Hospital personnel information sharing and health management collaborative decision-making system with privacy protection function

Abstract

The invention belongs to the technical field of hospital personnel information protection, and discloses a hospital personnel information sharing and health management collaborative decision-making system with a privacy protection function. Comprising a cross-mechanism data processing module, a privacy calculation sharing module, a personnel management optimization module, an internal sensitive information protection module, a personnel decision data acquisition module, a big data analysis module and a block chain supervision module. According to the method, multiple technologies such as federated learning and Shami r secret sharing are fused, multi-level protection is constructed, single encryption limitation is overcome, fine grading authorization and dynamic desensitization are matched, permission is adjusted in time, and the risk of data leakage is reduced; during data processing and sharing, data standards are unified, cross-mechanism integration cost and errors are reduced, internal and external data are comprehensively collected, and decision comprehensiveness is improved; the intelligent degree of personnel management is greatly improved, the functions of qualification renewal reminding, cross-department intelligent deployment and the like are achieved, manual operation is reduced, and efficiency and accuracy are improved.

Description

Technical Field

[0001] This invention belongs to the field of hospital personnel information protection technology, specifically a hospital personnel information sharing and health management collaborative decision-making system with privacy protection functions. Background Technology

[0002] In the wave of digital transformation in the healthcare industry, the importance of collaborative decision-making in hospital personnel information management and health management is becoming increasingly prominent. Currently, with the continuous development of related technologies, some systems use database technology to achieve basic storage and retrieval of personnel information, and network communication technology to achieve data transmission within a certain range. Some systems have also introduced simple data analysis functions to generate routine reports. At the same time, to ensure data security, encryption technology is also applied in some scenarios. These technologies provide certain support for collaborative hospital personnel information management and health management, and promote the informatization process of the healthcare industry.

[0003] However, existing technologies still have shortcomings in practical use: they often rely on single technologies, such as encryption alone, which offers limited protection against complex attacks; their access control is rudimentary and untimely, making them prone to data leaks; data standards vary across hospitals, requiring extensive conversion and cleaning for cross-institutional integration, which is costly and error-prone; data collection is incomplete, and the lack of external data integration affects the comprehensiveness of decision-making; the level of intelligence is low, with qualification management and personnel allocation relying on manual processes, which are inefficient and prone to errors; data analysis is simplistic, lacking in-depth analysis and failing to provide accurate information for management, making it difficult to predict changes in talent and resources; data traceability relies on manual processes and traditional databases, which are inefficient and easily tampered with; qualification verification relies on manual communication with authoritative institutions, which is cumbersome, time-consuming, prone to omissions, and poses security risks. Therefore, improvements are needed. Summary of the Invention

[0004] The purpose of this invention is to provide a hospital personnel information sharing and health management collaborative decision-making system with privacy protection functions, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a hospital personnel information sharing and health management collaborative decision-making system with privacy protection functions, comprising: a cross-institutional data processing module, a privacy computing and sharing module, a personnel management optimization module, an internal sensitive information protection module, a personnel decision data collection module, a big data analysis module, and a blockchain supervision module, wherein,

[0006] The cross-institutional data processing module is used to import data from different medical institutions, check the data structure, type, distribution and potential problems, and perform unified and standardized processing on core data such as the qualifications, practice status and performance evaluation of medical staff.

[0007] The privacy computing sharing module adopts a privacy computing scheme that combines federated learning and Shamir secret sharing to ensure the security of privacy data when sharing data across institutions;

[0008] The personnel management optimization module is used to link the validity period of medical staff's professional certificates and training record data, provide timely reminders for the renewal of medical staff's qualifications, and quickly match qualifications and available status through intelligent screening algorithms when transferring across departments, thus shortening the transfer response time.

[0009] The internal sensitive information protection module uses hierarchical authorization and dynamic data desensitization technology to protect sensitive fields such as ID card number and salary information to prevent internal information leakage.

[0010] The personnel decision data acquisition module is used to collect real-time dynamic personnel data of hospitals within the region.

[0011] The big data analysis module is used to perform big data analysis on the collected hospital personnel dynamic data in the region, generate a heat map of medical and nursing resource supply and demand and a talent loss early warning report, intuitively display the distribution and demand of medical and nursing resources, provide early warning of talent loss risks, and provide accurate data support for the formulation of policies such as targeted training and resource allocation, so as to make health management decisions more scientific and reasonable.

[0012] The blockchain supervision module introduces blockchain technology to achieve full lifecycle traceability of personnel data. Combined with smart contracts, it automatically verifies the validity of professional qualifications, ensuring that every link of personnel data from generation to use is traceable and queryable. The smart contracts can automatically verify the professional qualifications of medical personnel and promptly handle violations.

[0013] As a preferred embodiment of the present invention, the cross-institutional data processing module is divided into a data cleaning and preprocessing unit and a data standardization and transformation unit. The data cleaning and preprocessing unit includes:

[0014] Unified data format: Use ETL tools to convert heterogeneous data from different hospitals into a unified format and define standard fields;

[0015] Missing value handling: Missing data are filled by mean imputation, median imputation, or prediction model based on similar medical staff (KNN algorithm);

[0016] Outlier detection and correction: IQR statistical methods are used to identify outliers, a threshold range is set, and data that exceeds the range is identified and corrected through box plot analysis or clustering algorithms;

[0017] Duplicate data processing: Remove duplicate records using unique identifiers or similarity comparisons;

[0018] Data discretization and binning: dividing continuous data into discrete intervals to reduce noise;

[0019] Data quality verification: Check data quality after each processing step to ensure that the cleaning operation does not introduce new problems;

[0020] The data standardization and transformation unit includes:

[0021] Standardization of qualification data: Establish a medical professional classification system and map various qualification certificates to a unified coding table;

[0022] Standardization of professional practice status: Define status codes and use regular expressions to match and transform raw data;

[0023] Performance appraisal standardization: Z-score standardization or Min-Max normalization methods are used to unify the scoring systems of different hospitals to the 0-1 range.

[0024] As a preferred embodiment of the present invention, the privacy computing sharing module is divided into a federated learning unit and a Shamir secret sharing application unit, wherein the federated learning unit includes:

[0025] Model training collaboration: Each medical institution trains its own model using its own medical staff data locally, and transmits the model parameters to the central node through an encrypted communication protocol. The central node aggregates and updates the received parameters using the secure aggregation algorithm FedAvg to generate a global model, and then feeds back the updated model parameters to each medical institution.

[0026] Encrypted communication guarantee: During the transmission of model parameters, the TLS 1.3 protocol is used for encrypted transmission, and homomorphic encryption technology is used to perform encrypted calculations on the model parameters to ensure the security of data during transmission;

[0027] Privacy protection mechanism: Differential privacy and other technologies are used to add Laplacian noise during model training to further protect data privacy and prevent the original data from being inferred from the model output;

[0028] The Shamir secret sharing application unit includes:

[0029] Data segmentation and storage: Sensitive data of medical staff is segmented into n shares according to the Shamir secret sharing algorithm, and it is set that at least k shares (k≤n) are needed to restore the original data. These shares are stored in different participating medical institutions or security nodes respectively.

[0030] Data restoration and access control: When it is necessary to access the original data, only when the participants who meet the preset conditions cooperate and provide their respective storage shares can the original data be restored through the algorithm;

[0031] Security Audit and Tracking: Detailed audit records are kept of the data segmentation, storage, and restoration processes. Zero-knowledge proofs are used to verify the legality of operations. Once abnormal data access behavior is detected, it is tracked and investigated through audit records to ensure the security and compliance of the data sharing process.

[0032] As a preferred embodiment of the present invention, the specific process of the Shamir secret sharing application unit is as follows:

[0033] Encryption process:

[0034] Construct the polynomial: Select t-1 random numbers and construct the polynomial f(x) = a0 + a1x + ... + a t-1 x i-1 , where a0 = S (secret);

[0035] Generate shares: Take n numbers x1, x2, ..., x n Calculate f(x1), f(x2), ..., f(x) n ), will (x i ,f(x i Distribute to n servers;

[0036] Decryption process:

[0037] Collect t shares: Obtain (x1, y1), ..., (x1, y1) from any t servers. t ,y t );

[0038] Solving the polynomial: Solving for the coefficients a0, a1, ..., a2 using matrix operations. t-1 Construct a polynomial f(x);

[0039] To recover the secret: Substitute x = 0 into the polynomial to obtain the original secret S = a0.

[0040] As a preferred embodiment of the present invention, the personnel management optimization module is divided into a qualification renewal reminder unit and an inter-departmental transfer intelligent screening unit, wherein the qualification renewal reminder unit includes:

[0041] Validity monitoring: Real-time linking of medical staff's professional certificate information, scanning certificate expiration dates through scheduled tasks, and setting three levels of early warning: 60 days, 30 days, and 15 days in advance, to ensure that relevant personnel are notified in a timely manner before the certificate expires;

[0042] Multiple reminder channels: Employing various reminder methods ensures that medical staff receive reminder messages in a timely manner;

[0043] Renewal process guidance: In addition to informing you that the certificate is about to expire, the reminder message also provides guidance on the renewal process;

[0044] The cross-departmental allocation intelligent screening unit includes:

[0045] Intelligent screening algorithm: Based on the job requirements of the target department, detailed qualification matching rules are formulated. Through the rule engine Drools, the qualification information of medical staff is quickly matched and qualified personnel are screened out.

[0046] Idle Status Inquiry: Real-time synchronization with the hospital's scheduling system to obtain the work arrangements and idle status information of medical staff. During the screening process, medical staff who are currently on duty are excluded to ensure that the assigned personnel can be in place in a timely manner.

[0047] Comprehensive score ranking: Combining the work performance of medical staff and patient evaluation factors, the weights are calculated using the analytic hierarchy process (AHP) to generate a comprehensive score for each eligible medical staff member. The scores are then ranked from highest to lowest to generate a matching ranking list. Based on the ranking results, the optimal allocation plan is recommended.

[0048] As a preferred embodiment of the present invention, the internal sensitive information protection module is divided into a hierarchical authorization management unit and a dynamic data desensitization unit, wherein the hierarchical authorization management unit specifically includes:

[0049] Access permissions are divided into multiple levels based on the job responsibilities and needs of different positions within the hospital: general staff level, department head level, and hospital management level. Each level corresponds to different data access scopes and operation permissions.

[0050] Permission application and approval: Employees apply for corresponding permissions according to work needs, fill in the permission application form, explain the reason and scope of the permission application, and the superior or relevant department approves the application and dynamically updates the permissions to ensure the rationality and security of permission allocation.

[0051] Dynamic permission adjustment: When an employee's position changes or work requirements change, permission changes are automatically triggered.

[0052] The dynamic data desensitization unit specifically includes:

[0053] Customized anonymization rules: Define regular expression anonymization rules based on different data types and access scenarios: For ID card numbers, use partial hiding, only displaying the first and last few digits, with asterisks used to represent the middle digits; For salary data, anonymize the data proportionally.

[0054] Real-time data masking: During data display or transmission, requests are intercepted through the MySQL Proxy layer. The system dynamically masks the data and returns it based on the user's permissions and masking rules, ensuring that the user can only see the masked data and cannot obtain the complete sensitive information.

[0055] Desensitization effectiveness evaluation: Regularly conduct sampling checks on the data desensitization effectiveness to check for incomplete or excessive desensitization. Based on the evaluation results, adjust the desensitization rules in a timely manner to ensure the rationality and effectiveness of data desensitization.

[0056] As a preferred embodiment of the present invention, the personnel decision data acquisition module is divided into a multi-channel data acquisition unit and a data quality monitoring unit, wherein the multi-channel data acquisition unit includes:

[0057] Hospital Information System Integration: Through API interfaces, we integrate with the internal information systems of various hospitals to obtain real-time data from HR systems, attendance systems, and training systems, and collect relevant data from medical staff.

[0058] Regular data reporting: For data that cannot be obtained through system integration, regular data reporting will be adopted. Each hospital will report the relevant data to the system according to the prescribed format and time requirements.

[0059] External data integration: Collect external data related to hospital personnel and extract key indicators using NLP technology;

[0060] The data quality monitoring unit includes:

[0061] Data accuracy verification: The accuracy of the collected data is verified by comparing it with the original data or relevant authoritative data, verifying data consistency through hash algorithms, and checking for errors or inconsistencies in the data.

[0062] Data integrity check: Define the rules for required fields, check whether the collected data is complete, whether there are missing fields or records, and trigger an alarm when the missing data exceeds 10%, and notify the relevant hospitals to supplement or correct it;

[0063] Data timeliness monitoring: Set a data update threshold (within 24 hours). If the data is not updated within the time limit, it is marked as abnormal. For abnormal data, an alarm will be issued to ensure that personnel decisions are based on the latest data.

[0064] As a preferred embodiment of the present invention, the big data analysis module is divided into a medical resource supply and demand heat map generation unit and a talent loss early warning report generation unit. The medical resource supply and demand heat map generation unit includes:

[0065] Data aggregation and analysis: The collected hospital personnel dynamic data within the region is aggregated and analyzed according to different regions, departments and time dimensions. The Spark cluster is used to process the regional data and calculate the number of medical staff, workload and patient demand indicators for each region and department.

[0066] Heatmap visualization: Based on the analysis results, generate a heatmap of supply and demand for medical resources using ECharts or Tableau, using different colors or densities to represent the differences in supply and demand for resources;

[0067] Trend prediction and early warning: Combining historical data and current development trends, the LSTM neural network is used to predict the supply and demand of medical resources in the next 3 months. When a shortage or surplus of resources is predicted in a certain region or department, an early warning message is issued to provide a basis for resource allocation decisions.

[0068] The talent loss early warning report generation unit includes:

[0069] Multi-dimensional data analysis: constructing feature engineering from multiple dimensions such as work performance, career development, compensation and benefits, and work environment of medical staff;

[0070] Early warning model construction: A talent loss early warning model is constructed using machine learning algorithms. Based on historical talent loss data and related influencing factors, the XGBoost algorithm is used to train the talent loss early warning model and determine the early warning threshold. When the model detects that the relevant indicators of a medical staff member reach or exceed the early warning threshold, the staff member is marked as a potential employee to leave.

[0071] Report Generation and Interpretation: Based on the early warning results, a talent loss early warning report is generated. The report details the basic information of potential employees who may leave, the reasons for the early warning, and the risk level. At the same time, the report is interpreted to provide targeted suggestions and measures for hospital management to reduce the risk of talent loss.

[0072] As a preferred technical solution of the present invention, the blockchain supervision module is divided into a personnel data full lifecycle traceability unit and a smart contract automatic verification unit, wherein the personnel data full lifecycle traceability unit includes:

[0073] On-chain data storage: The hash value of the data is written to the blockchain to record creation, modification, deletion, transmission and usage operations;

[0074] Source tracing query interface: Provides a source tracing query interface, through which hospital management and relevant departments can query the full life cycle information of personnel data. By entering the key identifier of the data, they can obtain all operation records and related information of the data from its generation to its current status.

[0075] Data integrity verification: Utilizing the hash algorithm and chain structure of blockchain, the integrity of personnel data is verified. By comparing the hash value of the data on the blockchain with the hash value stored locally, it is ensured that the data has not been tampered with during transmission and storage.

[0076] The smart contract automatic verification unit includes:

[0077] Qualification verification rule setting: Pre-set the professional qualification verification rules for medical personnel in the smart contract, including certificate type, validity period, and professional scope;

[0078] Automated verification execution: Connect to an authoritative database via Oracle service (Chainlink) to verify qualification status in real time. When the qualification information of medical personnel changes or the verification time arrives, the smart contract automatically triggers the verification process. By connecting with the database of relevant authoritative institutions, it verifies in real time whether the qualifications of medical personnel are still valid.

[0079] Handling of violations: When the smart contract verifies that there are problems with the qualifications of medical staff, it will automatically take corresponding measures and record the violation on the blockchain as a basis for subsequent management and punishment.

[0080] The beneficial effects of this invention are as follows:

[0081] This invention integrates multiple technologies such as federated learning and Shamir secret sharing to construct a multi-layered protection system, overcoming the limitations of single encryption. Combined with fine-grained hierarchical authorization and dynamic desensitization, it allows for timely adjustment of permissions and reduces the risk of data leakage. In data processing and sharing, it unifies data standards, reducing cross-organizational integration costs and errors, and comprehensively collects internal and external data to improve the comprehensiveness of decision-making. The intelligence level of human resource management is significantly improved, with functions such as qualification renewal reminders and intelligent cross-departmental allocation reducing manual operations and improving efficiency and accuracy. In decision support, big data analysis deeply mines data, generating intuitive heatmaps and early warning reports, providing management with precise data for predicting changes in talent and resources. At the regulatory level, blockchain is used to achieve full lifecycle data traceability and automatic smart contract verification, improving efficiency, preventing tampering, simplifying qualification verification processes, and eliminating security risks. Attached Figure Description

[0082] Figure 1 This is a system block diagram of the present invention. Detailed Implementation

[0083] 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.

[0084] like Figure 1As shown, this embodiment of the invention provides a hospital personnel information sharing and collaborative decision-making system for health management with privacy protection functions, including: a cross-institutional data processing module, a privacy computing and sharing module, a personnel management optimization module, an internal sensitive information protection module, a personnel decision data collection module, a big data analysis module, and a blockchain supervision module.

[0085] The cross-institutional data processing module is used to import data from different medical institutions, check the data structure, type, distribution and potential problems, and perform unified and standardized processing of core data such as medical staff qualifications, practice status and performance evaluation.

[0086] The privacy-preserving computation sharing module employs a privacy-preserving computation scheme that combines federated learning with Shamir secret sharing to ensure the security of privacy data when sharing data across organizations.

[0087] The personnel management optimization module is used to link the validity period of medical staff's professional certificates and training record data, provide timely reminders for the renewal of medical staff's qualifications, and quickly match qualifications and available status through intelligent screening algorithms when transferring across departments, thus shortening the transfer response time.

[0088] The internal sensitive information protection module uses hierarchical authorization and dynamic data desensitization technology to protect sensitive fields such as ID card numbers and salary information to prevent internal information leakage.

[0089] The personnel decision data acquisition module is used to collect real-time personnel dynamic data of hospitals within the region;

[0090] The big data analysis module is used to perform big data analysis on the collected hospital personnel dynamic data in the region, generate heat maps of medical and nursing resource supply and demand and talent loss early warning reports, intuitively display the distribution and demand of medical and nursing resources, provide early warning of talent loss risks, and provide accurate data support for the formulation of policies such as targeted training and resource allocation, so as to make health management decisions more scientific and reasonable.

[0091] The blockchain supervision module introduces blockchain technology to achieve full lifecycle traceability of personnel data. Combined with smart contracts, it automatically verifies the validity of professional qualifications, ensuring that every link of personnel data from generation to use is traceable and queryable. The smart contracts can automatically verify the professional qualifications of medical personnel and promptly handle violations.

[0092] The above solution can effectively address the problems of data silos, high risk of privacy leaks, lagging decision support, and low regulatory efficiency in the current hospital personnel information management system.

[0093] The cross-institutional data processing module is divided into a data cleaning and preprocessing unit and a data standardization and transformation unit. The data cleaning and preprocessing unit includes:

[0094] Unified data format: Use ETL tools to convert heterogeneous data from different hospitals into a unified format and define standard fields;

[0095] Missing value handling: Missing data are filled by mean imputation, median imputation, or prediction model based on similar medical staff (KNN algorithm);

[0096] Outlier detection and correction: IQR statistical methods are used to identify outliers, a threshold range is set, and data that exceeds the range is identified and corrected through box plot analysis or clustering algorithms;

[0097] Duplicate data processing: Remove duplicate records using unique identifiers or similarity comparisons;

[0098] Data discretization and binning: dividing continuous data into discrete intervals to reduce noise;

[0099] Data quality verification: Check data quality after each processing step to ensure that the cleaning operation does not introduce new problems;

[0100] The data standardization and transformation unit includes:

[0101] Standardization of qualification data: Establish a medical professional classification system and map various qualification certificates to a unified coding table;

[0102] Standardization of professional practice status: Define status codes and use regular expressions to match and transform raw data;

[0103] Performance appraisal standardization: Z-score standardization or Min-Max normalization methods are used to unify the scoring systems of different hospitals to the 0-1 range.

[0104] Heterogeneous data includes XML, CSV, database tables, etc. Standard fields include medical staff ID and practice status code, while defined status codes include "01-On-the-job", "02-Resigned", "03-On Leave", etc. Different hospitals use different database formats and field naming rules when recording medical staff qualifications, practice status and performance evaluation data. For example, some hospitals use "doctor number" to record, while others use "employee ID". The cross-institutional data processing module will unify the naming of these fields with standard names, such as "unique identifier code for medical staff", and unify the data types to ensure that the data formats of numeric and character types are consistent.

[0105] The privacy-sharing module is divided into a federated learning unit and a Shamir secret-sharing application unit. The federated learning unit includes:

[0106] Model training collaboration: Each medical institution trains its own model using its own medical staff data locally, and transmits the model parameters to the central node through an encrypted communication protocol. The central node aggregates and updates the received parameters using the secure aggregation algorithm FedAvg to generate a global model, and then feeds back the updated model parameters to each medical institution.

[0107] Encrypted communication guarantee: During the transmission of model parameters, the TLS 1.3 protocol is used for encrypted transmission, and homomorphic encryption technology is used to perform encrypted calculations on the model parameters to ensure the security of data during transmission;

[0108] Privacy protection mechanism: Differential privacy and other technologies are used to add Laplacian noise during model training to further protect data privacy and prevent the original data from being inferred from the model output;

[0109] The Shamir Secret Sharing application unit includes:

[0110] Data segmentation and storage: Sensitive data of medical staff is segmented into n shares according to the Shamir secret sharing algorithm, and it is set that at least k shares (k≤n) are needed to restore the original data. These shares are stored in different participating medical institutions or security nodes respectively.

[0111] Data restoration and access control: When it is necessary to access the original data, only when the participants who meet the preset conditions cooperate and provide their respective storage shares can the original data be restored through the algorithm;

[0112] Security Audit and Tracking: Detailed audit records are kept of the data segmentation, storage, and restoration processes. Zero-knowledge proofs are used to verify the legality of operations. Once abnormal data access behavior is detected, it is tracked and investigated through audit records to ensure the security and compliance of the data sharing process.

[0113] The existence of encrypted communication guarantees ensures that only authorized central nodes and medical institutions can decrypt data, thereby preventing data from being stolen or tampered with during transmission. During federated learning, medical institutions cannot obtain the original data from other institutions, but can only improve model performance through the interaction of model parameters.

[0114] The specific process of the Shamir secret sharing application unit is as follows:

[0115] Encryption process:

[0116] Construct the polynomial: Select t-1 random numbers and construct the polynomial f(x) = a0 + a1x + ... + a t-1 x i-1 , where a0 = S (secret);

[0117] Generate shares: Take n numbers x1, x2, ..., x n Calculate f(x1), f(x2), ..., f(x) n ), will (x i ,f(x i Distribute to n servers;

[0118] Decryption process:

[0119] Collect t shares: Obtain (x1, y1), ..., (x1, y1) from any t servers. t ,y t );

[0120] Solving the polynomial: Solving for the coefficients a0, a1, ..., a2 using matrix operations. t-1 Construct a polynomial f(x);

[0121] To recover the secret: Substitute x = 0 into the polynomial to obtain the original secret S = a0.

[0122] The Shamir Secret Sharing Application Unit ensures that no single organization or group of organizations can obtain complete sensitive information alone, thus guaranteeing data privacy and security.

[0123] The personnel management optimization module is divided into a qualification renewal reminder unit and a cross-departmental transfer intelligent screening unit. The qualification renewal reminder unit includes:

[0124] Validity monitoring: Real-time linking of medical staff's professional certificate information, scanning certificate expiration dates through scheduled tasks, and setting three levels of early warning: 60 days, 30 days, and 15 days in advance, to ensure that relevant personnel are notified in a timely manner before the certificate expires;

[0125] Multiple reminder channels: Employing various reminder methods ensures that medical staff receive reminder messages in a timely manner;

[0126] Renewal process guidance: In addition to informing you that the certificate is about to expire, the reminder message also provides guidance on the renewal process;

[0127] The cross-departmental intelligent screening unit includes:

[0128] Intelligent screening algorithm: Based on the job requirements of the target department, detailed qualification matching rules are formulated. Through the rule engine Drools, the qualification information of medical staff is quickly matched and qualified personnel are screened out.

[0129] Idle Status Inquiry: Real-time synchronization with the hospital's scheduling system to obtain the work arrangements and idle status information of medical staff. During the screening process, medical staff who are currently on duty are excluded to ensure that the assigned personnel can be in place in a timely manner.

[0130] Comprehensive score ranking: Combining the work performance of medical staff and patient evaluation factors, the weights are calculated using the analytic hierarchy process (AHP) to generate a comprehensive score for each eligible medical staff member. The scores are then ranked from highest to lowest to generate a matching ranking list. Based on the ranking results, the optimal allocation plan is recommended.

[0131] Multiple reminder methods are provided, including in-system message push, SMS reminders, and email notifications; the renewal process guide includes the required materials, processing locations, and methods, thus facilitating medical staff to smoothly complete the certificate renewal procedures; and optimal personnel allocation solutions are provided for administrators, including the allocation of personnel to surgical departments, requiring the matching of medical staff with surgical professional qualifications and certain surgical experience.

[0132] The internal sensitive information protection module is divided into a hierarchical authorization management unit and a dynamic data desensitization unit. The hierarchical authorization management unit specifically includes:

[0133] Access permissions are divided into multiple levels based on the job responsibilities and needs of different positions within the hospital: general staff level, department head level, and hospital management level. Each level corresponds to different data access scopes and operation permissions.

[0134] Permission application and approval: Employees apply for corresponding permissions according to work needs, fill in the permission application form, explain the reason and scope of the permission application, and the superior or relevant department approves the application and dynamically updates the permissions to ensure the rationality and security of permission allocation.

[0135] Dynamic permission adjustment: When an employee's position changes or work requirements change, permission changes are automatically triggered.

[0136] The dynamic data anonymization unit specifically includes:

[0137] Customized anonymization rules: Define regular expression anonymization rules based on different data types and access scenarios: For ID card numbers, use partial hiding, only displaying the first and last few digits, with asterisks used to represent the middle digits; For salary data, anonymize the data proportionally.

[0138] Real-time data masking: During data display or transmission, requests are intercepted through the MySQL Proxy layer. The system dynamically masks the data and returns it based on the user's permissions and masking rules, ensuring that the user can only see the masked data and cannot obtain the complete sensitive information.

[0139] Desensitization effectiveness evaluation: Regularly conduct sampling checks on the data desensitization effectiveness to check for incomplete or excessive desensitization. Based on the evaluation results, adjust the desensitization rules in a timely manner to ensure the rationality and effectiveness of data desensitization.

[0140] For example, when an employee is promoted to a higher position, they are granted corresponding higher permissions; when an employee is transferred from their original position, permissions that are no longer needed are revoked; regarding salary data, it is anonymized according to a certain ratio, such as displaying it as 5000-8000 yuan.

[0141] The personnel decision-making data collection module is divided into a multi-channel data collection unit and a data quality monitoring unit. The multi-channel data collection unit includes:

[0142] Hospital Information System Integration: Through API interfaces, we integrate with the internal information systems of various hospitals to obtain real-time data from HR systems, attendance systems, and training systems, and collect relevant data from medical staff.

[0143] Regular data reporting: For data that cannot be obtained through system integration, regular data reporting will be adopted. Each hospital will report the relevant data to the system according to the prescribed format and time requirements.

[0144] External data integration: Collect external data related to hospital personnel and extract key indicators using NLP technology;

[0145] The data quality monitoring unit includes:

[0146] Data accuracy verification: The accuracy of the collected data is verified by comparing it with the original data or relevant authoritative data, verifying data consistency through hash algorithms, and checking for errors or inconsistencies in the data.

[0147] Data integrity check: Define the rules for required fields, check whether the collected data is complete, whether there are missing fields or records, and trigger an alarm when the missing data exceeds 10%, and notify the relevant hospitals to supplement or correct it;

[0148] Data timeliness monitoring: Set a data update threshold (within 24 hours). If the data is not updated within the time limit, it is marked as abnormal. For abnormal data, an alarm will be issued to ensure that personnel decisions are based on the latest data.

[0149] Data that cannot be obtained through system integration includes career development plans and satisfaction survey results for medical staff; external data related to hospital personnel, such as industry talent market dynamics and policy and regulatory changes, can be obtained by cooperating with relevant data providers or government departments and integrated into the system, which can provide a more comprehensive reference for personnel decisions.

[0150] The big data analysis module is divided into a heat map generation unit for medical resource supply and demand and a talent loss early warning report generation unit. The heat map generation unit for medical resource supply and demand includes:

[0151] Data aggregation and analysis: The collected hospital personnel dynamic data within the region is aggregated and analyzed according to different regions, departments and time dimensions. The Spark cluster is used to process the regional data and calculate the number of medical staff, workload and patient demand indicators for each region and department.

[0152] Heatmap visualization: Based on the analysis results, generate a heatmap of supply and demand for medical resources using ECharts or Tableau, using different colors or densities to represent the differences in supply and demand for resources;

[0153] Trend prediction and early warning: Combining historical data and current development trends, the LSTM neural network is used to predict the supply and demand of medical resources in the next 3 months. When a shortage or surplus of resources is predicted in a certain region or department, an early warning message is issued to provide a basis for resource allocation decisions.

[0154] The talent loss early warning report generation unit includes:

[0155] Multi-dimensional data analysis: constructing feature engineering from multiple dimensions such as work performance, career development, compensation and benefits, and work environment of medical staff;

[0156] Early warning model construction: A talent loss early warning model is constructed using machine learning algorithms. Based on historical talent loss data and related influencing factors, the XGBoost algorithm is used to train the talent loss early warning model and determine the early warning threshold. When the model detects that the relevant indicators of a medical staff member reach or exceed the early warning threshold, the staff member is marked as a potential employee to leave.

[0157] Report Generation and Interpretation: Based on the early warning results, a talent loss early warning report is generated. The report details the basic information of potential employees who may leave, the reasons for the early warning, and the risk level. At the same time, the report is interpreted to provide targeted suggestions and measures for hospital management to reduce the risk of talent loss.

[0158] When displaying heat maps, red can be set to indicate resource shortages and green to indicate resource surpluses. Through heat maps, hospital management can intuitively see the distribution of medical and nursing resources in different regions and departments. Multi-dimensional data analysis includes analyzing whether the work performance of medical and nursing staff has declined and whether there are frequent expressions of resignation intentions. After interpreting the report, targeted suggestions and measures are provided, including conducting retention talks and adjusting salaries and benefits.

[0159] The blockchain supervision module is divided into a personnel data full lifecycle traceability unit and a smart contract automatic verification unit. The personnel data full lifecycle traceability unit includes:

[0160] On-chain data storage: The hash value of the data is written to the blockchain to record creation, modification, deletion, transmission and usage operations;

[0161] Source tracing query interface: Provides a source tracing query interface, through which hospital management and relevant departments can query the full life cycle information of personnel data. By entering the key identifier of the data, they can obtain all operation records and related information of the data from its generation to its current status.

[0162] Data integrity verification: Utilizing the hash algorithm and chain structure of blockchain, the integrity of personnel data is verified. By comparing the hash value of the data on the blockchain with the hash value stored locally, it is ensured that the data has not been tampered with during transmission and storage.

[0163] The smart contract automatic verification unit includes:

[0164] Qualification verification rule setting: Pre-set the professional qualification verification rules for medical personnel in the smart contract, including certificate type, validity period, and professional scope;

[0165] Automated verification execution: Connect to an authoritative database via Oracle service (Chainlink) to verify qualification status in real time. When the qualification information of medical personnel changes or the verification time arrives, the smart contract automatically triggers the verification process. By connecting with the database of relevant authoritative institutions, it verifies in real time whether the qualifications of medical personnel are still valid.

[0166] Handling of violations: When the smart contract verifies that there are problems with the qualifications of medical staff, it will automatically take corresponding measures and record the violation on the blockchain as a basis for subsequent management and punishment.

[0167] When data is stored on the blockchain, such as the onboarding information, job change records, and training information of medical staff, it is stored in the distributed ledger of the blockchain in an immutable manner. Specific measures for handling violations include restricting their professional authority, issuing alerts to hospital management, and freezing accounts.

[0168] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0169] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A hospital personnel information sharing and collaborative decision-making system for health management with privacy protection features, characterized in that: It includes: a cross-organizational data processing module, a privacy computing and sharing module, a human resources management optimization module, an internal sensitive information protection module, a human resources decision-making data collection module, a big data analysis module, and a blockchain supervision module. The cross-institutional data processing module is used to import data from different medical institutions, check the data structure, type, distribution and potential problems, and perform unified and standardized processing on core data such as the qualifications, practice status and performance evaluation of medical staff. The privacy computing sharing module adopts a privacy computing scheme that combines federated learning and Shamir secret sharing to ensure the security of privacy data when sharing data across institutions; The personnel management optimization module is used to link the validity period of medical staff's professional certificates and training record data, provide timely reminders for the renewal of medical staff's qualifications, and quickly match qualifications and available status through intelligent screening algorithms when transferring across departments, thus shortening the transfer response time. The internal sensitive information protection module uses hierarchical authorization and dynamic data desensitization technology to protect sensitive fields such as ID card number and salary information to prevent internal information leakage. The personnel decision data acquisition module is used to collect real-time dynamic personnel data of hospitals within the region. The big data analysis module is used to perform big data analysis on the collected hospital personnel dynamic data in the region, generate a heat map of medical and nursing resource supply and demand and a talent loss early warning report, intuitively display the distribution and demand of medical and nursing resources, provide early warning of talent loss risks, and provide accurate data support for the formulation of policies such as targeted training and resource allocation, so as to make health management decisions more scientific and reasonable. The blockchain supervision module introduces blockchain technology to achieve full lifecycle traceability of personnel data. Combined with smart contracts, it automatically verifies the validity of professional qualifications, ensuring that every link of personnel data from generation to use is traceable and queryable. The smart contracts can automatically verify the professional qualifications of medical personnel and promptly handle violations.

2. The hospital personnel information sharing and collaborative decision-making system for health management with privacy protection function as described in claim 1, characterized in that: The cross-institutional data processing module is divided into a data cleaning and preprocessing unit and a data standardization and transformation unit. The data cleaning and preprocessing unit includes: Unified data format: Use ETL tools to convert heterogeneous data from different hospitals into a unified format and define standard fields; Missing value handling: Missing data are filled by mean imputation, median imputation, or prediction model based on similar medical staff (KNN algorithm); Outlier detection and correction: IQR statistical methods are used to identify outliers, a threshold range is set, and data that exceeds the range is identified and corrected through box plot analysis or clustering algorithms; Duplicate data processing: Remove duplicate records using unique identifiers or similarity comparisons; Data discretization and binning: dividing continuous data into discrete intervals to reduce noise; Data quality verification: Check data quality after each processing step to ensure that the cleaning operation does not introduce new problems; The data standardization and transformation unit includes: Standardization of qualification data: Establish a medical professional classification system and map various qualification certificates to a unified coding table; Standardization of professional practice status: Define status codes and use regular expressions to match and transform raw data; Performance appraisal standardization: Z-score standardization or Min-Max normalization methods are used to unify the scoring systems of different hospitals to the 0-1 range.

3. The hospital personnel information sharing and collaborative decision-making system for health management with privacy protection function according to claim 1, characterized in that: The privacy-sharing computation module is divided into a federated learning unit and a Shamir secret-sharing application unit. The federated learning unit includes: Model training collaboration: Each medical institution trains its own model using its own medical staff data locally, and transmits the model parameters to the central node through an encrypted communication protocol. The central node aggregates and updates the received parameters using the secure aggregation algorithm FedAvg to generate a global model, and then feeds back the updated model parameters to each medical institution. Encrypted communication guarantee: During the transmission of model parameters, the TLS 1.3 protocol is used for encrypted transmission, and homomorphic encryption technology is used to perform encrypted calculations on the model parameters to ensure the security of data during transmission; Privacy protection mechanism: Differential privacy and other technologies are used to add Laplacian noise during model training to further protect data privacy and prevent the original data from being inferred from the model output; The Shamir secret sharing application unit includes: Data segmentation and storage: Sensitive data of medical staff is segmented into n shares according to the Shamir secret sharing algorithm, and it is set that at least k shares (k≤n) are needed to restore the original data. These shares are stored in different participating medical institutions or security nodes respectively. Data restoration and access control: When it is necessary to access the original data, only when the participants who meet the preset conditions cooperate and provide their respective storage shares can the original data be restored through the algorithm; Security Audit and Tracking: Detailed audit records are kept of the data segmentation, storage, and restoration processes. Zero-knowledge proofs are used to verify the legality of operations. Once abnormal data access behavior is detected, it is tracked and investigated through audit records to ensure the security and compliance of the data sharing process.

4. The hospital personnel information sharing and collaborative decision-making system for health management with privacy protection function according to claim 3, characterized in that: The specific process of the Shamir secret sharing application unit is as follows: Encryption process: Construct the polynomial: Select t-1 random numbers and construct the polynomial f(x) = a0 + a1x + ... + a t-1 x i-1 , where a0 = S (secret); Generate shares: Take n numbers x1, x2, ..., x n Calculate f(x1), f(x2), ..., f(x) n ), will (x i ,f(x i Distribute to n servers; Decryption process: Collect t shares: Obtain (x1, y1), ..., (x1, y1) from any t servers. t ,y t ); Solving the polynomial: Solving for the coefficients a0, a1, ..., a2 using matrix operations. t-1 Construct a polynomial f(x); To recover the secret: Substitute x = 0 into the polynomial to obtain the original secret S = a0.

5. The hospital personnel information sharing and collaborative decision-making system for health management with privacy protection function according to claim 1, characterized in that: The personnel management optimization module is divided into a qualification renewal reminder unit and a cross-departmental transfer intelligent screening unit. The qualification renewal reminder unit includes: Validity monitoring: Real-time linking of medical staff's professional certificate information, scanning certificate expiration dates through scheduled tasks, and setting three levels of early warning: 60 days, 30 days, and 15 days in advance, to ensure that relevant personnel are notified in a timely manner before the certificate expires; Multiple reminder channels: Employing various reminder methods ensures that medical staff receive reminder messages in a timely manner; Renewal process guidance: In addition to informing you that the certificate is about to expire, the reminder message also provides guidance on the renewal process; The cross-departmental intelligent screening unit includes: Intelligent screening algorithm: Based on the job requirements of the target department, detailed qualification matching rules are formulated. Through the rule engine Drools, the qualification information of medical staff is quickly matched and qualified personnel are screened out. Idle Status Inquiry: Real-time synchronization with the hospital's scheduling system to obtain the work arrangements and idle status information of medical staff. During the screening process, medical staff who are currently on duty are excluded to ensure that the assigned personnel can be in place in a timely manner. Comprehensive score ranking: Combining the work performance of medical staff and patient evaluation factors, the weights are calculated using the analytic hierarchy process (AHP) to generate a comprehensive score for each eligible medical staff member. The scores are then ranked from highest to lowest to generate a matching ranking list. Based on the ranking results, the optimal allocation plan is recommended.

6. The hospital personnel information sharing and collaborative decision-making system for health management with privacy protection function according to claim 1, characterized in that: The internal sensitive information protection module is divided into a hierarchical authorization management unit and a dynamic data desensitization unit. The hierarchical authorization management unit specifically includes: Access permissions are divided into multiple levels based on the job responsibilities and needs of different positions within the hospital: general staff level, department head level, and hospital management level. Each level corresponds to different data access scopes and operation permissions. Permission application and approval: Employees apply for corresponding permissions according to work needs, fill in the permission application form, explain the reason and scope of the permission application, and the superior or relevant department approves the application and dynamically updates the permissions to ensure the rationality and security of permission allocation. Dynamic permission adjustment: When an employee's position changes or work requirements change, permission changes are automatically triggered. The dynamic data desensitization unit specifically includes: Customized anonymization rules: Define regular expression anonymization rules based on different data types and access scenarios: For ID card numbers, use partial hiding, only displaying the first and last few digits, with asterisks used to represent the middle digits; For salary data, anonymize the data proportionally. Real-time data masking: During data display or transmission, requests are intercepted through the MySQL Proxy layer. The system dynamically masks the data and returns it based on the user's permissions and masking rules, ensuring that the user can only see the masked data and cannot obtain the complete sensitive information. Desensitization effectiveness evaluation: Regularly conduct sampling checks on the data desensitization effectiveness to check for incomplete or excessive desensitization. Based on the evaluation results, adjust the desensitization rules in a timely manner to ensure the rationality and effectiveness of data desensitization.

7. The hospital personnel information sharing and collaborative decision-making system for health management with privacy protection function according to claim 1, characterized in that: The personnel decision data acquisition module is divided into a multi-channel data acquisition unit and a data quality monitoring unit. The multi-channel data acquisition unit includes: Hospital Information System Integration: Through API interfaces, we integrate with the internal information systems of various hospitals to obtain real-time data from HR systems, attendance systems, and training systems, and collect relevant data from medical staff. Regular data reporting: For data that cannot be obtained through system integration, regular data reporting will be adopted. Each hospital will report the relevant data to the system according to the prescribed format and time requirements. External data integration: Collect external data related to hospital personnel and extract key indicators using NLP technology; The data quality monitoring unit includes: Data accuracy verification: The accuracy of the collected data is verified by comparing it with the original data or relevant authoritative data, verifying data consistency through hash algorithms, and checking for errors or inconsistencies in the data. Data integrity check: Define the rules for required fields, check whether the collected data is complete, whether there are missing fields or records, and trigger an alarm when the missing data exceeds 10%, and notify the relevant hospitals to supplement or correct it; Data timeliness monitoring: Set a data update threshold (within 24 hours). If the data is not updated within the time limit, it is marked as abnormal. For abnormal data, an alarm will be issued to ensure that personnel decisions are based on the latest data.

8. The hospital personnel information sharing and collaborative decision-making system for health management with privacy protection function according to claim 1, characterized in that: The big data analysis module is divided into a heat map generation unit for medical resource supply and demand and a talent loss early warning report generation unit. The heat map generation unit for medical resource supply and demand includes: Data aggregation and analysis: The collected hospital personnel dynamic data within the region is aggregated and analyzed according to different regions, departments and time dimensions. The Spark cluster is used to process the regional data and calculate the number of medical staff, workload and patient demand indicators for each region and department. Heatmap visualization: Based on the analysis results, generate a heatmap of supply and demand for medical resources using ECharts or Tableau, using different colors or densities to represent the differences in supply and demand for resources; Trend prediction and early warning: Combining historical data and current development trends, the LSTM neural network is used to predict the supply and demand of medical resources in the next 3 months. When a shortage or surplus of resources is predicted in a certain region or department, an early warning message is issued to provide a basis for resource allocation decisions. The talent loss early warning report generation unit includes: Multi-dimensional data analysis: constructing feature engineering from multiple dimensions such as work performance, career development, compensation and benefits, and work environment of medical staff; Early warning model construction: A talent loss early warning model is constructed using machine learning algorithms. Based on historical talent loss data and related influencing factors, the XGBoost algorithm is used to train the talent loss early warning model and determine the early warning threshold. When the model detects that the relevant indicators of a medical staff member reach or exceed the early warning threshold, the staff member is marked as a potential employee to leave. Report Generation and Interpretation: Based on the early warning results, a talent loss early warning report is generated. The report details the basic information of potential employees who may leave, the reasons for the early warning, and the risk level. At the same time, the report is interpreted to provide targeted suggestions and measures for hospital management to reduce the risk of talent loss.

9. The hospital personnel information sharing and collaborative decision-making system for health management with privacy protection function according to claim 1, characterized in that: The blockchain supervision module is divided into a personnel data full lifecycle traceability unit and a smart contract automatic verification unit. The personnel data full lifecycle traceability unit includes: On-chain data storage: The hash value of the data is written to the blockchain to record creation, modification, deletion, transmission and usage operations; Source tracing query interface: Provides a source tracing query interface, through which hospital management and relevant departments can query the full life cycle information of personnel data. By entering the key identifier of the data, they can obtain all operation records and related information of the data from its generation to its current status. Data integrity verification: Utilizing the hash algorithm and chain structure of blockchain, the integrity of personnel data is verified. By comparing the hash value of the data on the blockchain with the hash value stored locally, it is ensured that the data has not been tampered with during transmission and storage. The smart contract automatic verification unit includes: Qualification verification rule setting: Pre-set the professional qualification verification rules for medical personnel in the smart contract, including certificate type, validity period, and professional scope; Automated verification execution: Connect to an authoritative database via Oracle service (Chainlink) to verify qualification status in real time. When the qualification information of medical personnel changes or the verification time arrives, the smart contract automatically triggers the verification process. By connecting with the database of relevant authoritative institutions, it verifies in real time whether the qualifications of medical personnel are still valid. Handling of violations: When the smart contract verifies that there are problems with the qualifications of medical staff, it will automatically take corresponding measures and record the violation on the blockchain as a basis for subsequent management and punishment.

Images