Masks in the community are an effective strategy: Author's response to Haslam et al (2020)

Abstract

Dear Editor We write in response to the letter from A Haslam (Haslam, 2020) about our paper (MacIntyre, Chughtai 2020). Haslam presents a range of arguments as to why community mask use is ineffective. We note that in doing so, Haslam gives more weight to sub-analysis data on unvalidated clinical case definitions than data on laboratory confirmed infections to support her statements. The sum of the evidence cited in our review shows that face masks are protective in the community and consistent with the findings of the individual studies. Most of the studies only examined influenza or clinical case definitions of influenza as outcomes. Influenza is a seasonal disease, and varies from year to year, with high incidence in some years and extremely low incidence in others (Lipsitch and Viboud 2009). Further, the short incubation period of influenza and the possibility of pre-symptomatic transmission make it complex to study. This makes it important to consider the findings of the RCTs in the sub-analyses such as by timing (36 hours) and compliance. Haslam suggests we have misinterpreted the data. The MacIntyre 2009 RCT (MacIntyre, Cauchemez et al. 2009) showed that adherence to mask use was associated with a significantly reduced risk of ILI-associated infection (MacIntyre, Cauchemez et al. 2009). Most community mask trials demonstrated low adherence, including this one (25-30% by day 5) (MacIntyre, Cauchemez et al. 2009) which likely explains the non-significant results in the Intention-To-Treat analysis. Adherence has been shown to be related to risk-perception, and would likely increase during the COVID-19 pandemic compared to seasonal influenza. As such analysis of adherent participants is justified and valid evidence. Cowling et al conclude that “Hand hygiene and facemasks seemed to prevent household transmission of influenza virus when implemented within 36 hours of index patient symptom onset.” This is consistent with our interpretation. We agree that the intention-to-treat analysis were not significantly different between the two intervention arms. However, Haslam is incorrect in concluding that ILI was higher (and she extrapolates form this that influenza transmission is higher) in the mask group based on single digit outcomes – these are not statistically significant differences. The statistically significant findings were on early use of interventions (within 36 hours) (Cowling, Chan et al. 2009). The authors show that for RT-PCR confirmed influenza, which is the most important and validated outcome, there were “fewer infections among participants using facemasks plus hand hygiene (adjusted odds ratio, 0.33 [95% CI, 0.13 to 0.87])”. This is shown in Table 5, and the only significant outcome. Hand hygiene alone is not significant in protecting against this validated outcome. The authors final conclusion is consistent with ours - that “our results suggest that hand hygiene and facemasks can reduce influenza virus transmission if implemented early after symptom onset in an index patient.” This probably relates to the relatively short latent period and incubation period of influenza (Cowling, Chan et al. 2009). Aeillo 2010 and 2012: The Aiello 2010 trial compared masks, masks + hand hygiene and control. The 2010 trial found that the mask only arm and the mask plus hand hygiene arm had significant protection during the later weeks of the trial (Table 4), (Aiello, Murray et al. 2010) probably reflecting a larger sample size, as recruitment continued and a there was a late influenza season that year (see discussion of the 2010 paper) (Aiello, Murray et al. 2010). The results state, “After the participant enrolment ended (ie, week 3 onward), significant reductions in ILI incidence were observed in the mask and hand hygiene group (weeks 4–6) and in the face mask–only group (weeks 3–5) compared with the control group. After covariate adjustment, ILI incidence was significantly lower among the mask and hand hygiene group compared with the control group from week 4 onward” (Aiello, Murray et al. 2010). The authors further state that the study year was a mild influenza season, which would reduce statistical power. The final conclusion of the authors is consistent with ours - that masks and handwashing were protective. They also state: “It is important to note that handwashing habits were the same in both the face mask-only and control groups at baseline and over the study period, which suggests that mask use alone may provide a reduction in respiratory illnesses regardless of handwashing practices” (Aiello, Murray et al. 2010). Regarding Aiello et al., 2012, the authors conclude:“Face masks and hand hygiene combined may reduce the rate of ILI and confirmed influenza in community settings. These non-pharmaceutical measures should be recommended in crowded settings at the start of an influenza pandemic.” Table 3 shows significant reduction of ILI from weeks 3-5 of the study. The results states, “At week 3 and onward, significantly reduced ILI rates were observed in the face mask and hand hygiene group compared to the control in adjusted models (see Table 3) (Aiello, Perez et al. 2012). Larson et al., 2010: While the rates of URI were not significantly different between groups, the Larson trial showed lower rates of more specific outcomes, confirmed influenza and ILI in the Masks + hand hygiene group (Larson, Ferng et al. 2010). Regarding hand hygiene, Table 4 of the 2010 paper shows the secondary attack rate was lowest for Masks + hand hygiene. For confirmed influenza, the Mask + hand hygiene group had the lowest number of confirmed influenza cases and the lowest secondary attack rate (Table 4) (Larson, Ferng et al. 2010). There is no statistical significance of the hand hygiene finding, and the authors conclude: “Consistent with our findings, Cowling et al. found a modest but nonsignificant impact of hand hygiene on viral respiratory transmission” (Larson, Ferng et al. 2010). Table 5 in Larson et al shows that the only statistically significant protection was the Hand Sanitizer and Face Mask group (OR 0.82, 95% confidence intervals 0.70, 0.97). Hand hygiene alone was not significant - OR 1.01 95% confidence intervals 0.85 - 1.2 (Larson, Ferng et al. 2010). Suess et al., 2012. The conclusions of Suess et al (abstract) are: “When analysing only households where intervention was implemented within 36 h after symptom onset of the index case, secondary infection in the pooled M and MH groups was significantly lower compared to the control group (adjusted odds ratio 0.16, 95% CI, 0.03-0.92)” (Suess, Remschmidt et al. 2012). This is what is reflected in our Table 1 and is in agreement with the author's conclusions. Please see Table 4 of the Suess paper as well (Suess, Remschmidt et al. 2012). In summary, it remains true as stated in our review (MacIntyre, Chughtai 2020) that there are more, and larger, RCTs of mask use in the community for well people than there are for use by sick people. The conclusion of our review (that there is evidence that masks protect well people in the community) does not differ from the authors of the included papers, as shown by direct quotes from the relevant papers, above. Whilst Intention-to-treat analysis did not show significance in many of these trials, the sum of the evidence shows effectiveness of mask use in community settings when accounting for early use (reflecting the short incubation period and latent period of influenza, which is the outcome of interest in most studies) and adherence (which was low in many studies). It is important to note that the available RCT evidence on community mask use has been interpreted very inconsistently by different expert groups, with the US recommending universal face mask use in the community, and the WHO, UK and others, not recommending it (Greenhalgh, Schmid et al. 2020). During the COVID-19 pandemic, the general community are often left to fend for themselves, especially in cities or countries where the incidence is high. The use of cheap devices like masks has very low risk and possible public health benefits in slowing transmission (Greenhalgh, Schmid et al.). It is therefore important to evaluate the available evidence, allowing for factors such as adherence and timing, to inform potential use of masks in the community. Finally, a WHO-commissioned study has shown that masks reduce the risk of infection with beta-coronaviruses by 85%, and are equally protective in community and healthcare settings (Chu, Akl, et al 2020). On this basis, the WHO changed its position in June 2020 to also recommend face masks in the community. Uncited References: (MacIntyre and Chughtai, 2020, Cowling et al., 2008, Haslam et al., 2020)