RODS: Real-Time Emergency Department Chief Complaints for Syndrome and Outbreak Detection

Real-Time Outbreak and Disease Surveillance (RODS) collected free-text emergency department chief complaints from hospitals in Pennsylvania, Utah, Atlantic City, and Ohio. A Bayesian text classifier assigned each complaint to one of eight syndrome categories for time-series analysis and alert generation. The system prospectively detected a cluster of carbon monoxide exposures within 4 hours of the first case presentation.

Use Cases

• Training Bayesian text classifiers for medical syndrome categorization based on free-text chief complaints.
• Evaluating time-series algorithms for outbreak detection using categorized syndrome data.
• Developing real-time public health dashboards with timelines and maps for authorized user review.
• Conducting retrospective experiments to measure case-classification and outbreak-detection accuracy.

Strengths

• Demonstrated real-time collection and analysis of chief-complaint data from multiple regional deployments.
• Bayesian classifier performance was adequate to support accurate and timely detection of seasonal disease outbreaks.
• Prospective evaluation showed detection of a specific carbon monoxide exposure cluster within 4 hours.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and sample data are unknown, which may limit suitability assessment.

Provenance

Source

University of Pittsburgh (Michael M. Wagner) via the Real-Time Outbreak and Disease Surveillance (RODS) project.

Collection Method

Real-time transmission of chief-complaint data as Health Level 7 messages from hospitals to a regional data center.

Geography

Pennsylvania, Utah, Atlantic City, and Ohio.