Data Correlation Engine for Distributed Clinical Data Federation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data warehousing solutions, such as the Common Health Framework, are inadequate for combining and querying patient demographic and clinical data from multiple independent sources, limiting access to epidemiological data and failing to provide comprehensive correlations, such as the percentage of patients with specific diagnoses who missed appointments.
Innovation Solution
A data correlation engine that combines data from multiple independent stores, allows clinicians to select and run correlations, and presents results in user-defined formats, with options for scheduling and alerting, using a web-based interface that can be customized and translated for international data sources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a central store contains only patient identification information and links to local databases (as in the Common Health Framework), then security and confidentiality of local data is maintained, but access to epidemiological data becomes very limited and difficult
Solution Approach 1:
The system segments data storage and processing across multiple independent data stores rather than centralizing all data. Each data store maintains its own data locally, and the federation architecture allows queries to be distributed across these segments, enabling both security maintenance and improved data access capability.
Solution Approach 2:
The patent introduces a federation intermediary that acts as a mediator between users and multiple independent data stores. This intermediary translates user queries into appropriate data retrieval operations across distributed stores, enabling easy access to epidemiological data without compromising local data security or requiring data centralization.
2Reliability
If data is stored in multiple independent data stores, then data security and local control are maintained, but the ability to run correlations across datasets is severely limited
Solution Approach 1:
The federation system provides universal query capability that can operate across multiple independent data stores with different formats and structures. The system translates various query types into operations that can be executed on heterogeneous data sources, enabling comprehensive data correlation while maintaining the independence and security of each data store.
Solution Approach 2:
The system dynamically changes query parameters and translation strategies based on the specific characteristics of each data store being accessed. This allows the federation intermediary to optimize data retrieval and correlation operations for different data sources, improving overall productivity without requiring standardized data formats or centralized storage.
3Ease of operation
If patient identification information is used as the sole portal to access data, then individual patient queries are enabled, but epidemiological analysis and population-level correlations become impossible
Solution Approach 1:
The patent adds a new dimension to data access by enabling queries that operate at the population level rather than only at the individual patient level. The federation system supports both granular patient-specific queries and aggregate epidemiological analyses, allowing users to switch between different levels of data aggregation and analysis without being constrained to a single access model.
Data Source
AI summary
Data is combined from multiple independent data stores, and is then queried using a data correlation engine. Contemplated engines preferably keep track of previously run correlations, and then makes those correlations available to clinicians for their own use. For example, a preferred system might provide a listing of correlations run by other clinicians in a particular medical specialty, or a particular community, whether geographic or otherwise. In another example, a preferred system might provide a listing of correlations sorted by popularity, so that the most frequently accessed correlations appear near the top of the list. In any case a clinician could simply view the list, and check off which correlations he/she would like to have run for his/her practice, or practice community.