Method for intelligent management of medical health data based on platform

Abstract

The application relates to the technical field of medical data processing, and discloses a full-life-cycle intelligent management method based on a medical health data middle platform, which comprises preferentially caching hot storage data, monitoring hot data access heat change in combination with a digital twin model, dynamically adjusting hot data cache resources, and releasing the resources to a cold data storage link; the dynamic storage strategy based on multi-dimensional heat analysis can accurately identify hot data and cold data, high-frequency access data is preferentially stored in a cache to guarantee business response speed, low-frequency cold data is timely migrated out, invalid occupation of high-performance storage resources is significantly reduced, and the overall cost structure is optimized. On this basis, an adaptive compression method for cold data automatically distinguishes key and non-key information through clinical natural language processing and a knowledge graph, lossless compression of key medical entities is realized, and the integrity and compliance of diagnosis and treatment semantics are ensured.

Description

Technical Field

[0001] This invention relates to the field of medical data processing technology, and in particular to a method for intelligent governance of the entire lifecycle based on a medical and health data platform. Background Technology

[0002] With the continuous development of healthcare informatization, medical and health data is experiencing explosive growth, with data sources covering numerous channels such as hospital information systems, electronic medical records, medical imaging, wearable devices, and public health monitoring. While the convergence of massive amounts of heterogeneous data from multiple sources brings rich analytical value, it also poses serious challenges to the storage, management, and governance of data platforms.

[0003] Existing data platforms generally suffer from the following shortcomings when processing healthcare data: First, they lack effective hierarchical identification and dynamic adjustment mechanisms for hot and cold data. All data is stored indiscriminately in a unified storage layer, resulting in high-speed storage resources being occupied by a large amount of infrequently accessed cold data, while frequently accessed hot data cannot obtain sufficient caching resources, leading to a severe imbalance between storage costs and access efficiency. Second, in the data compression and archiving stage, a simple full compression strategy is usually adopted, failing to distinguish between critical and non-critical information in the data, and thus failing to achieve the highest possible compression ratio while ensuring the integrity of critical business information.

[0004] To address the aforementioned technical deficiencies, a solution is proposed. Summary of the Invention

[0005] The purpose of this invention is to address the problem that storing all data indiscriminately in a unified storage layer leads to high-speed storage resources being occupied by a large amount of low-frequency accessed cold data, while high-frequency accessed hot data cannot obtain sufficient caching resources, resulting in a serious imbalance between storage costs and access efficiency. At the same time, in the data compression and archiving stage, a simple full compression strategy is usually adopted, which does not distinguish between key and non-key information in the data, greatly reducing storage efficiency.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a full-lifecycle intelligent governance method based on a medical and health data platform, comprising the following steps: Step 1: Collect multi-source medical and health data and perform standardized preprocessing. At the same time, build a digital twin model covering the entire data lifecycle to map the data status and storage resource status in real time. Step 2: Based on the digital twin model, obtain popularity data from the database, perform analysis and calculation to obtain the popularity value of the data to be stored, and divide the data to be stored into hot storage data and cold storage data, automatically triggering the dynamic storage strategy of medical and health data. Step 3: Prioritize caching of hot-storage data, monitor changes in the access frequency of hot data using a digital twin model, dynamically adjust hot data caching resources, and allocate the released resources to the cold data storage stage. Step 4: Identify key information in the cold data and use an adaptive compression algorithm to process the cold data, performing lossless compression on key information and lossy compression with a high compression ratio on non-key information. Step 5: Utilize the digital twin model to monitor the quality of the entire data process and establish a data security governance mechanism to ensure data security; Step Six: Based on the governed data, conduct value mining, and transmit the application feedback back to the digital twin model to optimize the governance strategy, forming a governance closed loop.

[0007] Furthermore, the calculation process for the data popularity value of the data to be stored is as follows: S11. By acquiring and analyzing popularity data, the popularity data includes data access frequency, data freshness, and business importance weight. The data access frequency is obtained by counting the number of times the medical and health data is queried or accessed within a unit time window. The data freshness is obtained by taking the reciprocal of the time interval between the last access time of the medical and health data and the current time; the more recent the access, the higher the freshness. The business importance weight is quantified by the medical knowledge graph information associated with the medical and health data, such as the diagnosis level and medication type. The larger the value of the business importance weight, the higher the importance of the corresponding business. S12. Calculate the data popularity value of the data to be stored according to the following formula. : in, For the frequency of access to medical and health data, For the time freshness of medical and health data, Weighting healthcare data according to its business importance. The preset standard frequency of access to medical and health data. To set the standard for the freshness of medical and health data, The preset frequency weighting coefficients, The data popularity value of the data to be stored is used to reflect the usage popularity of the medical and health data to be stored, with a preset freshness weighting coefficient. S13. Obtain the preset data popularity threshold. The data popularity value of the data to be stored Comparative analysis, when If the data is in high demand, it indicates that the medical and health data to be stored is frequently used and will be classified as hot storage data. S14, when If the usage rate is low, it indicates that the medical and health data to be stored is low and will be classified as cold storage data.

[0008] Furthermore, the dynamic storage strategy for medical and health data is as follows: hot-stored data enters the high-speed cache layer for priority caching, while cold-stored data enters the subsequent compression and archiving process. This is used to realize a dynamic storage strategy that prioritizes the caching of hot data and combines it with the compression and archiving of cold data.

[0009] Furthermore, the process of dynamically adjusting hot data cache resources is as follows: S21. Objects identified as hot storage data are loaded into the cache area, while the digital twin model continuously tracks the access behavior of each piece of hot data and updates its heat value. S22. Calculate the change in the heat of stored data according to the following formula. : ,in, To store the current popularity value of the data, To store the popularity value of the data object at the end of the previous monitoring period; S23. Obtain the preset heat degradation tolerance range. When the heat value of the hot storage data is lower than the set degradation threshold for multiple consecutive monitoring cycles, and When the value remains negative, the data is remarked as cold storage data and removed from the cache. The freed-up cache resources are then dynamically allocated to the storage of cold storage data to expand the concurrent processing capability of the compression and archiving task. S24. The allocation of hot cache resources is based on the proportion of the current popularity value of each hot storage data in the total popularity of all hot storage data. The basic cache quota allocated to the j-th hot data is... It can be represented as: ,in, This represents the total capacity of the hot buffer. This represents the total number of current hot data points. Let j be the heat value of the j-th hot storage data. This represents the total heat value of all hot-stored data within the hot buffer.

[0010] Furthermore, the process of lossless compression of critical information and lossy compression of non-critical information with high compression ratios is as follows: S31. Utilizing a parsing engine based on clinical natural language processing and knowledge graphs, automatically extract key medical entities from cold-stored data, and accordingly mark data blocks as key blocks and non-key blocks, defining binary identifiers. If data block If identified as containing a key medical entity, then ,otherwise ; S32, Start Adaptive Compression: For For critical data blocks, lossless compression is used to ensure complete preservation of medical semantic information; for Non-critical data blocks are processed using lossy compression with a high compression ratio; S33. Calculate the compression ratio of lossy compression according to the following formula. : ,in, This is the preset maximum compression ratio for lossy compression. This is the preset minimum compression ratio for lossy compression. The data importance score is a dynamic variable calculated in real time using natural language processing and a rule engine. Natural language processing technology scans cold-stored data content to identify key and non-key information features. Based on a pre-defined medical knowledge graph, different weights are assigned to different features. The weights of the identified features are then weighted and summed to obtain the data importance score for the current data block. The highest priority score is preset; S34. Perform adaptive compression based on the compression ratio of lossy compression, and write the data after hybrid compression to a large-capacity cold storage medium for archiving, while updating the storage location and occupied space information in the digital twin model.

[0011] Furthermore, end-to-end data quality monitoring includes calculating data integrity, accuracy, and consistency indicators at each node of data flow, and data security governance mechanisms include data anonymization, fine-grained access control, and operational auditing.

[0012] Furthermore, application feedback includes query patterns, access hotspots, and governance effectiveness indicators revealed by data analysis tasks.

[0013] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: This intelligent governance method based on the entire lifecycle of a healthcare data platform employs a dynamic storage strategy based on multi-dimensional heat analysis to accurately identify hot and cold data. It prioritizes high-frequency access data for high-speed caching to ensure business response speed, while promptly migrating low-frequency cold data, significantly reducing the ineffective use of high-performance storage resources and optimizing the overall cost structure. Furthermore, an adaptive compression method for cold data automatically distinguishes between critical and non-critical information through clinical natural language processing and knowledge graphs. This achieves lossless compression of critical medical entities to ensure the integrity and compliance of diagnostic semantics, while simultaneously applying high-compression-ratio lossy compression to non-critical information using a dynamic compression ratio based on data importance scoring. This significantly saves archival storage space and improves storage efficiency while ensuring the security and availability of core business data. In addition, the end-to-end data quality monitoring and fine-grained security governance mechanism jointly enhance the integrity, accuracy and consistency of data, and effectively prevent the risk of sensitive information leakage. Finally, by continuously transmitting application feedback back to the digital twin model to form a governance closed loop, the data heat judgment, resource allocation strategy and compression algorithm parameters can be adaptively optimized, and then continuously iterate to achieve a long-term, efficient, secure and low-cost operation of the medical data platform. Attached Figure Description

[0014] Figure 1 A schematic diagram of the method flow of the present invention is shown. Detailed Implementation

[0015] 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.

[0016] Example: like Figure 1 As shown, based on the intelligent governance method of the entire lifecycle of medical and health data platform, firstly, multi-source medical and health data is collected and standardized preprocessed. At the same time, a digital twin model covering the entire lifecycle of data is constructed to map the data status and storage resource status in real time. It should be noted that the standardized preprocessing includes cleaning, denoising and standardizing terminology mapping of multi-source medical and health data to form a standardized data representation. The digital twin model describes the storage location, access characteristics and resource consumption of medical data entities in real time through data status mapping functions.

[0017] Then, based on the digital twin model, the heat data is obtained from the database, analyzed and calculated to obtain the heat value of the data to be stored, and the data to be stored is divided into hot storage data and cold storage data, which automatically triggers the dynamic storage strategy of medical and health data. The calculation process for the data popularity value of the data to be stored is as follows: S11. By acquiring and analyzing popularity data, the popularity data includes data access frequency, data freshness, and business importance weight. The data access frequency is obtained by counting the number of times the medical and health data is queried or accessed within a unit time window. The data freshness is obtained by taking the reciprocal of the time interval between the last access time of the medical and health data and the current time; the more recent the access, the higher the freshness. The business importance weight is quantified by the medical knowledge graph information associated with the medical and health data, such as the diagnosis level and medication type. The larger the value of the business importance weight, the higher the importance of the corresponding business. S12. Calculate the data popularity value of the data to be stored according to the following formula. : in, For the frequency of access to medical and health data, For the time freshness of medical and health data, Weighting healthcare data according to its business importance. The preset standard frequency of access to medical and health data. To set the standard for the freshness of medical and health data, The preset frequency weighting coefficients, The preset freshness weighting coefficient is used to reflect the usage popularity of the medical and health data to be stored. The larger the value of the popularity value, the higher the usage popularity of the medical and health data to be stored. The smaller the value of the popularity value, the lower the usage popularity of the medical and health data to be stored. S13. Obtain the preset data popularity threshold. The data popularity value of the data to be stored Comparative analysis, when If the data is in high demand, it indicates that the medical and health data to be stored is frequently used and will be classified as hot storage data. S14, when If the usage rate is low, it indicates that the medical and health data to be stored is low and will be classified as cold storage data.

[0018] The dynamic storage strategy for medical and health data is as follows: hot-stored data enters the high-speed cache layer for priority caching, while cold-stored data enters the subsequent compression and archiving process. This is used to implement a dynamic storage strategy that prioritizes the caching of hot data and combines it with the compression and archiving of cold data.

[0019] Then, hot-stored data is cached first, and the hot data access frequency is monitored by combining the digital twin model. The hot data cache resources are dynamically adjusted, and the released resources are allocated to the cold data storage stage. The process of dynamically adjusting hot data cache resources is as follows: S21. Objects identified as hot storage data are loaded into a high-speed cache built from solid-state drives or memory. Meanwhile, the digital twin model continuously tracks the access behavior of each piece of hot data and updates its heat value. S22. Calculate the change in the heat of stored data according to the following formula. : ,in, To store the current popularity value of the data, To store the popularity value of the data object at the end of the previous monitoring period; S23. Obtain the preset heat degradation tolerance range. When the heat value of the hot storage data is lower than the set degradation threshold for multiple consecutive monitoring cycles, and When the value remains negative, the data is remarked as cold storage data and removed from the cache. The freed-up cache resources are then dynamically allocated to the storage of cold storage data to expand the concurrent processing capability of the compression and archiving task. S24. The allocation of hot cache resources is based on the proportion of the current popularity value of each hot storage data in the total popularity of all hot storage data. The basic cache quota allocated to the j-th hot data is... It can be represented as: ,in, This represents the total capacity of the hot buffer. This represents the total number of current hot data points. Let j be the heat value of the j-th hot storage data. This represents the total popularity value of all hot-stored data within the hot cache. This method ensures that data with higher popularity receives more cache space, effectively guaranteeing the performance of high-frequency access in medical services.

[0020] Next, key information in the cold data is identified, and an adaptive compression algorithm is used to process the cold data, performing lossless compression on key information and lossy compression with a high compression ratio on non-key information. The process of lossless compression of critical information and lossy compression of non-critical information with high compression ratio is as follows: S31. Utilize a parsing engine based on clinical natural language processing and knowledge graphs to automatically extract key medical entities (such as diagnostic conclusions, drug names, surgical procedures, pathology results, and allergy information) from cold-stored data, and mark data blocks as key blocks and non-key blocks accordingly, defining binary identifiers. If data block If identified as containing a key medical entity, then ,otherwise ; S32, Start Adaptive Compression: For For critical data blocks, lossless compression is used to ensure complete preservation of medical semantic information; for For non-critical data blocks, high compression ratio lossy compression is used to process them, which significantly reduces the data volume while meeting the acceptable error requirements of the business. S33. Calculate the compression ratio of lossy compression according to the following formula. : ,in, This is the preset maximum compression ratio for lossy compression. This is the preset minimum compression ratio for lossy compression. Data importance scoring is a dynamic variable calculated in real time using natural language processing and a rule engine. Natural language processing technology scans cold-stored data content to identify key and non-key information features (such as routine examinations, background descriptions, and log entries). Based on a pre-defined medical knowledge graph, different weights are assigned to different features. The weighted sum of these identified feature weights yields the data importance score for the current data block. The highest priority score is preset; S34. Perform adaptive compression based on the compression ratio of lossy compression, and write the data after hybrid compression to a large-capacity cold storage medium for archiving, while updating the storage location and occupied space information in the digital twin model.

[0021] Finally, relying on the digital twin model, data quality assessment probes are set up in each stage of data flow—collection, transmission, storage, processing, and application—to calculate data integrity indicators (such as the proportion of non-empty fields), accuracy indicators (such as the proportion of values ​​with compliant ranges), and consistency indicators (such as the consistency of related records) in real time. When a quality indicator is detected to deviate from the preset baseline, an alarm is triggered and an automatic repair or manual intervention workflow is initiated.

[0022] In terms of security governance, sensitive data is anonymized using K-anonymization and differential privacy protection. An attribute-based access control model is established to ensure that only roles with the appropriate permissions can access specific data. All operations are audited and recorded, forming an immutable log record. The security posture is also uniformly presented through a digital twin model.

[0023] Simultaneously, value mining is conducted based on the governed data, and the application feedback is fed back to the digital twin model to optimize the governance strategy, forming a governance closed loop. It should be noted that the application feedback includes query patterns, access hotspots, and governance effectiveness indicators shown by the data analysis tasks.

[0024] This invention employs a dynamic storage strategy based on multidimensional heat analysis to accurately identify hot and cold data. This prioritizes high-frequency access data in the high-speed cache to ensure business response speed, while promptly migrating low-frequency cold data, significantly reducing the ineffective use of high-performance storage resources and optimizing the overall cost structure. Furthermore, an adaptive compression method for cold data automatically distinguishes between critical and non-critical information using clinical natural language processing and knowledge graphs. This achieves lossless compression of critical medical entities to ensure the integrity and compliance of diagnostic semantics. Simultaneously, a dynamic compression ratio based on data importance scoring is used to apply high-compression-ratio lossy compression to non-critical information. This significantly saves archival storage space and improves storage efficiency while ensuring the security and availability of core business data. In addition, end-to-end data quality monitoring and fine-grained security governance mechanisms jointly strengthen data integrity, accuracy, and consistency, effectively preventing the risk of sensitive information leakage. Finally, by continuously feeding application feedback back to the digital twin model to form a governance closed loop, data heat assessment, resource allocation strategies, and compression algorithm parameters are adaptively optimized, continuously iterating to achieve a long-term, efficient, secure, and low-cost operating state for the medical data platform.

[0025] The size of the interval and threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by those skilled in the art for each set of sample data; as long as it does not affect the ratio between the parameter and the quantized value.

[0026] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation. The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A smart governance method based on the entire lifecycle of a medical and health data platform, characterized in that: Includes the following steps: Step 1: Collect multi-source medical and health data and perform standardized preprocessing. At the same time, build a digital twin model covering the entire data lifecycle to map the data status and storage resource status in real time. Step 2: Based on the digital twin model, obtain popularity data from the database, perform analysis and calculation to obtain the popularity value of the data to be stored, and divide the data to be stored into hot storage data and cold storage data, automatically triggering the dynamic storage strategy of medical and health data. Step 3: Prioritize caching of hot-storage data, monitor changes in the access frequency of hot data using a digital twin model, dynamically adjust hot data caching resources, and allocate the released resources to the cold data storage stage. Step 4: Identify key information in the cold data and use an adaptive compression algorithm to process the cold data, performing lossless compression on key information and lossy compression with a high compression ratio on non-key information. Step 5: Utilize the digital twin model to monitor the quality of the entire data process and establish a data security governance mechanism to ensure data security; Step Six: Based on the governed data, conduct value mining, and transmit the application feedback back to the digital twin model to optimize the governance strategy, forming a governance closed loop.

2. The intelligent governance method based on the entire lifecycle of a medical and health data platform according to claim 1, characterized in that, The calculation process for the data popularity value of the data to be stored is as follows: S11. By acquiring and analyzing popularity data, the popularity data includes data access frequency, data freshness, and business importance weight. The data access frequency is obtained by counting the number of times the medical and health data is queried or accessed within a unit time window. The data freshness is obtained by taking the reciprocal of the time interval between the last access time of the medical and health data and the current time; the more recent the access, the higher the freshness. The business importance weight is quantified by the medical knowledge graph information associated with the medical and health data, such as the diagnosis level and medication type. The larger the value of the business importance weight, the higher the importance of the corresponding business. S12. Calculate the data popularity value of the data to be stored according to the following formula. : in, For the frequency of access to medical and health data, For the time freshness of medical and health data, Weighting healthcare data according to its business importance. The preset standard frequency for accessing medical and health data. To set the standard for the freshness of medical and health data, The preset frequency weighting coefficients, The data popularity value of the data to be stored is used to reflect the usage popularity of the medical and health data to be stored, with a preset freshness weighting coefficient. S13. Obtain the preset data popularity threshold. The data popularity value of the data to be stored Comparative analysis, when If the data is in high demand, it indicates that the medical and health data to be stored is frequently used and will be classified as hot storage data. S14, when If the usage rate is low, it indicates that the medical and health data to be stored is low and will be classified as cold storage data.

3. The intelligent governance method based on the entire lifecycle of a medical and health data platform according to claim 1, characterized in that, The dynamic storage strategy for medical and health data is as follows: hot-stored data enters the high-speed cache layer for priority caching, while cold-stored data enters the subsequent compression and archiving process. This is used to implement a dynamic storage strategy that prioritizes the caching of hot data and combines it with the compression and archiving of cold data.

4. The intelligent governance method based on the entire lifecycle of a medical and health data platform according to claim 1, characterized in that, The process of dynamically adjusting hot data cache resources is as follows: S21. The objects identified as hot storage data are loaded into the cache area, while the digital twin model continuously tracks the access behavior of each piece of hot data and updates its heat value. S22. Calculate the change in the heat of stored data according to the following formula. : ,in, To store the current popularity value of the data, To store the popularity value of the data object at the end of the previous monitoring period; S23. Obtain the preset heat degradation tolerance range. When the heat value of the hot storage data is lower than the set degradation threshold for multiple consecutive monitoring cycles, and When the value remains negative, the data is remarked as cold storage data and removed from the cache. The freed-up cache resources are then dynamically allocated to the storage of cold storage data to expand the concurrent processing capability of the compression and archiving task. S24. The allocation of hot cache resources is based on the proportion of the current popularity value of each hot storage data in the total popularity of all hot storage data. The basic cache quota allocated to the j-th hot data is... It can be represented as: ,in, This represents the total capacity of the hot buffer. This represents the total number of current hot data points. Let j be the heat value of the j-th hot storage data. This represents the total heat value of all hot-stored data within the hot buffer.

5. The intelligent governance method based on the entire lifecycle of a medical and health data platform according to claim 1, characterized in that, The process of lossless compression of critical information and lossy compression of non-critical information with high compression ratio is as follows: S31. Utilizing a parsing engine based on clinical natural language processing and knowledge graphs, automatically extract key medical entities from cold-stored data, and accordingly mark data blocks as key blocks and non-key blocks, defining binary identifiers. If data block If identified as containing a key medical entity, then ,otherwise ; S32, Start Adaptive Compression: For For critical data blocks, lossless compression is used to ensure complete preservation of medical semantic information; for Non-critical data blocks are processed using lossy compression with a high compression ratio; S33. Calculate the compression ratio of lossy compression according to the following formula. : ,in, This is the preset maximum compression ratio for lossy compression. This is the preset minimum compression ratio for lossy compression. The data importance score is a dynamic variable calculated in real time using natural language processing and a rule engine. Natural language processing technology scans cold-stored data content to identify key and non-key information features. Based on a pre-defined medical knowledge graph, different weights are assigned to different features. The weighted sum of these identified feature weights yields the data importance score for the current data block. The highest priority score is preset; S34. Perform adaptive compression based on the compression ratio of lossy compression, and write the data after hybrid compression to a large-capacity cold storage medium for archiving, while updating the storage location and occupied space information in the digital twin model.

6. The intelligent governance method based on the entire lifecycle of a medical and health data platform according to claim 1, characterized in that, End-to-end data quality monitoring includes calculating data integrity, accuracy, and consistency indicators at each node of data flow. Data security governance mechanisms include data anonymization, fine-grained access control, and operation auditing.

7. The intelligent governance method based on the entire lifecycle of a medical and health data platform according to claim 1, characterized in that, Application feedback includes query patterns, access hotspots, and governance effectiveness indicators revealed by data analysis tasks.

Images