People watching pharma companies and the drug development process (this includes investors/VCs, academia and the general public) from the outside will seldom think about this unmet need.
If you’re a scientist, you don’t participate in the upside but you participate a lot in the downside in the case a drug fails to make it past trials and onto the market.. This isn’t exactly a secret (see interview: Vivek Ramaswamy, Roivant Sciences CEO), but is something casual biology enthusiasts or scientists in academia may not appreciate. Simply put, the pharmaceutical company’s management (CEO, CSO, CMO, etc.) receive the large stock grants and subsequent payouts in the event of success, while scientists generally do not.
Because of this, if you’re a scientist, you may be incentivized to do stuff you’re not “supposed to” in order to preserve job security. Generally, halfway through the experimental process, you can get a good idea of where the study is going. Thus, if you have less faith in a compound you’re involved in testing, you may design experiments that don’t quite get you the answer, but postpone getting the answer by a few months or years, so that you can keep your job.
If you think about the inner politics of a drug company, this is a totally rigged system. However, there are few insights that can be gleaned from scenarios like this:
drug companies should have a way to be more confident about a candidate or lead before proceeding with experiments and human trials
- On this front, uprooting the pharmaceutical R&D process makes sense. Leveraging AI/analytics is one solution that is worth betting on at this point, and there are already companies going after this (Schroedinger, Compugen, Recursion Pharma,…)
drugs that don’t work for one indication can be re-purposed for another
- All sorts of drugs can be repurposed. One of the best examples is in antibiotics, a drug class that has seen almost no innovation in recent years. However, existing antibiotics can be mixed and matched with each other into cocktails, addressing potential antibiotic resistance issues.
- In the commercial landscape, companies like Roivant specialize in things like this. Finding new indications for distressed assets can add multiples to a parent company’s bottom line especially if the distressed asset is panic sold for cheap. This is a contrarian move, but an important thing to consider is that even if a drug at a big pharma company fails, there ALWAYS is that one person that strongly believes in it and is willing to give it another shot to push it to the finish line, even if it’s for another indication.
how do you prevent these scientists from rigging the system to “preserve their job security” (among other things)?
- Interesting forms of manipulation happen all over the place. Here’s a recent example of where the founders of a company found out the clinical trial failed, and subsequently changed the name of the drug and tried to pretend it didn’t happen! Not to mention the countless other times where clinical trial researchers find excuses to exclude patients (“this patient didn’t respond to the drug, but she showed up 1 minute late to her appointment…let’s take her out of the study!”)
If you look at a market map of tech and AI in the healthcare space, you’d notice that most of the companies are working on the payer and provider side. Perhaps it’s because the time to ROI is quicker considering average reimbursement cycles. Perhaps it’s sexier for entrepreneurs to tell people “we’re helping millions of Americans lose weight and fend off diabetes with our mobile app!”.
How to uproot the pharmaceutical R&D process with disruptive technology
It’s no secret that tech and/or AI can really uproot the process for the better. Bessemer Venture Partners recently wrote a thought-provoking piece on this.
In my mind, the best way to think about this is to think about a (generic) value chain:
- candidate discovery
- lead generation
- lead prioritization
- animal studies
- human studies
- regulatory (FDA, etc)
- marketing, selling, distribution, profit-reaping
A startup going after any of these portions of the value chain can be successful. My prediction is that when considering exit-potential, there will be competitive large player acquisitions for target lead development. Most early candidates still in pre-clinical testing, but leveraging AI to prioritize 50 candidates from a pool of 1000 yields tremendous value to a large pharmaceutical company. It saves them 95% of the resources required to prioritize candidates via experimentation - this is a huge bottom line impact. Given a set of 50 candidates, a large pharmaceutical company has the resources to take a candidate through the subsequent parts of the value chain and put a drug on the market.