Related Experiment Videos
Methods for knowledge extraction from a clinical database on liver diseases
S Chowdhury1, G Bodemar, P Haug
1Department of Medical Informatics, Linköping University, Sweden.
Computers and Biomedical Research, an International Journal
|December 1, 1991
Summary
Exploratory data analysis (EDA) reveals that statistical methods can effectively extract knowledge from hospital information system (HIS) data for liver disease. A few variables achieved similar classification accuracy to many, optimizing decision support systems.
Area of Science:
- Medical Informatics
- Data Science
- Biostatistics
Background:
- Hospital information systems (HIS) generate vast amounts of patient data.
- Effective utilization of this data is crucial for improving patient care through decision support systems.
- Knowledge extraction from retrospective clinical data remains a challenge.
Purpose of the Study:
- To demonstrate the potential of statistical techniques for knowledge extraction from HIS databases.
- To support the creation and refinement of rules within medical knowledge bases.
- To enhance the development and updating of decision support systems.
Main Methods:
- Exploratory data analysis (EDA) was applied to liver disease data.
- Stepwise discriminant analysis was employed for disease class discrimination.
- Statistical and artificial intelligence methods were used for handling missing and atypical values.
Main Results:
- Stepwise discriminant analysis effectively discriminated between different liver disease classes.
- Classification accuracy using a small subset of variables (3) was comparable to using all available variables (19).
- Both statistical and AI approaches showed promise in estimating missing values.
Conclusions:
- Statistical approaches are viable for knowledge extraction from retrospective clinical data.
- A parsimonious set of variables can be sufficient for accurate disease classification.
- Addressing missing and atypical values is critical for reliable data analysis in healthcare.