Wrote this in Sweden today. As always, excuse for the quick Google translate solution:
"We do not know the absolute figures for how many people have received 30% respectively 50% pain relief after six weeks of treatment in the last cancer pain trial. We have though the results from the two 4 weeks of treatment studies.
We do not know when NP-1 will stop giving the patient no more pain relief after each added therapy weeks. We do not know the slope of the curve.
The placebo effect is a subject in itself among statisticians, and even more so when looking at the perceived pain. Here we study the same diagnosis, but with different cause (PHN, cancer or diabetes) that have given injured and trapped nerve which causes pain.
As a statistician, you should always be humble in that each placebo effect is unique. Having said that management can already do a multivariate / meta-analysis of all the three surveys to check this.
At the margin, it is all about if the placebo effect between the three surveys of the same disease would vary so strongly that the effect of each new week would only apply to neuropathic pain in cancer patients, but not, for example, in diabetic neuropathic pain. Company management can analys the data from the two previous trials to see if such an increase in effect for each week were there too. They may also perform a multivariat / meta analysis. Of course they already know the answer to the question: Do the result build up over time, after each added week, in the two first trials as well as in the cancer pain study? Talley stated on the last webcast that NP-1 have had a "breakthrough" concerning efficacy given the results in the last trial.
Neuropathic pain of cancer is about 2 million patients in the United States. Neuropathic pain in the general diagnosis includingdiabetes pain has 15 million patients. The latter broad diagnosis have 12 weeks in the FDA guidance for the design of phase III-trials. One approach that seems to fit the NP-1 cream very good?"
>As a statistician, you should always be humble in that each placebo effect is unique ...management can already do a multivariate / meta-analysis .. of course they already know the answer to the question<
Yes, these boys always know already anwers though they have no idea what the question is.You're trying to nebulize here in all length, with a ppor superficial knowlege, that these guys missed the right design