thanks MC, that is an excellent point and great find to clarify our current situation. I agree with you that the only way of interpreting Kurabayashi as prior valid teaching in this case is if you have a difference between groups (as there are both taking estriol). This then weights in favor of Amarin as well in my mind
This is an excellent point and should be sent to IR and Covington. Also, there was no ss at any point in the trial except at baseline to 48 weeks. All other time intervals had no significance. Lastly, the reduction in apo B is not mentioned anywhere in the discussion or conclusion of the article. Another error by Judge DU to falsely interpret the data. Great find
Table 3 of Kurabayashi has been cropped, removing its legend that distinguishes the two different definitions of P values: - P of differences in changes over time, vs - P of intergroup differences
With the cropped table, a reader would not readily identify this incorrect statement made by Judge Du on page 30:
That is clearly erroneous in my opinion because Kurabayashi definitely does not state that there exist statistically significant differential effects reported between the EPA and control groups. On the contrary, Kurabayashi explicitly states the inverse––that there is no statistically significant difference between the EPA group and the control group. (See the parent of this post for the details.)
A few people have asked me to comment on this post. I don't have too much to add. mc1988 is exactly right-- there is no statistical change in ApoB levels reported relative to the control group. This is quite clear in Table 3 of Kurabayashi (pasted below).
It is interesting to note that Kurabayashi reports:
"The apolipoprotein B level in the eicosapentaenoic acid group was significantly lower at week 48 compared with the baseline level, but there was no significant difference between the groups."
So to be clear, Kurabayashi found NO statistically significant change at any of the timepoints between the control and treatment groups. But they did find a difference between treatment and baseline at week 48.
So, oddly, the flavor of error is similar to that made in Mori.
However, again, the correct comparison is clearly BETWEEN groups here (and is even more important, IMO). Why? Because the treatment group received E3 plus EPA and the control group received E3 only. So it is critical to compare between groups, because you have to disentangle the effects of E3 from the effects of EPA.
This error on Du's part seems a little more disingenuous than the Mori error, since Mori did not do the between group comparison (as they were not testing that hypothesis), whereas here, Kurabayashi absolutely did the comparison and reported it. It's almost as if she were cherry picking the paper to support her conclusion.
Edit: I am NOT suggesting Du was not acting with honorable intent. I find it likely that she was. What I am suggesting is that she made up her mind that the patent was invalid and was looking for reasons to support this conclusion. I am not a lawyer, but it is my understanding that the method that she went about her reasoning is important, and I think this mistake may add some color to that methodology.