Wednesday, April 16, 2008

RODS Dataset for OGSA-DAI

Today, I provided a synthetic dataset of emergency department visits from Allegheny County Pennsylvania to the OGSA-DAI group. The dataset contains the same table and fields used by RODS to store emergency department visit information. The fields include date of visit, age, gender, zipcode, and chief complaint. It comprises 53020 records for a period of 1 month.

Unlike the OpenMRS datasets, the chief complaint/reason for visit field is not coded (i.e., uses a vocabulary) and is free text. This is the way that most emergency department surveillance systems collect their chief complaint data (i.e., non-coded free text). This presents interesting challenges for federation.

At RODS, we've solved the free text chief complaint challenge before using text indexing and preclassification of chief complaints into syndrome categories. We'll need to use similar solutions but this time under a federated api.

No comments: