Fall Event Identification via Code Segmentation and Lexical Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current healthcare data analytics systems struggle to accurately identify fall events among the elderly population due to undercoding of fall-related diagnosis codes in medical claims data.
Innovation Solution
A method involving a computing system that receives medical claims data and associated medical notes, using inclusionary and exclusionary codes (such as W-series and V-series ICD-10 codes) to determine if a fall event has occurred, and performing lexical searches for fall-related words to confirm the event.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional medical claims data analysis is used to identify fall events, then the process is simple, but the identification accuracy is low due to undercoding of fall-related diagnosis codes
Solution Approach 1:
The patent segments the fall event identification process into multiple stages: first filtering claims with inclusionary injury codes, then excluding claims with exclusionary codes, and finally performing lexical searches on medical notes. This multi-stage segmentation approach improves identification accuracy by systematically addressing undercoding issues at each stage rather than relying on a single simple filter.
Solution Approach 2:
The patent introduces an intermediary lexical search process that bridges the gap between coded data and unstructured medical notes. By using fall-related word searches as an intermediary step between code-based filtering and final event confirmation, the system recovers fall events that were undercoded in the original claims data.
2Measurement precision
If comprehensive code checking is performed on all medical claims data, then fall event identification accuracy improves, but computational resources and processing time increase
Solution Approach 1:
The patent divides the large-scale data processing task into smaller segments: first applying inclusionary code filters to narrow down the dataset, then applying exclusionary code filters, and finally performing lexical searches only on the filtered subset. This segmentation reduces the computational burden compared to checking all claims data comprehensively.
Solution Approach 2:
The patent performs preliminary filtering using inclusionary and exclusionary codes before conducting the more resource-intensive lexical searches on medical notes. This preliminary action reduces the volume of data requiring detailed analysis, thereby improving overall processing efficiency while maintaining accuracy.
Data Source
AI summary
A method performed by one or more processors includes: receiving at least one first data object and at least one second data object for an entity; determining if the at least one first data object includes at least one inclusionary code associated with a fall related injury; determining if the at least one first data object excludes at least one exclusionary code associated with a non-fall related injury; and when it is determined that the at least one first data object includes the at least one inclusionary code and excludes the at least one exclusionary code: determining if the at least one second data object includes at least one fall-related word; and when it is determined that the at least one second data object includes the at least one fall-related word, marking the at least one first data object as being indicative of an actual fall event.