Intelligent management method and system for medical data

Abstract

The invention discloses an intelligent management method and system for medical data, and relates to the technical field of artificial intelligence, and the method comprises the steps: enabling heterogeneous medical data to generate unified data; after the unified data of each mechanism is encrypted through a homomorphic encryption function, a privacy protection aggregation result of the unified data is calculated, and a privacy protection aggregation result after noise is added is calculated; a matching degree score of the unified data is obtained by calculating the ratio of the number of matched fields to the number of total extracted fields, if the matching degree score is smaller than a set threshold value, abnormal data are marked, and the influence range of the abnormal data on a downstream module is calculated; obtaining an optimal solution of monthly storage cost, average query delay and privacy intensity; based on the influence range and the optimal solution, the storage position and the encryption strength are dynamically selected during operation of the deep Q network, so that intelligent management of the medical data is realized, and the problem that intelligent management of the medical data cannot be realized in the prior art is solved.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to an intelligent management method and system for medical data. Background Technology

[0002] Current healthcare information systems often employ heterogeneous architectures, with data scattered across independent systems, creating data silos and semantic fragmentation. Different systems operate using different coding standards, requiring manual mapping for cross-system data queries, resulting in extremely low efficiency. Furthermore, data quality monitoring and privacy compliance audits rely on manual rule configuration, which is ineffective in handling dynamic medical scenarios such as emergency care and multidisciplinary consultations. In cross-institutional data sharing, collaborations within medical alliances and clinical trials are often conducted through centralized data lakes, which pose single points of failure risks and suffer from ambiguous data sovereignty definitions, hindering efficient governance, secure sharing, and intelligent application of medical data.

[0003] There is an urgent need for an intelligent management method for medical data to break through the limitations of the existing architecture and achieve data unification, privacy protection, and intelligent scheduling. Summary of the Invention

[0004] In view of this, the present invention proposes an intelligent management method and system for medical data, which can realize intelligent management of medical data.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A method for intelligent management of medical data includes: Heterogeneous medical data is mapped to a unified triple structure, and terminological differences in the triple structure are eliminated through vectorized metrics to generate unified data. After encrypting the unified data of each institution using a homomorphic encryption function, the privacy-preserving aggregated result of the unified data is calculated, and the privacy-preserving aggregated result after adding noise is also calculated to achieve differential privacy protection. The matching score of the unified data is obtained by calculating the ratio of the number of matched fields to the total number of extracted fields. If the matching score is less than a set threshold, abnormal data is marked, and the impact range of the abnormal data on downstream modules is calculated. To obtain the optimal solution for monthly storage cost, average query latency, and privacy strength; Based on the aforementioned scope of influence and the optimal solution, the storage location and encryption strength are dynamically selected during the operation of the deep Q-network to achieve intelligent management of medical data.

[0006] Based on the above technical solution, the present invention can be further improved as follows: Optionally, the step of mapping heterogeneous medical data into a unified triple structure and eliminating terminological differences in the triple structure through vectorized metrics to generate unified data includes: Heterogeneous medical data are mapped into a unified triplet structure using formula (1); Formula (1); In the formula, As the main body, As a predicate, For objects; The semantic matching degree between the head entity, relation, and tail entity is calculated using formula (2); Formula (2); In the formula, This represents the degree of semantic matching between the head entity, relation, and tail entity. For the head entity, For the relationship, It is a tail entity.

[0007] Optionally, the step of encrypting the unified data of each institution using a homomorphic encryption function, calculating the privacy-preserving aggregated result of the unified data, and calculating the privacy-preserving aggregated result after adding noise includes: The privacy-preserving aggregation result is calculated using formula (3); Formula (3); In the formula, Aggregate results to protect privacy. The total number of organizations participating in the data aggregation. For institutions Local counts, It is a homomorphic encryption function. Addition to ciphertext; The privacy-preserving aggregation result after adding noise is calculated using formula (4); Formula (4); In the formula, For privacy-preserving aggregation results with added noise, It is Laplace noise. To improve query sensitivity, Budget for privacy.

[0008] Optionally, the step of obtaining the matching score of the unified data by calculating the ratio of the number of matching fields to the total number of extracted fields includes: The matching score of the unified data is calculated using formula (5); Formula (5); In the formula, To score the matching degree, To match the number of fields, This represents the total number of fields extracted.

[0009] Optionally, calculating the impact range of the abnormal data on downstream modules includes: The influence range of node v on downstream modules is calculated using formula (6); Formula (6); In the formula, The impact range of node v on downstream modules is defined, and node v is determined based on the abnormal data. For all nodes, For the set of reachable nodes, For node weights, Score based on business criticality.

[0010] Optionally, the step of dynamically selecting the storage location and encryption strength based on the influence range and the optimal solution during the deep Q-network operation to achieve intelligent management of medical data includes: The state is calculated using formula (7). Next action Q value; Formula (7); In the formula, For state Next action Q value, For state vectors, For a set of actions, The expectation is determined based on the optimal solution. For instant rewards, This is the discount factor.

[0011] Optionally, the intelligent management method for medical data further includes: The kkk-th block is calculated using formula (8); Formula (8); In the formula, For the kk-th block, For SHA-256 digest, For detailed operation records, For timestamps; Blocks of the blockchain are constructed based on the formula (8) to ensure that the data is traceable and tamper-proof.

[0012] An intelligent management system for medical data includes: The data standardization module is used to map heterogeneous medical data into a unified triple structure, and eliminate the terminology differences of the triple structure through vectorized metrics to generate unified data. The privacy protection module is used to encrypt the unified data of various institutions using a homomorphic encryption function, calculate the privacy protection aggregation result of the unified data, and calculate the privacy protection aggregation result after adding noise, so as to achieve differential privacy protection. The data quality monitoring module is used to obtain a matching score for the unified data by calculating the ratio of the number of matching fields to the total number of extracted fields. If the matching score is less than a set threshold, abnormal data is marked, and the impact range of the abnormal data on downstream modules is calculated. The optimization decision module is used to obtain the optimal solution for monthly storage cost, average query latency, and privacy strength; The intelligent scheduling module is used to dynamically select the storage location and encryption strength during the operation of the deep Q network based on the influence range and the optimal solution, so as to realize intelligent management of medical data.

[0013] An electronic device includes a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the method described herein.

[0014] A non-transitory computer-readable storage medium having a computer program stored thereon, the computer program implementing the steps of the method when executed by a processor.

[0015] The present invention has the following advantages: The intelligent management method for medical data in this invention breaks down semantic barriers between heterogeneous data through a unified triplet structure and vectorization, achieving efficient integration; the combination of homomorphic encryption and differential privacy ensures the security of data aggregation while balancing privacy and sharing; matching degree detection accurately locates anomalies, clarifies the scope of impact, and improves data quality; the optimal solution for storage, latency, and privacy, along with deep Q-network scheduling, balances cost, efficiency, and security while meeting the needs of medical scenarios, facilitating intelligent, compliant, and efficient management of medical data and solving problems such as data silos and collaboration bottlenecks. Attached Figure Description

[0016] For illustrative purposes and not limiting, the present invention will now be described in conjunction with embodiments and accompanying drawings, wherein: Figure 1 This is a flowchart illustrating the intelligent management method for medical data in an embodiment of the present invention. Figure 2 This is a schematic diagram of the main components of the intelligent management system for medical data in an embodiment of the present invention; Figure 3 This is a schematic diagram of the physical structure of the electronic device provided by the present invention. Detailed Implementation

[0017] To enable those skilled in the art to better understand the present invention, 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 a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0018] It should be noted that the terms "first," "second," etc., in the specification and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be used interchangeably where appropriate for the embodiments of the invention described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0019] It should be noted that, where there is no conflict, the embodiments and features of the present invention can be combined with each other. The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0020] Figure 1 This is a flowchart illustrating the intelligent management method for medical data according to an embodiment of the present invention, such as... Figure 1 As shown, the intelligent management method for medical data provided in this embodiment of the invention includes the following steps S101 to S105.

[0021] S101 maps heterogeneous medical data into a unified triple structure, eliminates terminological differences in the triple structure through vectorized metrics, and generates unified data.

[0022] Heterogeneous messages such as HL7, FHIR, and DICOM are mapped to unified RDF triples, providing standard input for subsequent ontology pairing and semantic queries.

[0023] Heterogeneous medical data are mapped into a unified triplet structure using formula (1); Formula (1); In the formula, The main body (e.g., Patient#123). As a predicate (e.g., hasID). For objects (e.g., 00002345); Heterogeneous field mapping accuracy ≥99%, real-time synchronization latency ≤100 ms. The relationship between calculation results and subsequent steps. The generated Triple set is used as TransE training samples, and the embedding vectors h, r, t are calculated.

[0024] Eliminate terminological differences in triplet structures by using vectorized metrics to generate unified data.

[0025] The semantic matching degree between the head entity, relation, and tail entity is calculated using formula (2); Formula (2); In the formula, This represents the degree of semantic matching between the head entity, relation, and tail entity. For head entities (such as "High Blood Pressure"), For relations (such as "equalTo"), For tail entities (such as "Hypertension").

[0026] If f(h, r, t) < τ (threshold τ = 0.1), then it is considered equivalent.

[0027] The accuracy rate of terminology matching using formula (2) is 98%, and the amount of manual review is reduced by 85%.

[0028] S102, after encrypting the unified data of each institution through a homomorphic encryption function, calculate the privacy-preserving aggregate result of the unified data, and calculate the privacy-preserving aggregate result after adding noise, so as to achieve differential privacy protection.

[0029] The privacy-preserving aggregation result is calculated using formula (3); Formula (3); In the formula, Aggregate results to protect privacy. The total number of organizations participating in the data aggregation. For institutions Local counts, It is a homomorphic encryption function. Addition to ciphertext; The privacy-preserving aggregation result after adding noise is calculated using formula (4); Formula (4); In the formula, For privacy-preserving aggregation results with added noise, It is Laplace noise. To improve query sensitivity, Budget for privacy.

[0030] At ε = 0.8, the statistical error is <2% and the leakage probability is <0.1%.

[0031] S103: By calculating the ratio of the number of matched fields to the total number of extracted fields, a matching score for the unified data is obtained. If the matching score is less than a set threshold, abnormal data is marked, and the impact range of the abnormal data on downstream modules is calculated.

[0032] The matching score of the unified data is calculated using formula (5); Formula (5); In the formula, To score the matching degree, To match the number of fields, This represents the total number of fields extracted.

[0033] If S < 0.8, then mark it as an anomaly.

[0034] The influence range of node v on downstream modules is calculated using formula (6); Formula (6); In the formula, The impact range of node v on downstream modules is defined, and node v is determined based on the abnormal data. For all nodes, For the set of reachable nodes, For node weights, Score based on business criticality.

[0035] Formula (6) reduces the average tracing time from 30 min to 2 min.

[0036] S104, obtain the optimal solution for monthly storage cost, average query latency, and privacy strength.

[0037] ; Monthly storage cost (¥), The average query latency (ms) is... For privacy protection.

[0038] Availability ≥ 99.9%, ε ≤ 1.0.

[0039] Pareto frontier reduces costs by 28% and latency by 35%.

[0040] S105 dynamically selects the storage location and encryption strength during deep Q-network operation based on the scope of influence and the optimal solution, so as to achieve intelligent management of medical data.

[0041] The state is calculated using formula (7). Next action Q value; Formula (7); In the formula, For state Next action Q value, For the state vector <load, risk level, available budget>, For the action set <Migrate, Encrypt↑, Encrypt↓>, The expectation is determined based on the optimal solution. For instant rewards, The discount factor is 0.95.

[0042] The intelligent management method for medical data also includes: The kkk-th block is calculated using formula (8); Formula (8); In the formula, For the kk-th block, For SHA-256 digest, For detailed operation records, For timestamps; By using formula (8), 100% of the audit logs are uploaded to the blockchain, and the tracing time is less than 1 second.

[0043] Blocks of the blockchain are constructed based on the formula (8) to ensure that the data is traceable and tamper-proof.

[0044] Example 1: Multicenter Clinical Trial Data Platform Scenario: A multinational pharmaceutical company is collaborating with 10 hospitals to conduct a Phase III drug trial.

[0045] Implementation results: The data weaving layer integrates heterogeneous medical data such as EHR, laboratory data, and patient reports, reducing the semantic conflict rate from 15% to 2%.

[0046] The federal query engine now calculates adverse event rates in 5 minutes, down from 2 hours, without exposing individual data.

[0047] The system automatically detected 35,000 abnormal data entries (such as incorrect dosage units) and located LIS interface configuration issues in two hospitals through bloodline tracing.

[0048] Example 2: Emergency Department Multidisciplinary Consultation System Scenario: Real-time access to imaging, laboratory, and medical history data in the emergency department of a top-tier hospital.

[0049] Implementation results: The dynamic desensitization engine displays different data granularities based on the roles of the consultation team (such as attending physicians and interns).

[0050] Figure 2 This is a schematic diagram illustrating the main components of the intelligent management system for medical data in an embodiment of the present invention. Figure 2 As shown, the intelligent management system 1 for medical data provided in this embodiment of the invention includes a data standardization module 10, a privacy protection module 20, a data quality monitoring module 30, an optimization decision-making module 40, and an intelligent scheduling module 50.

[0051] The data standardization module 10 is used to map heterogeneous medical data into a unified triple structure, eliminate terminological differences in the triple structure through vectorized metrics, and generate unified data. The privacy protection module 20 is used to encrypt the unified data of each institution through a homomorphic encryption function, calculate the privacy protection aggregation result of the unified data, and calculate the privacy protection aggregation result after adding noise, so as to achieve differential privacy protection. The data quality monitoring module 30 is used to obtain a matching score of the unified data by calculating the ratio of the number of matching fields to the total number of extracted fields. If the matching score is less than a set threshold, abnormal data is marked, and the impact range of the abnormal data on downstream modules is calculated. The optimization decision module 40 is used to obtain the optimal solution for monthly storage cost, average query latency, and privacy strength; The intelligent scheduling module 50 is used to dynamically select the storage location and encryption strength during the operation of the deep Q network based on the influence range and the optimal solution, so as to realize intelligent management of medical data.

[0052] Figure 3 This is a schematic diagram of the physical structure of an electronic device provided in an embodiment of the present invention, such as... Figure 3 As shown, the electronic device 60 includes: a processor 601, a memory 602, and a bus 603; The processor 601 and the memory 602 communicate with each other via the bus 603. The processor 601 is used to call program instructions in the memory 602 to execute the methods provided in the above-described method embodiments, and to execute the methods provided in the embodiments of the present invention.

[0053] This embodiment provides a non-transitory computer-readable storage medium that stores computer instructions, which cause a computer to execute the method provided in this embodiment of the invention.

[0054] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various storage media capable of storing program code, such as ROM, RAM, magnetic disk, or optical disk.

[0055] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for intelligent management of medical data, characterized in that, include: Heterogeneous medical data is mapped to a unified triple structure, and terminological differences in the triple structure are eliminated through vectorized metrics to generate unified data. After encrypting the unified data of each institution using a homomorphic encryption function, the privacy-preserving aggregated result of the unified data is calculated, and the privacy-preserving aggregated result after adding noise is also calculated to achieve differential privacy protection. The matching score of the unified data is obtained by calculating the ratio of the number of matched fields to the total number of extracted fields. If the matching score is less than a set threshold, abnormal data is marked, and the impact range of the abnormal data on downstream modules is calculated. To obtain the optimal solution for monthly storage cost, average query latency, and privacy strength; Based on the aforementioned scope of influence and the optimal solution, the storage location and encryption strength are dynamically selected during the operation of the deep Q-network to achieve intelligent management of medical data.

2. The intelligent management method for medical data according to claim 1, characterized in that, The process of mapping heterogeneous medical data into a unified triple structure, and eliminating terminological differences in the triple structure through vectorized metrics to generate unified data, includes: Heterogeneous medical data are mapped into a unified triplet structure using formula (1); Formula (1); In the formula, As the main body, As a predicate, For objects; The semantic matching degree between the head entity, relation, and tail entity is calculated using formula (2); Formula (2); In the formula, This represents the degree of semantic matching between the head entity, relation, and tail entity. For the head entity, For the relationship, It is a tail entity.

3. The intelligent management method for medical data according to claim 2, characterized in that, After encrypting the unified data of each organization using a homomorphic encryption function, the privacy-preserving aggregated result of the unified data is calculated, and the privacy-preserving aggregated result after adding noise is also calculated, including: The privacy-preserving aggregation result is calculated using formula (3); Formula (3); In the formula, Aggregate results to protect privacy. The total number of organizations participating in the data aggregation. For institutions Local counts, It is a homomorphic encryption function. Addition to ciphertext; The privacy-preserving aggregation result after adding noise is calculated using formula (4); Formula (4); In the formula, For privacy-preserving aggregation results with added noise, It is Laplace noise. For query sensitivity, Budget for privacy.

4. The intelligent management method for medical data according to claim 3, characterized in that, The process of calculating the matching score of the unified data by measuring the ratio of the number of matched fields to the total number of extracted fields includes: The matching score of the unified data is calculated using formula (5); Formula (5); In the formula, To score the matching degree, To match the number of fields, This represents the total number of fields extracted.

5. The intelligent management method for medical data according to claim 4, characterized in that, The calculation of the impact range of the abnormal data on downstream modules includes: The influence range of node v on downstream modules is calculated using formula (6); Formula (6): In the formula, The impact range of node v on downstream modules is defined, and node v is determined based on the abnormal data. For all nodes, For the set of reachable nodes, For node weights, Score based on business criticality.

6. The intelligent management method for medical data according to claim 5, characterized in that, The method of dynamically selecting storage location and encryption strength based on the influence range and the optimal solution during the deep Q-network operation to achieve intelligent management of medical data includes: The state is calculated using formula (7). Next action Q value; Formula (7): In the formula, For state Next action Q value, For state vectors, For a set of actions, The expectation is determined based on the optimal solution. For instant rewards, This is the discount factor.

7. The intelligent management method for medical data according to claim 1, characterized in that, The intelligent management method for medical data also includes: The kkk-th block is calculated using formula (8); Formula (8): In the formula, For the kk-th block, For SHA-256 digest, For detailed operation records, For timestamps; Blocks of the blockchain are constructed based on the formula (8) to ensure that the data is traceable and tamper-proof.

8. A system for intelligent management of medical data, characterized in that, include: The data standardization module is used to map heterogeneous medical data into a unified triple structure, and eliminate the terminology differences of the triple structure through vectorized metrics to generate unified data. The privacy protection module is used to encrypt the unified data of various institutions using a homomorphic encryption function, calculate the privacy protection aggregation result of the unified data, and calculate the privacy protection aggregation result after adding noise, so as to achieve differential privacy protection. The data quality monitoring module is used to obtain a matching score for the unified data by calculating the ratio of the number of matching fields to the total number of extracted fields. If the matching score is less than a set threshold, abnormal data is marked, and the impact range of the abnormal data on downstream modules is calculated. The optimization decision module is used to obtain the optimal solution for monthly storage cost, average query latency, and privacy strength; The intelligent scheduling module is used to dynamically select the storage location and encryption strength during the operation of the deep Q network based on the influence range and the optimal solution, so as to realize intelligent management of medical data.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 7.

Images