Proprietary Algorithm to Predict Likelihood of FDA Approval
Hello everyone,
I founded a company that is developing a proprietary algorithm that predicts likelihood of FDA approval based on public clinical data and other factors. We're interviewing hundreds of physicians, biostatisticians and former regulatory professionals to make the algorithm as precise as possible and the results we're getting are very encouraging.
I was wondering how valuable would the probabilities of FDA approval generated by this algorithms be to hedge funds and other institutional investors?
I am thinking of setting it up as subscription service where subscribers would receive on a weekly basis the probabilities generated by the algorithm for all upcoming FDA decisions for a given time range.
How much would HFs be willing to pay (after a 1-month free trial) for this subscription service?
Thanks!
1) the more people who have it, the less useful it is.
2) you will not be talking to doctors about specific approvals right? Just to build the algo? Want to make sure you don’t create compliance risk for funds.
3) given the probabilities and moves on these outcomes, funds will want to see a strong track record.
Personally, I’m a bit skeptical of the product having commercial success given the number of moving parts associated with getting a drug through clinical trials. For example, how would you factor in the FDA randomnly shifting the goal post on what they consider for bar of approval? How would you account for trials that may not have hit statistical signifiance but could get approved based on FDA’s leniency? How would you account for companies massaging open-label data that would inevitably miss in a placebo-controlled trial? There’s a long list of occurences that are more common that not that would be counterintuitive to the generalist investor or everyday person, which could be tough to assess.
I would say most of the biotech/healthcare-dedicated funds would have no interest in this type of technology because they simply wouldn’t trust it and instead have their own PhDs, statisticians, and doctors to calculate the PoS. I second the point above that there’s a reason these funds have dedicated KOLs and scientific advisory boards, typically under exclusivity. I also think you underestimate how subjective healthcare investing is - there are many drugs that had subpar data, yet got approved through different means (i.e. patient advocacy groups); “75% PoS” on its own means little without the context, so I think your best bet would be someone more quant-oriented as opposed to your fundamental investors.
I once heard a lecture by the former CEO of Quantopian who said someone asks a variant of your question ("How much would a hedge fund pay for my algo that predicts XYZ?") every week and he always says the answer is nothing, nothing, nothing at all!!!!
First, nobody would buy your model unless you can prove it works, and it's very hard to prove it works without giving them the source code (and still hard to prove even then). You mention a 1-month free trial, but there aren't very many interesting FDA approvals in a given month, so a one month trial won't convince anybody.
But more importantly -- even if you can prove your algorithm really does work with, say, 99% accuracy, ***that alone is not enough to justify anyone paying for it*** cause you haven't yet proven this leads to more profits than I'm already getting on my own. Hedge funds are already making their own predictions of this and they've gotten good at it. Their predictions are already X% accurate. Your model may be, say, 99% accurate; but do you know the accuracy of what funds are already doing without your model? Are you sure their manual process isn't already nearly as good as your model? Cause if it is, then why would they need your model? You need to show not only that you are good, but also that you are significantly better than what they already have. And that is much much harder.
Tempore placeat ex cumque dolorem veniam praesentium. Praesentium qui qui in. Nulla rerum voluptatem earum cupiditate dolorem eos aspernatur laborum. Saepe ipsa ipsa rerum quia.
Ipsam delectus ut eum accusamus. Accusantium perspiciatis eligendi sunt et aut itaque tenetur.
Magnam fugiat ut occaecati dolorem suscipit odio qui. Adipisci aut voluptatem architecto.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...