RESUMEN
extensive protective measures, SARS-CoV-2 widely circulates within healthcare facilities, posing a significant risk to both patients and healthcare workers. Several control strategies have been proposed; however, the global efficacy of local measures implemented at the ward level may depend on hospital-level organizational factors. We aimed at better understanding the role of between-ward interactions on nosocomial outbreaks and their control in a multiward psychiatric hospital in Western France. We built a stochastic compartmental transmission model of SARS-CoV-2 in the 24-wards hospital, accounting for the various infection states among patients and staff, and between-ward connections resulting from staff sharing. We first evaluated the potential of hospital-wide diffusion of local outbreaks, depending on the ward they started in. We then assessed control strategies, including a screening area upon patient admission, an isolation ward for COVID-19 positive patients and changes in staff schedules to limit between-ward mixing. Much larger and more frequent outbreaks occurred when the index case originated in one of the most connected wards with up to four times more transmissions when compared to the more isolated ones. The number of wards where infection spreads was brought down by up to 53 % after reducing staff sharing. Finally, we found that setting up an isolation ward reduced the number of transmissions by up to 70 %, while adding a screening area before admission seemed ineffective. Significance StatementHospital acquired COVID-19 poses a major problem to many countries. Despite extensive protective measures, transmission within hospitals still occurs regularly and threatens those essential to the fight against the pandemic while putting patients at risk. Using a stochastic compartmental model, we simulate the spread of SARS-CoV-2 in a multi-ward hospital, assessing the effect of different scenarios and infection control strategies. The novelty of our method resides in the consideration of staff sharing data to better reflect the field reality. Our results highlight the poor efficiency of implementing a screening area before hospital admission, while the setting up of an isolation ward dedicated to COVID-19 patients and the restriction of healthcare workers movements between wards significantly reduce epidemic spread.
RESUMEN
BACKGROUND: In Northern Botswana, rural communities, livestock, wildlife and large numbers of mosquitoes cohabitate around permanent waters of the Okavango Delta. As in other regions of sub-Saharan Africa, Rift Valley Fever (RVF) virus is known to circulate in that area among wild and domestic animals. However, the diversity and composition of potential RVF mosquito vectors in that area are unknown as well as the climatic and ecological drivers susceptible to affect their population dynamics. METHODS: Using net traps baited with carbon dioxide, monthly mosquito catches were implemented over four sites surrounding cattle corrals at the northwestern border of the Okavango Delta between 2011 and 2012. The collected mosquito species were identified and analysed for the presence of RVF virus by molecular methods. In addition, a mechanistic model was developed to assess the qualitative influence of meteorological and environmental factors such as temperature, rainfall and flooding levels, on the population dynamics of the most abundant species detected (Culex pipiens). RESULTS: More than 25,000 mosquitoes from 32 different species were captured with an overabundance of Cx. pipiens (69,39 %), followed by Mansonia uniformis (20,67 %) and a very low detection of Aedes spp. (0.51 %). No RVF virus was detected in our mosquito pooled samples. The model fitted well the Cx. pipiens catching results (ρ = 0.94, P = 0.017). The spatial distribution of its abundance was well represented when using local rainfall and flooding measures (ρ = 1, P = 0.083). The global population dynamics were mainly influenced by temperature, but both rainfall and flooding presented a significant influence. The best and worst suitable periods for mosquito abundance were around March to May and June to October, respectively. CONCLUSIONS: Our study provides the first available data on the presence of potential RVF vectors that could contribute to the maintenance and dissemination of RVF virus in the Okavango Delta. Our model allowed us to understand the dynamics of Cx. pipiens, the most abundant vector identified in this area. Potential predictions of peaks in abundance of this vector could allow the identification of the most suitable periods for disease occurrence and provide recommendations for vectorial and disease surveillance and control strategies.