Preprint Short Note Version 3 This version is not peer-reviewed

The Ongoing COVID-19 Epidemic Curves Indicate Initial Point Spread in China With Log-Normal Distribution of New Cases per Day With a Predictable Last Date of the Outbreak Version 3: Test for When After a Peak in Daily Cases the Predicted Equation Becomes Reliable and Use of the Derivative of the Equation to Detect Time of Key Changes Determining the Length of the Outbreak

Version 1 : Received: 4 March 2020 / Approved: 5 March 2020 / Online: 5 March 2020 (02:58:51 CET)
Version 2 : Received: 11 March 2020 / Approved: 12 March 2020 / Online: 12 March 2020 (05:04:52 CET)
Version 3 : Received: 25 March 2020 / Approved: 27 March 2020 / Online: 27 March 2020 (02:22:11 CET)
Version 4 : Received: 16 April 2020 / Approved: 19 April 2020 / Online: 19 April 2020 (08:15:10 CEST)

How to cite: Olsson, S.; Zhang, J. The Ongoing COVID-19 Epidemic Curves Indicate Initial Point Spread in China With Log-Normal Distribution of New Cases per Day With a Predictable Last Date of the Outbreak Version 3: Test for When After a Peak in Daily Cases the Predicted Equation Becomes Reliable and Use of the Derivative of the Equation to Detect Time of Key Changes Determining the Length of the Outbreak. Preprints 2020, 2020030077 (doi: 10.20944/preprints202003.0077.v3). Olsson, S.; Zhang, J. The Ongoing COVID-19 Epidemic Curves Indicate Initial Point Spread in China With Log-Normal Distribution of New Cases per Day With a Predictable Last Date of the Outbreak Version 3: Test for When After a Peak in Daily Cases the Predicted Equation Becomes Reliable and Use of the Derivative of the Equation to Detect Time of Key Changes Determining the Length of the Outbreak. Preprints 2020, 2020030077 (doi: 10.20944/preprints202003.0077.v3).

Abstract

During an epidemic outbreak it is useful for planners and responsible authorities to be able to plan ahead to estimate when an outbreak of an epidemic is likely to ease and when the last case can be predicted in their area of responsibility. Theoretically this could be done for a point source epidemic using epidemic curve forecasting. The extensive data now coming out of China makes it possible to test if this can be done using MS Excel a standard spreadsheet program available to most offices. The available data is divided up for whole China and the different provinces. This and the high number of cases makes the analysis possible. Data for new confirmed infections for Hubei, Hubei outside Wuhan, China excluding Hubei as well as Zhejiang and Fujian provinces all follow a log-normal distribution that can be used to make a rough estimate for the date of the last new confirmed cases in respective areas. In the version 2 continuation work, 9 additional days were added for the Chinese data to evaluate the previous predictions. The extra data then available from China follows the previous predicted trend supporting the usefulness of this simple technique. In the version 2 we also tested the feasibility for a non-specialist to make similar predictions using additional data from S Korea now available. In this third continuation the predictions for Version 2 are evaluated for S Korea and fits well the beginning of the decline but it seems to be difficult to bring down numbers of cases per day under about 100 new cases per day, potential reasons for this is discussed. To further evaluate when in a prediction becomes reliable the Chinese data was used to evaluate to make predictions for each day around the peak in number of cases and after2-3 consecutive days of decreasing new cases per day the prediction becomes reliable. In version 3 data for Italy just reaching this point was used to make further predictions for that country. A second new analysis was also added to use the fitted equation to detect when the acceleration of new cases per day stopped increasing exponentially. In the Chinese case this measured point coincides with the date of the complete Hubei lockdown and in the new Italian analysis it coincides with the mandatory Italian lockdown. Predicted dates for the end of the Italian outbreak is also added.

Subject Areas

epidemiology; COVID-19

Comments (1)

Comment 1
Received: 27 March 2020
Commenter: Stefan Olsson
Commenter's Conflict of Interests: Author
Comment: In this further extended work, we evaluated the predictions previously made for S Korea and found that predictions were valid but inflow from other countries and some minor outbreaks and/or not strong enough measures might make it difficult get the outbreak to completely disappear. We have in V3 added an analysis for when the prediction using our method becomes reliable and that happens a few days after the peak where number of infections per day start a reliable decline. We also in this version added an analysis of when new infections pre day stopped accelerating with the same rate. We then found that it was rather early in the China case and only some days later for Italy. To our big surprise Both these dates coincide with the days mandatory lockdown took effect in both countries. The few days later in the outbreak lockdown in Italy is then also a probable cause of the higher peak and is predicted to result in 2 times longer outbreak with 3 times higher numbers of total cases than was the case for China. Thus a few days hesitation taking the lockdown decisions appears to have had huge effects.
+ Respond to this comment

We encourage comments and feedback from a broad range of readers. See criteria for comments and our diversity statement.

Leave a public comment
Send a private comment to the author(s)
Views 0
Downloads 0
Comments 1
Metrics 0


×
Alerts
Notify me about updates to this article or when a peer-reviewed version is published.