Bayesian-network-based method and system for detection of clinical-laboratory errors using synthetic errors

• US 7,783,582 B2
• Filed: 07/10/2007
• Issued: 08/24/2010
• Est. Priority Date: 07/10/2006
• Status: Expired due to Fees

• Current Assignee: University of Washington
• Original Assignee: University of Washington
• Inventors: Doctor, Jason N., Strylewicz, Gregory B.
• Primary Examiner(s): VINCENT, DAVID ROBERT
• Application Number:
  US11/827,449
• Publication Number:
  US 20080222058A1
• Time in Patent Office: 1,141 Days
• Field of Search: 706/12, 706/14, 706/20, 706/45, 706/62
• US Class Current: 706/12
• CPC Class Codes: G16H 10/40 for data related to laboratory analysis, e.g. patient specimen analysis

Abstract

Embodiments of the present invention include methods and systems for analyzing clinical-laboratory results and data in order to detect erroneous clinical-laboratory results. Embodiments of the present invention employ Bayesian networks and modified Bayesian networks that are constructed using cleaned clinical-laboratory results into which various types of synthetic errors have been introduced and that are optimized using different, cleaned clinical-laboratory results into which synthetic errors have been introduced.

Claims

1. A method for detecting errors in a clinical-laboratory-analysis-results data set, the method comprising:

receiving a cleaned clinical-laboratory-analysis-results data set;

introducing synthetic errors into the cleaned clinical-laboratory-analysis-results data set to produce a training data set and a test data set;

inferring a Bayesian network or modified Bayesian network, based on the training data set, for detecting errors;

evaluating and optimizing the Bayesian network or modified Bayesian network, using the test data set to produce a final Bayesian network or modified Bayesian network; and

using the final Bayesian network or modified Bayesian network to detect errors in subsequently received clinical-laboratory-analysis-results data sets.