Masks. Should we wear them? Is the science robust in their efficacy?
Recently, I took it upon myself to thoroughly critique (elsewhere) a video that was doing the rounds:
Some of you will love this video because it will support your conclusions with lots of data and a calm voice seemingly of reason. But it hides a heck of a lot of methodological bias and issue (though there are some good points in it too). I won’t bore you with my criticisms.
One of the core issues I had concerned cherry-picking of data and this happened most with his reliance on Xiao et al for mask-wearing efficacy. A rabidly conservative commenter here with absolutely no interest in checking his sources, not least to see if they are correct or if there are better, more recent ones out there, posted this as a mask gotcha the other day on Bert’s article:
Although mechanistic studies support the potential effect of hand hygiene or face masks, evidence from 14 randomized controlled trials of these measures did not support a substantial effect on transmission of laboratory-confirmed influenza. [His emphases]
Xiao et al, the study he quoted from, concluded that masks had no substantial effect on transmission in influenza.
I am not about to write a huge post here about masks. what I will do is express the need to be better if you want to comment here or if you want to be slightly more informed about…anything.
I may be wrong in my conclusions, but I would advocate mask wearing for a couple of reasons:
- The top scientists and advisors in pretty much every country in the world advise this. On a probability analysis, this would be staggering if there was absolutely no data to support this.
- The Xiao et al study has been superseded by a larger, more recent meta-analysis: “Efficacy of face mask in preventing respiratory virus transmission: A systematic review and meta-analysis” (Liang et al). The Royal Society in the UK also has a good analysis. The most recent corroborating study came out only the other day.
It is on this second point that I will briefly dwell.
What the commenter didn’t quote you from the Xiao paper itself was:
…most studies were underpowered because of limited sample size, and some studies also reported suboptimal adherence in the face mask group.
Indeed, the study did not distinguish estimates by the type of mask but did examine masks in combination with hand hygiene. Let me emphasise – there was no distinguishing between types of mask.
The more recent meta-analysis (Liang et al), on the other hand, concluded (with the Xiao et al study included in their own analysis):
A total of 21 studies met our inclusion criteria. Meta-analyses suggest that mask use provided a significant protective effect (OR = 0.35 and 95% CI = 0.24–0.51). Use of masks by healthcare workers (HCWs) and non-healthcare workers (Non-HCWs) can reduce the risk of respiratory virus infection by 80% (OR = 0.20, 95% CI = 0.11–0.37) and 47% (OR = 0.53, 95% CI = 0.36–0.79). The protective effect of wearing masks in Asia (OR = 0.31) appeared to be higher than that of Western countries (OR = 0.45). Masks had a protective effect against influenza viruses (OR = 0.55), SARS (OR = 0.26), and SARS-CoV-2 (OR = 0.04). In the subgroups based on different study designs, protective effects of wearing mask were significant in cluster randomized trials and observational studies.
This study adds additional evidence of the enhanced protective value of masks, we stress that the use masks serve as an adjunctive method regarding the COVID-19 outbreak.
Indeed, the introduction admitted that the data appears contradictory, not least when looking at the Xiao et al paper:
Facemasks are recommended for diseases transmitted through droplets and respirators for respiratory aerosols, yet recommendations and terminology vary between guidelines. The concepts of droplet and airborne transmission that are entrenched in clinical practice recently are more complex than previously thought. The concern is now increasing in the face of the Coronavirus Disease 2019 (COVID-19) pandemic . The spread of respiratory viral infections (RVIs) occurs primarily through contact and droplet routes. And new evidence suggests severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) can remain viable and infectious in aerosols for hours . Therefore, the use of masks as appropriate personal protective equipment (PPE) is often considered when preventing the spread of respiratory infections. Experimental data shows that the micropores of mask block dust particles or pathogens that are larger than the size of micropores . For example, the micropores of N95 masks materials are only 8 μm in diameter, which can effectively prevent the penetration of virions [4,5].
Although the aforementioned studies support the potential beneficial effect of masks, the substantial impact of masks on the spread of laboratory-diagnosed respiratory viruses remains controversial . Smith et al. indicated that there were insufficient data to determine definitively whether N95 masks are superior to surgical masks in protecting healthcare workers (HCWs) against transmissible acute respiratory infections in clinical settings . Another meta-analysis suggested that facemask provides a non-significant protective effect (OR = 0.53, 95% CI 0.16–1.71, I 2 = 48%) against the 2009 influenza pandemic . Xiao et al. concluded that masks did not support a substantial effect on the transmission of influenza from 7 studies . On the contrary, Jefferson et al. suggested that wearing masks significantly decreased the spread of SARS (OR = 0.32; 95% CI 0.25–0.40; I2 = 58.4%) . Up to date, existing evidence on the effectiveness of the use of masks to prevent respiratory viral transmission contradicts each other.
Therefore, we performed a systematic review and meta-analysis to evaluate the effectiveness of the use of masks to prevent laboratory-confirmed respiratory virus transmission.
A little bit of discussion was had concerning Xiao et al:
4.3. Protective effects against influenza, SARS, and COVID-19
The risk of influenza, SARS, and COVID-19 infection were reduced by 45%, 74%, and 96% by wearing masks, respectively, which were consistent with previous meta-analyses during the SARS outbreaks [9,53]. The previous systematic review from Xiao et al., though reporting non-significant protection of masks against the influenza virus , did not strictly follow the PRISMA statement and represented merely non-aggregated data. For example, there was one study  included by Xiao et al. did not report a significantly protection by wearing masks. However, it should be noted that the result of this study was not convincing because the H1N1 pandemic broke out during the study period, and the national hygiene campaign implemented at this time influenced all participants to wear masks .
To conclude: you have to work harder to be more accurate. Be more accurate, put in some time or don’t say anything at all. (Also, I am happy to be wrong on my own claims if more recent and better data shows me to be wrong.)
Stay in touch! Like A Tippling Philosopher on Facebook: