A Meta-analysis Method and System for Analyzing and Computable Input of Outcome Indicators in RCTs

Abstract

This invention discloses a method and system for analyzing and computable input of outcome indicators in meta-analysis RCTs, relating to the fields of medical statistics and literature data processing. The method first organizes a set of statistical elements with outcome indicators, detection time points, and comparison relationships as their scopes; then, it abstracts the meta-analysis formula into statistical element slot templates and performs slot matching within the outcome scope; when a target slot is missing, it is filled in based on preset substitution rules; after computability verification, a standardized meta-analysis input structure is generated. This invention solves the problems of existing technologies where meta-analysis data preparation is highly dependent on manual processes, computability verification is difficult to automate, and missing elements cannot be automatically filled in. It transforms manual judgment and calculation into a reusable structured process, significantly improving the efficiency, consistency, and data utilization of large-scale RCT outcome data processing. The output standardized data can be directly used by existing meta-analysis software.

Description

Technical Field

[0001] This invention belongs to the field of medical statistics and literature data processing technology, specifically relating to a method and system for analyzing and calculable input of meta-analysis RCT outcome indicators. Background Technology

[0002] In the preparation of existing meta-analysis data, researchers typically need to manually read RCT literature, identify the statistical elements corresponding to each outcome indicator, and determine whether they meet the calculation conditions for meta-analysis. However, the outcome indicator data in RCT literature are expressed in complex and diverse forms, possibly including mean, standard deviation, standard error, confidence interval, p-value, number of events, sample size, hazard ratio, odds ratio, risk difference, hazard ratio and its confidence interval, etc. Some key statistical elements are not directly given in the literature, but exist indirectly in a form that can be derived from other statistical elements.

[0003] Existing related technologies mainly fall into the following categories: The first category is traditional manual meta-analysis data extraction methods. Researchers manually extract outcome index data from included literature and then judge and convert data such as mean, standard deviation, number of events, and confidence intervals according to the Cochrane manual or statistical methods. The reliability of this method depends on the researcher's experience and is difficult to process in batches or in a standardized manner. The second category is automatic extraction technology from medical literature. Some existing systems can use natural language processing models, large language models, or rule extraction methods to identify PICO information, baseline information, outcome index names, and some statistical elements from the literature. However, this type of technology usually focuses on "extracting information from the literature" rather than further judging whether these statistical elements can form the complete computational structure required for a meta-analysis formula. The third category is general data transformation or ETL technology. This type of technology can convert data from different sources and in different formats into a unified structure, such as field mapping, format conversion, and enumeration value conversion. However, RCT outcome index data is not a simple field conversion problem, but involves the combination relationship between statistical elements, formula input conditions, substitution derivation relationships, and meta-analysis computability judgment. General-purpose data transformation tools struggle to directly address structural closure issues at the statistical semantic level. The fourth category comprises statistical software or meta-analysis software, such as RevMan, Stata, and R packages. This type of software typically requires users to have pre-prepared input data in a specified format, such as mean, standard deviation, sample size, or the number of events or total cases. They generally do not automatically determine whether a particular outcome indicator is calculable from heterogeneous statistical expressions, nor do they automatically identify whether a missing statistical element can be derived by substitution from other elements.

[0004] Therefore, while existing technologies cover literature extraction, data transformation, and meta-analysis calculations, they lack a structured parsing and computability verification mechanism between the "results of literature statistical element extraction" and the "inputs for meta-analysis statistical calculations." Existing technologies mainly suffer from the following drawbacks: First, existing literature extraction techniques typically only obtain scattered statistical elements, making it difficult to determine whether these data belong to the same outcome indicator, the same comparison relationship, or the same detection time point. This is because traditional extraction methods often process information at the field or text fragment level, lacking a statistical element organization mechanism with outcome indicators as the scope. Second, existing technologies lack a slot verification mechanism for meta-analysis formulas. Even if data such as the mean, standard deviation, sample size, and number of events have been extracted, it is difficult to automatically determine whether these data meet the input structures for meta-analysis such as MD, SMD, RR, OR, and HR. This is because existing systems typically do not abstract meta-analysis formulas into a verifiable structure composed of several statistical element slots. Third, existing methods struggle to handle situations where "key statistical elements do not appear directly but can be derived from other statistical elements." For example, when the standard deviation is not directly reported, it may be calculated from the standard error and sample size; when the effect size lacks a standard error, it may be derived from the confidence interval. Existing automated systems often simply treat these situations as missing, leading to the omission of potentially usable data. This is because existing technologies lack calculable substitution rules between combinations of target statistical elements and alternative statistical elements. Fourth, existing statistical software generally assumes that the input data has already been processed and does not handle the closure verification of the data structure before input. This is because meta-analysis software focuses on calculation and merging analysis, rather than solving the problem of converting the statistical expressions of RCT literature to the standardized input structure. Fifth, in existing processes, formula selection, data conversion, and usability judgment heavily rely on human experience, which can easily lead to inconsistencies in judgment among different personnel. Especially in large-scale RCT database construction scenarios, manual judgment of each article will result in high time and quality control costs.

[0005] Existing technologies struggle to automatically determine whether a given outcome indicator, at a specific comparison relationship and detection time point, possesses all the input conditions required for a meta-analysis statistical formula. They also struggle to automatically perform computability verification and standardized input construction when key statistical components are missing but alternative combinations exist. This results in a high reliance on manual judgment and calculations for meta-analysis data preparation, leading to low efficiency, poor consistency, high error rates, and difficulty in scaling up processing. Summary of the Invention

[0006] To address the challenge of automatically converting statistical elements of outcome indicators in randomized controlled trials (RCTs) into standardized input data suitable for meta-analysis, specifically by structurally analyzing heterogeneous statistical expressions corresponding to outcome indicators in RCTs and automatically verifying their computational viability based on predefined meta-analysis formula slot requirements, and supplementing target statistical elements with predefined substitution rules when necessary, thereby generating standardized input data usable for meta-analysis, this application designs a method and system for meta-analysis of RCT outcome indicator structure analysis and computable input. This method organizes outcome indicator-related statistical elements into a set of statistical elements within a unified scope, and verifies the computability of the statistical element set based on predefined meta-analysis formula slots and substitution rules, thereby automatically generating standardized input data usable for meta-analysis.

[0007] This invention does not take the literature extraction model itself or the selection of the final meta-analysis statistical strategy as its core innovation. Instead, it focuses on solving the problem of automatic conversion and verification between the "acquired set of statistical elements of outcome indicators" and the "computable input structure for meta-analysis".

[0008] A method for analyzing and computable inputs of outcome indicators in meta-analysis RCTs includes the following steps: Step S1: Obtain and organize the set of statistical elements for outcome indicators: Receive the set of statistical elements associated with outcome indicators in RCT literature, and organize the statistical elements with outcome indicators, detection time points, and comparison relationships as the scope; Step S2: Construct a preset Meta-analysis formula slot template: Represent the preset Meta-analysis statistical formula as a formula template composed of several statistical element slots; Step S3: Perform formula slot matching within the outcome scope: For a set of statistical elements corresponding to a certain outcome indicator, detection time point, and comparison relationship, select a formula slot template for matching and determine whether there are directly usable statistical elements in each slot. Step S4: Fill in missing slots based on substitution rules: When a target slot is missing, query the preset substitution rules to determine whether there is a combination of substitution statistical elements that can be used to fill in the statistical element corresponding to the target slot. If there is, the slot is filled in. Step S5: Perform computability verification: After completing direct slot matching and necessary substitution completion, determine whether all necessary slots of the selected formula template meet the preset input conditions. Step S6: Generate standardized Meta-analysis input structure: For formulas that have passed computability verification, generate standardized Meta-analysis input structure.

[0009] Preferably, in step S1, each set of statistical elements includes at least an outcome indicator identifier, an outcome indicator name, a detection time point, information about each group, statistical elements corresponding to each group, and comparison relationship information; the comparison relationship information is used to represent the relationship between the groups to be compared. For three-arm or higher RCTs, multiple comparison units are formed through multiple comparison structures, and each comparison unit enters the subsequent formula slot verification process.

[0010] Preferably, in step S2, each formula slot template includes at least the formula type, the statistical element slot required by the formula, the applicable outcome data type, the alternative rules that can be used when the target statistical element is missing, and the standardized meta-analysis input structure of the output.

[0011] Preferably, in step S3, the slot matching process specifically includes: determining the treatment group and the control group based on the comparison relationship; retrieving the treatment group slot required for the formula from the statistical elements of the treatment group; retrieving the control group slot required for the formula from the statistical elements of the control group; and for non-grouped effect size data, retrieving the effect size and its precision information in the corresponding outcome domain.

[0012] Preferably, in step S4, the substitution rule is used to describe the computable relationship between the target statistical element and other combinations of statistical elements; the substitution completion process preferentially uses direct statistical elements, and only uses combinations of substitution statistical elements when direct statistical elements do not exist; the substitution verification is performed according to a preset level and does not perform open-ended infinite reasoning.

[0013] Preferably, in step S5, the preset input conditions include: the statistical element values ​​are not empty; the sample size meets the requirements; the key statistical elements meet the valid value requirements; the statistical elements in the same formula belong to the same outcome indicator, the same detection time point, and the same comparison relationship; and the corresponding group relationship is clear.

[0014] Preferably, in step S6, the generated standardized meta-analysis input structure includes at least the outcome indicator identifier, outcome indicator name, detection time point, comparison relationship identifier, formula type, meta-input data, and slot status; the slot status is used to mark whether each statistical element is obtained directly or through substitution completion, and if it is substitution completion, the source statistical element and the rule used are recorded.

[0015] Preferably, in step S3, the system supports selecting multiple formula slot templates simultaneously for matching and verification, generating standardized Meta-analysis input structures corresponding to different formula types; the final formula selection is determined by user configuration or an external Meta-analysis strategy module.

[0016] Preferably, in step S1, the sources of the statistical element set include the extraction results of the full text of the RCT by the large language model, the identification results of statistical elements in the table or picture by the multimodal model, OCR or rule parsing tools, manual input, and other literature data standardization modules.

[0017] Based on the above design, this application also designs a meta-analysis RCT outcome index structure analysis and computable input system, including: The statistical element organization module is used to receive a set of statistical elements associated with outcome indicators in RCT literature, and organizes the statistical elements by outcome indicators, detection time points, and comparison relationships. The formula template management module, connected to the statistical element organization module, is used to build and manage preset meta-analysis formula slot templates, representing preset meta-analysis statistical formulas as formula templates composed of several statistical element slots. The slot matching module is connected to the statistical element organization module and the formula template management module respectively. It is used to select a formula slot template for matching based on a set of statistical elements corresponding to a certain outcome indicator, detection time point and comparison relationship, and to determine whether there are directly usable statistical elements in each slot. The missing slot completion module is connected to the slot matching module. When a target slot is missing, it queries the preset substitution rules to determine whether there is a combination of substitute statistical elements that can be used to complete the statistical element corresponding to the target slot. If it exists, the slot completion is completed. The computability verification module, connected to the missing slot completion module, is used to determine whether all necessary slots of the selected formula template meet the preset input conditions after completing direct slot matching and necessary substitution completion. The standardized output module, connected to the computability verification module, is used to generate a standardized meta-analysis input structure for formulas that have passed computability verification.

[0018] The advantages and effects of this application are as follows: 1. This application presents a method for analyzing and computable input of outcome indicators in meta-analysis RCTs. It employs a statistical element organization mechanism based on outcome scope, which organizes statistical elements belonging to the same outcome indicator, the same detection time point, and the same comparison relationship in the same RCT literature into a set of statistical elements under a unified scope. This solves the problem of incorrect combination between data from different outcomes, different time points, or different groups in the prior art, and lays an accurate data foundation for subsequent computability verification.

[0019] 2. This application presents a method for analyzing and computable input of meta-analysis RCT outcome indicators. It adopts a technical solution based on the computability verification mechanism of meta-analysis formula slots. The meta-analysis formula is abstracted into a formula template composed of multiple statistical element slots, and slot matching verification is performed on the set of statistical elements. This solves the problem that existing technologies cannot automatically determine whether a certain outcome indicator has the corresponding meta-analysis input conditions, and realizes the automatic determination of meta-analysis input conditions.

[0020] 3. The Meta-analysis RCT outcome index structure analysis and computable input method designed in this application adopts a slot completion mechanism based on the combination of alternative statistical elements. When the target statistical element is missing, the system detects whether there is a combination of alternative statistical elements that can be used to generate the target statistical element based on preset substitution rules, and completes slot completion when the conditions are met. This solves the problem of the prior art misjudging potential computable data as unusable and significantly improves the utilization rate of potential computable statistical elements.

[0021] 4. The present application proposes a method for parsing and computable input of meta-analysis RCT outcome index structure. It adopts a technical solution that supports the processing of multiple comparison relationships. The specific comparison relationship is represented by the comparison structure, which supports the processing of multiple comparison relationships in two-arm and three-arm or more RCT studies. This solves the problem that the existing technology is difficult to process multi-arm RCT data and expands the applicability of the system.

[0022] 5. The present application proposes a method for analyzing and calculable input of meta-analysis RCT outcome indicators. It adopts a structured verification process to transform the manual judgment and calculation steps in the preparation of meta-analysis data into a reusable and traceable structured verification process. This solves the problem of existing technologies being highly dependent on human experience and having poor consistency, and improves the consistency and automation of large-scale RCT outcome data processing.

[0023] 6. The Meta-analysis RCT outcome index structure analysis and computable input method designed in this application adopts a standardized output structure technical solution, which outputs a standardized Meta-analysis input structure containing slot status markers. This solves the problem of inconsistent output data formats and difficulty in direct calling by subsequent Meta-analysis modules in the existing technology, and achieves seamless integration with existing Meta-analysis software.

[0024] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the preferred embodiments of this application are described in detail below with reference to the accompanying drawings.

[0025] The above and other objects, advantages and features of this application will become more apparent to those skilled in the art from the following detailed description of specific embodiments in conjunction with the accompanying drawings. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In all drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0027] Figure 1 This is the overall flowchart of the Meta-analysis RCT outcome index structure analysis and computable input method of the present invention. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. In the following description, specific details such as specific configurations and components are provided merely to help fully understand the embodiments of this application. Therefore, those skilled in the art should understand that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this application. In addition, for clarity and brevity, descriptions of known functions and structures are omitted in the embodiments.

[0029] It should be understood that the phrase "an embodiment" or "this embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "an embodiment" or "this embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments.

[0030] Furthermore, reference numerals and / or letters may be repeated in different examples within this application. Such repetition is for the purpose of simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or settings discussed.

[0031] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, and A and B exist simultaneously. The term " / and" in this article describes another type of relationship between related objects, indicating that two relationships can exist. For example, A / and B can mean: A exists alone, and A and B exist alone. In addition, the character " / " in this article generally indicates that the related objects before and after it are in an "or" relationship.

[0032] In this article, the term "at least one" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, "at least one of A and B" can mean: A exists alone, A and B exist simultaneously, or B exists alone.

[0033] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion.

[0034] Example 1, Reference Figure 1 The present invention discloses a method for analyzing and computable input of outcome indicators in meta-analysis RCTs, the overall process of which is as follows: Step S1: Obtain and organize the set of statistical components for the outcome indicators; The system acquires a structured set of statistical elements. This set of statistical elements can come from: the extraction results of the full text of the RCT by a large language model; the recognition results of statistical elements in tables or images by a multimodal model; OCR or rule parsing tools; manual input; and other literature data standardization modules. This application does not limit the method of extracting statistical elements.

[0035] Each outcome indicator corresponds to one or more sets of statistical elements. Each set includes at least: outcome_id: Outcome indicator; outcome_name: The name of the outcome metric; detection_time_point: The detection time point; Group information: including group_id and group_name; components: Statistical elements under this group, such as mean, sd, se, ci, p_value, event, total, hr, rr, or, etc. Comparison information: used to indicate the relationship between groups being compared, such as the correspondence between the treatment group and the control group; The example structure is as follows: { "outcome_id":"OUT_001", "outcome_name":"HbA1cchange", "detection_time_point":"12weeks", "groups":[ { "group_id":"G1", "group_name":"DrugA", "components":{ "mean":-1.2, "sd":0.4, "n":50 } }, { "group_id":"G2", "group_name":"Placebo", "components":{ "mean":-0.3, "sd":0.6, "n":52 } } ], "comparisons":[ { "comparison_id":"C1", "treatment_group_id":"G1", "control_group_id":"G2" } ] } After the statistical elements are organized, the system will perform a preliminary scope check to check whether there are logical conflicts among the statistical elements in the same scope (such as two different means appearing at the same time point in the same group). For statistical elements with conflicts, they will be marked as suspicious and prompted for manual review to ensure that the scope of statistical elements entering the subsequent process is correctly assigned.

[0036] The purpose of this step is to establish the scope of outcome indicators, preventing the incorrect combination of statistical elements between different outcome indicators, different time points, or different groups. For three-armed or longer RCTs, this application uses a comparison structure to represent specific comparison relationships. For example, in a three-armed study, multiple comparison units can be formed, with each comparison unit entering the subsequent formula slot validation process.

[0037] In one embodiment, the system processes a three-arm RCT study comprising Drug A, Drug B, and Placebo groups, reporting HbA1c changes at week 12. The system first constructs the outcome scope based on the extracted statistical element objects, identifying the corresponding outcome_id, detection_time_point, and group information. Specifically, Drug A reports the mean, standard deviation, and sample size; Drug B reports the mean, standard error, and sample size; and Placebo reports the mean, standard deviation, and sample size. Subsequently, the system generates multiple comparison units based on preset comparison relationships, including comparison units between Drug A and Placebo, and between Drug B and Placebo, assigning each a unique comparison_id. For the Drug A vs. Placebo comparison unit, the system maps the two sets of statistical elements to slots in the SMD formula template object. Since the required mean, standard deviation, and sample size are readily available, the corresponding slot status is updated to "matched" and verified through the formula template. For the comparison unit of Drug B against Placebo, the system finds that the slot corresponding to treatment_sd is missing during slot mapping. Therefore, it calls the substitution rule object associated with the standard deviation, executes a preset generation expression based on the existing standard error and sample size of the Drug B group, generates the corresponding standard deviation statistical element object, and updates the slot status to substituted. Simultaneously, it records the derivation path, source statistical element, and rule identifier. After completing the missing slot completion, the system continues to perform formula template status verification. When all necessary slots meet the preset input conditions, the corresponding formula template object status is updated to validated, and two standardized meta-input structure objects are generated for subsequent meta-analysis module calls. Through this method, the system can automatically complete comparison unit generation, slot mapping, substitution completion, and computability verification in multi-arm RCT scenarios, achieving automatic conversion of complex outcome indicator statistical expressions into standardized meta-analysis input structures.

[0038] Step S2: Construct a preset Meta-analysis formula slot template; The system has a pre-defined set of statistical formulas for meta-analysis. Each formula is represented as a formula slot template. Each formula slot template includes at least: formula_type: Formula type, such as MD, SMD, RR, OR, HR; required_slots: The number of component slots required by the formula; applicable_data_type: Applicable data type; substitution_rules: Substitution rules that can be used when the target statistical element is missing; output_schema: The normalized meta-analysis input structure of the output; For example, an SMD formula slot template may include: { "formula_type":"SMD", "required_slots":[ "treatment_mean", "treatment_sd", "treatment_n", "control_mean", "control_sd", "control_n ] } The RR formula slot template for binary classification outcomes can include: { "formula_type":"RR", "required_slots":[ "treatment_event", "treatment_total", "control_event", "control_total" ] } HR formula slot templates can include: { "formula_type":"HR", "required_slots":[ "hr", "ci_lower", "ci_upper ] } This application does not require an exhaustive list of all formula pools, but should include representative formula templates that illustrate the technical mechanism.

[0039] Step S3: Perform formula slot matching within the outcome scope; The system selects one or more formula slot templates for matching based on the set of statistical elements corresponding to a given outcome_id, detection_time_point, and comparison_id. The matching process includes: 1. The treatment group and the control group were determined based on comparison_id; 2. Retrieve the required treatment group slot from the treatment group components; 3. Retrieve the required control group slot from the control group components; 4. For non-grouped effect size data, such as HR and its confidence interval, retrieve the effect size and its precision information in the corresponding outcome domain; 5. Determine if there are any directly usable statistical elements in each slot.

[0040] If all the slots required by the formula are satisfied by direct statistical elements, then the outcome index is determined to meet the corresponding formula input conditions at the comparison relationship and time point. The system supports selecting multiple formula slot templates for matching and verification simultaneously, generating standardized meta-analysis input structures corresponding to different formula types; the final formula selection is determined by user configuration or an external meta-analysis strategy module.

[0041] Step S4: Complete missing slots based on substitution rules; When a target slot is missing, the system does not immediately determine that the formula is unavailable, but instead queries a preset substitution rule. The substitution rule describes the computable relationship between the target statistical element and other combinations of statistical elements. For example: Target statistical element: SD; Alternative statistical element combinations: SE+n; Completion rule: SD = SE × √n; When the system detects a missing treatment_sd required for an SMD formula, it checks if treatment_se and treatment_n exist under the same outcome_id, detection_time_point, and treatment_group_id. If they exist, the system generates a treatment_sd based on preset rules and marks the slot as "alternative completion". Similarly, if control_sd is missing, it can be completed using control_se and control_n.

[0042] In this application, the substitution completion process preferentially uses direct statistical elements; only when direct statistical elements are unavailable is a combination of substitution statistical elements used. The system employs a derivation control mechanism to constrain the execution process of the substitution rules. The derivation control mechanism includes: Maximum derivation depth limit; Derivation of path records; Accessed statistical element detection; When the derivation depth reaches a preset threshold or a statistical element is accessed repeatedly, the expansion of the substitution path stops. For statistical elements generated through substitution rules, the system records their derivation path, derivation depth, and source statistical element information.

[0043] Step S5: Perform computability verification; After completing direct slot matching and necessary substitution completion, the system determines whether all necessary slots of the selected formula template meet the preset input conditions. These input conditions include not only the presence of statistical elements, but also: The statistical element values ​​are not empty and are valid values; The standard deviation is greater than zero; The sample size is a positive integer; The number of events is no greater than the total number of cases; The lower limit of the confidence interval is less than the upper limit; Statistical elements within the same formula template have consistent outcome_id, detection_time_point, and comparison_id. For target statistical elements generated through substitution rules, all source statistical elements must satisfy the above data validity constraints. If all conditions are met, the system determines that the outcome indicator can generate the corresponding meta-analysis input structure under the comparison relationship and time point. If there are still necessary slots that cannot be filled, the system determines that the formula template is not satisfied under this scope and outputs the reason for the missing information.

[0044] Step S6: Generate standardized meta-analysis input structure; For formulas that pass computability verification, the system generates a standardized meta-analysis input structure. An example is shown below: { "outcome_id":"OUT_001", "outcome_name":"HbA1cchange", "detection_time_point":"12weeks", "comparison_id":"C1", "formula_type":"SMD", "meta_input":{ "treatment_mean":-1.2, "treatment_sd":0.4, "treatment_n":50, "control_mean":-0.3, "control_sd":0.6, "control_n":52 }, "slot_status":{ "treatment_mean":"direct", "treatment_sd":"direct", "treatment_n":"direct", "control_mean":"direct", "control_sd":"direct", "control_n":"direct" } } If a slot is obtained through substitution completion, record its source: { "control_sd":{ "status":"substituted", "source_components":["control_se","control_n"], "rule":"SD=SE×sqrt(n)" } } This step does not require the system to perform the final meta-analysis or to decide which statistical effect size to use. The final formula selection can be determined by user configuration or an external meta-analysis strategy module.

[0045] The present invention also discloses a system for implementing the above method, comprising: a statistical element organization module, a formula template management module, a slot matching module, a missing slot completion module, a computability verification module, and a standardized output module.

[0046] The above description is merely a preferred embodiment of the present invention and does not limit the scope of protection of the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any changes, modifications, substitutions, integrations, and parameter alterations to these embodiments within the spirit and principles of the present invention, achieved through conventional substitutions or by achieving the same function without departing from the principles and spirit of the present invention, fall within the scope of protection of the present invention.

Claims

1. A Meta-analysis RCT outcome indicator structure analysis and computable input method, characterized in that, Includes the following steps: Step S1: Obtain and organize the set of statistical elements for outcome indicators: Receive the set of statistical elements associated with outcome indicators in RCT literature, and organize the statistical elements with outcome indicators, detection time points, and comparison relationships as the scope; In step S1, each set of statistical elements includes at least the outcome indicator identifier, outcome indicator name, detection time point, group information, statistical elements under that group, and comparison relationship information; The comparison relationship information is used to represent the group relationship to be compared. For RCTs with three arms or more, multiple comparison units are formed through multiple comparison structures, and each comparison unit enters the subsequent formula slot verification process. Step S2: Construct a preset Meta-analysis formula slot template: Represent the preset Meta-analysis statistical formula as a formula template composed of several statistical element slots; Step S3: Perform formula slot matching within the outcome scope: For a set of statistical elements corresponding to a certain outcome indicator, detection time point, and comparison relationship, select a formula slot template for matching and determine whether there are directly usable statistical elements in each slot. Step S4: Complete missing slots based on substitution rules: When a target slot is missing, query the preset substitution rules to determine whether there is a combination of substitution statistical elements that can be used to complete the target statistical element. If there is, complete the slot completion. In step S4, the substitution rule is used to describe the computable relationship between the target statistical element and other combinations of statistical elements; the substitution completion process prioritizes the use of direct statistical elements, and only uses the combination of substitution statistical elements when the direct statistical element does not exist; the system adopts a derivation control mechanism to constrain the execution process of the substitution rule. The derivation control mechanism includes: maximum derivation depth limit; derivation path recording; detection of visited statistical elements; and stopping the expansion of alternative paths when the derivation depth reaches a preset threshold or when a repeatedly visited statistical element appears. Step S5: Perform computability verification: After completing direct slot matching and necessary substitution completion, determine whether all necessary slots of the selected formula template meet the preset input conditions. Step S6: Generate standardized Meta-analysis input structure: For formulas that have passed computability verification, generate standardized Meta-analysis input structure.

2. The Meta-analysis RCT outcome indicator structure analysis and computable input method according to claim 1, characterized in that, In step S2, each formula slot template includes at least the formula type, the required statistical element slot for the formula, the applicable outcome data type, the alternative rules that can be used when the target statistical element is missing, and the standardized meta-analysis input structure for the output.

3. The Meta-analysis RCT outcome indicator structure analysis and computable input method according to claim 1, wherein, In step S3, the slot matching process specifically includes: determining the treatment group and the control group based on the comparison relationship; retrieving the treatment group slot required for the formula from the statistical elements of the treatment group; retrieving the control group slot required for the formula from the statistical elements of the control group; and for non-grouped effect size data, retrieving the effect size and its precision information in the corresponding outcome domain.

4. The method for analyzing and computable input of meta-analysis RCT outcome index structure according to claim 1, characterized in that, In step S4, the substitution verification is performed at a preset level, without open-ended infinite reasoning.

5. The method for analyzing and computable input of meta-analysis RCT outcome index structure according to claim 1, characterized in that, In step S5, the preset input conditions include: the statistical element values ​​are not empty; the sample size meets the requirements; the key statistical elements meet the valid value requirements; the statistical elements in the same formula belong to the same outcome index, the same detection time point, and the same comparison relationship; and the corresponding group relationship is clear.

6. The method for analyzing and computable input of meta-analysis RCT outcome index structure according to claim 1, characterized in that, In step S6, the generated standardized meta-analysis input structure includes at least the outcome indicator identifier, outcome indicator name, detection time point, comparison relationship identifier, formula type, meta-input data, and slot status; the slot status indicates whether each statistical element is obtained directly or through substitution completion, and if it is substitution completion, the source statistical element and the rule used are recorded.

7. The method for analyzing and computable input of meta-analysis RCT outcome index structure according to claim 1, characterized in that, In step S3, the system supports selecting multiple formula slot templates for matching and verification simultaneously, generating standardized meta-analysis input structures corresponding to different formula types; the final formula selection is determined by user configuration or an external meta-analysis strategy module.

8. The method for analyzing and computable input of meta-analysis RCT outcome index structure according to claim 1, characterized in that, In step S1, the sources of the statistical element set include the extraction results of the full text of the RCT by the large language model, the recognition results of statistical data in tables or pictures by the multimodal model, OCR or rule parsing tools, manual input, and other literature data standardization modules.

9. A meta-analysis RCT outcome index structure analysis and computable input system, characterized in that, include: The statistical element organization module is used to perform step S1 as described in claim 1, organizing the set of statistical elements with the scope of outcome indicators, detection time points, and comparison relationships. The formula template management module is used to execute step S2 as described in claim 1, and to construct and manage preset Meta analysis formula slot templates; The slot matching module is used to execute step S3 as described in claim 1, performing formula slot matching within the scope of the outcome. The missing slot completion module is used to perform step S4 as described in claim 1, and to complete the missing slots based on a preset substitution rule; The computability verification module is used to execute step S5 as described in claim 1, perform computability verification, and output the verification result; The standardized output module is used to perform step S6 as described in claim 1 to generate a standardized meta-analysis input structure.

Images