Therapy Line Identification Architecture for Claims Data Integration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing healthcare data analysis systems inefficiently and inaccurately process unstructured healthcare data, particularly health insurance claims data, which is crucial for understanding patient treatments and biological conditions, leading to potential detrimental impacts on individual health.
Innovation Solution
A computing architecture that integrates and analyzes structured health insurance claims data with genomic and molecular data to identify lines of therapy, using data integration and analysis systems to generate an integrated data repository, apply data pipelines, and determine therapy gaps and lines of therapy through timeline data structures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing healthcare data analysis systems process unstructured healthcare data, then data processing can be performed, but the processing is inefficient and inaccurate
Solution Approach 1:
The patent segments unstructured healthcare data into structured components using standardized data models (e.g., FHIR resources, OMOP CDM). Data is divided into discrete entities such as patients, encounters, medications, procedures, and outcomes, each with defined schemas. This segmentation enables efficient querying and analysis while maintaining data accuracy through consistent structuring across diverse data sources.
Solution Approach 2:
The patent transforms unstructured data parameters into standardized structured parameters through mapping processes. Unstructured free-text fields are converted to coded values using standardized terminologies (ICD-10, SNOMED, RxNorm). This parameter transformation enables accurate measurement and comparison while improving processing efficiency through standardized data formats that can be efficiently queried and analyzed.
2Loss of information
If health insurance claims data is analyzed to understand patient treatments, then treatment understanding is improved, but processing inefficiency and inaccuracy occur
Solution Approach 1:
The patent introduces structured data models and standardized terminologies as intermediaries between unstructured claims data and analysis systems. These intermediaries (data models, vocabularies, and mapping layers) preserve complete treatment information while enabling efficient processing. The intermediary layer translates diverse claims formats into a unified structured representation that maintains information completeness without sacrificing processing speed.
Solution Approach 2:
The patent performs preliminary structuring and standardization of claims data before analysis. Data is pre-processed into standardized formats with predefined schemas, codes, and relationships established in advance. This preliminary action ensures that all treatment information is captured and structured correctly before analysis begins, eliminating the need for complex real-time processing while maintaining information completeness.
3Measurement precision
If pharmacy transaction data and medical procedure data are integrated, then therapy identification is improved, but system complexity increases
Solution Approach 1:
The patent implements a universal structured data model that handles multiple data types (pharmacy transactions, medical procedures, patient demographics, clinical outcomes) through a single unified framework. This multi-functional data model uses consistent schemas, relationships, and terminologies across all data types, enabling accurate therapy identification without requiring separate complex integration systems for each data source.
Solution Approach 2:
The patent adds a structural dimension to diverse healthcare data by organizing it into hierarchical relationships (patients → encounters → services → products). This dimensional organization transforms unstructured multi-source data into a structured multi-dimensional framework where therapies can be identified by traversing relationships across different data dimensions, simplifying integration while improving identification accuracy.
Data Source
AI summary
A computing machine accesses a pharmacy transaction data set and a medical procedure transaction data set. The computing machine filters, from the pharmacy data set and the medical procedure transaction data set, transactions relevant to a biological condition. The computing machine identifies one or more lines of therapy from the filtered pharmacy data set and the filtered medical procedure transaction data set.