In a new study, One Health Trust researchers developed an agent-based model of an intensive care unit (ICU) to assess the effects of uncertainties on the nosocomial transmission dynamics of hospital-acquired infections (HAIs). They aimed to evaluate the effect of inherent uncertainty, or stochasticity, on the ability to detect the seasonality of HAI acquisition rates and admission prevalence. They found that when the seasonality strength and baseline acquisition rate decreased past a specific threshold, there was a 20 to 30 percent chance that the seasonality effect would be undetectable, due to the presence of “noise” — caused by stochasticity — blurring the seasonality signal. Given an admission prevalence of five to 10 percent, the minimum sample size of ICUs needed to detect the effect of seasonality was 50 (assuming a seasonality strength of 90% or more). These findings reveal that the highly stochastic nature of HAI transmission can drastically complicate the effects of environmental changes, such as seasonal transmission. The researchers also developed a methodology that allows for the assessment of interventions in a healthcare facility and the calculation of the probability distribution of detection given a specific number of facilities involved.

Read the full article, published in Royal Society Open Science, here.