PHOSITA and the Federal circuit court. The way to go. Same conclusion as Jomama
A PHOSITA (A person having ordinary skill in the art) is central in the determination of obviousness and has been a reference in many cases at the level of the Federal Circuit and even the Supreme court and was referenced by judge DU.
Judge Du didn’t apply law correctly regarding the necessary skills of the PHOSITA. Since judge Du conclusions about obviousness stems from the Mori and Kurabayashi trials and from the p value reported as a differential effect, her implicit fictitious character has to be cognizant of biostatistics. It is a prerequisite to not wrongly interpret the two trials.
Judge Du stated in her ruling:
“In light of the statistically-significant differential effects reported between the EPA and control groups, a POSA would have attributed the reduction in Apo-B to EPA.”
A PHOSITA wouldn’t have taken p value at face value! Judge Du didn’t mention all the shortcomings of the two trials (Mori Kurabayashi) which a person of ordinary biostatistical skills would have noticed and didn’t correctly address them. Her PHOSITA missed all the following points or didn’t report them correctly.
A quick recap of what I developed in my earlier post and Jomama has established:
1 Randomization not done properly
2 dissimilarity of studied populations
They are two different population. The Amarin’s patent applies to a US population not a Japanese one. I think this consideration is important by itself
3 Difference in Triglycerides at baseline: must be done with a control group
4 Population size not important enough to diminish the possibility of unknown imbalances
5 Different EPA dosing 1.8g as compared to 2.4g
6 Molecule tested not the same as in the patent since it is a combination of drugs.
7 Independent review not present with high possibility of bias
8 P value of the Kurabayashi trial: no Bonferroni adjustment to counteract the problem of multiple comparisons. "p" value for each test must be equal to alpha divided by the number of tests.
9 Meta-analysis not correctly performed by judge DU when compiling the conclusions from the Mori and Kurabayashi trials.
This list gives a count of no less than 9 missing requirements for scientific accuracy (this is a non-exhaustive list and Jesse or jomama could add a lot to this one) that a PHOSITA would have noticed. The PHOSITA would have concluded the p value was worthless to determine the effect of EPA on a US population.
Since it is up to the defendants to prove the obviousness, the PHOSITA should have addressed the possible irrelevance of these shortcomings in the determination of the outcome, which he did not. This can be considered an error. No new evidence is necessary to nullify the ruling.
There is a further proof that the repeatability of these experiments could only have been performed in hindsight (courtesy of AZN):
Let’s compare two trials Kurabayashi and Jelis as harbingers of future results to assess the certainty of their findings. The interesting point is that they are testing the same drug but in combination (blurring the interpretation of the results) and for different measurements.
Kurabayashi and Jelis
Randomization
K not performed correctly
Jelis performed properly according to age gender etc.…
Similarity of studied population
Both populations are Japanese with the possibility that the experiment cannot be repeated in a US population. Applying results
Population size
K 69
Jelis 16000 (231 times more patients!)
EPA dosing
Both trials 1.8 g daily but different from the dosing in the ensuing confirmatory trials
4 g in the Marine trial
2.4 g in the Strength trial
Molecule tested not the same since it is a combination of drugs
Combination of EPA and estradiol in the Kurabayashi trial and only EPA in the Marine trial
Only EPA in the Jelis trial but a combination of EPA and DHA in the Strength trial
Independent review
Jelis Independent monitoring committee? I think so but don’t know for sure
K no external monitoring
Endpoint
Only one primary endpoint or JELIS
Many different endpoints with no subsequent Bonferroni adjustment for the Kurabayashi trial.
The two trials had their limitations but obviously one, JELIS, had much fewer imperfections. It would be reasonable for a PHOSITA without any hindsight to consider a Jelis confirmatory rial, the STRENGTH trial to be much more likely to succeed than the Marine trial following in the steps of Kurabayashi…after all AZN did bet a few hundred million dollars expecting a positive outcome.
Since STRENGTH, the “confirmatory trial with the less shortcomings” aka the one with the greatest chances of success did fail, the prior art based on the Kurabayashi trial, the one with least expectations of repeatability, could not have provided a reasonable expectation of success.
When a trial does not proceed according to the requirements for scientific exactitude a PHOSITA with an average understanding of biostatistics knows a p value is worthess to determine the repeatability of such findings in another context.
I think an appeal based on the skills of a PHOSITA has merit its own right.
PS For instance, because of the inadequacies of the Raoult trial (marvelous p value of 0,0001) a PHOSITA cannot determine in advance chloroquine’s possibility of success, hence this raging debate. P value has to be used with caution!
AI
