Abstract
Background An accurate estimate of the distribution of the incubation period for COVID-19 is the foundational building block for modelling the spread of the SARS COV2 and the effectiveness of mitigation strategies on affected communities. Initial estimates were based on early infections, the aim of this study was to provide an updated estimate and meta-analysis of the incubation period distribution for COVID-19.
Methods The review was conducted according to the PRISMA Scoping Review guidelines. Five databases were searched; CINAHL, MEDLINE, PUBMED, EMBASE, ASSIA, and Global Index Medicus for studies published between 1 January 2020 - 27 July 2020.
Results A total of 1,084 articles were identified through the database searches and 1 article was identified through the reference screening of retrieved articles. After screening 64 articles were included. The studies combined had a sample of 45,151 people. The mean of the incubation periods was 6.71 days with 95% CIs ranging from 1 to 12.4 days. The median was 6 days and IQR ranging from 1.8 to 16.3. The resulting parameters for a Gamma Distribution modelling the incubation period were Γ(α, λ) = Γ(2.810,0.419) with mean, μ = α/λ.
Conclusion Governments are planning their strategies on a maximum incubation period of 14 days. While our results are limited to primarily Chinese research studies, the findings highlight the variability in the mean period and the potential for further incubation beyond 14 days. There is an ongoing need for detailed surveillance on the timing of self-isolation periods and related measures protecting communities as incubation periods may be longer.
INTRODUCTION
To define the incubation period ECDC (2012) provides a simple overview of infectious disease epidemiology definitions at the different stages of infection according to time. The period between exposure and onset of clinical symptoms is called the ‘incubation period’. The host may become infectious (i.e. able to transmit the pathogen to other hosts) at any moment of the infection and this moment will vary per pathogen1
Early estimates of the incubation period of the emerging SARS COV2 virus were reported by Backer et al in Wuhan, China on the 27th January 20202. Authors reported that within the cases observed the incubation period was estimated to be 6.4 days. Following on from this early estimate authors Li et al in Wuhan reported a mean incubation period of 5.2 days3 and Lauer et al in China reported the median of the incubation period of 5.1 days4. Further individual studies in Europe by Bohmer et al have shown that the median period was 4 days with an Interquartile period of 2 days5. Within Saudi Arabia it has been shown that the incubation period can be 6 days6. Since these early studies, numerous further individual cases notifications and modelling studies have emerged estimating the period to range from 1 to 34 days2-65. However, to date no scoping review or meta-analysis has been conducted on the results of these individual studies with the primary objective of quantifying the mean and standard deviation of the incubation period for the purposes of modelling it’s distribution.
In the early stages of a new epidemic where no vaccine is available all persons are susceptible particularly some communities of older and vulnerable people. As the epidemic progresses and the number of infectious individuals increases the number of susceptible individuals will decrease. However, when an epidemic can produce both asymptomatic and symptomatic cases the identification of the numbers infected becomes more challenging. Mathematical modelling of the effective reproductive number R of an epidemic, the spread of infection, and the subsequent decisions on planning and mitigation is fully dependent on accurate and up to date estimates of the key modelling parameters. To aid this process and given the passage of time and improvements in surveillance and monitoring since the first cases, we conducted a scoping review and meta-analysis of the parameters of the global incubation period of COVID-19. Our ultimate aim was to provide an up to date estimate of the incubation period distribution to aid ongoing modelling estimates of the effective reproductive number R and subsequent estimates of the hidden prevalence of asymptomatic cases.
METHODS
The study design was a scoping review with a meta-analysis. Daudt et al66in their work on enhancing the scoping review methodology based on the seminal work on scoping reviews by Askey and O’Malley (2005) retain the original definition of a scoping review and state that such reviews, “map rapidly the key concepts underpinning a research area and the main sources and types of evidence available, and can be undertaken as standalone projects in their own right, especially where an area is complex or has not been reviewed comprehensively before”66p.2. Within this review we are seeking to rapidly uncover, ‘the main sources and types of evidence available in an area that has not been reviewed comprehensively before’. In addition, the framework Askey and O’Malley (2005) developed expanded on this definition by identifying four main reasons for conducting a scoping study: (1) to examine the extent, range and nature of research activity; (2) to determine the value of undertaking a full systematic review; (3) to summarise and disseminate research findings; and (4) to identify research gaps in the existing literature. Within this review aims three and four are considered66. Levac et al (2010) highlight that scoping studies differ from systematic reviews because authors do not typically assess the quality of included studies67. Scoping studies also differ from narrative or literature reviews in that the scoping process can require analytical reinterpretation of the literature. Grant et al in their typology of reviews describe scoping reviews as providing a preliminary assessment of the potential size and scope of available research in the literature, with the aim of identifying the nature and extent of research evidence including ongoing research68. The completeness of the search in a scoping review is also determined by time and scope constraints. They share several characteristics of the systematic review in attempting to be systematic, transparent and replicable however their lack of a quality assessment and their rapid duration are recognised limitations.
Search strategy and selection criteria
This review was conducted according to the PRISMA Scoping Review guidelines69. Five databases were searched; CINAHL, MEDLINE, PUBMED, EMBASE, ASSIA, and Global Index Medicus for studies published between 1 January 2020 and 27 July 2020. No language restrictions were applied, and non-English publications were translated. Search terms used included “coronavirus”, “covid-19”, AND “Incubation.” An example of search terms used for CINAHL was (((MH “Coronavirus+”) or coronavirus*) AND (wuhan or beijing or shanghai or “2019-nCoV” or “Covid-19” or “SARS-CoV-2”)) OR ((((“novel coronavirus*” AND ((MH “China”) or China)) OR TI coronavirus*) OR (((MH pneumonia) or pneumonia) AND Wuhan OR ((“Covid-19” or “2019-nCoV” or “SARS-CoV-2” or (MH Coronavirus Infections))))) AND PY 2020) AND (MH Incubation).
The database search was supplemented by screening the references of retrieved articles. Data was imported into EndNote version x9 and Excel was also used for data extraction. Duplicates were removed using EndNote and another duplicate check was conducted manually. Studies were included if they presented incubation data on COVID-19. Study designs included retrospective case studies, multicentre studies, case series, prospective cohort studies, web data mining, cross-sectional studies, and individual case reports. Studies were excluded if they did not include findings on incubation period of COVID-19 or new information on the incubation period, or if they were reviews, no other study designs were excluded. Studies which assumed values of the incubation period based on other studies were also excluded at full text screening stage. Grey literature sources and unpublished studies were also assessed but trial registries were not.
One reviewer performed the search and the screening and queries were resolved with the second reviewer. Data were extracted based on the publication and no individual patient-level data were requested from the study authors, summary of estimates were included.
Data analysis
Data extraction was conducted and summarised by PB. Duplicates were removed automatically using EndNote and a manual check was conducted. Extracted variables included first author, country, study design, sample size, and incubation period data with the primary outcomes being the mean, standard deviation and 95% confidence interval for the mean of the period and secondary outcomes being the minimum, maximum, range, median, interquartile range, and any additional reported confidence intervals. Using the Central Limit Theorem, the pooled mean was computed, and a pooled estimate of the variance and standard deviation were provided using standard formulas. For example, if we have 30 relevant studies we have,
The standard deviation was available for four studies giving a pooled estimate described by,
The median of the medians and the ranges in all other outcomes were provided. Forest plots were prepared to display the mean and median values and their respective 95% confidence intervals and inter-quartile ranges.
To model the distribution of incubation times in an epidemic a Gamma Probability Density Function is recommended41.This is described by
where f Is the probability of an incubation period of duration x given a mean incubation period of mean, ¼= ±/ »Parameter values for the Gamma Distribution describing the incubation period distribution were prepared based on the review findings.
As this was a scoping review a quality assessment of the studies was not carried out as is in keeping with the rapid nature and the protocol for this review type.
RESULTS
A total of 1,084 articles were identified through the database searches and 1 article was identified through the reference screening of retrieved articles. Duplicates were removed (n= 227) and 858 articles were screened based on abstract and title. After the screening 689 articles were excluded and the remaining 169 articles were assessed for eligibility. The total number of full-text articles excluded with reasons was 105. A breakdown is provided in Figure 1.
The 64 included studies had a combined sample size of 45,151 (Table 1). Table 2 provides a summary of the incubation period from the 64 studies included in this scoping review. The age ranged between 0 and 90 years, with a median of 43 years and mean of 43.9 years. The minimum incubation period ranged between 1 and 9 days. The maximum incubation period ranged between 4 and 34 days. The mean incubation period was 6.71 (31 studies), and the median was 6 days (58 studies).
Detailed Results from the 64 Articles
Summary of Incubation Period
Forest plot illustrating the spread in the mean values, and their respective confidence intervals are provided in Figures 2. We can see from Figure 2 that 30 studies provided information on the mean value of the observed incubation period. These thirty studies represented a combined sample size of 43,132 individuals. Only 12 of the 30 studies provided estimates for the 95% confidence interval of the mean incubation period. A total of 58 studies provided details on the median incubation period observed and 18 of these provided some details on the interquartile range. Given the pooled estimates for the mean and the standard deviation and given the Gamma Distribution where 
and 
describes the incubation period distribution we found that the shape parameter was given α=2.81 and λ=0.42 was the estimated rate or scale parameter.
DISCUSSION
This scoping review finds that based on studies to date, the mean incubation period for COVID-19 is given by 6.71 days with a standard deviation of 4.00 days and 95% CIs ranging from 1.8 to 12.4 days. These findings provide for the first time the parameter estimates required for a Gamma distribution to model the period.
The importance of accurately modelling the incubation period on the spread of an epidemic has long been identified from the earliest work of Italian physician Girolamo Fracastro (1478-1553) on rabies 70 to the problem of modelling the incubation period of infectious diseases such as measles, poliomyelitis, smallpox and SARS 70,71. The importance of accurately measuring and modelling this period for SARS COV2 was identified early in this pandemic by Backer et al 2 when the mean incubation period was estimated to be 6.4 days and by Li et al 3 where a mean incubation period of 5.2 days was reported. Our updated review finds the mean may be longer than previously estimated and in addition provides the first estimates for a gamma distribution describing the period.
A limitation of our results is however the lack of published data on the variability in mean times. Of the 64 papers included in the review only 30 reported mean incubation period times and only 4 reported standard deviations or variance. This lack of information on the variability in mean times will undoubtedly introduce a level of uncertainty and bias into the Gamma distribution parameter estimates. Furthermore, the range in mean values reported across the 30 studies varied greatly. Clearly the quality of some studies was lacking, and this assessment of quality was by definition outside the stated objectives of a scoping g review. None the less, this scoping review has provided the first meta-analysis of 64 studies published from January to July 2020.
In spite of the provision of 95% confidence intervals for the mean incubation period it was not clear from the studies how these intervals were computed and estimates of the variance and standard deviation could not be computed from them using standard formulae. Given the lack of clarity on the variance and hence standard deviation it is recommended that in future research greater attention is given to the computation details and formula used are provided.
To conclude, for the planning and provision of mitigation strategies and advice on isolation periods for the general public, it is essential that an accurate and update estimate of the incubation period is maintained as the SARS COV2 virus spreads. It is equally essential that the mitigation strategies on affected communities and hospital planning requirements that are determined from modelling scenarios are based on reliable and accurate global estimates of the incubation period. There is an ongoing need for detailed surveillance on the timing of self-isolation periods and related measures protecting communities as incubation periods may be longer. This international scoping review provides this estimate based on the findings to the end of July 2020.
What is already known on this subject
▸ Early estimates of the incubation period of the emerging SARS COV2 virus were reported by Backer et al in Wuhan, China on the 27th January 2020.
▸ Following on from this early estimate authors Li et al in Wuhan reported a mean incubation period of 5.2 days and Lauer et al in China reported the median of the incubation period of 5.1 days.
▸ Since these early studies, numerous further individual cases notifications and modelling studies have emerged estimating the period to range from 1 to 34 days.
What this study adds
▸ This scoping review finds that based on studies to date, the mean incubation period for COVID-19 is given by 6.71 days with a standard deviation of 4.00 days and 95% CIs ranging from 1.8 to 12.4 days.
▸ These findings provide for the first time the parameter estimates required for a Gamma distribution to model the period.
REFERENCES
Subject Area
Reviews and Context
0
Comment
0
TRIP Peer Reviews
0
Community Reviews
0
Automated Services
5
Blogs/Media
Author Videos