Impact of spatiotemporal heterogeneity in COVID-19 disease surveillance on epidemiological parameters and case growth rates

• DOI: 10.1016/j.epidem.2022.100627
• Publication Date: 2022-09-05T22:29:55Z
• Authors: Rhys P.D. Inward; Felix Jackson; Abhishek Dasgupta; Graham Lee; Anya Lindström Ba...; Kris V. Parag; Moritz U.G. Kraemer
• Keywords: 0303 health sciences; SARS-CoV-2; COVID-19; Infectious and parasitic diseases; RC109-216; Article; Epidemic models; Disease Outbreaks; 3. Good health; Hospitalization; 03 medical and health sciences; Outbreak surveillance; Humans; Brazil

SARS-CoV-2 case data are primary sources for estimating epidemiological parameters and modelling the dynamics of outbreaks. Understanding biases within case-based used in analyses is important as they can detract from value these rich datasets. This raises questions how variations surveillance affect estimation such growth rates. We use standardised line list COVID-19 Argentina, Brazil, Mexico Colombia to estimate delay distributions symptom-onset-to-confirmation, -hospitalisation -death well hospitalisation-to-death at high spatial resolutions throughout time. Using estimates, we model introduced by symptom-onset-to-confirmation on national state level rates (rt) using an adaptation Richardson-Lucy deconvolution algorithm. find significant heterogeneities through time space with difference up 19 days between epochs level. Further, that changing scale, estimates rate vary 0.13 d−1. Lastly, states a variance and/or mean symptom-onset-to-diagnosis also have largest rt estimated raw deconvolved counts highlight importance high-resolution understanding disease reporting be avoided adjusting numbers based empirical distributions. Code openly accessible reproduce presented here available.