Replies to post #3478 on Repros Therapeutics Inc (RPRX)

[corpstrat]

Anemia prob. Now we're talking, exwannabe! Thanks. Good to break it down into component risks like this. Looks a bit pessimistic, to this layperson, on the efficacy, chance of FDA acceptance and unknown unknowns (Rummy, not Cheney, as it happens - the one with brains). Safety might be more of a question, since FDA is worried about chronic use and a lot more data will have to be collected than has been reported so far. By showing the several things that have smallish chances of going wrong, you end up cumulatively with a higher handicap than my gut had given this. Good exercise.

It'll be interesting to see if any others chime in.

So we'll know the efficacy and ex-US answers by late this year. Not the rest? Would we be able to clear up the ex-US risk by someone asking JP rather than speculating?

Of course, valuing Proellex requires prob of the two big indications, too. Anyone who holds or decides not to hold the stock is making an implicit judgment on this. Prob went up a little bit with this IND, but the market's broken.

[io_io]

<Efficiency (sic). Let's guess that 90% is the true power. That yeilds .9*9 = .81>

The power is based upon an assumed advantage, but for the same drug and same trial, the true power (not of achieving the target advantage, but of achieving statistical significance) is always higher, bank on it. Therefore figure maybe 0.95*0.95 = 0.90 approx.

<Safety. This seams fairly clear. I will give it a .9>

It's barely life-threatening ? I would figure 0.75.

<The ex US arm is screwed up. .9 >

I think 0.99 because this only happens in cancer trials, and it's due to doctors ignoring protocols to help certain patients. Well why I dont know, but does anyone have evidence of a foreign screw-up* in something non-life-threatening ?

* = I just remember I think it was the Croatians who screwed up a PTIE pain drug trial or something ? And this was down to a language issue ? Well times have improved, and we can even speak Bulgarian now.

[DewDiligence]

Re: Probability of success in anemia

There are three main ways that clinical trials such as these can fail:

1. The true efficacy of the drug is less than what was observed in phase-2. In general, this is more likely to happen than many investors realize. The reason is program-survival bias — in the aggregate, the efficacy seen in phase-2 trials of programs that are advanced to phase-3 is biased high relative to the true efficacy of the drugs in question. (This is trivially because drugs which perform badly in phase-2 generally don’t get advanced to phase-3.)

2. An excessive number of dropouts and protocol violations occurs. The statistical power of a trial, reported to investors as 85% in the case of RPRX’s anemia trials, is highly sensitive to modeled assumptions about dropouts and protocol violations; consequently, the reported power (85% in this case) is not especially illuminating without knowing how dropouts and protocol violations have been modeled. Other posters have commented on the heightened risk of protocol violations at trial sites in Third World countries. Offsetting this to some degree is a relatively low risk of excessive dropouts in the anemia trials compared to a typical phase-3 trial, IMO.

3. A bizarre but serious side effect shows up. The side effect need not even be plausibly related to the investigational drug to wreak havoc. A case in point is CEPH’s phase-3 trial for Sparlon; the FDA rejected the Sparlon NDA because of a single case of Stevens-Johnson Syndrome even though there is no plausible link between Sparlon and SJS and CEPH did not even agree that SJS had been observed in the trial.