Basic patterns and information of different regions
We collected disease information from the total country of China and 18 provincial regions up to the March 11, 2020 (Supplemental able Table 2). The infection rate, death rate based on PIBA, and daily numbers of patients and deaths are calculated from 15 of these regions (Supplemental Table 3). The data from three regions, Guangdong, Shanghai and Sichuan were not usable for the analyses because of missing data on the total number of persons who were in close contact with the disease careers. In addition, these calculations did not include the patients who came to China after they were infected outside of China. Based on this information, we divided these regions into four groups (Figure 1A):
1). Hubei, a single province as a group because of its large patient population and disease pandemic situation. The initial date of disease has been estimated at early December 2019 or earlier. From the December 1, 2019 through March 11, 2020, the COVID-19 has been an pandemic in Hubei for 95 days (Fig. 1C). More importantly, currently new patients are still identified every day.
2). The three regions in which the pandemic of COVID-19 has not been ended include Heilongjiang, Beijing and Shandong (Fig. 1D).
3). The four regions that the disease is potentially ending because no new patients were last found in more than 10 days, but not yet to 14 days (Fig. 1D).
4). There are 7 regions that in which the COVID-19 has stopped its status as an pandemic (Fig. 1F). The peak period of disease pandemics of these four groups are similar, with a period of 3 to 4 weeks, approximately from January 20, 2020 to the middle of February 2020 (Fig. 1C-1F). The total number of patients from each region were logicized to its 10th [9] so that Hubei’s data can be listed together with others. However, because of its extreme large number of patients, its number in the analysis is over-weighted and was not used in the correlation analysis. The correlation analysis indicated that the disease duration has no correlation with the number of patients, with r = 0.184 and P = 1.7987E-09. In addition, the number of deaths was not correlated to the disease duration, with r = 0.242 and P = 7.43019E-09.
Infection rate in different groups of pandemic duration of COVID-19.
Based on the collected data, we first analyzed the infection rate. Due to the difference in the availability of data in these provinces, we are counting from February 7, 2020. From this date, every province and city posted available data that we can use to calculate the daily infection rate and mortality. Our results show that the infection rates of these four groups are very different from each other (Fig. 2A). Hubei Province has the highest infection rate (Fig. 2B), followed by provinces where the pandemic is continuing (Fig. 2C). The lowest infection rates are in the groups that the pandemic has potentially ended or ended (Figure 2D-2E). T-test showed that the infection rate of Hubei is significantly different from other three groups, with P values of 1.126E-27, 7.51E-32, 7.43E-23, to not ending, potentially ending and ended groups, respectively. The infection rate of the not ending group is also significantly different from the potentially and ended groups, with P values of 1.69E-31, and 8.62E-26, respectively.
Daily death rate in different disease development groups
In view of the differences in the daily infection rates among these four groups, we further observed whether the cumulative daily mortality rate was different among these four groups. Daily mortality is simply the number of deaths per day divided by the total number of patients. Similarly, the calculation was started on Feb 7, 2020.
As we predicted, these four groups also have different mortality rates (Fig. 3A). Hubei has the highest mortality rate (Fig. 3B), followed by the group that the COVID-19 pandemic has not ended yet (Fig. 3C). The next group is likely to end in the near future (Fig. 3D). Finally, it is the group whose pandemic has ended (Fig. 3E). The current average mortality rates for these four groups are 0.0392, 0.0178, 0.0075 and 0.0069, respectively. The P values between Hubei and not ending, potentially ending and ended groups are 1.733E-23, 1.61E-24, 7.77E-22, respectively. The not ending group was also different from the potential and ending groups, with P values of 6.73E-25 and 3.48E-25, respectively.
We noticed that the death rate in Hubei province has been dropped from 4% at the beginning to 2% at the end of pandemic. However, the death rate of rest of the country is much lower and stable, mostly lower than 2% (Fig. 3A). Death rate in regions with shortest pandemic period were below 1%.
Initial infection and death rate in different groups
We next compared the infection and death rate of different groups at the beginning of the COVID-19 pandemic. Based on the PIBA method, the maximum days of a patient from inpatients to the death μ + 2σ is 25 days. We therefore used the data from 25 days at the beginning to calculate the initial infection and death rate in different groups. We used the average death rate of the first 25 days. Thus, Dir = ∑(d1~d25)/∑(t1~t25), where Dir is the initial death rate, d1~d25 = number of deaths from day 1 to day 25. T1~t25, total number of inpatients from day 1 to day 25.
Our analysis indicated that the initial infection and death rate in Wuhan and the group of not ending of disease are higher than that of other two groups (Figure 4). These data confirm the analysis on the infection and death rate of different groups in which the Wuhan and not ended groups have higher infection and death rates when the potentially ended and ended groups.
Prediction of disease development in different countries.
Based on the differences in infection rate and death rate of different groups, we further investigated the relationship between the days of disease durations and the infection and death rates (Figure 5). Our analysis showed that there is a positive correlation between the days of disease duration and infection rate, with a R = 0.626 (Fig. 5A). Furthermore, there is a strong positive correlation between the disease duration and total death rate, with a R = 0.707 (Fig. 5B). These data together with the data above clearly demonstrate that the high infection rate and death rate predict a long pandemic period of COVID-19. However, we realize that, due to the differences in sizes of tested populations and availability and variations in the testing technologies, the mathematic equation for the relationship between infection rate and pandemic duration from China may not be equivalent to other countries.
As to the relation between the death rate and pandemic duration, it may be possible to apply this approach to other countries or regions, not only because of their strong correlation, but also because of the similar criteria and credibility of the death rate among different countries. Accordingly, we attempted to predict the pandemic durations of some countries using their death rates. As the pandemic of disease of COVID-19 is still ongoing in many countries, we question whether the initial data of death rate can be a predictor for the pandemic duration. As mentioned above, based on PIBA method we collected the death rate of (μ + 2σ) 25 days of these regions in China and examined the relationship between the death rate and pandemic duration. We obtained a positive relationship with a R value of 0.597. A line regress(?) formula is obtained from their data (Fig. 5C). We then obtained the death rate of first 25 days of 8 countries. We calculated the pandemic days of these countries using the regression formula (Fig. 5D). Because China took an extreme measure on the social isolation, while many other countries used the soft measure of social distance, we assume there could be a longer time of pandemic period in other countries. Therefore, we also provide secondary days which are derived by using calculated days to multiple 1.5, as the potential maximum days of pandemic duration. Our calculation suggested that the pandemic days of these countries ranges from 37 to 76 days (Fig. 5D) based on direct calculation and from 56 to 114 days based on the multiplication of 1.5.