A logistic growth model for COVID-19 proliferation: experiences from China and international implications in infectious diseases.

Shen, Christopher Y

OBJECTIVE As the COVID-19 global pandemic continues its proliferation globally, this paper shares the findings and experiences of modelling the outbreak in China and its provinces and proposes implications for the study of infectious diseases and the COVID-19 pandemic in other nations. METHODS An NLS (non-linear least square) method estimated the parameters of a differentiated logistic growth function of new COVID-19 cases in multiple regions in China and nations with a large amount of cases based on training data from Feb. 20 to Mar. 13. A restriction test was subsequently employed to test whether a designated parameter was identical among regions or countries. We conducted diagnosis of the residuals and also its goodness of fit with testing data from Mar. 13 to Apr. 18. FINDINGS We find that the model fits time series data exceedingly well in China, its provinces, and two other nations, and we provide estimates of key parameters. We reject the null hypothesis that the growth rate of the outbreak is the same among ten selected provinces as well as between South Korea and Iran. We also find that the model does not provide reliable estimates in countries that are in the early stages of outbreak. We further find that the R2 value be varied and misleading when compared among different curves of the same nonlinear model, as well as the existence of heteroskedasticity and positive serial correlation within residuals in some provinces and nations. CONCLUSION We believe that there is potential for our model to contribute to better public health policy combatting COVID-19 by providing a simple logistical framework for retrospectively analyzing the outbreak in regions that have already experienced maximal proliferation in cases. However, based upon our statistical findings, we outline certain pitfalls in modelling and their implications on results.