Related Experiment Videos
A foundational model of time for heterogeneous clinical databases
1Section on Medical Informatics, Standford University School of Medicine, California 94305-5479, USA. das@smi.stanford.edu
Summary
Temporal mismatches in clinical data hinder database integration and decision-support applications. We introduce a foundational time model to standardize timestamp representations for improved clinical data analysis.
Area of Science:
- Biomedical Informatics
- Data Science
- Database Management
Background:
- Database representation differences impede clinical data integration.
- Existing research lacks focus on temporal mismatches in clinical data timestamping.
- Temporal data heterogeneity limits clinical decision-support applications.
Purpose of the Study:
- Analyze temporal mismatches in clinical databases.
- Formalize timestamping approaches with a foundational time model.
- Enable uniform temporal representation for decision support.
Main Methods:
- Developed a foundational model of time to define temporal semantics.
- Encoded temporal dimensions of clinical data using the model.
- Transformed heterogeneous data into a uniform temporal representation.
- Extended the Chronus system with the foundational model.
Main Results:
- Identified and analyzed various types of temporal mismatches in clinical databases.
- Established a formal model for understanding timestamping variations.
- Created a uniform temporal representation for clinical data.
- Integrated the model into the Chronus system for decision support.
Conclusions:
- A foundational time model is crucial for resolving temporal data heterogeneity.
- Standardized temporal representations enhance clinical data integration and decision-support.
- The Chronus system extension facilitates temporal pattern matching in clinical databases.