Epic Launches County-Level Health Alerts to Detect Rising Illness Trends

Developed by Epic Research, the tool applies statistical models to patient records to identify when the rate of a specific health condition exceeds expected levels.

Electronic health record company Epic has launched a new Health Alerts feature designed to monitor disease trends at the county level and flag elevated rates of illness based on real-world clinical data.

Developed by Epic Research, the tool applies statistical models to patient records to identify when the rate of a specific health condition exceeds expected levels. Each alert undergoes manual review by a team of clinicians and data scientists to assess clinical relevance before being published, according to the company.

The alerts are generated using data from Epic’s Cosmos platform, which includes records from approximately 300 million patients across 2,067 hospitals and 47,000 clinics. The system evaluates population-level trends rather than complete case counts, focusing on identifying unusual increases in disease activity.

Initial alerts have identified elevated rates of conditions such as acute bronchiolitis, acute tonsillitis, measles, strep throat, and viral gastroenteritis in select counties across Illinois, Missouri, Tennessee, Arkansas, and South Carolina.

The detection process relies on monitoring diagnosis rates using ICD-10-CM codes. Conditions are flagged when they show a year-over-year increase adjusted for seasonal patterns, along with accelerating growth in reported cases. These signals are further validated using the Farrington improved algorithm, a method commonly used in public health surveillance.

According to Epic Research, the model analyzes three years of historical data to establish baseline expectations, accounting for seasonal variation and long-term trends. It then determines whether observed increases are statistically significant and unlikely to be due to random variation.

The Health Alerts feature is designed to prioritize acute conditions, communicable diseases, and unexpected spikes in illness rates. Chronic conditions and predictable seasonal trends are less likely to trigger alerts.

Users can access active alerts through an online dashboard and subscribe to receive notifications via email.

Stay tuned for more such updates on Digital Health News