select search filters
roundups & rapid reactions
before the headlines
Fiona fox's blog

expert reaction to review study looking at whether evidence suggests face coverings could lead to a false sense of security

A study, published in BMJ Analysis, looked at whether evidence suggests face coverings could lead to a false sense of security.


Dr Antonio Lazzarino, Honorary Research Associate, Institute of Epidemiology and Healthcare, UCL, and Director of EPISTATA (agency for clinical research and medical statistics), said:

“This article does not add anything new to the existing evidence, as it is a review of previous evidence and is in my view non-systematic. The paper does not look at risk compensation in COVID-19, it gives some examples of past research. The paper does not produce any quantifications of the effects it describes, nor does it assess the quality of the reported studies.

“As far as I know, there are a few ongoing studies aiming at measuring those effects. We must wait and see if they are run correctly, in an unbiased way, and what are the results they give. These studies are meant to quantify the net effects of masks (and therefore they will incorporate the risk compensation effect), so there is no point in keeping publishing opinions that are based on past studies, which were conducted for different reasons and which were not designed to give the answers we need. We need a quantification of the net effect of face masks in COVID-19, i.e. taking the pros and cons into account. If as seems to be becoming clear airborne transmission plays an important role in COVID-19, rather than just salivary droplets, then it’s likely that the masks may turn out to give very little protection and social distancing would remain the most important measure to reduce transmission.”


Prof Susan Michie, Director of UCL Centre for Behaviour Change, UCL, said:

“The authors rightly point to evidence that ‘risk compensation’ is not a universal phenomenon but, as they themselves recognise, there may be contexts in which it can occur.  So the idea of ‘burying it’ is as potentially dangerous as assuming that it will occur.  The appropriate approach is to consider it as a possibility and rapidly investigate it to determine whether it is an issue and if so how significant, in whom, in what contexts and with what alternative protective behaviours.

“The review does not present a methodology so it is not possible to evaluate the findings presented but it does appear that it makes the error of pooling study findings in very different contexts rather than considering in detail the ways in which the contexts, populations and behaviours may influence the results.

“It is important that researchers resist the temptation to use headline-grabbing language and oversimplifications when presenting findings.”


Prof Patricia Riddell, Professor of Applied Neuroscience, University of Reading, said:

“The article by Mantzari et al has systematically scanned for all of the available evidence that relates directly to the effects of mask wearing on other risk taking behaviours.  Unfortunately, the research in this area is limited and so most research has only investigated the effect of mask wearing on hand-sanitising.  Additional research on the effect on other risk taking behaviours, for instance, social distancing, would be needed to confirm this finding more robustly.

“At the start of the Covid19 pandemic, it was argued that starting lockdown too soon might not be effective if people were not willing to maintain this behaviour for long enough to reduce transmission rates.  This concept, referred to in behavioural science as temporal discounting, was used to justify a delay in implementing social distancing measures in the UK though it is not clear that this was supported by relevant scientific experts on the SAGE committee.  Another concept from behavioural science, the idea of risk compensation, is now being used to suggest that wearing a face mask might not be an effective measure.  The argument made is that individuals are prepared to take a certain level of risk and therefore, if they act to reduce the risk in one way, they will change their behaviour in another way to increase the risk.  Examples this come from driving where increased car safety (e.g. introduction of ABS braking systems) is thought to have resulted in riskier driving (shorter distance between cars) such that the number of accidents has stayed the same rather than reducing.

“Mantzari and colleagues have searched the literature for evidence that wearing face masks might result in changes in other behaviours that would increase risk of infection.  They show that there is no decrease (and even a potential increase) in hand sanitising in groups who are required to wear face masks in comparison to control groups who do not wear masks.  They argue from this that there is no risk compensation, and therefore that the concept of risk compensation could be a threat to public health if it reduces the likelihood that governments will mandate the wearing of face masks in public spaces.

“In situations where there is some dispute about the effectiveness of particular measures for public health, it would seem circumspect, especially when lives are at risk, to follow not only the advice of scientists but the scientific method.  It would be possible to set up experiments that can be used to test whether masks are effective in reducing transmission rates or whether people maintain a constant level of risk when asked to wear a mask by, for instance, decreasing social distancing or hand sanitising behaviours.  Limited experiments would provide data to answer these questions in relatively short time frames thus allowing governments to implement strategy based on evidence quickly.

“There is a difference between testing a particular strategy directly and depending on either theory or evidence from other contexts to back political strategy.  When no direct evidence in favour of a strategy is available, the only alternative is to fall back on theory and evidence from similar situations.  While strategies based on indirect evidence can still be claimed to be science-led, it is more likely that counter evidence can be found that would support an alternative strategy.  It is therefore harder to use indirect evidence to support the strong and consistent messaging that is needed to convince the majority of the public that an initiative should be followed.”


Dr Julian Tang, Honorary Associate Professor in Respiratory Sciences, University of Leicester, said:

“This is essentially a qualitative (or focused) review of the literature looking at how different interventions may or may not have led to more risky behaviour, including the use of helmets in cycling and skiing, pre-exposure prophylaxis and circumcision to reduce the transmission of HIV, and HPV (human papillomavirus) vaccination to reduce the risks of cervical cancer.  No evidence of enhanced associated risky behaviour was found by the authors of these particular studies.

“The authors of this article then assess 22 reviews on the effects of mask-wearing, and focus on six experimental studies – none of which showed any evidence of reduced hand-washing during the study period.  Indeed, two of the studies were highlighted as showing an increase in hand-washing behaviour along with the wearing of masks.

“In fact, within the mask-wearing context, the authors point out that one precautionary behaviour (like mask-wearing or hand-washing) may actually remind the wearer about the need to perform the other related behaviours – thus reinforcing the whole suite of behaviours that combine to reduce transmission.

“Most importantly, the authors conclude that the such unproven statements on increased risk compensation behaviour can be used to argue against potentially beneficial public health interventions – such as can be seen in this European CDC document1 which includes “The use of face masks may provide a false sense of security leading to suboptimal physical distancing, poor respiratory etiquette and hand hygiene – and even not staying at home when ill.”

“Whilst the authors accept that individual variation is possible, the overall impact of mask-wearing to the population is more likely to be more beneficial than not.  Given the benefits already seen with universal mask wearing earlier in Southeast Asian countries that have experienced SARS 20032,3 which is now mandated globally, it is important for policy-makers to identify such unproven risk compensation behaviour ‘myths’ (and those who present them with loud voices in the absence of evidence) before they perpetuate for too long – thus depriving the population of potentially beneficial interventions that could reduce SARS-CoV-2 infection and therefore COVID-19 deaths in their populations.

“This is an important article to raise awareness of this ‘myth’ of risky behaviour compensation for mask-wearing – an intervention which, if mandated earlier, may have helped reduce the overall number of SARS-CoV-2 infections and therefore COVID-19-related deaths.  (Some preprints have estimated the numbers of lives which may have been saved in the US if universal masking was practiced earlier4, and the effectiveness of hair-stylist/client-masking to stop SARS-CoV-2 transmission5.)”







Dr Nilu Ahmed, Lecturer in Social Sciences, University of Bristol, said:

“This study addresses the important concern that face coverings will reduce existing precautionary behaviour around hand washing.  Pulling together data from over 20 studies across the world on face coverings both in the current context of Covid-19 and previous studies, the authors provide the most comprehensive and up to date evidence on the link between face coverings and hand washing.  The findings are encouraging, showing people don’t compensate for wearing masks by reducing hand washing, and in fact hand washing may increase as a result of the new behaviour.  Given the findings reported here, it is hopeful that emphasising the importance of clean hands when putting on, and before and after removing the mask will help to codify these rules as we build face coverings into daily outdoor routines.”


Dr Flavio Toxvaerd, University Lecturer at the Faculty of Economics, University of Cambridge, who specializes in the economics of infectious diseases and economic epidemiology, said:

“The study is interesting primarily for what it tells us about the effects of people’s mask wearing on their other protective behaviour.  But it’s less clear what they tell us about the aggregate effects of widespread mask wearing in the population.

“At the centre of risk compensation theory is the idea that an intervention that reduces infection risk will make people less cautious, for example by socially distancing less.  The problem with applying this theory to face masks is that masks do not per se provide much protection to the wearer of the mask.  Mask wearing primarily protects others against infection from the wearer.  If this is widely understood by people, we should not expect to see any decrease in additional protective measures such as hand washing or social distancing by mask wearers.

“For these reasons, the most important question is not whether wearing a mask will make one less cautious, but rather whether widespread mask wearing by others will make one less cautious.  The study does not provide firm evidence on this issue.

“Overall, I agree with the authors that risk compensation is theoretically unfounded, but there are alternative theories, such as behavioural disinhibition, that have similar observational outcomes.  In a recently published article (F. Toxvaerd; Rational Disinhibition and Externalities in Prevention, International Economic Review, 2019, 60(4), 1737-1755), I show that behavioural disinhibition may make everyone worse off.  Thus, there is strong theoretical backing for this possibility.  Whether it is empirically true for COVID-19 and the widespread use of face masks remains an open question.”


Dr David Strain, Senior Clinical Lecturer, University of Exeter, said:

“These findings come from a world renowned team are the result of a well conducted systematic review.  It collates all of the available evidence from multiple different sources, and includes reports from previous coronaviral outbreaks of SARS and MERS, as well as other respiratory conditions such as influenza.  The main limitation is the relative lack of information directly from COVID-19, as these data from the current pandemic are still being collated.

“The report strongly suggests that, instead of withholding a preventative tool, it should be accompanied by clear and accurate messaging that presents the cumulative benefit of multiple preventative measures.  This paper adds to the weight of evidence that the immediate and comprehensive implementation of face masks, in combination with enhanced hand hygiene and physical distancing, would display trust in the general public’s ability to act responsibly and empower citizens.

“The fear of risk compensation is well established as a reason for political inertia.  The same argument was made for motorcycle helmets, seat belts and ski or snowboarding head protection.  In each case, the benefit to the population tremendously outweighed the additional and very modest risk compensation that occurred in some individuals.  Indeed, in alpine skiing and snowboarding, wearing a helmet was generally associated with risk reduction oriented-behaviour, suggesting safety devices are both compatible with and perhaps encourage safety-oriented behaviour, reducing head injuries more than can be accounted for by the helmet alone.

“At the height of the 2009 influenza epidemic in Mexico City, it was found that mandatory mask requirements increased compliance compared to voluntary recommendations such as hand washing and social distancing.  Voluntary compliance was strongly influenced by public perception regarding the effectiveness of the recommended measures.”


Prof Robert Dingwall, Professor of Sociology, Nottingham Trent University, said:

“This appears to be a very narrow approach to the study of risk compensation, which has been widely studied by a range of academic disciplines beyond behavioural science and public health.  There has been a persistent problem in Covid-19 research of ignoring work that is not thrown up by systematic reviews limited to the traditional databases in biomedical and behavioural science.  The paper does not provide any details about how the studies reviewed were identified and selected for consideration.  Is this, in other words, an unsystematic review of systematic reviews that do not, as the authors concede, directly address the issue of risk compensation?

“As Carl Heneghan and colleagues at the Centre for Evidence-Based Medicine in Oxford have pointed out, there is a lack of rigorous research on face covering in the community and few studies that are taking advantage of the pandemic to remedy this.  Similarly, as Bakhit and colleagues from Australia have noted in a preprint, few of the observational studies have had the impartiality that one would normally expect between harms and benefits: in other interventions, such as new drugs, investigators look as carefully for adverse effects as they do for beneficial ones.  This has not been the case with the literature on masks and face coverings.  The lack of evidence for risk compensation may then simply mean that no-one has seriously looked for it.

“The authors do us all a service by bringing together a body of work in a convenient form.  Merely piling up fragile evidence does not, though, necessarily make it any stronger.  It certainly seems premature to claim that risk compensation theory has been ‘laid to rest’ or that the concern should not be taken seriously in better-designed and more balanced studies of this intervention.”



‘Is risk compensation threatening public health in the covid-19 pandemic?’ by Eleni Mantzari et al. was published in BMJ Analysis at 00:01 UK time on Monday 27 July 2020.

DOI: 10.1136/bmj.m2913


All our previous output on this subject can be seen at this weblink:


Declared interests

None received.

in this section

filter RoundUps by year

search by tag