Careful approach to single studies needed (mask effectiveness)

I've too often seen (and might have done it myself) seizing on a single study that appears to prove something. Case in point:

Study Comparing Surgical and N95 Masks Sparks Concern

• the fit-testing protocol wasn't defined
• results may not apply to other countries because of differences in treatment effects.
• wide confidence intervals indicating a high degree of uncertainty,
• differences in self-reported adherence
• baseline SARS-CoV-2 antibody status
• between-country differences in vaccination coverage
• dominant circulating variants
• HCWs in Canada were allowed to make their own decisions about whether to don a mask or N95, regardless of which intervention they were assigned to.
• 81% of N95 users reported using them all the time.
• Most HCWs not infected in hospital
  About a quarter of participants reported never caring for COVID patients yet were still included in the analysis
  relied on self-reported household and community exposures, which he said are unreliable because many infections go unrecognized.
  Study participants were required to get tested for infection only if they developed symptoms
• They use a figure of '>50% efficacy' for excluding vaccines, but do not present the estimates of efficacy for Sinopharm and other vaccines, or any information on how they calculate this,
• A control group was also needed,

Some of the bullet points come from here: Study on masks vs N95 respirators for health workers spurs concerns | CIDRAP