Lies, Gosh Darn Lies, and Not Enough Good Statistics: Why Epidemic Model Parameter Estimation Fails

Daniel E Platt; Laxmi E Parida; Pierre Zalloua

Infection 1,2 , transcription and replication 3,4 by SARS-COV-2 involves a number of rate limiting interactions with host cells that are likely to be modulated by mutations in cellular as well as viral genes. At the same time, phylogenetic analysis shows geographic specificity 5, 6 , indicating that geographic regions may show specific exposure to distinctive SNP combinations, or viral haplotypes, in SARS-COV-2. This specificity suggests a benefit to exploring relationships between duration of the prodromic phase, proportions of asymptomatic cases 7 , proportions of severe cases, rates of recovery, among other infection attributes 8 , that define temporal progression of compartmental epidemic models, starting with SIR (Susceptible-Infected-Recovered) models 9 . Beside host and viral genetic impacts, other aspects driving SARS-COV-2 rates are population specific and demic, such as the impact of age on both asymptomatic and mild cases, as well as the proportion of severe and critical cases. Other aspects include normal social distance, and how effectively social-distancing rules have been followed. Hospital survival may also reflect impacts of some genetic susceptibility, presence of comorbidities (Hypertension, Diabetes, Asthma, lung disease, obesity and others yet to be identified) as well as the level of stress on the region's medical facilities and medical staff.