Significance boundary for total deaths from 82 to

Significance boundary for total deaths from 82 to 91

We don’t know whether post-28-day deaths are included in the 87 (and/or the mid-term 45) number, so I did a table. For a range, I added 4 for the last week, while studies indicate 10% is a reasonable fraction to be outside 28 days.

I calculated the minimum deaths in the control group to attain significance at each total death number. This ranges from 35 (26.7%) to 38 (29.0%), quite low versus other control groups discussed here.

Relative risk at the boundary ranges from 0.671 to 0.697, or a 30.7% to 32.9% relative reduction.

Absolute death reduction at the boundary ranges from 8.8 to 9.5 lives saved per hundred treated, reflecting a number needed to treat to save a life of 10.5 to 11.4.


Total Control Treatment P(2 sided)
. 82 . . 35/131 . . 47/262 . . .0421
. 85 . . 36/131 . . 48/262 . . .0447
. 88 . . 37/131 . . 51/262 . . .0472
. 91 . . 38/131 . . 53/262 . . .0497

I’ve switched to MedCalc’s relative risk calculator - it seems about .01 more significant than the difference on proportions test, but it’s the test others are using. https://www.medcalc.org/calc/relative_risk.ph

Why am I posting this? We have a long weekend. How about everyone make their final guess on deaths in each group and we'll see who nails it best? Is the award called the Ohmie?