Privacy-Preserving Record Linkage via Honest Broker
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The proliferation of electronic records across various entities creates challenges in record linkage due to differing formats, typographical errors, outdated information, and lack of access, making it difficult to merge and link data effectively, especially in privacy-sensitive areas like electronic health records.
Innovation Solution
A system and method for privacy-preserving record linkage involving a control center server that generates configuration files for data sites to manipulate and normalize private information, using hash keys to encrypt and hash data, which is then linked by an honest broker without revealing clear-text linkage data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If records from multiple entities are merged and linked, then data integration and analysis capability are improved, but privacy leakage and security risks increase
Solution Approach 1:
The patent introduces an honest broker server as an intermediary that receives hashed data from multiple data sites and performs linkage operations without ever accessing the original private information. This mediator enables data integration while maintaining privacy by ensuring that no single entity possesses both the hashed data and the corresponding hash keys needed to reverse the hashing process.
Solution Approach 2:
The patent transforms the data from its original state by applying hashing functions and generating derived values that preserve the ability to perform linkage operations while fundamentally changing the data representation. The hash keys are generated through cryptographic processes that convert original data into irreversible hashed forms, changing the parameter state from readable to encrypted while maintaining functional equivalence for linkage purposes.
2Measurement precision
If data is normalized and cleaned to improve matching accuracy, then linkage precision is improved, but processing time and computational resources increase
Solution Approach 1:
The patent performs data normalization, cleaning, and hashing operations as preliminary actions before the actual linkage process. By pre-processing the data at each data site to generate clean hashed values and derived attributes, the system reduces the computational burden during the linkage phase, improving overall efficiency while maintaining high matching accuracy.
Solution Approach 2:
The patent divides the data processing workflow into distinct segments: local data cleaning and hashing at each data site, key generation at the keymaster server, and linkage operations at the honest broker server. This segmentation allows parallel processing of multiple data sets simultaneously, reducing total processing time while maintaining precision through specialized processing at each stage.
3Reliability
If hash keys are generated and stored securely, then data security is improved, but system complexity and key management overhead increase
Solution Approach 1:
The patent introduces a keymaster server as a specialized intermediary dedicated solely to key generation and management. This separate entity handles the complex cryptographic operations of generating hash keys and managing key distribution without involving the data sites or honest broker in key management, thereby simplifying the overall system architecture while maintaining high security standards through dedicated key management infrastructure.
Data Source
AI summary
A record linking platform having a computer system with a processor, multiple databases each having records, such as private health information. Each of the multiple databases are not in direct communication with one another. A control center is in communication with the different entities, the control center configured to: generate a site configuration file, a key configuration file, and a broker configuration file. A keymaster configured to: receive the key configuration file, generate a hash key, and transmit the hash key to the databases. The databases configured to: receive the site configuration file and the hash key; process the records to generate evaluation records data that conform to a data model specified in the site configuration file; and encrypt the evaluation records using the hash key. An honest broker configured to: receive the broker configuration file and the encrypted evaluation records; and link the encrypted evaluation records without decrypting.