protocol according to the logical order of the protocol generation path, thereby ensuring the integrity and coherence of the clinical trial protocol, making the connection between the trial steps more reasonable, and facilitating the smooth progress of the trial process.

[0138] Specifically, according to the scheme verification path, the initial clinical trial scheme is verified and optimized through the second channel model in the preset dual-channel scheme optimization model to obtain an optimized clinical trial scheme; the second channel model comprehensively tests and optimizes the initial clinical trial scheme based on the key verification elements in the scheme verification path; the second channel model includes machine learning models such as random forest models and decision tree models, and a large amount of clinical trial schemes and their verification result data are used to train the second channel model to obtain a pre-trained second channel model, and the initial clinical trial scheme and scheme verification path are input into the pre-trained second channel model, and the model outputs the verification results of the initial clinical trial scheme and the optimized clinical trial scheme; through comprehensive verification of the initial clinical trial scheme, potential problems in the scheme can be discovered and solved in a timely manner, the trial risk can be reduced, the probability of trial success can be increased, it can be ensured that the clinical trial scheme can achieve the expected trial goals, the reliability and credibility of the trial results can be improved, and high-quality scheme support can be provided for medical research and drug development.

[0139] Furthermore, according to the protocol verification path, the initial clinical trial protocol is verified and optimized by the second channel model in the preset dual-channel protocol optimization model to obtain an optimized clinical trial protocol, so as to automatically optimize the clinical trial protocol, including:

[0140] S901. According to the protocol verification path, the initial clinical trial protocol is verified using the second channel model in the preset dual-channel protocol optimization model to obtain a protocol verification result.

[0141] S902: Searching for path nodes that need to be optimized in the solution verification path according to the solution verification result to obtain a set of path nodes to be optimized;

[0142] S903: Based on the updated solution graph, searching for the optimization node with the highest correlation with each node in the set of nodes on the path to be optimized;

[0143] S904. Optimize the initial clinical trial plan according to the optimization node to obtain an optimized clinical trial plan, so as to automatically optimize the clinical trial plan.

[0144] In this embodiment, first, according to the protocol verification path, the initial clinical trial protocol is verified through the second channel model in the preset dual-channel protocol optimization model to obtain the protocol verification result; each node and the corresponding connection relationship in the protocol verification path are analyzed to determine the verification content and standard represented by each node, and each link of the initial clinical trial protocol is compared one by one with the corresponding requirements in the protocol verification path through the second channel model to check whether the protocol meets the requirements. During the comparison process, the consistency or non-compliance of the initial protocol with the protocol verification path requirements is recorded. For the part that meets the requirements, it is marked as passed the verification; for the part that does not meet the requirements, the existing problems and deviations are recorded to obtain the protocol verification result; the initial clinical trial protocol is comprehensively evaluated through the protocol verification path to ensure that the protocol can be tested in all key aspects and avoid potential risks caused by the lack of attention to some links.

[0145] Specifically, based on the obtained solution verification results, the parts of the initial solution that do not meet the solution verification path requirements are identified, and these parts are matched to the nodes in the solution verification path as path nodes that need to be optimized. The path nodes that need to be optimized are combined to obtain a set of path nodes to be optimized. By identifying the parts that need to be optimized in the initial solution and locating the corresponding path nodes, the path nodes corresponding to the key links with problems in the initial solution are accurately located, providing goals for the optimization process, avoiding blind optimization, and improving the efficiency and accuracy of the optimization work.

[0146] Specifically, according to each node in the set of path nodes to be optimized, the node with the highest correlation with the node in the set of path nodes to be optimized is searched in the update scheme map as the optimization node; according to the characteristics of the clinical trial and the nature of the path nodes to be optimized, indicators for evaluating node correlation are formulated, including the effectiveness of the path node content in solving the problem to be optimized, compatibility with existing schemes, feasibility of optimization, etc., and corresponding scoring criteria are formulated for each indicator. According to the scoring criteria, each node in the set of path nodes to be optimized is scored on each indicator, and the score of each indicator is added to obtain the correlation score of each node. The node with the highest correlation score is selected as the optimization node, and the corresponding path node to be optimized is optimized according to each optimization node; by matching each path node to be optimized with the corresponding optimization node with the highest correlation, high-quality optimization direction can be provided for the optimization of the experimental scheme, and the quality and feasibility of the optimized scheme can be improved.

[0147] Specifically, the initial clinical trial plan is optimized based on the selected optimization nodes, the content of the optimization nodes is added to the initial clinical trial plan, and the overall order of the plan is adjusted and optimized accordingly to obtain an optimized clinical trial plan; the content of the optimization nodes is added to the corresponding part of the initial clinical trial plan according to its association with the path node to be optimized. After the optimized content is added to the plan, the continuity and feasibility of the plan are verified to ensure that the logic between each test step is reasonable and that the newly added plan content does not cause logical contradictions or process confusion in the plan; at the same time, the optimized plan is verified to check the integrity and feasibility of the optimized plan, and ensure that the corresponding problems to be optimized are solved to obtain a complete optimized clinical trial plan. By optimizing the clinical trial plan through the optimization nodes, the corresponding problems in the initial clinical trial plan can be solved, the rationality and feasibility of the test plan can be improved, and the successful implementation of the clinical trial can be guaranteed.

[0148] Example 2

[0149] In this embodiment, if Figure 6 , provides a clinical trial protocol automatic optimization system for implementing the aforementioned clinical trial protocol automatic optimization method, comprising:

[0150] The protocol map construction module constructs a protocol map based on pre-acquired clinical trial data by analyzing the correlation between trial steps and corresponding parameters;

[0151] A solution map updating module reorders the structure of the solution map by analyzing the node priorities of the solution map to obtain an updated solution map;

[0152] A path search module, in response to a clinical trial requirement input by a user, performs a path search in the updated protocol map to obtain a protocol generation path and a protocol verification path;

[0153] The clinical trial protocol optimization module generates and optimizes the clinical trial protocol based on the protocol generation path and the protocol verification path through a preset dual-channel protocol optimization model to obtain an optimized clinical trial protocol, so as to automatically optimize the clinical trial protocol. In the dual-channel protocol optimization model, the first channel generates a corresponding initial clinical trial protocol according to the protocol generation path, and the second channel verifies and optimizes the initial clinical trial protocol according to the protocol verification path.

[0154] In this embodiment, the protocol map construction module analyzes the clinical trial data based on pre-acquired clinical trial data, extracts the test step nodes and parameter nodes in the test process, and constructs a protocol map by analyzing the association relationship between each test step node and the corresponding parameter node. By establishing a connection between the test step node and the parameter node based on the association relationship, a reference is provided for the generation and optimization of clinical trial protocols by constructing the protocol map. The protocol map update module calculates the priority of each node by analyzing the association relationship between each node (including the test step node and the parameter node) in the protocol map and the effect of the trial protocol, updates the structure of the protocol map based on the node priority, places the node with a high priority at the center of the map, and obtains an updated protocol map by adjusting the node position. By updating the protocol map, the core node and the corresponding core path can be quickly searched when generating the clinical trial protocol, and the clinical trial protocol generated based on the core path has a better trial effect.

[0155] Specifically, the path search module searches for a solution generation path that meets user needs in the updated solution map based on user needs, including key experimental steps and parameters that meet customer needs, and matches the corresponding solution verification path in combination with the solution generation path to verify and optimize the clinical trial solution. By searching the solution generation path and the solution verification path, an effective and feasible clinical trial solution can be generated while meeting user needs; the clinical trial solution optimization module is based on the solution generation path and the solution verification path searched by the path search module. In the preset dual-channel solution optimization model, the first channel integrates the experimental steps and parameters in the path according to the solution generation path to quickly generate an initial clinical trial solution, and the second channel verifies and optimizes the initial solution according to the solution verification path; through the collaborative work of the two channels, the clinical trial solution can be optimized in real time to obtain an optimized clinical trial solution, realize automatic optimization of the clinical trial solution, improve the feasibility and effectiveness of the clinical trial solution, and ensure that the clinical trial is carried out smoothly and achieves the expected goals.

[0156] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for automatic optimization of clinical trial protocols, characterized in that: include: Based on pre-acquired clinical trial data, a protocol map is constructed by analyzing the correlation between the trial steps and the corresponding parameters; Analyze the influence relationship between each node in the scheme map and the effect of the experimental scheme to obtain the node priority of each node, wherein the nodes include experimental step nodes and parameter nodes; Sort the test step nodes and parameter nodes according to the node priority from high to low, and obtain the test step node priority order and the corresponding parameter node priority order; According to the priority order of the test step nodes, the test step nodes whose order is before the preset first order threshold are used as core test step nodes; According to the parameter node priority order, the parameter node whose order is before the preset second order threshold is used as the core parameter node of the corresponding test step node; Arrange the core test step nodes from the center of the graph outward according to the priority order of the corresponding test step nodes from high to low to obtain a core step area; Arrange the core parameter nodes of each test step node from the center of the test step node outward according to the corresponding parameter node priority order from high to low to obtain the core parameter area; In combination with the core step region and the core parameter region, the structure of the solution map is updated to obtain an updated solution map; In response to the clinical trial requirements input by the user, a path search is performed in the updated protocol map to obtain a protocol generation path and a protocol verification path; Based on the protocol generation path and the protocol verification path, a clinical trial protocol is generated and optimized through a preset dual-channel protocol optimization model to obtain an optimized clinical trial protocol, so as to automatically optimize the clinical trial protocol. In the dual-channel protocol optimization model, the first channel generates a corresponding initial clinical trial protocol according to the protocol generation path, and the second channel verifies and optimizes the initial clinical trial protocol according to the protocol verification path.

2. A method for automatic optimization of clinical trial protocols according to claim 1, characterized in that: The method of constructing a protocol map based on pre-acquired clinical trial data by analyzing the correlation between the trial steps and corresponding parameters includes: Identify trial step nodes and corresponding parameter nodes from pre-acquired clinical trial data; Analyze the association relationship between each test step node and the corresponding parameter node to obtain the corresponding node association relationship; According to the node association relationship, connections are established between corresponding test step nodes and parameter nodes to construct a solution map.

3. The method for automatic optimization of clinical trial protocols according to claim 1, characterized in that: In response to the clinical trial requirements input by the user, a path search is performed in the updated protocol map to obtain a protocol generation path and a protocol verification path, including: According to the clinical trial requirements input by the user, the corresponding required test step node set is extracted; According to the required test step node set, a corresponding solution generation path is searched in the update solution map; Combined with the solution generation path, the corresponding solution verification path is matched in the update solution map.

4. A method for automatic optimization of clinical trial protocols according to claim 3, characterized in that: According to the required test step node set, a corresponding solution generation path is searched in the update solution map, including: Each node in the requirement test step node set is used as the starting node; According to the starting node, in the core step area of ​​the update solution map, according to the connection relationship between the nodes, a first search path set is searched; According to the first search path set, searching in a non-core step area of ​​the update solution map to obtain a second search path set; By analyzing the similarities between the paths, the paths in the first search path set and the second search path set are merged to obtain a first candidate path set and a second candidate path set; Combining the paths in the first candidate path set with the paths in the second candidate path set to obtain a candidate solution generation path set; By calculating the correlation between each path in the candidate solution generation path set and the experimental solution effect, the path with the highest correlation is selected as the solution generation path.

5. The method for automatic optimization of clinical trial protocols according to claim 3, characterized in that: Combined with the solution generation path, the corresponding solution verification path is matched in the update solution map, including: According to each path node in the solution generation path, by expanding the associated nodes, the corresponding experimental knowledge associated node set is matched in the updated solution graph; By analyzing the correlation between each node in the test knowledge associated node set and the corresponding path node, the nodes having the correlation greater than a preset third threshold are screened out as the candidate test knowledge associated node set; Connecting the nodes in the candidate test knowledge associated node set according to the connection relationship between the nodes in the update solution graph to obtain a candidate solution verification path set; By analyzing the correlation between each path in the candidate solution verification path set and the experimental solution effect, the path with the highest correlation is selected as the solution verification path.

6. A method for automatic optimization of clinical trial protocols according to claim 1, characterized in that: Based on the protocol generation path and the protocol verification path, a clinical trial protocol is generated and optimized through a preset dual-channel protocol optimization model to obtain an optimized clinical trial protocol, thereby automatically optimizing the clinical trial protocol, including: According to the protocol generation path, the initial clinical trial protocol is generated through the first channel model in the preset dual-channel protocol optimization model; According to the protocol verification path, the initial clinical trial protocol is verified and optimized through the second channel model in the preset dual-channel protocol optimization model to obtain an optimized clinical trial protocol, so as to automatically optimize the clinical trial protocol.

7. A method for automatic optimization of clinical trial protocols according to claim 6, characterized in that: According to the protocol verification path, the initial clinical trial protocol is verified and optimized by the second channel model in the preset dual-channel protocol optimization model to obtain an optimized clinical trial protocol, so as to automatically optimize the clinical trial protocol, including: According to the protocol verification path, the initial clinical trial protocol is verified through the second channel model in the preset dual-channel protocol optimization model to obtain the protocol verification results; According to the solution verification result, searching for path nodes that need to be optimized in the solution verification path to obtain a set of path nodes to be optimized; Based on the update solution graph, searching for the optimization node with the highest correlation with each node in the set of path nodes to be optimized; The initial clinical trial plan is optimized according to the optimization nodes to obtain an optimized clinical trial plan, so as to automatically optimize the clinical trial plan.

8. A clinical trial protocol automatic optimization system, characterized by: A method for automatically optimizing a clinical trial protocol according to any one of claims 1 to 7, comprising: The protocol map construction module constructs a protocol map based on pre-acquired clinical trial data by analyzing the correlation between trial steps and corresponding parameters; A solution map updating module reorders the structure of the solution map by analyzing the node priorities of the solution map to obtain an updated solution map; A path search module, in response to a clinical trial requirement input by a user, performs a path search in the updated protocol map to obtain a protocol generation path and a protocol verification path; The clinical trial protocol optimization module generates and optimizes the clinical trial protocol based on the protocol generation path and the protocol verification path through a preset dual-channel protocol optimization model to obtain an optimized clinical trial protocol, so as to automatically optimize the clinical trial protocol. In the dual-channel protocol optimization model, the first channel generates a corresponding initial clinical trial protocol according to the protocol generation path, and the second channel verifies and optimizes the initial clinical trial protocol according to the protocol verification path.

Images