Blockchain-based medical data traceability and privacy protection auditing method and system

Abstract

The application discloses a medical data traceability and privacy protection auditing method and system based on a block chain, relates to the technical field of medical data traceability, and specifically comprises the following steps: step one, constructing an alliance chain medical data system, calculating the initial quality value of a node, and iteratively obtaining the final quality value of the node; step two, setting three verification rules for abnormal data identification, setting a topological potential value according to the final quality value of each node, and setting a region division rule according to the topological potential value; and step three, locating the source of abnormal data through single-source traceability calculation and multi-source node determination in a subnetwork, and formulating a privacy protection auditing method. According to the region division rule of the topological potential value, the application can clarify a complex block chain medical network according to the influence relationship between nodes, perform single-source traceability calculation in the subnetwork, and determine the sources of multiple abnormal data in the block chain medical network.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of blockchain auditing, and particularly relates to a medical data traceability and privacy protection auditing method and system based on a blockchain. BACKGROUND

[0002] With the rapid development of medical informatization, medical data presents the characteristics of multi-source, massiveness and high sensitivity. Medical information such as medical records, test reports and image data generated by different medical institutions, medical devices and departments has different formats and complex sources, which brings great challenges to the unified management and sharing of data. In the traditional medical data management mode, data is mostly stored in the local databases of each institution, and there is a data island phenomenon, making it difficult to share and trace data across institutions. Even if a medical data sharing platform is established in some areas, the centralized management mode still has the risk of single point failure, and the operation records of data are difficult to trace throughout the process, which makes it difficult to accurately locate the source and identify responsibility when there is incorrect data or maliciously injected data.

[0003] The blockchain technology provides a new solution for medical data management due to its characteristics of non-tamperability, traceability and distributed storage. However, there are still many problems in applying blockchain to the medical data field. Firstly, medical data is generally injected from multiple sources, and different hospitals, devices and departments may generate incorrect information or malicious data. Traditional single-source traceability methods cannot cope with multi-source positioning, and the traceability accuracy is low and the error positioning distance is large. Secondly, there is a circular dependency problem in node quality evaluation, and it is difficult to objectively measure the credibility and influence of each node in data propagation. Thirdly, there is a lack of unified standards and precise verification mechanisms for identifying abnormal data, which affects the efficiency of traceability.

[0004] Therefore, there is an urgent need for a medical data traceability and privacy protection auditing method and system based on a blockchain to solve the above problems. SUMMARY

[0005] The purpose of the present application is to provide a medical data traceability and privacy protection auditing method and system based on a blockchain to solve the technical problems that the traditional single-source traceability method cannot cope with multi-source positioning, and the traceability accuracy is low and the error positioning distance is large in the prior art.

[0006] In order to achieve the above purpose, the present application adopts the following technical solutions:

[0007] The medical data traceability and privacy protection auditing method based on a blockchain comprises:

[0008] Step one, constructing an alliance chain medical data system, calculating the initial quality value of the node, and iteratively obtaining the final quality value of the node;

[0009] Step two, set the triple verification rule of abnormal data identification, set the topological potential value by synthesizing the final quality value of each node, and set the regional division rule through the topological potential value;

[0010] Step three, locate the source of abnormal data through single-source tracing calculation and multi-source node determination of sub-network, and develop a privacy protection audit method.

[0011] Further, a medical data system based on a consortium chain is constructed, and the specific method is as follows:

[0012] The consortium chain architecture is adopted, the nodes cover medical institutions, regulatory departments and third-party audit institutions, different sources and different formats of medical data are converted into a unified standard format, an encryption algorithm is used to encrypt the medical data, the encrypted data is stored on the blockchain, distributed data management is realized through the consensus mechanism, only operation meta information of the medical data is stored on the chain, the original data is stored in the local database of each institution, the data integrity is verified by using the hash value on the chain, the SM2 algorithm is used to encrypt the communication between the blockchain nodes, and when verifying the data correlation, only the data meets the index requirements, and the original data is not exposed.

[0013] Further, the initial quality value of the node is calculated, and the specific method is as follows:

[0014] The qualification level of the medical institution is determined, and the initial value of the institution is quantified, the authentication level of the medical equipment is determined, and the initial value of the authentication is quantified, the professional title of the doctor is determined, and the initial value of the professional title is quantified, the historical data accuracy rate of the department is determined, and the initial value of the department is quantified, the initial quality value of node i in T time length is obtained by weighted summation of the initial value of the institution, the initial value of the authentication, the initial value of the professional title and the initial value of the department of node i in T time length.

[0015] Further, the final quality value of the node is obtained by iteration, and the specific method is as follows:

[0016] Based on the initial quality value, the node quality formula is used for iterative calculation until convergence;

[0017] In the first iteration, the quality value of node i is calculated, and the initial quality value of other nodes is used, that is:

[0018] ;

[0019] Wherein, Qi represents the quality value of node i in the first time, Qi represents the influence weight of node i, the reciprocal of the initial value of the institution is used to represent the influence weight of node i, Qj represents the total number of personal information of node j in T time length, Qj represents the initial quality value of node j, and J represents the total number of nodes, represents the association degree value of node i and node j in T time length, when j = i, is equal to 0.

[0020] In the tth iteration, the node quality value of the current iteration is updated using the node quality value calculated in the last round:

[0021] ;

[0022] wherein, represents the quality value of node i in the tth time, represents the quality value of node j in the t-1th time.

[0023] When the quality value change of all nodes meets , the iteration is stopped, and the quality value of the node at this time is , which represents the final quality value of the node, is a preset change threshold.

[0024] Further, a three-verification rule for identifying abnormal data is set, and the specific method is:

[0025] Based on the tamper-proof and timestamp characteristics of the blockchain, a basic verification rule is set, which includes whether the data format conforms to the medical industry standard, whether the data hash value matches the original hash of the blockchain record, and whether the data generation timestamp is earlier than the associated preceding data;

[0026] Deepen verification combined with node interaction features, analyze the abnormal performance of data in the propagation path that passes the basic verification rule, including whether the same data is consistent in different nodes, whether the data propagation range matches the node permission, and whether the data is abnormally high frequency forwarded;

[0027] Introduce a dynamic threshold mechanism, set an abnormal judgment threshold according to the data type, and perform abnormal judgment on different types of data.

[0028] After the above three verifications, the marked abnormal data will enter the traceability process and serve as the basis for calculating the "amount of abnormal data of associated nodes" and "number of abnormal data blocks" in single-source traceability calculation.

[0029] Further, the topology potential value is set by comprehensively considering the final quality value of each node, and the specific method is:

[0030] A preset association degree threshold A is determined to determine the association degree value between different nodes. If the association degree value between node i and node j is greater than or equal to the threshold A, they are associated nodes;

[0031] The topology potential value is set by comprehensively considering the final quality value of each node, the hash chain distance, and the associated nodes, and the specific calculation formula is:

[0032] ;

[0033] in, Let represent the topological potential of the i-th node, where i represents the i-th node. This represents the associated node of the i-th node. Indicates that the i-th node has a total of One associated node, Indicates the relationship between the i-th node and the i-th node. The hash chain distance between two nodes, i.e., the number of hops between the two nodes in the blockchain. Represents a node The final quality value, Represents the relationship between node i and node T within a time duration T. The degree of correlation.

[0034] Furthermore, the region division rules are set through topological potential values, specifically as follows:

[0035] A node whose topological potential is greater than that of all its associated nodes is identified as a peak node. With each peak node as the core, the region formed by all associated nodes around the peak node is called the interaction region.

[0036] There are some cross-regional shared associated nodes. These associated nodes belong to the interaction area of ​​multiple peak nodes at the same time. These associated nodes are recorded as overlapping nodes.

[0037] Nodes whose topological potential is less than a preset threshold f are recorded as isolated nodes and exist alone.

[0038] Furthermore, the specific method for locating the source of abnormal data through single-source tracing calculation of sub-networks and determination of multi-source nodes is as follows:

[0039] Each interactive region formed around a peak node constitutes a sub-network.

[0040] In the subnetwork, prior and posterior estimates are calculated, and their multiplication yields the overall probability value. Specifically, the ratio of anomalous data in the associated nodes to the total data in the associated nodes is calculated first. This ratio is then raised to the power of the total number of nodes in the subnetwork. Finally, the result is multiplied by the final quality value of the node to obtain the prior estimate. Alternatively, the total number of nodes in the subnetwork is obtained, and its factorial is calculated. Then, for each node in the subnetwork, the number of anomalous data blocks propagating from that node in the direction away from the target node is counted. The reciprocal of each reciprocal is taken, and all these reciprocals are multiplied together. Finally, the product is multiplied by the factorial of the total number of nodes in the subnetwork to obtain the posterior estimate. Multiplying the prior and posterior estimates yields the overall probability value of the node as a source of anomalous data. A higher overall probability value indicates a greater likelihood that the node is a source of anomalous data in the subnetwork.

[0041] In each sub-network, the node with the maximum comprehensive possibility value is selected as the abnormal data source of the sub-network;

[0042] For the coincident nodes, the comprehensive possibility value is calculated in each sub-network to which the node belongs, and if the sub-network source determination condition is met, the node is determined as the abnormal data source of the corresponding multiple sub-networks;

[0043] The isolated node does not participate in the determination of the multi-source node.

[0044] The application also provides a medical data traceability and privacy protection audit system based on a blockchain, which is applied to the medical data traceability and privacy protection audit method based on the blockchain and comprises:

[0045] A final quality value calculation module for constructing a medical data system of the alliance chain, calculating an initial quality value of the node, and iteratively obtaining a final quality value of the node;

[0046] A region division module for setting a three-verification rule for abnormal data identification, setting a topological potential value based on the final quality value of each node, and setting a region division rule based on the topological potential value;

[0047] An abnormal data source positioning module for accurately positioning the source of the abnormal data through sub-network single-source traceability calculation and multi-source node determination, and formulating a privacy protection audit method.

[0048] As described above, due to the adoption of the above technical solutions, the application has the following beneficial effects:

[0049] 1. The application realizes the standardized processing and distributed management of multi-source heterogeneous medical data by constructing a medical data system of the alliance chain, and the mode of storing only operation meta information on the chain and storing original data locally, which breaks the data island and reduces the leakage risk of centralized storage.

[0050] 2. The application filters abnormal data from three dimensions of data format, propagation path and dynamic threshold through the three-verification rule, which covers basic problems such as format error and tampering, and can also identify complex abnormalities such as cross-node inconsistency and permission mismatch, thereby significantly reducing the misjudgment rate.

[0051] 3、The present application runs privacy protection audit method throughout the whole link of data management, through access log review, transmission encryption verification, desensitization effect evaluation and node system review, builds four protection mechanisms of preventing unauthorized access, preventing transmission leakage, preventing desensitization failure and preventing operation violation, especially the key audit of multi-source abnormal nodes, which not only ensures that the traceability process does not leak patient privacy, but also balances data sharing and privacy protection through strict verification of desensitization processing and permission control. BRIEF DESCRIPTION OF DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0053] Figure 1 The step diagram of the medical data traceability and privacy protection audit method based on blockchain of the present application is shown;

[0054] Figure 2 The step diagram of the final quality value iteration calculation method of the node is shown;

[0055] Figure 3 The module diagram of the medical data traceability and privacy protection audit system based on blockchain of the present application is shown. DETAILED DESCRIPTION

[0056] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0057] Embodiment one, the medical data traceability and privacy protection audit method based on blockchain as shown in the specific steps include: Figure 1

[0058] Step one, build a consortium chain medical data system, calculate the initial quality value of the node, and iterate to get the final quality value of the node.

[0059] ​The alliance chain architecture is adopted, the nodes cover medical institutions, regulatory departments and third-party audit institutions, different sources and different formats of medical data are converted into a unified standard format, encryption algorithms are used for encryption processing of medical data, at the same time, the encrypted data is stored on the block chain, and distributed data management is realized through the consensus mechanism (such as practical Byzantine fault tolerance algorithm), only the operation meta information (ID, timestamp, data hash value, purpose label, etc.) of the medical data is stored on the chain, the original data (such as medical records, images, test reports) is stored in the local database of each institution, the data integrity is verified by using the hash value on the chain, the communication between the nodes of the block chain uses SM2 algorithm encryption, when verifying the data correlation, only the data meets the index requirements, and the original data is not exposed;

[0060] The propagation of medical data in the block chain depends on the transaction verification between nodes and the block chain relationship (such as the previous hash and the subsequent block), the nodes recording the role information (medical institutions, medical equipment, doctors / department) and the medical data theme (such as medical records, test reports, image data meta information) of medical data are recorded as medical nodes, according to the inherent properties of the medical nodes, the initial quality value is directly assigned;

[0061] Further, according to Figure 2 , the medical institution qualification level is determined and the initial value of the institution is quantified (such as the initial value of a third-grade hospital = 9, a second-grade hospital = 7, and a primary clinic = 5), the medical equipment certification level and the historical failure rate are determined and the certification initial value is quantified (such as the initial value of FDA certified equipment = 8, only domestic certification = 6, and the failure rate of continuous working T length is more than 5% every time 2 is deducted), the doctor title is determined and the title initial value is quantified (such as the initial value of a chief physician = 8.5, and a resident physician = 6), the historical data accuracy rate of the department is determined and the department initial value is quantified (such as the initial value is equal to 1 when the historical data accuracy rate is 100, and 0.01 is deducted every 1%), the initial quality value of node i in T length is obtained by weighted summation of the initial value of the institution, the certification initial value, the title initial value and the department initial value of node i in T length.

[0062] Based on the initial quality value, the node quality formula is used for iterative calculation until convergence, and the specific process is as follows:

[0063] The first iteration: when calculating the quality value of node i, the initial quality value of other nodes is used, that is:

[0064] ;

[0065] Among them, indicates the first quality value of node i, indicates the influence weight of node i, which is represented by the reciprocal of the initial value of the institution, indicates the total number of personal information of node j in T length, Qi,j represents the initial quality value of node j, J represents the total number of nodes, Qi,j represents the association degree value of node i and node j within T time length, when j = i, Qi,j = 0.

[0066] The tthiteration: update the current node quality value using the node quality value calculated in the last round:

[0067] ;

[0068] wherein, Qi represents the tthquality value of node i, Qi-1,j represents the t-1thquality value of node j.

[0069] Further, the personal information refers to a medical data set generated by a patient in the whole process of registration-examination-treatment-departure in different medical institutions, and the number of personal information sent / received between node i and node j is equal to the association degree value of node i and node j .

[0070] When the quality value change amount of all nodes is less than a preset change threshold, the iteration is stopped, wherein is the final quality value of the node.

[0071] Step two, set the three verification rules for abnormal data identification, set the topological potential value by comprehensively considering the final quality value of each node, and set the regional division rule through the topological potential value.

[0072] Based on the tamper-proofing and timestamp characteristics of the block chain, set the basic verification rules, the basic verification rules include whether the data format conforms to the medical industry standard (such as missing medical record field, test report unit error), whether the data hash value matches the original hash recorded in the block chain (mismatch indicates that the data is tampered), and whether the data generation timestamp is earlier than the associated previous data (such as test report time earlier than sampling time, there is a logical contradiction).

[0073] Secondly, combined with the node interaction characteristics, the abnormal performance of the data passing through the basic verification rule in the propagation path is analyzed, including whether the same data recorded in different nodes is consistent (such as the blood pressure value recorded in the node of hospital A and the synchronous record in the node of hospital B deviates more than the clinical reasonable range), whether the data propagation range matches the node authority (such as the ordinary clinic node propagates high-level medical record data with limited authority of three A hospitals), and whether the data is abnormally high frequency forwarded (unnecessary nodes are forwarded multiple times in a short time, which may be malicious diffusion).

[0074] ​Finally, a dynamic threshold mechanism is introduced to set the abnormality judgment threshold according to the data type. For example, physiological indicator data (such as heart rate, blood sugar) uses the clinical normal range as the threshold, and data beyond the range is marked as potential abnormality. Text data (such as diagnosis conclusion) is compared with historical similar data through natural language processing technology, and if the semantic deviation exceeds the preset threshold (such as 30%), it is determined as abnormal. Image data is verified through an AI-assisted diagnosis system, and if the matching degree with the professional physician's annotation result is less than the threshold (such as 80%), it is included in the abnormal candidate set.

[0075] After the above three verifications, the marked abnormal data will enter the traceability process as the statistical basis for "correlation node abnormal data volume" and "abnormal data block number" in single-source traceability calculation.

[0076] A preset correlation degree threshold A is determined to determine the correlation degree value between different nodes. If the correlation degree value between node i and node j is greater than or equal to the threshold A, then node i and node j are each other's correlation nodes. The final quality value, hash chain distance and correlation node of each node are set to topological potential value, which is used to measure the influence of nodes in the propagation of blockchain medical data. The specific calculation formula of topological potential value is:

[0077] ;

[0078] Wherein, represents the topological potential value of the i-th node, i represents the i-th node, represents the correlation node of the i-th node, represents that the i-th node has correlation nodes, represents the hash chain distance between the i-th node and the j-th node, i.e. the hop number of the two nodes in the blockchain, represents the final quality value of the node represents the correlation degree value between node i and node in T time length.

[0079] ​​The node with a topological potential value greater than all its associated nodes is identified as a peak node. For example, in a regional medical alliance chain, large tertiary hospitals have high node quality, extensive data interaction, and key positions in the blockchain, and often have high topological potential values. Each peak node is taken as the core, and all associated nodes around the peak node form a region, which is called an interaction region. There are some associated nodes that share across regions. These associated nodes belong to multiple peak node interaction regions and have a significant impact on the topological potential field in different interaction regions. These associated nodes are recorded as overlapping nodes. Some nodes have low quality and sparse data interaction with other nodes, and their topological potential values are less than a preset threshold f. Their topological potential values are much lower than those of their surrounding nodes. These nodes are not included in any main community and exist as isolated nodes.

[0080] Based on the regional division rule of topological potential value, the influence relationship between nodes is clarified to sort out the complex blockchain medical network.

[0081] Step three, through single-source tracing calculation and multi-source node determination in the subnetwork, the source of abnormal data is located, and a privacy protection audit method is developed.

[0082] Each interaction region formed with a peak node as the core constitutes a subnetwork. Single-source tracing calculation is performed in the subnetwork to determine the nodes that may be the source of abnormal data in the subnetwork. The specific calculation involves prior estimate value and posterior estimate value, and finally the comprehensive likelihood value is obtained.

[0083] The determination of the prior estimate value requires the combination of the final quality value of the node and the abnormal data situation of the associated nodes. Specifically, the ratio of the amount of abnormal data in the associated nodes to the total amount of data in the associated nodes is calculated first, then this ratio is raised to the power of the total number of subnetwork nodes, and finally the result is multiplied by the final quality value of the node to obtain the prior estimate value. This process reflects the influence of the node's own quality and the proportion of abnormal data in its associated nodes on the initial possibility of the node becoming an abnormal source.

[0084] The determination of the posterior estimate value requires tracing the propagation path of abnormal data based on the characteristics of the blockchain's hash chain and timestamp. First, the total number of nodes in the subnetwork is obtained, and its factorial is calculated. Then, for each node in the subnetwork, the number of abnormal data blocks propagating away from the target node is counted, and the reciprocals of these numbers are multiplied. Finally, the product obtained is multiplied by the factorial of the total number of subnetwork nodes to obtain the posterior estimate value. The posterior estimate value reflects the influence of the propagation of abnormal data by other nodes away from the target node on the possibility of the target node as a source.

[0085] The prior estimate value and the posterior estimate value are multiplied to obtain a comprehensive likelihood value of the node as an abnormal data source. The higher the comprehensive likelihood value, the greater the likelihood of the node as an abnormal data source in the subnetwork. After completing the single-source tracing calculation of each subnetwork, the multi-source node is determined according to the comprehensive likelihood value. The specific rules are as follows:

[0086] Subnetwork source determination: In each subnetwork, the node with the highest comprehensive likelihood value is selected and determined as the abnormal data source of the subnetwork. For example, in a subnetwork with a certain large third-grade hospital as a peak node, after calculation, if a certain department node of the hospital has the highest comprehensive likelihood value, the department node is identified as the abnormal data source of the subnetwork.

[0087] Coincidence node processing: For a coincidence node, since it belongs to multiple subnetworks at the same time, it needs to calculate the comprehensive likelihood value in each subnetwork it belongs to. If the comprehensive likelihood value of the coincidence node is high in multiple subnetworks and meets the judgment conditions of the subnetwork source, it is determined as the abnormal data source of the corresponding multiple subnetworks.

[0088] Isolated node exclusion: Isolated nodes are not involved in the determination of multi-source nodes and are not identified as the abnormal data source of any subnetwork because their topological potential value is small, and their own quality is low and data interaction is sparse.

[0089] Through the above method, the sources of multiple abnormal data in the blockchain medical network can be accurately determined, providing a reliable basis for medical data auditing, error correction, and responsibility identification, and effectively improving the security and accuracy of blockchain medical data management.

[0090] After determining the multi-source node, privacy protection audit is needed to ensure that the patient's privacy information is effectively protected during the tracing and management process. The specific method is as follows:

[0091] Access the access log of all nodes in the blockchain to medical data, including access time, access node, accessed data content, and access permission information. Check whether unauthorized nodes have accessed data containing privacy information, and whether authorized nodes have accessed data beyond their permission range. For example, if a certain primary clinic node accesses a patient's detailed medical record that can only be viewed by a third-grade hospital node, it is a violation of access;

[0092] Verify whether a secure encryption algorithm (such as AES, RSA, etc.) is used to encrypt privacy information during medical data transmission. By checking the encryption records and encryption strength of data transmission, it is ensured that data cannot be stolen or tampered with during transmission to cause privacy leakage;

[0093] For medical data that needs to be made public or shared, check whether it has undergone effective privacy desensitization processing, such as partially hiding the patient's ID number (displaying it as 110********1234) and blurring sensitive descriptions of illness in the medical record. Assess whether the desensitized information can still identify the specific patient. If it can, it means that the desensitization effect is not good and there is a risk of privacy leakage.

[0094] Review whether the privacy protection measures and systems of each node (such as medical institution nodes, equipment nodes, etc.) are sound, including the node's privacy protection training records for employees and emergency plans for privacy leaks. For multi-source nodes that are the source of abnormal data, focus on auditing whether they comply with privacy protection regulations during data generation and processing, and whether improper operation has led to the leakage of privacy information.

[0095] Example 2, as follows Figure 3 The blockchain-based medical data traceability and privacy protection audit system shown specifically includes:

[0096] The final quality value calculation module for nodes is used to build a consortium blockchain medical data system, calculate the initial quality value of nodes, and iteratively obtain the final quality value of nodes.

[0097] The region partitioning module sets up a triple verification rule for abnormal data identification, sets a topological potential value based on the final quality value of each node, and sets the region partitioning rule based on the topological potential value.

[0098] The abnormal data source location module accurately locates the source of abnormal data through single-source tracing calculation and multi-source node determination in the sub-network, and formulates a privacy protection audit method.

[0099] 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.

[0100] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to specific implementations. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A blockchain-based method for medical data traceability and privacy protection auditing, characterized in that: include: Step 1: Construct a consortium blockchain medical data system, calculate the initial quality value of the nodes, and iterate to obtain the final quality value of the nodes; The initial quality value of node i within the time period T is obtained by determining the qualification level of medical institutions and quantifying the initial value of institutions, determining the certification level and historical failure rate of medical equipment and quantifying the initial value of certification, determining the professional title of doctors and quantifying the initial value of professional title, determining the historical data accuracy rate of departments and quantifying the initial value of departments, and by weighted summing the initial values ​​of institutions, certification, professional title and department of node i within the time period T. Step 2: Set up a triple verification rule for abnormal data identification, set the topological potential value by combining the final quality values ​​of each node, and set the region division rule by using the topological potential value; Based on the immutability and timestamp characteristics of blockchain, basic verification rules are set. These basic verification rules include whether the data format conforms to medical industry standards, whether the data hash value matches the original hash of the blockchain record, and whether the data generation timestamp is earlier than its associated previous data. By combining node interaction characteristics to deepen verification, we analyze the abnormal behavior of data that passes the basic verification rules in the propagation path, including whether the records of the same data are consistent in different nodes, whether the data propagation range matches the node permissions, and whether the data is forwarded abnormally frequently. A dynamic threshold mechanism is introduced, which sets anomaly detection thresholds based on data type to detect anomalies in different types of data. Step 3: Locate the source of abnormal data and formulate a privacy protection audit method by performing single-source tracing calculations and determining multi-source nodes in the sub-network; Each interactive region formed around a peak node constitutes a sub-network. Calculate the prior and posterior estimates in the subnetwork, and multiply them to obtain the combined probability value; In each sub-network, the node with the highest overall probability value is selected as the source of abnormal data for that sub-network; For overlapping nodes, calculate the comprehensive probability value in each of their respective sub-networks. If the sub-network source determination conditions are met, then the node is simultaneously identified as the abnormal data source of multiple corresponding sub-networks. Isolated nodes do not participate in the determination of multi-source nodes.

2. The blockchain-based medical data traceability and privacy protection auditing method according to claim 1, characterized in that, The specific methods for building a consortium blockchain medical data system are as follows: Adopting a consortium blockchain architecture, with nodes covering medical institutions, regulatory authorities, and third-party auditing agencies, it converts medical data from different sources and in different formats into a unified standard format. Encryption algorithms are used to encrypt the medical data, which is then stored on the blockchain. Distributed data management is achieved through a consensus mechanism. Only operational metadata of the medical data is stored on the chain; the original data remains in each institution's local database. Data integrity is verified using on-chain hash values. Communication between blockchain nodes uses the SM2 algorithm for encryption. When verifying data correlation, only data compliance with indicator requirements is proven, without exposing the original data.

3. The blockchain-based medical data traceability and privacy protection auditing method according to claim 1, characterized in that, The final quality value of a node is obtained through iteration, specifically by: Based on the initial mass value, the node mass formula is used for iterative calculation until convergence. In the first iteration, the mass of node i is calculated, and other nodes use their initial mass values. In the t-th iteration, the current node mass is updated using the node mass calculated in the previous round. The process continues until the mass changes of all nodes satisfy the given conditions. Stop iterating at this time. This is represented as the final quality value of the node. This is a preset threshold for change.

4. The blockchain-based medical data traceability and privacy protection auditing method according to claim 1, characterized in that, The topological potential is set by combining the final mass values ​​of each node. The specific method is as follows: A preset correlation threshold A is used to determine the correlation value between different nodes. If the correlation value between node i and node j is greater than or equal to the threshold A, then they are related nodes. The topological potential is set by combining the final quality value of each node, the hash chain distance, and the associated nodes.

5. The blockchain-based medical data traceability and privacy protection auditing method according to claim 4, characterized in that, The region partitioning rules are set using topological potential values. The specific method is as follows: A node whose topological potential is greater than that of all its associated nodes is identified as a peak node. With each peak node as the core, the region formed by all associated nodes around the peak node is called the interaction region. There are some cross-regional shared associated nodes. These associated nodes belong to the interaction area of ​​multiple peak nodes at the same time. These associated nodes are recorded as overlapping nodes. Nodes whose topological potential is less than a preset threshold f are recorded as isolated nodes.

6. A blockchain-based medical data traceability and privacy protection auditing system, applied to the blockchain-based medical data traceability and privacy protection auditing method described in any one of claims 1-5, characterized in that, Specifically, it includes: The final quality value calculation module for nodes is used to build a consortium blockchain medical data system, calculate the initial quality value of nodes, and iteratively obtain the final quality value of nodes. The region partitioning module sets up a triple verification rule for abnormal data identification, sets a topological potential value based on the final quality value of each node, and sets the region partitioning rule based on the topological potential value. The abnormal data source location module accurately locates the source of abnormal data through single-source tracing calculation and multi-source node determination in the sub-network, and formulates a privacy protection audit method.

Images