Detection of errors in the inference engine of a clinical decision support system

摘要

An electronic clinical decision support system (CDSS) (10, 12) comprises: an inference engine (20, 22) configured to generate clinical decision recommendations for a patient based on information pertaining to the patient, the inference engine comprising rules (16) developed by a plurality of medical experts (14) and codified into software; an electronic outliers detector (52) configured to detect outlier cases that are probative of a potential flaw in the inference engine; an outliers database (60) configured to collect information pertaining to the outlier cases detected by the electronic outliers detector; and an outliers report generator (62) configured to generate a report (64) on the outlier cases detected by the electronic outliers detector, the generated report containing at least some information collected in the outliers database.

主权项

1. An electronic clinical decision support system (CDSS) comprising:
an inference engine configured to generate clinical decision recommendations for a patient based on information pertaining to the patient, the inference engine comprising rules developed h a plurality of medical experts and codified nun software, the inference engine includes a software error comprising a sound rule incorrectly codified into software; an electronic outliers detector configured to detect outlier cases that are probative of a potential flaw in the inference engine; an outliers database configured to collect information pertaining to the outlier cases detected by the electronic outliers detector; an outliers report generator configured to generate a report on the outlier cases detected by the electronic outliers detector, the generated report containing at least some information collected in the outliers database; and a revision release configured to update the inference engine, the revision release comprising update software generated by (i) identifying the software based on review of the report on outlier cases, and (ii) configuring the update software to remove the software error from the inference engine; wherein the inference engine, the electronic outliers detector, and the outliers report generator comprise a digital device executing software.