Biotech Values

[DewDiligence]

Re: ALTH hypothetical trial design:

>> It seems that if you could choose a couple of likely subgroups in the context of the larger trial, you might have some additional filing expenses, but you might save the cost of additional trials. <<

If you wanted to run a single trial with two prospectively defined subgroups (lung, breast) while retaining the option of analyzing data on the full patient pool consisting of both subgroups, then the FDA would probably require you to allocate the 0.05 of p-value among the three analysis groups.

E.g., you might be able to negotiate a p-value threshold of 0.015 for each of the two subgroups and a threshold of 0.02 for the full patient pool. (.015+.015+.02 sum to the usual .05.) What you would effectively be doing with this approach would be trading off a lower expense (relative to conducting three distinct trials) for a higher statistical hurdle in each analyzed group. This strikes me as an unfavorable tradeoff, but that's JMHO and others might think the opposite.

If, on the other hand, you were not interested in data on the combined group of lung and breast patients, then you would effectively have two separate trials –one for lung and one for breast—and no special handling would be needed.