Background: Norovirus causes a significant disease burden of 20 million cases per year in the United States. Hospitals and long-term care facilities constitute the most commonly reported settings for noroviral outbreaks and clusters and thus represent a critically important site for prevention. Our institutional surveillance and response system identified 10–14 clusters or outbreaks of gastrointestinal viral disease per year, predominantly affecting staff. We sought to develop a compartmental mathematical model to examine the potential efficacy of various infection control practices in the management of noroviral clusters. Methods: We developed a set of parallel compartments representing both patient and staff categories (nursing, nurse assistants, etc) involved in a prototypical outbreak, using a 38-bed mixed medium- and high-acuity medical unit as the model basis. A susceptible–exposed–infected–recovered/immune (SEIR) model structure was used (Fig. 1). We conducted interviews with infection preventionists and nursing management to parameterize the model with data on (1) staff-to-patient ratios, (2) staff-patient contact time, (3) staff-staff contact time, (4) spatial distribution of patient assignments, and (5) baseline and intraoutbreak infection prevention practices. With these data, we proceeded to develop submodels, building on the primary model, that examined the effects of additional parallel compartmentalization of granular groups of staff, including resident physicians, environmental services, and clinical nursing assistants. Model parameters for these subanalyses were informed by interviews with clinical experts and review of internal data. Results and Conclusions: An SEIR model was developed that allowed for examination of a modeled outbreak of norovirus and comparison with a known prior outbreak on the same modeled unit for fidelity. Submodeling was performed with more staffing detail, allowing for the addition of further parallel SEIR tracks that delineated more granular staffing patterns. Staff interviews proved critical in the parameterization of these submodels, allowing for a more faithful representation of real-world dynamics. This work, through modification of model parameterization, can be used to assess the efficacy of hypothetical infection control interventions (eg, earlier unit closure, longer staff furlough) in altering transmission dynamics during an outbreak.

Funding: None

Disclosures: None