Claim by Independent SAGE about Covid-19 infections in children is based on unclear evidence

9th Jun 2020


Recent UK and international data suggests that children are in fact as likely as adults to become infected and carry the virus.


We haven’t seen reliable and consistent evidence that backs up this claim. Reviews of the literature to date suggest that children are less likely to become infected and carry the virus. More evidence from antibody studies is needed to quantify the exact transmission risk from children.

“Recent UK and international data suggests that children are in fact as likely as adults to become infected and carry the virus”

Independent SAGE, 29 May 2020

Independent SAGE⁠—a group of experts set up with the aim of providing “robust, independent advice” to the UK Government during Covid-19—has tweeted that children are as likely to be infected as adults with the new coronavirus, according to recent data. 

The tweet is a picture of the Q&A section from Independent SAGE’s “Should Schools Reopen?” final report. A similar claim is also made earlier in the report.

This section of the report says: “Studies have shown that between 1% and 5% of diagnosed COVID-19 cases are children, but many children may be undiagnosed because up to a third of infected children never develop any symptoms (Ludvigsson, 2020). There are still some questions about the epidemiology of COVID-19 in children. Current UK data suggest that they are as likely as adults to become infected and carry the virus but they may be less likely than adults to transmit the virus because, for instance, adults are contagious for longer than children (Office for National Statistics, 2020).”

The “UK data” refers to a weekly survey published by the Office for National Statistics (ONS) in England which tests for the presence of Covid-19 among the population. The citation refers specifically to data that was published on 21 May 2020. However, experts have advised caution against over-interpreting these early stage results. In the 21 May release of the report, referenced by Independent SAGE, only 35 Covid-19 positive people were included in the analysis in total. 

This small sample size means no firm conclusions can be drawn about Covid-19 infection rates between different age groups. The margin of error associated with each result is too high.

More recent releases of the ONS statistics now include results from the COVID-19 Infection Survey rather than results from the most recent 14-day period (as was reported previously). Based on these results, the ONS writes “It is not possible to say with confidence that there is any difference in the proportion of individuals in different age groups testing positive for COVID-19”.

The ONS said the data showed “there is no evidence of differences in the proportions of individuals testing positive for COVID-19 between different age categories.” But saying that its data shows no evidence of differences, does not mean it shows that there are no differences, as Independent SAGE suggested.

We asked Independent SAGE for the international evidence they refer to in their claim that children are as likely to become as infected and carry the new coronavirus as adults, and they provided us with the references for two studies. They did add that there are other studies that show lower infection in children.

The first is a Chinese study that followed up the contacts of 391 SARS-CoV-2 (the coronavirus that causes Covid-19) positive cases between 14 January and 12 February. This study found that “the rate of infection in children younger than 10 years (7.4%) was similar to the population average (6.6%)”. However, as we have written about before, this study is countered by three other Chinese studies and one Japanese study that used similar methods and found infection rates in children were much lower than adults.

The second study, from Germany, looked at the amount of SARS-CoV-2 virus produced by an infected person (known as viral load) among patients in different age groups. It concluded that “children may be as infectious as adults”. Viral load, alongside other factors such as the duration and frequency of exposure, can help to understand how likely someone is to transmit a virus. The aim of the study was to “enable an estimate of infectivity in children”, rather than measure infection rates. This study (which has not undergone peer review) has also had mixed feedback from other academics, including criticism of the statistical methods used to detect differences by age. An independent re-analysis of the same data by Leonhard Held, Professor of Biostatistics at the University of Zurich, suggests that the data actually shows a moderate increase in viral load with age.

As we have written about before, reviews of the evidence to date are cautiously suggesting that children are less prone to catching the new coronavirus and play a minimal role in its transmission.

The Royal College of Paediatrics and Child Health writes that high-quality evidence from completed antibody studies (which can show whether someone has had Covid-19 rather than whether they currently have it) are needed to confirm the precise details regarding paediatric transmission.

Update 16 June 2020

We updated this article to include the final report from Independent SAGE rather than the interim report, and the international evidence they supplied us with after this article was initially published. We have updated the headline to clarify this article is looking at claims about infection rather than transmission, and updated the conclusion to better reflect the updated article.