Drug Development Shouldn't Start in the Lab. It Should Start with Patients.
I started my career behind a pharmacy counter. That’s where I learned that a medicine is only as good as the person who swallows it, tolerates it, refills it, and tells their doctor whether anything actually changed.
I think about that often when I read the latest figures on clinical trial failures. Roughly nine out of ten experimental drugs that begin human testing never make it to market, and the most cited analysis of why puts the dominant reason at the top of every list: the drug didn’t work in patients. Not in cells. Not in mice. In actual people. We’ve built an entire industry that treats this 90% washout as an unavoidable cost of doing business. I don’t accept that framing anymore, and I don’t think the next generation of biotech founders should either.
The lab is the wrong place to start
The standard sequence in pharma goes something like this. A scientist identifies an interesting molecular target. They test compounds against it in a dish. They move into animal models. Eventually, after years and many millions of dollars, the candidate enters human trials, and that’s where the patient finally walks onstage. Most of the important design decisions have already been made by that point.
I’ve sat through enough portfolio reviews and pitches at Propel Bio Partners to know how often this sequencing produces medicines that look brilliant on paper and disappoint at the bedside. We chase elegant biology and discover, late, that we never asked the people we were trying to help what mattered to them. What symptoms wreck their day. What side effects they’d accept. What outcomes they’d consider worth the trade.
The FDA has been quietly pushing the industry to flip that order for more than a decade. Its Patient-Focused Drug Development guidance series lays out, in four documents, how sponsors should systematically gather patient experience data and bring it into regulatory decisions. A 2024 review in *Frontiers in Medicine* argues the framework is now baked into the development paradigm. I’d argue we’re still at the beginning of figuring out what it means in practice.
What I look for in a patient-first company
When I evaluate a startup, I want to see evidence that the founder understood the patient before they understood the molecule. That’s not a checkbox. It shows up in the kinds of questions a team asks. It shows up in how they design endpoints, who they put on their advisory boards, and how willing they are to redesign a protocol when patient feedback contradicts a beautiful hypothesis.
Two of our portfolio companies sit on my board, and I’ll point to them because they make the abstract concrete.
Persephone Biosciences ran one of the largest infant gut microbiome studies in the United States and published its initial findings in Communications Biology, a Nature portfolio journal. The team found that 76% of U.S. infants are deficient in Bifidobacterium, a foundational beneficial microbe, and a quarter have none detectable at all. The study is a study because the patients (in this case, families and their pediatricians) needed an answer that didn’t exist. Persephone built the data, then built the product.
Inherent Biosciences is doing something similar in male reproductive health. For decades, “unexplained” infertility has been a label clinicians put on couples when the existing tools failed them. The science of male epigenetics had moved on. The clinic hadn’t. Inherent’s diagnostic exists because the patient experience was already telling us the standard workup was incomplete. The team listened.
The economics actually favor this approach
There’s a comfortable story that patient-centric development is more expensive and slower than the conventional path. I think the opposite is true once you account for the failure rate.
The Tufts Center for the Study of Drug Development has spent decades estimating the average cost of bringing a new prescription medicine to market, and the numbers run into the billions when capitalized over the years a candidate spends in development. When you build that math on a 90% failure rate, every avoidable Phase 3 collapse is enormous waste. I’ve watched programs blow up in late-stage trials because the endpoint we picked, often years earlier, didn’t actually correlate with anything patients could feel. That kind of mistake gets cheaper to avoid the earlier you make it.
It also matters who you’re testing in. The FDA’s June 2024 draft guidance on Diversity Action Plans tries to address a long-standing problem, that trials often look nothing like the populations a drug is supposed to serve. The agency is requiring sponsors of many late-stage studies to lay out enrollment goals broken down by sex, age, race, and ethnicity. A STAT News report on a 2024 inspector general review found that NIH-funded trials still consistently fall short of their own diversity goals. If you build a study population that doesn’t reflect the real world, you don’t get to be surprised when the medicine doesn’t either.
What this means for founders
If you’re a scientist-entrepreneur thinking about how to design your next company, here’s the version of the patient-first thesis I find myself repeating.
Start with the unmet need before the molecule. Talk to enough patients that you can describe their experience before you describe your science. Build advisory relationships with clinicians who see those patients every day. Design your trial endpoints around what those patients would call a meaningful change in their lives, not what’s easiest to measure. Make diversity an early design choice, not a regulatory afterthought.
I won’t pretend this is the only way to build a biotech company. Plenty of important medicines have come out of pure target-driven discovery. But when I look at the failure data, the regulatory direction, and the companies in my own portfolio that are working, the pattern keeps pointing the same way. The drugs that get to patients tend to be the ones that started there.
That’s the version of this work I keep choosing, and the longer I do it, the more I’m convinced it’s the version that will outlast the louder one.
___
Leen Kawas, Ph.D., is the co-founder and managing general partner of Propel Bio Partners, a Los Angeles–based life sciences venture firm. She serves on the boards of Inherent Biosciences and Persephone Biosciences.