Another open source health tools consortium

Electronic Medical Records are structured databases, but linking them can be difficult. Mayo bioinformatics expert Christopher Chute (right),  who previously worked on open source implementations of HL7, will be senior consultant for the effort.

The hope is that Natural Language Processing techniques will allow non-programmers to consult unstructured data in EMR databases and extract clinically useful data.

The big difficulty here involves the notes that clinicians write while seeing patients, and the pathology reports of tissue biopsies. This is unstructured data attached as addenda to structured EMR databases.

The hope is that improved search of the notes can let physicians find cases of rare conditions with treatments that worked, and that past studies of tissue samples may yield something similar to what they are looking at in a new culture.

Mayo already has a system called clinical Text Analysis and Knowledge Extraction System (cTAKES) to process clinical notes. IBM offers a medical knowledge analysis tool (medKAT) to extract structured data from unstructured sources, using techniques from an existing Apache project.