Collection and integration of clinical data for surveillance

Abstract

Objective: The syndromic surveillance project at Public Health-Seattle & King County incorporates several data sources, including emergency department and primary care visit data collected and normalized through an automated mechanism. We describe significant changes made in this "second generation" of our system to improve data quality while complying with privacy and state public health reporting regulations.

Methods/results: The system uses de-identified visit and patient numbers to assure data accuracy, while shielding patient identity. Presently, we have 124,000 basic visit records (used to generate stratified denominators), and 29,000 surveillance records, from four emergency departments and a primary care clinic network. The system is capable of producing syndrome-clustered data sets for analysis.

Discussion: We have incorporated data collection techniques such as automated querying, report parsing, and HL7 electronic data interchange. We are expanding the system to include greater population coverage, and developing an understanding how to implement data collections more rapidly at individual hospital sites, as well as how best to prepare the data for analysis.