Inference of COVID-19 epidemiological distributions from Brazilian hospital data

Hawryluk, Iwona; Mellan, Thomas A.; Hoeltgebaum, Henrique H.; Mishra, Swapnil; Schnekenberg, Ricardo P.; Whittaker, Charles; Zhu, Harrison; Gandy, Axel; Flaxman, Seth; Bhatt, Samir

Knowing COVID-19 epidemiological distributions, such as the time from patient admission to death, is directly relevant to effective primary and secondary care planning, and moreover, the mathematical modelling of the pandemic generally. Here we determine epidemiological distributions for patients hospitalised with COVID-19 using a large dataset (range $N=21{,}000-157{,}000$) from the Brazilian SIVEP-Gripe (Sistema de Informa\c{c}\~ao de Vigil\^ancia Epidemiol\'ogica da Gripe) database. We fit a set of probability distribution functions and estimate a symptom-onset-to-death mean of $15.2$ days for Brazil, which is lower than earlier estimates of 17.8 days based on early Chinese data. A joint Bayesian subnational model is used to simultaneously describe the $26$ states and one federal district of Brazil, and shows significant variation in the mean of the symptom-onset-to-death time, with ranges between $11.2-17.8$ days across the different states. We find strong evidence in favour of specific probability distribution function choices: for example, the gamma distribution gives the best fit for onset-to-death and the generalised log-normal for onset-to-hospital-discharge. Our results show that epidemiological distributions have considerable geographical variation, and provide the first estimates of these distributions in a low and middle-income setting. At the subnational level, variation in COVID-19 outcome timings are found to be correlated with poverty, deprivation and segregation levels, and weaker correlation is observed for mean age, wealth and urbanicity.