Obtaining reliable estimates for the distribution of COVID serial intervals is a crucial input in deciding the specific number of reproductions R 0 , which can demonstrate the magnitude of the measures needed to contain an epidemic 6.
This quantity can not however be deduced from the regular case count data alone 7. We have curated a detailed dataset of COVID translated from case reports posted online by 18 provincial health departments in mainland China outside of Hubei Province between January 20 and February 19, Appendix Table 1. Each report consists of a probable date of initiation of symptoms for both the infector and the infectee, as well as the probable locations of infection for both case-patients.
We obtained publicly available data from cities in mainland China on 9, reported COVID infection events, which were accessible online as of February 19, The data were collected from the websites of the provincial departments of public health and translated into English from Chinese Appendix Table 1. We then searched the data for clearly identified transmission events that consisted of: i a known infector and infectee, ii recorded infection locations for both cases, and iii documented symptom onset dates and locations for both cases.
We thereby obtained infector-infectee pairs identified via contact tracing in 86 Chinese cities between January 20, and February 19, Appendix Figure 1. The cases included distinct individuals, with 62 index cases infecting multiple individuals and 18 individuals occurring as both infector and infected individuals.
We range from 0 to 86 years of age, which include women and males. For each infector-infectee pari, we measured the number of days between the reported symptom onset date of the infector and the recorded symptom onset date of the infectee.
Negative values suggest that the infectee developed symptoms before the infector. We then used the fitdist function in Matlab 8 to fit a normal distribution to all observations. We applied the same method to estimate the means and standard deviations after stratifying by whether the index case was infected locally or imported.
We used maximum likelihood fitting and the Akaike information criterion AIC to evaluate four candidate models for the COVID serial interval distributions: normal, lognormal, Weibull and gamma.
The lognormal distribution provides the best fit for the truncated data followed closely by the gamma and Weibull. However, we do not believe there is cause for excluding the non-positive data and would caution against making assessments and projections based on the truncated data. The normal distribution provides the best fit for the full dataset shifted or not and thus is the distribution we recommend for future epidemiologic assessments and planning.
To facilitate interpretation and future analyses, we summarize key characteristics of the COVID infection report dataset. Of the unique cases in the dataset, 1. Across the transmission events, there were unique infectors. The mean number of transmission events per infector is 1. The transmission events were reported from 86 Chinese cities in 17 Chinese provinces and Tianjin Appendix Figure 3. There are 18 cities with at least five infection events and 68 cities with fewer than five infection events in the sample.
The maximum number of reports from a city is 36 for Xinyang, which reported cumulative cases as of February 19, In the csv file, there are 21 columns ordered by index infection, secondary infection and source. The format for this file is the following. City Chinese : initial entry of name of the city in Chinese, in which the index and secondary cases are reported. City English : initial entry of name of the city in English, in which the index and secondary cases are reported.
Index - infection location Chinese : initial entry of name of the city in Chinese, in which the index case is infected. Index-infection location English : initial entry of name of the city in English, in which the index case is infected. Secondary - infection location Chinese : initial entry of name of the city in Chinese, in which the secondary case is infected.
Secondary-infection location English : initial entry of name of the city in English, in which the secondary case is infected. Secondary description Chinese : description of secondary case in Chinese from data source. Other description Chinese : description of other information in Chinese from data source. Thirty-three of the cases suggest that the infectee had earlier symptoms than the infection.
Therefore, there may be presymptomatic transmission. Because of these negative serial intervals, the COVID serial intervals are better fit by normal distributions than the generally assumed gamma or Weibull distributions 10 , 11 , which are limited to positive values Appendix.
The mean serial interval is slightly longer when the index case is imported 5. Estimated serial interval distribution for novel coronavirus disease COVID based on reported transmission events in mainland China outside of Hubei Province from January 20 to February 19, Negative serial intervals left of the vertical dotted lines suggest the possibility of COVID transmission from asymptomatic or mildly symptomatic case-patients.
These estimates reflect the recorded dates of onset of symptoms for case-patients from 86 China cities, ranging from 0 to 86 years of age mean Although none of these studies indicate negative serial intervals before the infector in which the infectee had symptoms, 9.
We note in our estimates four possible causes of bias First, the data is restricted to online records of reported cases and thus could be skewed towards more serious cases in areas with high-functioning healthcare and public health systems.
The rapid isolation of these case-patients may have avoided longer serial intervals, possibly moving our estimate downward relative to the serial intervals that could be found in an uncontrolled outbreak.
Second, the distribution of serial intervals varies during an outbreak, with time contracting around the outbreak peak between successive cases A susceptible person would possibly get infected faster if they are surrounded by two rather than one infected person.
Since our projections are mainly based on transmission events recorded during the early stages of outbreaks, we do not specifically account for such fragmentation and view the figures at the start of an epidemic as simple serial intervals.
However, if any of the recorded infections occurred in the midst of increasing clusters of cases, our estimates may represent successful compressed serial intervals anticipated during an epidemic growth period. Third, each infector's identity and timing of the onset of symptoms is probably based on an individual memory of past events.
If the precision of the recall is impeded by time or trauma, case-patients may be more likely to relate infection over prior experiences longer serial intervals to recent experiences short serial interval. Skip to main content. This browser is no longer supported. Download Microsoft Edge More info. Contents Exit focus mode. Please rate your experience Yes No. Any additional feedback? Submit and view feedback for This product This page.
Appliance cannot generate manufacture dates for appliances made before the s. They recommend checking with a product knowledge specialist for help with dating older or antique appliances. Simple Science: Quantum Physics, Explained.
What Is the Federal Minimum Wage?
0コメント