An infectious disease model on empirical networks of human contact: bridging the gap between dynamic network data and contact matrices

• DOI: 10.1186/1471-2334-13-185
• Publication Date: 2013-04-23T14:22:02Z
• Authors: Anna Machens; Francesco Gesualdo; Caterina Rizzo; Alberto E Tozzi; Alain Barrat; Ciro Cattuto
• Keywords: Physics - Physics and Society; 0303 health sciences; [PHYS.PHYS.PHYS-SOC-PH]Physics [physics]/Physics [physics]/Physics and Society [physics.soc-ph]; [PHYS.PHYS.PHYS-COMP-PH] Physics [physics]/Physics [physics]/Computational Physics [physics.comp-ph]; Populations and Evolution (q-bio.PE); FOS: Physical sciences; Physics and Society (physics.soc-ph); Models, Theoretical; Computational Physics (physics.comp-ph); Hospitals, Pediatric; Communicable Diseases; Disease Outbreaks; 3. Good health; [PHYS.PHYS.PHYS-COMP-PH]Physics [physics]/Physics [physics]/Computational Physics [physics.comp-ph]; 03 medical and health sciences; Infectious Diseases; [PHYS.PHYS.PHYS-SOC-PH] Physics [physics]/Physics [physics]/Physics and Society [physics.soc-ph]; FOS: Biological sciences; Humans; Contact Tracing; Quantitative Biology - Populations and Evolution; Physics - Computational Physics; Algorithms; Research Article

Abstract Background The integration of empirical data in computational frameworks designed to model the spread infectious diseases poses a number challenges that are becoming more pressing with increasing availability high-resolution information on human mobility and contacts. This deluge has potential revolutionize efforts aimed at simulating scenarios, designing containment strategies, evaluating outcomes. However, highly detailed sources yields models less transparent general their applicability. Hence, given specific disease model, it is crucial assess which representations raw work best inform striking balance between simplicity detail. Methods We consider face-to-face interactions individuals pediatric hospital ward, obtained by using wearable proximity sensors. simulate this community an SEIR top different mathematical contact patterns. At most level, we take into account all contacts exact timing order. Then, build hierarchy coarse-grained patterns preserve only partially temporal structural available data. compare dynamics across these representations. Results show matrix contains average durations role classes fails reproduce size epidemic also identify at-risk classes. introduce probability distributions takes heterogeneity (and within) individuals, that, case study presented, representation good approximation spreading properties Conclusions Our results mark first step towards definition synopses dynamic networks, providing compact can correctly discover risk groups evaluate policies. typical structured population novel kind simulation quantitative features epidemics for management.