Discussion
We have proposed a methodology for the estimation of the key epidemiological parameters as well as the modelling and forecasting of the spread of the COVID-19 epidemic in Hubei, China by considering publicly available data from the 11th of January 2020 to the 10th of February 2020.
By the time of the acceptance of our paper, according to the official data released on the 29th of February, the cumulative number of confirmed infected cases in Hubei was ∼67,000, that of recovered was ∼31,300 and the death toll was ∼2,800. These numbers are within the lower bounds and expected trends of our forecasts from the 10th of February that are based on Scenario I. Importantly, by assuming a 20-fold scaling of the confirmed cumulative number of the infected cases and a 40-fold scaling of the confirmed number of the recovered cases in the total population, forecasts show a decline of the outbreak in Hubei at the end of February. Based on this scenario the case fatality rate in the total population is of the order of ∼0.15%.
At this point we should note that our SIRD modelling approach did not take into account many factors that play an important role in the dynamics of the disease such as the effect of the incubation period in the transmission dynamics, the heterogeneous contact transmission network, the effect of the measures already taken to combat the epidemic, the characteristics of the population (e.g. the effect of the age, people who had already health problems). Also the estimation of the model parameters is based on an assumption, considering just the first period in which the first cases were confirmed and reported. Of note, COVID-19, which is thought to be principally transmitted from person to person by respiratory droplets and fomites without excluding the possibility of the fecal-oral route [21] had been spreading for at least over a month and a half before the imposed lockdown and quarantine of Wuhan on January 23, having thus infected unknown numbers of people. The number of asymptomatic and mild cases with subclinical manifestations that probably did not present to hospitals for treatment may be substantial; these cases, which possibly represent the bulk of the COVID-19 infections, remain unrecognized, especially during the influenza season [22]. This highly likely gross under-detection and underreporting of mild or asymptomatic cases inevitably throws severe disease courses calculations and death rates out of context, distorting epidemiologic reality.
Another important factor that should be taken into consideration pertains to the diagnostic criteria used to determine infection status and confirm cases. A positive PCR test was required to be considered a confirmed case by China’s Novel Coronavirus Pneumonia Diagnosis and Treatment program in the early phase of the outbreak [14]. However, the sensitivity of nucleic acid testing for this novel viral pathogen may only be 30-50%, thereby often resulting in false negatives, particularly early in the course of illness. To complicate matters further, the guidance changed in the recently-released fourth edition of the program on February 6 to allow for diagnosis based on clinical presentation, but only in Hubei province [14].
The swiftly growing epidemic seems to be overwhelming even for the highly efficient Chinese logistics that did manage to build two new hospitals in record time to treat infected patients. Supportive care with extracorporeal membrane oxygenation (ECMO) in intensive care units (ICUs) is critical for severe respiratory disease. Large-scale capacities for such level of medical care in Hubei province, or elsewhere in the world for that matter, amidst this public health emergency may prove particularly challenging. We hope that the results of our analysis contribute to the elucidation of critical aspects of this outbreak so as to contain the novel coronavirus as soon as possible and mitigate its effects regionally, in mainland China, and internationally.