Onpode
Cover art for Scientists make breakthroughs in the lab—but getting them to patients remains the hard part

Scientists make breakthroughs in the lab—but getting them to patients remains the hard part

November 20, 2026 · 8 min

Zara Reyes & Megan Skiendel

Long week — I need this conversation, honestly. Same — but wait, I want to start with an image I cannot get out of my head since I went down this rabbit hole. Go. Picture a kitchen full of genuinely extraordinary chefs — Michelin-level people — perfecting recipes for years. And then none of the…

Dr Sam Barrell CBE is the Chief Executive Officer of LifeArc, a UK-based medical research charity focused on translating scientific discoveries into patient benefit.

0:007:41
Get the next episode on Science →

Follow it free — new episodes land in your feed.

Or make your own — any topic, in minutes

More Onpode episodes on Science →

About this episode

Dr Sam Barrell CBE is the Chief Executive Officer of LifeArc, a UK-based medical research charity focused on translating scientific discoveries into patient benefit.

Grounded in 7 sources
Lost in translation: the valley of death across preclinical and clinical divide – identification of problems and overcoming obstacles | Translational Medicine Communications | Springer Nature Link ↗ · link.springer.com
Gaps and paths forward in cancer pharmacology and translational research ↗ · pmc.ncbi.nlm.nih.gov
The Challenge of Balancing Scientific Discovery and Translation ↗ · ahajournals.org
Lost in Translation: Why Most Biomedical Research Never Reaches Patients | Cell Guidance Systems ↗ · cellgs.com
Bridging the gap between academic discovery and commercial translation | Drug Discovery News ↗ · drugdiscoverynews.com
New £6m fellowship programme to help rare disease discoveries reach patients faster  - LifeArc ↗ · lifearc.org
New boost for personalised rare disease treatments as the Rare Therapies Launch Pad becomes part of LifeArc - LifeArc ↗ · lifearc.org
Read transcript

Megan Skiendel: Long week — I need this conversation, honestly.

Zara Reyes: Same — but wait, I want to start with an image I cannot get out of my head since I went down this rabbit hole.

Megan Skiendel: Go.

Zara Reyes: Picture a kitchen full of genuinely extraordinary chefs — Michelin-level people — perfecting recipes for years. And then none of the food ever reaches a table. Like, the cooking is extraordinary. The eating never happens.

Megan Skiendel: That is — look, that is almost exactly how Dr Sam Barrell, who runs LifeArc, frames the translation gap in her News-Medical LifeSciences interview. And it's not poetry for her — she's describing a structural reality. Ninety percent failure rate in clinical development, with roughly half of that because the drugs just don't work in patients. Not unsafe. Not unmanufacturable. Just — ineffective in an actual human body.

Zara Reyes: No but — who is Barrell, specifically? Like why does this interview land differently than the hundred other people saying 'we need to fix drug development'?

Megan Skiendel: Because she's not selling a platform or a fund raise. LifeArc is a charity — UK-backed, rare disease focus — and she's the CEO making a pretty blunt argument: success in drug discovery cannot be measured by promising targets or datasets. It has to be measured by whether it actually improves patients' lives. That's the whole thing. That's her line.

Zara Reyes: So she's basically saying the metrics the sector brags about are the wrong metrics entirely.

Megan Skiendel: And that's exactly where it gets uncomfortable — because wrong metrics aren't just a PR problem. They're baked into how funding flows. Academic labs get rewarded for papers, for novelty, for a compound that looks exciting in a dish. Not for whether it survives contact with a human body. That's the actual mechanism underneath the 90% number.

Zara Reyes: Wait — so you're saying the selection process itself is broken? Like, the drugs that make it into trials are the ones that looked good on the wrong scorecard?

Megan Skiendel: Exactly that. And the 40–50% efficacy failure figure is the proof. Those aren't drugs that turned toxic. They're drugs that passed every preclinical gate and then — nothing. Because the preclinical models weren't capturing real human biology. That's a selection failure, not a science failure.

Zara Reyes: No but that distinction matters so much — like, if it were a safety problem you'd fix the chemistry. This is almost worse because it means the entire pipeline logic is off.

Megan Skiendel: And honestly, Hassan Ali at the Miller School of Medicine put it cleanly — good biology gets a program through the door, but only a credible development path seals a partnership. Which means the biology was never the bottleneck. The planning was.

Zara Reyes: Hold on — credible development path. That's the part Barrell is essentially arguing nobody builds at the discovery stage.

Megan Skiendel: Right — and that's what makes the ELRIG Drug Discovery 2026 keynote a moment worth watching. Because Barrell bringing this argument to that room, in front of that industry, isn't just a talk. It's a challenge to how the whole sector accounts for its own work.

Zara Reyes: So the valley of death isn't a gap between good scientists and slow regulators. It's a motivation problem dressed up as a methodology problem — and nobody wanted to say it out loud.

Megan Skiendel: And that motivation problem is exactly why the structural response has to be — I mean, it can't just be more funding for the same pipeline. LifeArc's actual answer is to fund accountability infrastructure. The £30 million collaboration with the UK Dementia Research Institute isn't 'here's money for more dementia science.' It's 'convert what we already know into tests and treatments.' The output is the accountability.

Zara Reyes: So the money is targeted at the translation step specifically — not the discovery.

Megan Skiendel: Right — and the Medicines Discovery Catapult piece runs parallel. That's biomarker work, specifically to demonstrate drug efficacy and safety in human trials before you're deep in Phase 2 and burning everything. You're essentially building the receipts earlier.

Zara Reyes: Lowkey that's a different thesis than 'we need better science.' That's 'we need better evidence that the science is real.'

Megan Skiendel: And then there's the PAD programme — Personalised and Adaptive Disease — which is Barrell pushing into tailored disease approaches. Plus the fellowships, which are designed to, actually the framing is specifically to help accelerate discoveries to patients faster. Not to produce more discoveries. Faster movement of the ones that exist.

Zara Reyes: Wait — so fellowships as a velocity mechanism, not a talent pipeline. That's a different reason to fund them.

Megan Skiendel: And Barrell's layering in adaptive regulatory pathways on top — frameworks built for small or tailored patient populations. Paediatric datasets, end-to-end trial frameworks. It's architecturally coherent. I'll give it that. But — and this is the honest caveat — neither the Dementia Research Institute deal nor the Medicines Discovery Catapult collaboration has confirmed clinical outputs yet. We don't know if any of this has actually moved a treatment.

Zara Reyes: So we have a structurally different model with no proof of delivery. That's the tension.

Megan Skiendel: And the part that makes this harder to hold — we haven't even gotten to what AI tools are doing to this picture, which frankly may be optimizing entry into the valley faster without fixing the crossing. It's genuinely uncomfortable.

Zara Reyes: And that's the irony that actually breaks my brain a little — AI and generative models come in positioned as the fix for exactly that translation problem, right? Faster screening, more candidates, compress the timeline. But if the selection methods were already missing real-world drug behavior, we've just built... a faster conveyor belt into the same broken zone.

Megan Skiendel: Honestly — that's the uncomfortable version nobody's putting in the press release.

Zara Reyes: So picture this: an oncology nurse in a hospital ward, and she's administering a drug that cleared every AI-optimized screen. Perfect pharmacokinetics on paper. And then it starts metabolizing unpredictably — in patients eating a completely normal diet. Food interactions. The exact variable no computational model had weighted.

Megan Skiendel: And that's not a hypothetical — that's the pharmacokinetic failure class that standard computational selection methods are already documented to overlook. AI doesn't fix that. It amplifies the volume of candidates reaching the gap.

Zara Reyes: More candidates, same gap.

Megan Skiendel: And look — Barrell's patient-centered design, the PAD programme, the fellowships, all of it — that work happens at discovery. Which is real. But regulatory pathways, payer decisions, actual health system adoption? Those sit downstream. Largely outside what LifeArc controls.

Zara Reyes: Wait — so you can embed patient relevance perfectly at the discovery stage and still hit a wall at NICE or at a payer board that just won't fund a small-population therapy.

Megan Skiendel: Adaptive regulatory pathways help at the margins — honestly, they do — but they're frameworks for small populations who already have fewer options. They don't move the payer calculus for a drug that's competing with a cheap generic. That's a different fight entirely.

Zara Reyes: So the question isn't whether Barrell's approach is right — it's whether it's sufficient. And AI speeds up the part that was never actually the bottleneck.

Megan Skiendel: And that's — honestly, that's the image stuck with me. A lab somewhere, today, right now. Someone has just named a promising compound. The patient-relevance checklist is filled in. The biomarker plan is filed with Medicines Discovery Catapult. The adaptive regulatory pathway is in the folder. And somewhere downstream, a payer committee is meeting to decide whether a small-population therapy pencils out — and nobody from that lab has a seat at that table. That's the gap Barrell is trying to close from one end. She can't close it from the other.

Zara Reyes: So Barrell's bet is basically — if you change what happens in that lab, you change the odds of what happens in that room. Maybe the evidence trail is cleaner. Maybe the payer case is easier to make. Maybe. We don't actually know yet.

Megan Skiendel: We don't. And look — the chefs metaphor you opened with. The extraordinary cooking that never reaches a table. I think Barrell's work is genuinely trying to build the dining room. But the payer committee is the maître d', and they haven't been invited to the kitchen either. That part's still open.