RE: ACTG Activ SNG0015 Oct 2022 12:50
Not a statistician, but the methodology is 'Fisher's exact test' which is appropriate for small sample sizes.
If we had one more patient hospitalised in the placebo group, or one fewer in the drug group, it would have tipped firmly into statistical significance with a P value of under 0.04. You could argue that the trial was undersized, but when it was designed, hospitalisation rates were much higher - same goes for our P2@home. As it is, P of 0.07 implies there is a 7% chance that the results were a fluke. As has been pointed out, if you pool the Activ2 results with our P2@home then those results would certainly be statistically significant. It is unlikely that regulators could accept pooling the results, but decision makers for further trials will be looking at all the available evidence.
I think we can reasonably say it is highly likely that the drug is highly effective in early stage use, but we haven't proved that to an acceptable standard yet. However, it would almost certainly not be economically justifiable in populations with such low rates of hospitalisation - you would be spending tens if not hundreds of thousands to prevent a single hospitalisation, and most hospitalisations are short and do not progress to ICU etc.
The hospitalisation rate and the rate of progression to severe illness or death may well change with different variants, and it ought to be possible to focus in on populations with a higher chance of hospitalisation.
My view is that there is a strong rationale for further trials for early use as well as for hospitalised patients in high risk populations.
Sign In
Follow the stocks
