VIEW: a framework for organization level interactive record linkage to support reproducible data science
This addresses the challenge of reproducible data science for researchers dealing with messy real-world data, though it is incremental as it builds on existing semi-automated linkage methods.
The authors tackled the problem of linking messy hospital data by developing VIEW, an interactive framework that combines algorithmic processing with human judgment tasks, achieving 100% precision in linking Texas hospitals from two databases.
Objective: To design and evaluate a general framework for interactive record linkage using a convenient algorithm combined with tractable Human Intelligent Tasks (HITs; i.e. micro tasks requiring human judgment) that can support reproducible data science. Materials and Methods: Accurate linkage of real data requires both automatic processing of well-defined tasks and human processing of tasks that require human judgment (i.e., HITs) on messy data. We present a reproducible, interactive, and iterative framework for record linkage called VIEW (Visual Interactive Entity-resolution Workbench). We implemented and evaluated VIEW by integrating two commonly used hospital databases, the American Hospital Association (AHA) Annual Survey of Hospitals and the Medicare Cost Reports for Hospitals from CMS. Results: Using VIEW to iteratively standardize and clean the data, we linked all Texas hospitals common in both databases with 100% precision by confirming 78 approximate linkages using HITs and manually linking 28 hospitals using HITs. Discussion: Similarities in hospital names and addresses and the dynamic nature of hospital attributes over time make it impossible to build a fully automated linkage system for hospitals that can be maintained over time. VIEW is a software that supports a reproducible semi-automated process that can generate and track HITs to be reviewed and linked manually for messy data elements such as hospitals that have been merged. Conclusion: Effective software that can support the interactive and iterative process of record linkage, and well-designed HITs can streamline the linkage processes to support high quality replicable research using messy real data.