Preliminary prediction of the basic reproduction number of the Wuhan novel coronavirus 2019‐nCoV

Tao Zhou, Quan-Hui Liu, Zimo Yang, Jingyi Liao, Kexin Yang, Wei Bai, Xin Lü, Wei ZhangJournal of Evidence-Based Medicine 2020 · DOI

TL;DR

This paper estimates the basic reproduction number (R₀) of the novel coronavirus (2019-nCoV) during its early outbreak in Wuhan. The findings suggest that 2019-nCoV has a transmissibility similar to or slightly higher than SARS, highlighting the urgent need for timely and effective control measures.

What Problem Does This Paper Address?

The rapid spread of the novel coronavirus (2019-nCoV) in Wuhan presented a critical challenge: understanding its infectiousness to inform public health responses. A key metric for this is the basic reproduction number (R₀), which represents the average number of secondary infections caused by a single infected individual in a completely susceptible population. Estimating R₀ early in an epidemic is crucial for predicting its trajectory, assessing the effectiveness of interventions, and designing appropriate control strategies. This paper aims to provide an early estimate of R₀ for 2019-nCoV, drawing parallels with the SARS outbreak to inform these efforts.

Methods

The authors employed a Susceptible-Exposed-Infected-Removed (SEIR) compartmental model to describe the early dynamics of 2019-nCoV transmission. This model tracks individuals through four states: susceptible, exposed (infected but not yet infectious), infectious, and removed (recovered or deceased). They assumed that cases with symptom onset before January 25, 2020, represented the "free propagation" phase, meaning transmission occurred without significant interventions.

The basic reproduction number (R₀) was estimated using the formula:

$R_0 = \lambda T_g \left(1 + \frac{\rho(1-\rho)}{\lambda T_g}\right)$

where:

  • $\lambda$ is the growth rate of the early exponential growth of infected cases.
  • $T_g$ is the generation time (the average time between an infection and the onset of infectiousness in the secondary cases).
  • $\rho$ is the ratio of the exposed period to the generation time.

To estimate these parameters, the researchers utilized real-time data on confirmed and suspected cases from various sources, including People's Daily reports and data from Northeastern University. They also incorporated epidemiological determinants learned from SARS, such as incubation periods and generation times, while acknowledging the need for specific data for 2019-nCoV. Sensitivity analyses were performed to assess the impact of key parameters like generation time and the proportion of infectious cases.

Datasets

The study utilized the following datasets:

  • People's Daily Reports: Real-time data on confirmed and suspected cases of 2019-nCoV.
  • DXY.cn: An online community for health care professionals, also providing data on the 2019-nCoV situation.
  • Northeastern University Reports: Estimated numbers of 2019-nCoV infections.

The paper also references data from previous studies on SARS, Zika virus, and MERS-CoV for comparative analysis.

Key Insights and Ideas

  • Early R₀ estimates are crucial but challenging: The paper underscores the difficulty of obtaining precise R₀ values early in an outbreak due to limited and evolving data.
  • 2019-nCoV transmissibility is comparable to SARS: The estimated R₀ values suggest that 2019-nCoV is at least as transmissible as SARS, and potentially slightly more so, necessitating a robust response.
  • Generation time is a critical parameter: The study highlights the sensitivity of R₀ estimates to the generation time ($T_g$), emphasizing the need for accurate, data-driven estimates of this parameter.
  • Data sources significantly impact estimates: Different data sources (e.g., People's Daily vs. Northeastern University reports) yielded a range of R₀ values, illustrating the importance of data quality and source.
  • Controllability depends on timely interventions: Despite its moderate-to-high transmissibility, the disease is considered controllable if swift and effective measures are implemented.

Key Results

The estimated basic reproduction number (R₀) for 2019-nCoV fell within the following ranges:

  • 2.8 to 3.3 based on People's Daily Reports.
  • 3.2 to 3.9 based on Northeastern University Reports.
  • A broader range of [2.3, 5.2] was observed during sensitivity analysis, with a more refined range of [2.8, 3.9] considered more likely.
  • A later update, incorporating new data and a shorter serial interval, suggested a range of 2.2 to 3.0.

The study concluded that the early transmission ability of 2019-nCoV is close to or slightly higher than SARS, classifying it as a disease with moderate-to-high transmissibility.

Limitations and Future Work

  • Data limitations: The authors acknowledge that the study was a preliminary estimation based on largely insufficient data, especially regarding the generation time. The number of confirmed cases obtained from different sources varied widely, and a significant proportion of patients might have had mild symptoms and not been officially confirmed.
  • Model assumptions: The SEIR model assumes that individuals in the exposed state have very low or no infectiousness, which might not hold true for 2019-nCoV, as suggested by some preliminary findings.
  • Need for precise epidemiological determinants: To improve R₀ estimation and epidemic prediction, more precise epidemiological determinants are needed.
  • Model refinement: Future work should consider a more complex model that accounts for diversity in susceptibilities, contact probabilities, different spreading mechanisms (e.g., in hospitals vs. communities), and the effects of regional population density and human mobility.
  • Microscopic survey: A microscopic survey about transmission processes is needed to further improve the quality of predictions, particularly regarding the generation time.
  • Updated data: The paper notes that with updated data and a shorter serial interval, the predicted R₀ became smaller, but this does not necessarily imply a more controllable situation due to faster propagation.

Summary updated

Abstract

The early transmission ability of 2019-nCoV is close to or slightly higher than SARS. It is a controllable disease with moderate to high transmissibility. Timely and effective control measures are needed to prevent the further transmissions.
BibTeX
@article{zhou2020preliminary,
  author = {Tao Zhou and Quan-Hui Liu and Zimo Yang and Jingyi Liao and Kexin Yang and Wei Bai and Xin Lü and Wei Zhang},
  title = {{Preliminary prediction of the basic reproduction number of the Wuhan novel coronavirus 2019‐nCoV}},
  journal = {Journal of Evidence-Based Medicine},
  year = {2020},
  doi = {10.1111/jebm.12376},
}
Figure