Big data based compliance document management system and method

Abstract

This invention discloses a compliance document management system and method based on big data, relating to the field of pharmaceutical data management technology. The invention completes the standardized collection and definition of metadata for pharmaceutical management documents, and, combining the drug lifecycle stage and document type, classifies documents into three initial popularity levels: hot, warm, and cold. Through three compliance checks—audit exemption locking, special drug control, and statutory retention period—the initial popularity results are forcibly corrected, and a differentiated hierarchical storage strategy is matched based on the final popularity. Subsequently, document access behavior is continuously monitored to achieve automated dynamic transitions and downgrades in popularity levels. Finally, a quantitative scoring model adapted to the entire lifecycle is constructed to complete the identification of key documents and mandatory popularity control, realizing intelligent hierarchical storage of pharmaceutical management documents. This ensures rapid response to critical documents while optimizing storage resources, meeting the compliance management requirements of the entire pharmaceutical lifecycle.

Description

Technical Field

[0001] This invention relates to the field of pharmaceutical data management technology, specifically to a compliance document management system and method based on big data. Background Technology

[0002] The research, development, production, distribution, and regulation of pharmaceuticals involve a large number of drug management documents, including Standard Operating Procedures (SOPs), Batch Production Records (BPR), Batch Inspection Records (BIR), deviation handling documents, Corrective and Preventive Action (CAPA) documents, research and development files, and registration application documents. These documents span the entire lifecycle of a drug, from research and development, registration, and production to withdrawal from the market, and are characterized by their diverse types, large quantity, long storage period, and high compliance requirements.

[0003] Currently, drug management documents are mostly stored using static, uniform storage strategies, such as dividing storage areas by file type or creation time. This lacks dynamic responsiveness to factors like file access frequency, business stage, and compliance retention requirements. As drug lifecycle stages change, the importance and frequency of file usage can change significantly. Static storage strategies are prone to risks such as wasted storage resources and decreased access efficiency. Therefore, there is an urgent need for a method that can dynamically adjust the storage of drug management documents by combining multi-dimensional data such as the drug life cycle, in order to improve the efficiency of storage resource utilization, ensure the access performance of critical documents, and meet the requirements of drug regulatory compliance. Summary of the Invention

[0004] The purpose of this invention is to provide a compliance document management system and method based on big data to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a compliance document management method based on big data, the method comprising the following steps: Step S101: Perform unified information sorting and standardized definition of all drug management documents, and collect metadata of all drug management documents. The metadata includes drug life cycle stage L, document type F, and last access time T. last The time T of existence of drug management documents current Drug expiration date T expiry compliant retention period T retention Special drug identification mark K, audit exemption mark A; the drug life cycle stage L includes research and clinical trial stage L1, registration application stage L2, market production stage L3, and production stoppage and withdrawal stage L4; The document type F includes core control document F1, batch record document F2, process document F3, R&D archive document F4, and registration application document F5; the core control document F1 refers to current and historical versions of SOPs, process procedures, and quality standard documents; the batch record document F2 refers to batch production record (BPR) and batch inspection record (BIR); the process document F3 refers to drug deviation handling documents and CAPA documents; when the special drug identifier K=K1, it refers to narcotic drugs, psychotropic drugs, vaccines, and medical toxic drugs; when the special drug identifier K=K0, it refers to ordinary drugs; when the audit exemption identifier A=A1, it means that the current document's corresponding product is under GMP inspection, registration supplementation, or audit concern; when the audit exemption identifier A=A0, it means that the above status does not exist. After data collection is complete, calculate the unaccessed time interval ΔT of the drug management file: ΔT = T current -T last This provides a reliable data foundation for subsequent hot and cold data stratification, compliance assessment, and dynamic adjustment; Step S102: Perform initial popularity level classification on the drug management documents to match business scenario requirements. Specifically, define the popularity level H for file type F, where the value of popularity level H is: hot data H1, warm data H2, and cold data H3; combine the metadata from step S101, and execute fixed initial popularity determination rules: When the drug life cycle stage L is the research and development clinical trial stage L1, the initial popularity of the core control document F1, research and development archive document F4, and registration application document F5 is determined as hot data H1, and the initial popularity of the batch record document F2 and process document F3 is determined as hot data H1. When the drug life cycle stage L is the registration application period L2, the initial popularity of the registration application document F5 and the core control document F1 is determined to be hot data H1, and the R&D archive document F4, batch record document F2, and process document F3 are determined to be warm data H2. When the drug life cycle stage L is the market production stage L3, the initial popularity of the registration application document F5, core control document F1, batch record document F2, and process document F3 is determined to be warm data H2, and the initial popularity of the R&D archive document F4 is determined to be cold data H3. When the drug life cycle stage L is the production stoppage and withdrawal period L4, the initial popularity of the registration application document F5 is determined to be warm data H2, and the initial popularity of all other document types F is determined to be cold data H3.

[0006] Step S201: Perform compliance verification and correction on the initial popularity determination result completed in step S102, and execute three correction operations in sequence: The first step is to perform an audit exemption lock verification. When the audit exemption identifier A corresponding to the drug management document is A1, the popularity of all document types F in the current drug management document is forcibly locked to hot data H1 until the value of the audit exemption identifier A changes to A0. The second step is to perform special drug control verification. When the special drug identifier K corresponding to the drug management document is K1, the heat of all file types F in the current drug management document will be forcibly locked to hot data H1. Thirdly, perform a retention period compliance verification and calculate the statutory compliance retention period T corresponding to the document. retention :T retention =T expiry +T extra In the formula, T extra The statutory retention period for current drug management documents; for the batch record file F2, if the current drug management document exists for a time T... current Less than the compliant retention period T retention The heat level of the drug management document shall not be lower than the temperature data H2.

[0007] Step S202: The file type F corresponding to the hot data H1 is stored on a high-performance storage medium, supporting second-level access response and high-frequency viewing and modification by multiple users concurrently, ensuring smooth use during peak business periods; the file corresponding to the warm data H2 is stored on a medium-performance storage medium, supporting millisecond-level access response and regular viewing and modification, meeting the stable needs of daily business; the file corresponding to the cold data H3 is stored on a low-performance storage medium, supporting on-demand retrieval.

[0008] Step S301: Continuously monitor the access behavior of the drug management documents to achieve automatic dynamic adjustment of the popularity level, so that the storage strategy can adapt to the changes in real business needs and achieve a dynamic balance between access efficiency and storage resources. Set fixed calculation parameters, including the access behavior statistics time window T. window Popularity threshold; the access behavior statistics time window T window The fixed statistical period used to calculate the average monthly access frequency of drug management documents is a fundamental parameter for determining dynamic shifts and downgrades in popularity. The popularity threshold includes the access frequency threshold f. th1 =a1 times / month, f th2 =a2 times / month, a1 < a2, inaccessible duration threshold ΔT th1 =b1 month, ΔT th2 =b2 months, b1 < b2; the threshold f th1 f th2 ΔT th1 ΔT th2 Set up by professionals; Calculate the access frequency f: ; In the formula, f represents the average monthly access frequency of the file, measured in times per month; N represents the statistical time window T. window The cumulative number of times users have accessed the current drug management documents within the system; Step S302: Based on the calculation results, perform dynamic popularity transition and downgrade determination, specifically: when the current popularity level of the file type F is cold data H3 and the average monthly access frequency f is greater than the access frequency threshold f th1 When the current file type F's popularity level is automatically changed to warm data H2; when the current popularity level of file type F is warm data H2 and the average monthly access frequency f is greater than the access frequency threshold f th2 When the time comes, the popularity level of the current file type F will automatically jump to hot data H1; When the current popularity level of the file type F is hot data H1 and the unaccessed time interval ΔT is greater than the unaccessed duration threshold ΔT th1 When the current file type F's popularity level is automatically downgraded to warm data H2; when the current popularity level of file type F is warm data H2 and the unaccessed time interval ΔT is greater than the unaccessed duration threshold ΔT th2 When this happens, the popularity level of the current file type F will be automatically downgraded to cold data H3.

[0009] Step S401: Based on metadata, file access frequency f, and popularity level H, construct a quantitative scoring model that adapts to different drug management document lifecycle stages, continuously monitor each drug management document at fixed intervals, and complete the quantitative scoring. The preset model has core quantization parameters, including the access frequency contribution weight W. f Popularity level matching weight W h Popularity level weight M N A set of lifecycle determination thresholds, wherein M N The weight values ​​correspond to different heat levels. When N=1, it corresponds to the weight of hot data H1; when N=2, it corresponds to the weight of warm data H2; and when N=3, it corresponds to the weight of cold data H3. The life cycle determination threshold set includes determination threshold S1 for the research and development clinical trial period L1, determination threshold S2 for the registration application period L2, determination threshold S3 for the market production period L3, and determination threshold S4 for the production stoppage and withdrawal period L4. After the parameter preset is completed, the window T... window As a statistical period, the score S of the drug management document is calculated. i : ; In the formula, S ic1 represents the importance score of the i-th drug life cycle stage L in the current drug management document, c2 represents the number of file type F in the hot data H1 in the current drug management document, c3 represents the number of file type F in the warm data H2 in the current drug management document, and c3 represents the number of file type F in the cold data H3 in the current drug management document. Step S402: Based on the score S i The drug management documents are judged and subject to mandatory popularity control to ensure the access efficiency of high-frequency and high-importance documents. When the drug lifecycle stage L in the drug management document is the research and clinical trial phase L1, if S i >S1, the current drug management document is identified as a key focus document, and the popularity level (H) of all documents under it is adjusted to H1; When the drug lifecycle stage L in the drug management document is the registration application period L2, if S i >S2, the current drug management document is identified as a key focus document, and the popularity level (H) of all document types under it is adjusted to H1; When the drug lifecycle stage L in the drug management document is the market production stage L3, if S i >S3, the current drug management document is identified as a key focus document, and the popularity level (H) of all document types under it is adjusted to H1; When the drug lifecycle stage L in the drug management document is the discontinuation and withdrawal period L4, if S i >S4, the current drug management document is identified as a key focus document, and the popularity level (H) of all document types under it is adjusted to H1.

[0010] The compliance document management system based on big data includes a data collection and initial classification module, a compliance verification and storage mapping module, a dynamic heat adjustment module, and a quantitative assessment and mandatory control module. The data acquisition and initial classification module is responsible for sorting out information and collecting metadata from all drug management documents, covering core fields such as drug life cycle stage, file type, timestamp, expiration date, compliance retention period and special identifiers. Based on the collected metadata, it calculates the unaccessed time interval and classifies the files into initial popularity levels according to preset rules, providing a basis for subsequent storage strategies. The data acquisition and initial classification module includes a metadata acquisition unit and an initial heat classification unit; The metadata collection unit performs unified information sorting and standardized definition of all drug management documents, specifically: collecting the drug lifecycle stage L, file type F, and file last access time T. last The time T of existence of drug management documents currentDrug expiration date T expiry compliant retention period T retention Special drug identifier K, audit exemption identifier A, calculate the unaccessed time interval ΔT after data collection is completed; The initial heat classification unit classifies drug management documents into initial heat levels, defining heat levels H as hot data H1, warm data H2, and cold data H3. It executes the initial heat judgment rules and assigns initial heat levels to various types of documents at different stages according to the correspondence between drug life cycle stage L and document type F, thereby achieving preliminary classification that matches the needs of business scenarios.

[0011] The compliance verification and storage mapping module performs compliance verification and correction on the initial heat determination results, and sequentially performs three correction operations: audit exemption lock, special drug control, and retention period compliance, to ensure that the heat classification meets regulatory requirements. The module maps the final determined heat level to the corresponding performance storage medium, realizing differentiated deployment of high-performance storage for hot data, medium-performance storage for warm data, and low-performance storage for cold data. The compliance verification and storage mapping module includes a compliance correction unit and a storage mapping unit; The compliance correction unit performs three correction operations on the initial heat level determination result: when the audit exemption identifier A is A1, it forcibly locks the heat level of all file types F in the current drug management documents to hot data H1; when the special drug identifier K is K1, it forcibly locks the heat level of all file types F to hot data H1; for batch record file F2, if the existence time is less than the compliance retention period, the heat level is set to be no lower than warm data H2. The storage mapping unit stores the file type F corresponding to hot data H1 on a high-performance storage medium, supporting second-level access response and high-frequency operation by multiple users; the file corresponding to warm data H2 is stored on a medium-performance storage medium, supporting millisecond-level access and regular browsing; the file corresponding to cold data H3 is stored on a low-performance storage medium, supporting on-demand retrieval, thereby achieving optimized configuration of storage resources and access efficiency.

[0012] The dynamic heat adjustment module continuously monitors the access behavior of drug management documents to achieve automatic dynamic adjustment of heat level. It sets access behavior statistical time window and heat jump threshold, calculates access frequency, and executes upward or downward heat jump based on the comparison relationship between the current heat level, access frequency, and non-access duration, so that the storage strategy fits the changes in real business needs. The dynamic heat adjustment module includes an access frequency calculation unit and a transition / degradation determination unit. The access frequency calculation unit is set with fixed calculation parameters, including the access behavior statistics time window T. window Access frequency threshold f th1 and fth2 Unaccessed duration threshold ΔT th1 and ΔT th2 Calculate the average monthly access frequency f to provide a quantitative basis for dynamic adjustment of popularity; The transition degradation determination unit performs dynamic adjustment of heat based on the calculation results, specifically: when f > f th1 When f > f, cold data H3 transitions to warm data H2; th2 When ΔT > ΔT, the temperature data H2 transitions to the thermal data H1; th1 When ΔT > ΔT, thermal data H1 is downgraded to temperature data H2; th2 At that time, the temperature data H2 is downgraded to the cold data H3, realizing the automatic dynamic adjustment of the heat level.

[0013] The quantitative assessment and mandatory control module constructs a quantitative scoring model adapted to different drug lifecycle stages. Based on metadata, file access frequency, and popularity level, it comprehensively scores drug management files at fixed intervals. According to the comparison between the scoring results and the judgment thresholds of each stage, it implements mandatory popularity control, identifies files that exceed the threshold as key files of concern, and forcibly adjusts the popularity of all file types under them to hot data H1, so as to ensure the access efficiency of high-frequency and high-importance files.

[0014] The quantitative assessment and mandatory control module includes a quantitative scoring unit and a mandatory control unit; The quantization scoring unit presets core quantization parameters for the model, including the access frequency contribution weight W. f Popularity level matching weight W h Popularity level weight M N A set of lifecycle determination thresholds, denoted by T window For the statistical period, calculate the file score S. i The life cycle determination threshold set includes S1, S2, S3, and S4; The mandatory control unit is based on a scoring system (S). i The drug management documents at each stage of the life cycle are determined as follows: when L is L1, S i When L is >S1, the popularity level of all document types F in the drug management documents is adjusted to H1; when L is L2, S i When L is >S2, the popularity level of all document types F in the drug management documents is adjusted to H1; when L is L3, S i When L is >S3, the popularity level of all document types F in the drug management documents is adjusted to H1; when L is L4, S i >S4, the F popularity level of all file types of the drug management documents is adjusted to H1 to ensure the efficiency of accessing high-frequency and high-importance documents.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. Improve business access efficiency and ensure response speed: By classifying and storing drug management documents as hot, warm, and cold data, key documents with high-frequency access, audit concerns, and special control are prioritized for deployment on high-performance storage media. This ensures second-level response in scenarios such as GMP inspections, registration supplements, and audit tracking, thereby improving business processing efficiency and regulatory response capabilities.

[0016] 2. Achieve dynamic optimization of storage resources and improve resource utilization: Based on metadata such as file access frequency, inaccessible duration, and lifecycle stage, build an automated dynamic adjustment mechanism for hotness, automatically migrate low-frequency access or historical files to low-cost storage media, avoid resource waste caused by "not distinguishing between hot and cold" files, and achieve a dynamic balance between storage cost and access performance.

[0017] 3. Strengthen data governance capabilities to support compliance management throughout the entire drug lifecycle: By combining multi-dimensional metadata such as drug lifecycle stages, document types, compliance retention periods, and special drug identification, a quantitative scoring model and a mandatory heat control mechanism are constructed to ensure that key documents are always stored at a reasonable level throughout the entire lifecycle, meeting the requirements of drug regulatory regulations for document integrity, traceability, and retention compliance, and improving the company's data governance level. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the application of the present invention to a big data-based compliance document management method; Figure 2 This is a schematic diagram of the structure of the present invention applied to a big data-based compliance document management system. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Example: Figures 1-2 As shown, this invention provides a technical solution. A pharmaceutical company, in order to meet the requirements of Good Manufacturing Practice (GMP) for pharmaceuticals and improve document management efficiency, decides to implement a big data-based compliance document management method for pharmaceutical management documents. The method includes the following steps: Step S101: Perform unified information sorting and standardized definition of all drug management documents, and collect metadata of all drug management documents. The metadata includes drug life cycle stage L, document type F, and last access time T. last The time T of existence of drug management documents current Drug expiration date T expiry compliant retention period T retention Special drug identification mark K, audit exemption mark A; the drug life cycle stage L includes research and clinical trial stage L1, registration application stage L2, market production stage L3, and production stoppage and withdrawal stage L4; The document type F includes core control document F1, batch record document F2, process document F3, R&D archive document F4, and registration application document F5; the core control document F1 refers to current and historical versions of SOPs, process procedures, and quality standard documents; the batch record document F2 refers to batch production record (BPR) and batch inspection record (BIR); the process document F3 refers to drug deviation handling documents and CAPA documents; when the special drug identifier K=K1, it refers to narcotic drugs, psychotropic drugs, vaccines, and medical toxic drugs; when the special drug identifier K=K0, it refers to ordinary drugs; when the audit exemption identifier A=A1, it means that the current document's corresponding product is under GMP inspection, registration supplementation, or audit concern; when the audit exemption identifier A=A0, it means that the above status does not exist. After data collection is complete, calculate the unaccessed time interval ΔT of the drug management file: ΔT = T current -T last This provides a reliable data foundation for subsequent hot and cold data stratification, compliance assessment, and dynamic adjustment; Step S102: Perform initial popularity level classification on the drug management documents to match business scenario requirements. Specifically, define the popularity level H for file type F, where the value of popularity level H is: hot data H1, warm data H2, and cold data H3; combine the metadata from step S101, and execute fixed initial popularity determination rules: When the drug life cycle stage L is the research and development clinical trial stage L1, the initial popularity of the core control document F1, research and development archive document F4, and registration application document F5 is determined as hot data H1, and the initial popularity of the batch record document F2 and process document F3 is determined as hot data H1. When the drug life cycle stage L is the registration application period L2, the initial popularity of the registration application document F5 and the core control document F1 is determined to be hot data H1, and the R&D archive document F4, batch record document F2, and process document F3 are determined to be warm data H2. When the drug life cycle stage L is the market production stage L3, the initial popularity of the registration application document F5, core control document F1, batch record document F2, and process document F3 is determined to be warm data H2, and the initial popularity of the R&D archive document F4 is determined to be cold data H3. When the drug life cycle stage L is the production stoppage and withdrawal period L4, the initial popularity of the registration application document F5 is determined to be warm data H2, and the initial popularity of all other document types F is determined to be cold data H3.

[0021] Step S201: Perform compliance verification and correction on the initial popularity determination result completed in step S102, and execute three correction operations in sequence: The first step is to perform an audit exemption lock verification. When the audit exemption identifier A corresponding to the drug management document is A1, the popularity of all document types F in the current drug management document is forcibly locked to hot data H1 until the value of the audit exemption identifier A changes to A0. The second step is to perform special drug control verification. When the special drug identifier K corresponding to the drug management document is K1, the heat of all file types F in the current drug management document will be forcibly locked to hot data H1. Thirdly, perform a retention period compliance verification and calculate the statutory compliance retention period T corresponding to the document. retention :T retention =T expiry +T extra In the formula, T extra The statutory retention period for current drug management documents; for the batch record file F2, if the current drug management document exists for a time T... current Less than the compliant retention period T retention The heat level of the drug management document shall not be lower than the temperature data H2.

[0022] Step S202: The file type F corresponding to the hot data H1 is stored on a high-performance storage medium, supporting second-level access response and high-frequency viewing and modification by multiple users concurrently, ensuring smooth use during peak business periods; the file corresponding to the warm data H2 is stored on a medium-performance storage medium, supporting millisecond-level access response and regular viewing and modification, meeting the stable needs of daily business; the file corresponding to the cold data H3 is stored on a low-performance storage medium, supporting on-demand retrieval.

[0023] Step S301: Continuously monitor the access behavior of the drug management documents to achieve automatic dynamic adjustment of the popularity level, so that the storage strategy can adapt to the changes in real business needs and achieve a dynamic balance between access efficiency and storage resources. Set fixed calculation parameters, including the access behavior statistics time window T. windowPopularity threshold; the access behavior statistics time window T window The fixed statistical period used to calculate the average monthly access frequency of drug management documents is a fundamental parameter for determining dynamic shifts and downgrades in popularity. The popularity threshold includes the access frequency threshold f. th1 =a1 times / month, f th2 =a2 times / month, a1 < a2, inaccessible duration threshold ΔT th1 =b1 month, ΔT th2 =b2 months, b1 < b2; the threshold f th1 f th2 ΔT th1 ΔT th2 Set up by professionals; Calculate the access frequency f: ; In the formula, f represents the average monthly access frequency of the file, measured in times per month; N represents the statistical time window T. window The cumulative number of times users have accessed the current drug management documents within the system; Step S302: Based on the calculation results, perform dynamic popularity transition and downgrade determination, specifically: when the current popularity level of the file type F is cold data H3 and the average monthly access frequency f is greater than the access frequency threshold f th1 When the current file type F's popularity level is automatically changed to warm data H2; when the current popularity level of file type F is warm data H2 and the average monthly access frequency f is greater than the access frequency threshold f th2 When the time comes, the popularity level of the current file type F will automatically jump to hot data H1; When the current popularity level of the file type F is hot data H1 and the unaccessed time interval ΔT is greater than the unaccessed duration threshold ΔT th1 When the current file type F's popularity level is automatically downgraded to warm data H2; when the current popularity level of file type F is warm data H2 and the unaccessed time interval ΔT is greater than the unaccessed duration threshold ΔT th2 When this happens, the popularity level of the current file type F will be automatically downgraded to cold data H3.

[0024] Step S401: Based on metadata, file access frequency f, and popularity level H, construct a quantitative scoring model that adapts to different drug management document lifecycle stages, continuously monitor each drug management document at fixed intervals, and complete the quantitative scoring. The preset model has core quantization parameters, including the access frequency contribution weight W. f Popularity level matching weight W h Popularity level weight M N A set of lifecycle determination thresholds, wherein M NThe weight values ​​correspond to different heat levels. When N=1, it corresponds to the weight of hot data H1; when N=2, it corresponds to the weight of warm data H2; and when N=3, it corresponds to the weight of cold data H3. The life cycle determination threshold set includes determination threshold S1 for the research and development clinical trial period L1, determination threshold S2 for the registration application period L2, determination threshold S3 for the market production period L3, and determination threshold S4 for the production stoppage and withdrawal period L4. After the parameter preset is completed, the window T... window As a statistical period, the score S of the drug management document is calculated. i : ; In the formula, S i c1 represents the importance score of the i-th drug life cycle stage L in the current drug management document, c2 represents the number of file type F in the hot data H1 in the current drug management document, c3 represents the number of file type F in the warm data H2 in the current drug management document, and c3 represents the number of file type F in the cold data H3 in the current drug management document. Step S402: Based on the score S i The drug management documents are judged and subject to mandatory popularity control to ensure the access efficiency of high-frequency and high-importance documents. When the drug lifecycle stage L in the drug management document is the research and clinical trial phase L1, if S i >S1, the current drug management document is identified as a key focus document, and the popularity level (H) of all documents under it is adjusted to H1; When the drug lifecycle stage L in the drug management document is the registration application period L2, if S i >S2, the current drug management document is identified as a key focus document, and the popularity level (H) of all document types under it is adjusted to H1; When the drug lifecycle stage L in the drug management document is the market production stage L3, if S i >S3, the current drug management document is identified as a key focus document, and the popularity level (H) of all document types under it is adjusted to H1; When the drug lifecycle stage L in the drug management document is the discontinuation and withdrawal period L4, if S i >S4, the current drug management document is identified as a key focus document, and the popularity level (H) of all document types under it is adjusted to H1.

[0025] The compliance document management system based on big data is characterized by the following: the system includes a data acquisition and initial classification module, a compliance verification and storage mapping module, a dynamic heat adjustment module, and a quantitative assessment and mandatory control module. The data acquisition and initial classification module is responsible for sorting out information and collecting metadata from all drug management documents, covering core fields such as drug life cycle stage, file type, timestamp, expiration date, compliance retention period and special identifiers. Based on the collected metadata, it calculates the unaccessed time interval and classifies the files into initial popularity levels according to preset rules, providing a basis for subsequent storage strategies. The data acquisition and initial classification module includes a metadata acquisition unit and an initial heat classification unit; The metadata collection unit performs unified information sorting and standardized definition of all drug management documents, specifically: collecting the drug lifecycle stage L, file type F, and file last access time T. last The time T of existence of drug management documents current Drug expiration date T expiry compliant retention period T retention Special drug identifier K, audit exemption identifier A, calculate the unaccessed time interval ΔT after data collection is completed; The initial heat classification unit classifies drug management documents into initial heat levels, defining heat levels H as hot data H1, warm data H2, and cold data H3. It executes the initial heat judgment rules and assigns initial heat levels to various types of documents at different stages according to the correspondence between drug life cycle stage L and document type F, thereby achieving preliminary classification that matches the needs of business scenarios.

[0026] The compliance verification and storage mapping module performs compliance verification and correction on the initial heat determination results, and sequentially performs three correction operations: audit exemption lock, special drug control, and retention period compliance, to ensure that the heat classification meets regulatory requirements. The module maps the final determined heat level to the corresponding performance storage medium, realizing differentiated deployment of high-performance storage for hot data, medium-performance storage for warm data, and low-performance storage for cold data. The compliance verification and storage mapping module includes a compliance correction unit and a storage mapping unit; The compliance correction unit performs three correction operations on the initial heat level determination result: when the audit exemption identifier A is A1, it forcibly locks the heat level of all file types F in the current drug management documents to hot data H1; when the special drug identifier K is K1, it forcibly locks the heat level of all file types F to hot data H1; for batch record file F2, if the existence time is less than the compliance retention period, the heat level is set to be no lower than warm data H2. The storage mapping unit stores the file type F corresponding to hot data H1 on a high-performance storage medium, supporting second-level access response and high-frequency operation by multiple users; the file corresponding to warm data H2 is stored on a medium-performance storage medium, supporting millisecond-level access and regular browsing; the file corresponding to cold data H3 is stored on a low-performance storage medium, supporting on-demand retrieval, thereby achieving optimized configuration of storage resources and access efficiency.

[0027] The dynamic heat adjustment module continuously monitors the access behavior of drug management documents to achieve automatic dynamic adjustment of heat level. It sets access behavior statistical time window and heat jump threshold, calculates access frequency, and executes upward or downward heat jump based on the comparison relationship between the current heat level, access frequency, and non-access duration, so that the storage strategy fits the changes in real business needs. The dynamic heat adjustment module includes an access frequency calculation unit and a transition / degradation determination unit. The access frequency calculation unit is set with fixed calculation parameters, including the access behavior statistics time window T. window Access frequency threshold f th1 and f th2 Unaccessed duration threshold ΔT th1 and ΔT th2 Calculate the average monthly access frequency f to provide a quantitative basis for dynamic adjustment of popularity; The transition degradation determination unit performs dynamic adjustment of heat based on the calculation results, specifically: when f > f th1 When f > f, cold data H3 transitions to warm data H2; th2 When ΔT > ΔT, the temperature data H2 transitions to the thermal data H1; th1 When ΔT > ΔT, thermal data H1 is downgraded to temperature data H2; th2 At that time, the temperature data H2 is downgraded to the cold data H3, realizing the automatic dynamic adjustment of the heat level.

[0028] The quantitative assessment and mandatory control module constructs a quantitative scoring model adapted to different drug lifecycle stages. Based on metadata, file access frequency, and popularity level, it comprehensively scores drug management files at fixed intervals. According to the comparison between the scoring results and the judgment thresholds of each stage, it implements mandatory popularity control, identifies files that exceed the threshold as key files of concern, and forcibly adjusts the popularity of all file types under them to hot data H1, so as to ensure the access efficiency of high-frequency and high-importance files.

[0029] The quantitative assessment and mandatory control module includes a quantitative scoring unit and a mandatory control unit; The quantization scoring unit presets core quantization parameters for the model, including the access frequency contribution weight W. f Popularity level matching weight Wh Popularity level weight M N A set of lifecycle determination thresholds, denoted by T window For the statistical period, calculate the file score S. i The life cycle determination threshold set includes S1, S2, S3, and S4; The mandatory control unit is based on a scoring system (S). i The drug management documents at each stage of the life cycle are determined as follows: when L is L1, S i When L is >S1, the popularity level of all document types F in the drug management documents is adjusted to H1; when L is L2, S i When L is >S2, the popularity level of all document types F in the drug management documents is adjusted to H1; when L is L3, S i When L is >S3, the popularity level of all document types F in the drug management documents is adjusted to H1; when L is L4, S i >S4, the F popularity level of all file types of the drug management documents is adjusted to H1 to ensure the efficiency of accessing high-frequency and high-importance documents.

[0030] Example: Taking file "BPR-2024-001" as an example, the metadata collected is set as follows: Drug life cycle stage L: L3 (marketing and production period), File type F: F2 (batch record file), File last access time T last The file existed on March 1, 2025, for a period of time T. current The expiration date of the drug is March 18, 2026. expiry : December 31, 2025, compliance retention period T retention :T expiry + 5 years (statutory retention period), special drug label K: K0 (ordinary drug), audit exemption label A: A0 (no special audit status); calculate the time interval without access: ΔT = T current - T last ≈12.5 months; Based on the file's lifecycle stage L3 and file type F2, the initial popularity determination rules are as follows: For L3 stage, F2 (batch record file) is initially determined as: warm data H2; Perform the following three correction operations in sequence: Set the document audit exemption flag A to A0, without triggering locking; set the document special drug flag K to K0, without triggering locking; and adjust the statutory retention period T. retention = 2025-12-31 + 5 years = 2030-12-31, set the current time T. current =2026-01-11, less than T retentionSince the document has entered the production suspension period and meets compliance requirements, its popularity must not be lower than the warm data H2, so H2 is maintained. The current popularity of document BPR-2024-001 is H2, stored on medium-performance storage medium, supporting millisecond-level access, which meets the needs of daily viewing and modification. Set the following fixed calculation parameters: statistical time window T window =3 months, access frequency threshold: f th1 = 2 times / month, f th2 =5 visits / month, inaccessibility time threshold: Δ Tth1 =6 months, Δ Tth2 =12 months; Calculate the average monthly access frequency of this file in the past 3 months: f = cumulative access count / 3 = 1 time / 3 = 0.33 times / month; The current popularity level is H2, and the average monthly visit frequency is f = 0.33 < f th2 The transition conditions are not met; the unvisited time interval ΔT = 12.5 months > ΔT th2 =12 months, meets the downgrade criteria, therefore the popularity level is automatically downgraded to cold data H3; Preset model parameters: W f =0.4, W h =0.6, M1 (H1 weight) = 5, M2 (H2 weight) = 3, M3 (H3 weight) = 1, lifecycle determination threshold: S1 = 20, S2 = 18, S3 = 15, S4 = 10; Taking files as units, statistically analyze the distribution of all file types under the current lifecycle stage L3, setting: c1 = 0, c2 = 5, c3 = 3; calculate S i =10.932; Current lifecycle stage is L3, S i =10.932<S3=15, forced heat boost was not triggered, and the file remains cold data H3.

[0031] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A compliance document management method based on big data, characterized by: The method includes the following steps: Step S100: Collect metadata of drug management documents, including drug lifecycle stage L and file type F, and calculate the unaccessed time interval ΔT; combine the metadata and, according to the matching of L and F, complete the initial heat level classification of hot data H1, warm data H2 and cold data H3 of file type F; Step S200: Perform three checks and corrections sequentially on the initial heat level classification results: audit exemption lock, special drug control, and retention period compliance; according to the corrected heat level, store the corresponding files on high-performance, medium-performance, and low-performance storage media respectively. Step S300: Set fixed calculation parameters and calculate the average monthly access frequency f of the file; based on the calculation results and a preset threshold, perform dynamic transition and downgrade of the file popularity level; Step S400: Preset quantitative parameters and calculate the importance score S of the drug management document corresponding to the life cycle stage at a fixed period. i Based on S i Based on the corresponding judgment threshold, determine the files to be of key concern, and forcibly adjust the file type F popularity level of the files of key concern to hot data H1.

2. The compliance document management method based on big data according to claim 1, characterized in that: Step S100 includes the following steps: Step S101: Collect metadata for all drug management documents, including drug lifecycle stage L, file type F, and last access time T. last The time T of existence of drug management documents current Drug expiration date T expiry compliant retention period T retention Special drug identification mark K, audit exemption mark A; the drug life cycle stage L includes research and clinical trial stage L1, registration application stage L2, market production stage L3, and production stoppage and withdrawal stage L4; The document type F includes core control documents F1, batch record documents F2, process documents F3, R&D archive documents F4, and registration application documents F5. Core control documents F1 refer to current and historical versions of SOPs, process procedures, and quality standard documents; batch record documents F2 refer to batch production records (BPR) and batch inspection records (BIR); process documents F3 refer to drug deviation handling documents and CAPA documents. When the special drug identifier K=K1, it refers to narcotic drugs, psychotropic drugs, vaccines, and medical toxic drugs; when the special drug identifier K=K0, it refers to ordinary drugs. When the audit exemption identifier A=A1, it indicates that the current document's corresponding product is under GMP inspection, undergoing registration supplementation, or is an audit concern; when the audit exemption identifier A=A0, it indicates that the aforementioned status does not exist. After data collection, the unaccessed time interval ΔT of the drug management documents is calculated: ΔT=T current -T last ; Step S102: Perform initial popularity level classification on the drug management documents, specifically: define the popularity level H of document type F, where the value of popularity level H is: hot data H1, warm data H2, and cold data H3; combine the metadata from step S101 to perform initial popularity determination: When the drug life cycle stage L is the research and development clinical trial stage L1, the initial popularity of the core control document F1, research and development archive document F4, and registration application document F5 is determined as hot data H1, and the initial popularity of the batch record document F2 and process document F3 is determined as hot data H1. When the drug life cycle stage L is the registration application period L2, the initial popularity of the registration application document F5 and the core control document F1 is determined to be hot data H1, and the R&D archive document F4, batch record document F2, and process document F3 are determined to be warm data H2. When the drug life cycle stage L is the market production stage L3, the initial popularity of the registration application document F5, core control document F1, batch record document F2, and process document F3 is determined to be warm data H2, and the initial popularity of the R&D archive document F4 is determined to be cold data H3. When the drug life cycle stage L is the production stoppage and withdrawal period L4, the initial popularity of the registration application document F5 is determined to be warm data H2, and the initial popularity of all other document types F is determined to be cold data H3.

3. The compliance document management method based on big data according to claim 1, characterized in that: Step S200 includes the following steps: Step S201: Correct the initial heat determination result completed in step S102 by performing three correction operations in sequence: The first step is to perform an audit exemption lock verification. When the audit exemption identifier A corresponding to the drug management document is A1, the popularity of all document types F in the current drug management document is forcibly locked to hot data H1 until the value of the audit exemption identifier A changes to A0. The second step is to perform special drug control verification. When the special drug identifier K corresponding to the drug management document is K1, the heat of all file types F in the current drug management document will be forcibly locked to hot data H1. Thirdly, perform a retention period compliance verification and calculate the statutory compliance retention period T corresponding to the document. retention :T retention =T expiry +T extra In the formula, T extra The statutory retention period for current drug management documents; for the batch record file F2, when the current drug management document exists for time T current Less than the compliant retention period T retention At that time, the heat level of the drug management document shall not be lower than the temperature data H2; Step S202: The file type F corresponding to the hot data H1 is stored in a high-performance storage medium; the file corresponding to the warm data H2 is stored in a medium-performance storage medium; and the file corresponding to the cold data H3 is stored in a low-performance storage medium.

4. The compliance document management method based on big data according to claim 1, characterized in that: Step S300 includes the following steps: Step S301: Continuously monitor the access behavior of the drug management documents, and set fixed calculation parameters, including the access behavior statistical time window T. window Popularity threshold; the popularity threshold includes the access frequency threshold f. th1 =a1 times / month, f th2 =a2 times / month, a1 < a2, inaccessible duration threshold ΔT th1 =b1 month, ΔT th2 =b2 months, b1 < b2; the threshold f th1 f th2 ΔT th1 ΔT th2 Set by professionals, based on the time window T window Calculate the access frequency f; Step S302: Based on the access frequency f, perform dynamic popularity transition and degradation determination, specifically: when the current popularity level of the file type F is cold data H3 and the average monthly access frequency f is greater than the access frequency threshold f th1 When the current file type F's popularity level is automatically changed to warm data H2; when the current popularity level of file type F is warm data H2 and the average monthly access frequency f is greater than the access frequency threshold f th2 When the time comes, the popularity level of the current file type F will automatically jump to hot data H1; When the current popularity level of the file type F is hot data H1 and the unaccessed time interval ΔT is greater than the unaccessed duration threshold ΔT th1 When the current file type F's popularity level is automatically downgraded to warm data H2; when the current popularity level of file type F is warm data H2 and the unaccessed time interval ΔT is greater than the unaccessed duration threshold ΔT th2 When this happens, the popularity level of the current file type F will be automatically downgraded to cold data H3.

5. The compliance document management method based on big data according to claim 1, characterized in that: Step S400 includes the following steps: Step S401: Based on the metadata, file access frequency f, and popularity level H, construct a quantitative scoring model adapted to different lifecycle stages of drug management documents to complete the quantitative scoring, specifically as follows: The preset model has core quantization parameters, including the access frequency contribution weight W. f Popularity level matching weight W h Popularity level weight M N A set of lifecycle determination thresholds, wherein M N The weight values ​​correspond to different heat levels. When N=1, it corresponds to the weight of hot data H1; when N=2, it corresponds to the weight of warm data H2; and when N=3, it corresponds to the weight of cold data H3. The life cycle determination threshold set includes determination threshold S1 for the research and development clinical trial period L1, determination threshold S2 for the registration application period L2, determination threshold S3 for the market production period L3, and determination threshold S4 for the production stoppage and withdrawal period L4. After the parameter preset is completed, the window T... window As a statistical period, the score S of the drug management document is calculated. i ; Step S402: Based on the score S i The aforementioned drug management documents are subject to judgment and mandatory popularity control, specifically as follows: When the drug lifecycle stage L in the drug management document is the research and clinical trial phase L1, if S i >S1, the current drug management document is identified as a key focus document, and the popularity level (H) of all document types (F) of the drug management document is adjusted to H1; When the drug lifecycle stage L in the drug management document is the registration application period L2, if S i >S2, the current drug management document is identified as a key focus document, and the popularity level (H) of all document types (F) of the drug management document is adjusted to H1; When the drug lifecycle stage L in the drug management document is the market production stage L3, if S i >S3, the current drug management document is identified as a key focus document, and the popularity level (H) of all document types (F) of the drug management document is adjusted to H1; When the drug lifecycle stage L in the drug management document is the discontinuation and withdrawal period L4, if S i >S4, the current drug management document is identified as a key focus document, and the popularity level (H) of all document types (F) of the drug management document is adjusted to H1.

6. A compliance document management system based on big data, characterized by: The system includes a data acquisition and initial classification module, a compliance verification and storage mapping module, a dynamic heat adjustment module, and a quantitative assessment and mandatory control module. The data acquisition and initial classification module is responsible for sorting out information and collecting metadata from all drug management documents, calculating the unaccessed time interval based on the collected metadata, and classifying the documents into initial popularity levels according to preset rules. The compliance verification and storage mapping module performs compliance verification and correction on the initial heat level determination result, and sequentially performs three correction operations: audit exemption lock, special drug control, and retention period compliance, and maps the final determined heat level to the corresponding performance storage medium. The dynamic heat adjustment module continuously monitors the access behavior of drug management documents, sets the access behavior statistical time window and heat jump threshold, calculates the access frequency, and executes the upward or downward heat jump based on the comparison relationship between the current heat level and the access frequency and the non-access time. The quantitative assessment and mandatory control module constructs a quantitative scoring model adapted to different drug life cycle stages. Based on metadata, file access frequency and popularity level, it comprehensively scores drug management documents at fixed intervals. Based on the comparison of the scoring results with the judgment thresholds of each stage, it implements mandatory popularity control.

7. The big data-based compliance document management system according to claim 6, characterized in that: The data acquisition and initial classification module includes a metadata acquisition unit and an initial heat classification unit; The metadata collection unit performs unified information sorting and standardized definition of all drug management documents, specifically: collecting the drug lifecycle stage L, file type F, and file last access time T. last The time T of existence of drug management documents current Drug expiration date T expiry compliant retention period T retention Special drug identifier K, audit exemption identifier A, calculate the unaccessed time interval ΔT after data collection is completed; The initial heat level classification unit classifies drug management documents into initial heat level categories, defining heat level H as hot data H1, warm data H2, and cold data H3. It executes the initial heat level judgment rule and assigns an initial heat level to various types of documents at different stages according to the correspondence between drug life cycle stage L and document type F.

8. The big data-based compliance document management system according to claim 6, characterized in that: The compliance verification and storage mapping module includes a compliance correction unit and a storage mapping unit; The compliance correction unit performs three correction operations on the initial heat level determination result: when the audit exemption identifier A is A1, it forcibly locks the heat level of all file types F in the current drug management documents to hot data H1; when the special drug identifier K is K1, it forcibly locks the heat level of all file types F to hot data H1; for batch record file F2, if the existence time is less than the compliance retention period, the heat level is set to be no lower than warm data H2. The storage mapping unit stores the file type F corresponding to hot data H1 in a high-performance storage medium; the file corresponding to warm data H2 in a medium-performance storage medium; and the file corresponding to cold data H3 in a low-performance storage medium.

9. The compliance document management system based on big data according to claim 6, characterized in that: The dynamic heat adjustment module includes an access frequency calculation unit and a transition / degradation determination unit. The access frequency calculation unit is set with fixed calculation parameters, including the access behavior statistics time window T. window Access frequency threshold f th1 and f th2 Unaccessed duration threshold ΔT th1 and ΔT th2 Calculate the average monthly access frequency f; The transition degradation determination unit performs dynamic adjustment of heat based on the calculation results, specifically: when f > f th1 When f > f, cold data H3 transitions to warm data H2; th2 When ΔT > ΔT, the temperature data H2 transitions to the thermal data H1; th1 When ΔT > ΔT, thermal data H1 is downgraded to temperature data H2; th2 At that time, the warm data H2 was downgraded to the cold data H3.

10. The big data-based compliance document management system according to claim 6, characterized in that: The quantitative assessment and mandatory control module includes a quantitative scoring unit and a mandatory control unit; The quantization scoring unit presets core quantization parameters for the model, including the access frequency contribution weight W. f Popularity level matching weight W h Popularity level weight M N A set of lifecycle determination thresholds, denoted by T window For the statistical period, calculate the file score S. i The life cycle determination threshold set includes S1, S2, S3, and S4; The mandatory control unit is based on a scoring system (S). i The drug management documents at each stage of the life cycle are determined as follows: when L is L1, S i When L is >S1, the popularity level of all document types F in the drug management documents is adjusted to H1; when L is L2, S i When L is >S2, the popularity level of all document types F in the drug management documents is adjusted to H1; when L is L3, S i When L is >S3, the popularity level of all document types F in the drug management documents is adjusted to H1; when L is L4, S i When >S4, the popularity level of all file types F of the drug management documents is adjusted to H1.

Images