Medical data leakage prevention system and method

Abstract

The invention provides a medical data leakage prevention system and method. The system comprises a data acquisition module, a data encryption module, a central processing module and an access control module which are connected in sequence, by constructing a closed-loop protection system consisting of a data acquisition module, an encryption module, a central processing module and an access control module, the whole-process safety management and control of medical data from generation to access is realized. And a dynamic key updating mechanism is adopted, and a quantum key distribution technology is combined, so that the anti-cracking capability of encrypted data is remarkably improved, and the key leakage risk is effectively coped with. The medical data is classified and the user behavior is analyzed through the artificial intelligence technology, the abnormal access mode can be recognized in real time, the secret key updating frequency and the desensitization strategy are dynamically adjusted, and intelligent self-adaptive protection is achieved. The modules of the system are coordinated and linked to form a closed-loop protection link, and quick response can be realized when the risk is detected, so that the anti-leakage capability of the medical data is comprehensively improved.

Description

Technical Field

[0001] This invention relates to the field of data security technology, and in particular to a medical data leakage prevention system and method. Background Technology

[0002] With the rapid development of medical informatization, medical institutions generate and store large amounts of sensitive data daily, including patient physiological parameters, diagnostic reports, and imaging data. This data has high privacy and security requirements; leakage could seriously damage patients' rights and the reputation of medical institutions. Existing medical data leakage prevention systems mostly employ static encryption and fixed access control mechanisms, which are insufficient to cope with increasingly complex internal unauthorized access and external malicious attacks. For example, traditional encryption systems have long key update cycles, making them vulnerable to brute-force attacks; access control often relies on single authentication methods, lacking dynamic response capabilities in abnormal access scenarios; furthermore, systems typically do not perform fine-grained desensitization based on data sensitivity levels and access contexts, leading to over-exposure or insufficient information in different application scenarios such as research and consultations. Therefore, there is an urgent need for a medical data leakage prevention system capable of dynamic encryption, intelligent behavior monitoring, and multi-dimensional access control to improve the security level of the entire data chain. Summary of the Invention

[0003] In view of this, in order to solve the problems existing in the technical background, the present invention proposes a medical data leakage prevention system and method. Specifically, it includes the following:

[0004] A medical data leakage prevention system includes a data acquisition module, a data encryption module, a central processing module, and an access control module connected in sequence. The data acquisition module is connected to the hospital's medical equipment and information system to collect patients' physiological parameters, diagnostic reports, and imaging data in real time. The data encryption module encrypts the collected medical data and includes a key management submodule. This submodule uses a dynamic key update mechanism based on chaos theory to automatically regenerate encryption keys at preset intervals and establishes a bidirectional communication connection with the central processing module. The central processing module includes an artificial intelligence-based data classification unit and a behavior analysis unit. The data classification unit classifies medical data security levels according to sensitivity, and the behavior analysis unit identifies abnormal access patterns using deep learning algorithms. The access control module integrates a biometric recognition component and a dynamic token verification component. The biometric recognition component verifies user identity using vein pattern recognition technology, and the dynamic token verification component generates an access password that changes over time.

[0005] Furthermore, the key management submodule works in collaboration with the behavior analysis unit of the central processing module, receiving the risk level signal output by the behavior analysis unit and dynamically adjusting the key update frequency; when the behavior analysis unit detects high-risk abnormal access, the key management submodule immediately triggers the key reset procedure and synchronously updates the authentication credentials of all associated systems, while storing the expired key in an independent security log database.

[0006] Furthermore, the behavior analysis unit further includes a user behavior modeling component and an environment perception component; the user behavior modeling component constructs a normal operation profile of authorized users through a long short-term memory neural network; the environment perception component monitors the source address, access time, and device fingerprint information of access requests in real time. When user operations deviate from the behavior profile and environmental parameters are abnormal, the environment perception component sends a secondary authentication command to the access control module, triggering a multi-factor authentication process that includes voiceprint verification and iris feature comparison.

[0007] Furthermore, the data encryption module is implemented based on the BB84 protocol, specifically as follows:

[0008] |ψ>=α|0>+β|1>;

[0009] Where |0> and |1> represent the ground states of the qubit, and α and β are complex probability amplitudes, satisfying |α| 2 +|β| 2 =1. During transmission, the sender randomly selects basis vectors {|0>,|1>} or {|+>,|->} to encode the quantum state, and the receiver randomly selects a measurement basis for measurement. The two parties compare the basis vector selection through a classical channel and finally generate a shared key.

[0010] Furthermore, the access control module is equipped with an emergency access channel, which is connected to the medical emergency response system and is automatically activated when a patient's critical vital signs are detected. When the emergency access channel is activated, a temporary access token is automatically generated and data operation permissions are restricted to read-only mode. At the same time, full operation auditing and tracking are initiated, and all data operations accessed through the emergency channel are marked as special events and uploaded to the medical supervision platform for record-keeping. The validity period of the temporary access token is dynamically adjusted according to the severity of the patient's critical value and automatically expires after the patient's vital signs return to normal.

[0011] Furthermore, the system also includes a data anonymization processing module, which is deployed between the central processing module and the access control module. This module contains a field-level anonymization rule base and a context-aware anonymization engine. The field-level anonymization rule base stores anonymization strategies corresponding to data with different security levels. The context-aware anonymization engine automatically selects anonymization algorithms based on the visitor's role permissions and the access scenario. For access requests for research purposes, it retains the statistical characteristics of the data but hides personal identifiers. For clinical consultation scenarios, it displays complete diagnostic information but obscures the patient's identity information. The data anonymization processing module works in conjunction with the behavior analysis unit, automatically increasing the anonymization intensity to the highest level when abnormal access patterns are detected.

[0012] A method for preventing medical data leakage, based on the aforementioned system, includes the following steps: A data acquisition module collects patient physiological parameters, diagnostic reports, and imaging data from medical devices and information systems in real time; a data encryption module encrypts the collected data using quantum key distribution technology, and a key management submodule dynamically updates the key based on chaos theory; a central processing module's data classification unit classifies medical data security levels according to sensitivity, and a behavior analysis unit uses deep learning algorithms to identify abnormal access patterns; an access control module integrates vein pattern recognition and dynamic token verification for identity authentication; the data acquisition module, data encryption module, central processing module, and access control module form a closed-loop data security protection link.

[0013] The above technical solution has the following beneficial effects:

[0014] This invention constructs a closed-loop protection system comprising data acquisition, encryption, central processing, and access control modules, achieving end-to-end security management of medical data from generation to access. Employing a dynamic key update mechanism combined with quantum key distribution technology significantly enhances the anti-cracking capability of encrypted data, effectively addressing the risk of key leakage. By using artificial intelligence technology to classify medical data and analyze user behavior, it can identify abnormal access patterns in real time and dynamically adjust the key update frequency and desensitization strategy, achieving intelligent adaptive protection. Integrating vein pattern recognition, dynamic tokens, and multi-factor biometric verification ensures convenient access for legitimate users while effectively preventing identity theft and unauthorized intrusion. The establishment of an emergency access channel and data desensitization processing module ensures timely data retrieval in critical situations and balances data availability and privacy protection in different application scenarios through context-aware desensitization. The collaborative operation of all system modules forms a closed-loop protection chain, enabling rapid response when risks are detected, thereby comprehensively improving the ability to prevent medical data leakage. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the use of a medical data leakage prevention system according to the present invention. Detailed Implementation

[0016] 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.

[0017] Example 1: A medical data leakage prevention system includes a data acquisition module, a data encryption module, a central processing module, and an access control module connected in sequence. The data acquisition module is connected to the hospital's medical equipment and information system to collect patient physiological parameters, diagnostic reports, and imaging data in real time. The data encryption module encrypts the collected medical data and includes a key management submodule. This submodule uses a dynamic key update mechanism based on chaos theory to automatically regenerate encryption keys at preset intervals and establishes a bidirectional communication connection with the central processing module. The central processing module includes an artificial intelligence-based data classification unit and a behavior analysis unit. The data classification unit classifies medical data security levels according to sensitivity, and the behavior analysis unit identifies abnormal access patterns using deep learning algorithms. The access control module integrates a biometric identification component and a dynamic token verification component. The biometric identification component verifies user identity using vein pattern recognition technology, and the dynamic token verification component generates an access password that changes over time. The data acquisition module is directly connected to the hospital's medical equipment and information system to collect multi-source medical data, including patient physiological parameters, diagnostic reports, and imaging data, in real time. This module acquires data from monitors, imaging equipment, and hospital information platforms through standardized interfaces. It performs preliminary cleaning and format standardization on the collected raw data to ensure data consistency in subsequent processing. During data acquisition, the module adds metadata such as timestamps and data source device identifiers to each data entry, providing contextual basis for data classification and behavior analysis by the central processing module.

[0018] Example 2, based on Example 1, describes a data encryption module that receives raw data from the acquisition module and encrypts it using quantum key distribution technology based on the BB84 protocol. Specifically, the encryption process involves preparing a quantum state |ψ>=α|0>+β|1>, where |0> and |1> are the ground states of the qubits, and α and β are complex probability amplitudes satisfying normalization conditions. The sender randomly selects basis vectors to encode the quantum state, and the receiver randomly selects a measurement basis for measurement. Both parties compare the basis vector selections through a classical channel to ultimately generate a shared key. This module also integrates a key management submodule, which dynamically updates the encryption key using a chaotic theory model. The initial update cycle is set to 24 hours, and it maintains bidirectional communication with the behavior analysis unit of the central processing module. When the behavior analysis unit outputs a high-risk signal, the key management submodule immediately triggers a key reset, synchronously updates the authentication credentials of all associated systems, and transfers expired keys to an independent security log database for future reference.

[0019] Example 3, based on Example 1, describes a central processing module comprising a data classification unit and a behavior analysis unit. The data classification unit, based on preset rules and machine learning models, categorizes medical data into multiple security levels according to their sensitivity. For example, patient identity information and genetic data are classified as the highest level, while anonymized physiological indicators are classified as lower levels. The behavior analysis unit constructs user behavior profiles using a long short-term memory neural network and, in conjunction with an environmental awareness component, monitors the source address, access time, and device fingerprint of access requests in real time. When user actions significantly deviate from the behavior profile and environmental parameters are abnormal, this unit sends a secondary authentication command to the access control module, initiating a multi-factor verification process including voiceprint and iris feature comparison. Simultaneously, the risk level signal output by the behavior analysis unit is fed back in real time to the key management submodule and the data desensitization processing module, enabling dynamic adjustment of the key update frequency and desensitization intensity.

[0020] The access control module integrates vein pattern recognition and dynamic token verification components for dual authentication of user identity. The vein pattern recognition component acquires images of the user's palm veins using near-infrared spectroscopy and matches them with pre-stored templates; the dynamic token verification component generates a constantly changing access password based on a time synchronization algorithm. This module also includes an emergency access channel that links with the medical emergency response system. When the system detects a patient's critical vital signs, the emergency channel is automatically activated, generating a temporary read-only access token and initiating full-process operation auditing. The token's validity period is dynamically set according to the patient's critical condition and automatically expires once the patient's vital signs return to normal. All relevant operation records are uploaded to the medical supervision platform for filing.

[0021] The data anonymization module is deployed between the central processing module and the access control module, and includes a field-level anonymization rule base and a context-aware anonymization engine. The rule base stores anonymization strategies corresponding to different security levels of data; for example, for research access, statistical characteristics are retained but personal identifiers are hidden, while for clinical consultations, complete diagnostic information is displayed but identity information is obscured. The context-aware anonymization engine automatically selects the anonymization algorithm based on the visitor's role and the context, and automatically increases the anonymization strength to the highest level when the behavior analysis unit detects anomalies, ensuring that data meets usage requirements while maximizing privacy protection.

[0022] The specific application is as follows: The system of this invention has been fully deployed and applied in the coronary care unit (CCU) of a tertiary-level Class A hospital. At 3:15 AM, the system's data acquisition module collected real-time data from the bedside monitor of 68-year-old patient Li, showing ventricular tachycardia and a blood oxygen saturation drop to 85%, indicating critical conditions. Simultaneously, it retrieved the patient's recent coronary angiography report and a list of long-term medications from the hospital information system. The data encryption module immediately used the BB84 protocol to perform quantum encryption on this data, generating an initial key pair of quantum states with α = 0.707 and β = 0.707 for encoding and transmission. At this time, the on-duty doctor, Zhang, initiated an access request to the patient's complete medical records at the doctor's workstation. The access control module first verified the palm vein characteristics using a vein pattern recognition device, and the dynamic token system simultaneously generated an access password valid for 60 seconds.

[0023] The behavior analysis unit of the central processing module detected two abnormal characteristics in this access: the source IP address did not match Zhang's usual workstation login, and the access time deviated from his normal shift schedule. The environmental awareness component immediately triggered a secondary authentication process, requiring Zhang to repeat the voice command and undergo an iris scan. Simultaneously, the system detected that the patient's vital signs had reached critical thresholds, automatically activating the emergency access channel and generating a temporary read-only token valid for two hours. This restricted Zhang to viewing only data related to the current emergency situation and initiated a full-process audit. The data anonymization module, based on the emergency scenario requirements, displayed complete ECG data and medication records, but obscured personal information such as the patient's ID number.

[0024] After Zhang successfully passed multi-factor authentication, the user behavior modeling component of the behavior analysis unit updated his behavioral profile, incorporating this emergency nighttime access into the normal mode. However, the system detected another access request from the administrative office area attempting to retrieve the patient's genetic testing report during the same period. Since this request lacked emergency access rights and significantly deviated from the administrative staff's routine operating pattern, the behavior analysis unit immediately raised the risk level to high risk. The key management submodule completed a global key reset within 32 seconds, and the data anonymization module automatically increased the anonymization strength to the highest level, returning only generalized statistical data types. All these operations were recorded in the security log database, forming a complete audit trail. The next day, the medical supervision platform verified that the incident handling complied with regulations. This practical application case fully demonstrates the system's comprehensive data security protection capabilities in real-world medical scenarios.

[0025] This invention constructs a closed-loop protection system comprising data acquisition, encryption, central processing, and access control modules, achieving end-to-end security management of medical data from generation to access. Employing a dynamic key update mechanism combined with quantum key distribution technology significantly enhances the anti-cracking capability of encrypted data, effectively addressing the risk of key leakage. By using artificial intelligence technology to classify medical data and analyze user behavior, it can identify abnormal access patterns in real time and dynamically adjust the key update frequency and desensitization strategy, achieving intelligent adaptive protection. Integrating vein pattern recognition, dynamic tokens, and multi-factor biometric verification ensures convenient access for legitimate users while effectively preventing identity theft and unauthorized intrusion. The establishment of an emergency access channel and data desensitization processing module ensures timely data retrieval in critical situations and balances data availability and privacy protection in different application scenarios through context-aware desensitization. The collaborative operation of all system modules forms a closed-loop protection chain, enabling rapid response when risks are detected, thereby comprehensively improving the ability to prevent medical data leakage.

[0026] The basic principles and main features of the present invention have been described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are only illustrative of the principles of the present invention. Various changes and modifications can be made to the present invention without departing from the spirit and scope of the present invention. All such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the invention is defined by the appended claims and their equivalents.

Claims

1. A medical data leak prevention system, characterized by, The system comprises a data acquisition module, a data encryption module, a central processing module and an access control module connected in sequence. The data acquisition module is connected with hospital medical equipment and information systems, and is used for collecting physiological parameters, diagnosis reports and image data of patients in real time; the data encryption module encrypts the collected medical data, and comprises a key management sub-module, which adopts a dynamic key update mechanism based on chaos theory, automatically regenerates an encryption key at a preset period, and establishes a bidirectional communication connection with the central processing module; the central processing module comprises a data classification unit based on artificial intelligence and a behavior analysis unit, the data classification unit divides medical data security levels according to sensitivity, and the behavior analysis unit identifies abnormal access modes through a deep learning algorithm; the access control module integrates a biological feature recognition component and a dynamic token verification component, the biological feature recognition component verifies the identity of a user through a vein pattern recognition technology, and the dynamic token verification component generates an access password changing with time.

2. The medical data leakage prevention system of claim 1, wherein, The key management sub-module cooperates with the behavior analysis unit of the central processing module, receives a risk level signal output by the behavior analysis unit, and dynamically adjusts the key update frequency; when the behavior analysis unit detects high-risk abnormal access, the key management sub-module immediately triggers a key reset program and synchronously updates the authentication credentials of all related systems, and stores the expired key in an independent secure log database.

3. The medical data leakage prevention system of claim 2, wherein, The behavior analysis unit further comprises a user behavior modeling component and an environment perception component; the user behavior modeling component constructs a normal operation portrait of an authorized user through a long short-term memory neural network; the environment perception component monitors the source address, access time and device fingerprint information of an access request in real time, and sends a secondary authentication instruction to the access control module when the user operation deviates from the behavior portrait and the environment parameters are abnormal, thereby triggering a multi-factor identity verification process comprising voiceprint verification and iris feature comparison.

4. The medical data leakage prevention system of claim 1, wherein, The data encryption module is implemented based on the BB84 protocol, and specifically: |Ψ>=α|0>+β|1>; where |0> and |1> represent the ground state of the quantum bit, and α and β are complex probability amplitudes satisfying |α| 2 +|β| 2 = 1, in the transmission process, the sender randomly selects the base vector {|0>, |1>} or {|+>, |->} to encode the quantum state, the receiver randomly selects the measurement base to measure, and the two parties compare the base vector selection through the classical channel to finally generate a shared key.

5. The medical data leakage prevention system of claim 1, wherein, The access control module is further provided with an emergency access channel connected with a medical emergency response system, which is automatically activated when critical values of patient vital signs are monitored; when the emergency access channel is enabled, a temporary access token is automatically generated and the data operation permission is limited to a read-only mode, and at the same time, whole-process operation audit tracking is started, and all data operations accessed through the emergency channel are marked as special events and uploaded to a medical supervision platform for archiving; the validity period of the temporary access token is dynamically adjusted according to the severity of the critical values of the patient, and the temporary access token is automatically invalidated when the vital signs of the patient return to normal.

6. The medical data leakage prevention system of claim 1, wherein, The system further comprises a data desensitization processing module, which is arranged between the central processing module and the access control module, and comprises a field-level desensitization rule library and a context-aware desensitization engine; the field-level desensitization rule library stores desensitization strategies corresponding to data of different security levels; the context-aware desensitization engine automatically selects a desensitization algorithm according to the role permission and access scene of an access person, retains data statistical features but hides personal identifiers for research use access requests, and displays complete diagnosis information but blurs patient identity information for clinical consultation scenes. The data desensitization processing module works with the behavior analysis unit to automatically increase the desensitization strength to the highest level when an abnormal access mode is detected.

7. A medical data leakage prevention method characterized by comprising: The system of any one of claims 1-6, comprising the following steps: collecting physiological parameters of patients, diagnosis reports and image data in medical devices and information systems in real time through a data acquisition module; encrypting the collected data using quantum key distribution technology through a data encryption module, and updating the dynamic key based on chaos theory through a key management sub-module; classifying the medical data security level according to the sensitivity through the data classification unit of the central processing module, and identifying the abnormal access mode through the deep learning algorithm of the behavior analysis unit; realizing identity authentication through the integration of vein pattern recognition and dynamic token verification through the access control module; the data acquisition module, the data encryption module, the central processing module and the access control module form a closed-loop data security protection link.

Images