Highlights
- CDC and the Florida Department of Health in Polk County worked together to detect increases in community norovirus activity and send timely alerts to long-term care facilities.
- Epidemic trends, based on the time-varying reproductive number (Rt), can be observed alongside surveillance data to help health departments identify disease transmission trends.

Background
Norovirus is a highly infectious virus that can cause severe gastrointestinal symptoms and dehydration. It is the leading cause of vomiting, diarrhea, and foodborne illness in the United States. It is a seasonal virus with peak activity between November to April. Though national trends are available for norovirus, including norovirus in long-term care facilities (LTCFs), there is limited visibility into these trends at the local level.
Nationally, over half of all norovirus outbreaks occur in LTCFs. Compared with the general population, norovirus illness can be more severe among LTCF residents.
Partnering for public health
In December 2024, the Florida Department of Health in Polk County ("Polk County") contacted CDC's Center for Forecasting and Outbreak Analytics (CFA) to explore modeling local norovirus activity to improve the timing of prevention messaging and alerts to LTCFs. Each year, Polk County alerts more than 60 local LTCFs when community norovirus activity begins to increase so they can prepare for and take measures to prevent potential outbreaks. This partnership with CFA led to the development of a weekly norovirus trend report for Polk County. In its first season, insights from the trend report supported Polk County's decision to alert LTCFs earlier than usual, when community activity began increasing before the start of the typical norovirus season.
Polk County had multiple data inputs from emergency department (ED) syndromic surveillance to inform the modeling. They created synthetic versions of each data set by adding random variation to the counts, preserving the overall trends while respecting data sharing limitations. Polk County shared approximately 10 years of historic, synthetic ED surveillance data with CDC and Emory University Insight Net partner Center for Infectious Disease Modeling & Analytics and Training Hub (CIDMATH); this included daily and weekly counts of discharge diagnosis codes for acute gastroenteropathy and viral intestinal infection, daily and weekly counts of chief complaints, and weekly positive norovirus lab reports from one large hospital for validation. Additionally, Polk County shared annotated timelines of six historic norovirus seasons indicating when they had alerted LTCFs and, in retrospect, when they would have preferred to send the alerts based on complete season data.
CFA, in consultation with norovirus subject-matter experts in the National Center for Immunization and Respiratory Diseases' (NCIRD) Viral Gastroenteritis Branch, generated time-varying reproductive number (Rt) estimates using Polk County's chief complaint and discharge diagnosis data from recent years as inputs. Separately, the CIDMATH team generated short-term forecasts for Polk County using a machine learning model and evaluated the model's performance.
The teams at CFA, NCIRD, and Polk County reviewed the outputs of CFA's Rt and CIDMATH's short-term forecast models to select the model best visually aligned with the annotated ideal timing for historic alerts. They found that all ideal alert timing annotations coincided with discharge-diagnosis-based Rt-estimated trends increasing for three weeks.
As a result, the Rt model using discharge diagnoses was selected for implementation by Polk County. With technical support from CFA to design a weekly report, Polk County successfully piloted the model during late 2025 and early 2026. During the pilot, Polk County identified multiple weeks with increasing norovirus Rt trends before the start of the typical norovirus season and decided to alert LTCFs earlier than usual to share norovirus prevention guidance. No subsequent norovirus outbreaks were identified within these LTCFs that season.
Future work
Weekly Rt-based metrics, alongside ED syndromic surveillance data, are now a routine piece of information to help identify increases in norovirus transmission in Polk County. Additional evaluation can assess how this approach improved alert timing and norovirus preparedness in LTCFs in Polk County, including any associated reductions in local norovirus outbreaks in LTCFs. This collaboration shows how CDC and local health departments can improve the timeliness of public health messaging by integrating modeling into routine local public health practice.
Acknowledgements
- NCIRD works to prevent disease, disability, and death through immunization and by surveillance and control of respiratory and related diseases.
- CFA aims to use data, modeling, and analytics to respond to outbreaks in real time to inform effective decision-making.
- CIDMATH advances infectious disease modeling and outbreak forecasting, emphasizing technological applications that are accessible and usable by public health practitioners.
- FDOH in Polk County: The mission of FDOH is to protect, promote, and improve the health of all people in Florida through integrated state, county, and community efforts.
This page was last updated on this date. Updates may include minor edits, image changes, or other modifications to page content.
The information on this page was last reviewed by subject matter experts to ensure accuracy.