Great question. I hope that I can shed some light.
The one-sided P-value is a result of the one-sided t-test, which basically means that the study was looking at the relationship in one direction - the direction of HER-Vaxx and Chemo. For example, is the survival in the HER-Vaxx and chemo group significantly different better or worse than the standard chemo treatment. A one-sided t-test provides more power to detect an effect because it is not concerned with the other direction, whether standard chemo is significantly better or worse than HER-Vaxx.
The p-value of 0.1 means that 90% of the measures gathered are in favor of HER-Vax + chemo being better than standard chemo. I had previously questioned why they have used a p-value of 0.1 instead of 0.05 and have made my own opinions about why this was done. As mentioned, because the study has used a one-sided t-test to determine the relationship between the two cohorts, it strengthens the relationship observed requiring a larger P-value to show efficacy.
A very important point to make is that it is very common for studies with a small cohort of people (like this one) to show no statistical significance (the fact that this has done the opposite adds
significant strength for the relationship observed). A P-value of 0.05 would have indicated that there was no significant difference between HER-Vax + chemo and chemo on it's own, however, the difference observed between the groups (90% of measures showing better outcomes than standard chemo) is clinically significant due to the efficacy, safety, and tolerability. I also assume that they have allowed more leniency with P-values, as these are interim results and have not yet reached their full data maturity. Let's say that this study stretches on for another 12, 18, or 24 months, you would expect to see that the statistical and clinical significance improves over time as the HER-Vax + chemo group performs better than the chemo group. Ultimately, this would mean that a P-value of 0.05 would be appropriate as the full data set may be used for analysis. Had this been a study of 270 people, the data would be
groundbreaking. In my opinion, it is important to not be discouraged by the P-value of 0.1.
All IMO
(20min delay)