Tweet
Hi. Some people were interested.
I just read it.
A few things I noticed:
-There is a variety of outcomes measured - swab test, antibodies, and PCR. So because different subjects are having different tests, the difference in test accuracy could contribute to differences in results.
-They did a power study for how big a group they need to study. To be able to determine the outcome they were measuring they needed 4636 subjects. 4862 completed the study so if there were a significant difference they had enough subjects to find it. They also included intention to treat analysis.
-Largely everyone wore the masks as recommended or predominantly as recommended.
-The primary outcome (Covid infection) occurred in 42 participants (1.8%) in the mask group and 53 (2.1%) in the control group. In an intention-to-treat analysis, the between-group difference was −0.3 percentage point (CI, −1.2 to 0.4 percentage point; P = 0.38) (odds ratio [OR], 0.82 [CI, 0.54 to 1.23]; P = 0.33) in favor of the mask group. When this analysis was repeated with multiple imputation for missing data due to loss to follow-up, it yielded similar results (OR, 0.81 [CI, 0.53 to 1.23]; P = 0.32).
What the statistics mean is as follows:
CI means 95% confidence interval. This means statistically there is a 90% chance that when you eliminate noise from the statistics the actual value is between this and that. So you might say 1% of mask wearers got it at 2% of non-mask wearers got it. Wow. So masks work. But when you try to filter out the statistic noise, the CI for masks might be 0-2 and for non-masks 1-3%, so as the intervals or confidence bars overlap, it means there is really no real difference between the groups.
In this case with the CI, they conclude that masks either help make it better or actually make it worse - despite a large study, they find no statistically significant difference between the groups.
The OR is the odds ratio. And odds ratio of 1 means the intervention results in no difference between the group. An odds ratio < 1 means the intervention results in decreased incidence and > 1 means greater incidence. Eg OR 0.5 means 50% less and 2.0 means double.
The OR measured is 0.82 which means the masks are protective. However the CI for the OR is 0.54-1.23 which means that they cannot say statistically it is protective. Rather they can say the OR might be as low as 0.54 (it's really protective) but could be as high as 1.23 (it's really harmful).
The long and short of it is, they got a large group of people, large enough to detect a difference between the groups, but could detect no statistically significant protective effect for people in the mask group.
-The main limitation of the study is that some results come from people using home antibody tests and they report if they are positive or negative. Doesn't seem like a big deal to me but that is what some who believe in masks describe as a flaw.
Seems good to me, but I didn't believe in masks before.
I just read it.
A few things I noticed:
-There is a variety of outcomes measured - swab test, antibodies, and PCR. So because different subjects are having different tests, the difference in test accuracy could contribute to differences in results.
-They did a power study for how big a group they need to study. To be able to determine the outcome they were measuring they needed 4636 subjects. 4862 completed the study so if there were a significant difference they had enough subjects to find it. They also included intention to treat analysis.
-Largely everyone wore the masks as recommended or predominantly as recommended.
-The primary outcome (Covid infection) occurred in 42 participants (1.8%) in the mask group and 53 (2.1%) in the control group. In an intention-to-treat analysis, the between-group difference was −0.3 percentage point (CI, −1.2 to 0.4 percentage point; P = 0.38) (odds ratio [OR], 0.82 [CI, 0.54 to 1.23]; P = 0.33) in favor of the mask group. When this analysis was repeated with multiple imputation for missing data due to loss to follow-up, it yielded similar results (OR, 0.81 [CI, 0.53 to 1.23]; P = 0.32).
What the statistics mean is as follows:
CI means 95% confidence interval. This means statistically there is a 90% chance that when you eliminate noise from the statistics the actual value is between this and that. So you might say 1% of mask wearers got it at 2% of non-mask wearers got it. Wow. So masks work. But when you try to filter out the statistic noise, the CI for masks might be 0-2 and for non-masks 1-3%, so as the intervals or confidence bars overlap, it means there is really no real difference between the groups.
In this case with the CI, they conclude that masks either help make it better or actually make it worse - despite a large study, they find no statistically significant difference between the groups.
The OR is the odds ratio. And odds ratio of 1 means the intervention results in no difference between the group. An odds ratio < 1 means the intervention results in decreased incidence and > 1 means greater incidence. Eg OR 0.5 means 50% less and 2.0 means double.
The OR measured is 0.82 which means the masks are protective. However the CI for the OR is 0.54-1.23 which means that they cannot say statistically it is protective. Rather they can say the OR might be as low as 0.54 (it's really protective) but could be as high as 1.23 (it's really harmful).
The long and short of it is, they got a large group of people, large enough to detect a difference between the groups, but could detect no statistically significant protective effect for people in the mask group.
-The main limitation of the study is that some results come from people using home antibody tests and they report if they are positive or negative. Doesn't seem like a big deal to me but that is what some who believe in masks describe as a flaw.
Seems good to me, but I didn't believe in masks before.
Comment