The Study You Skip Today Is the Clinical Hold You Face Tomorrow

Most nonclinical teams spend a lot of time thinking about which studies to run. Almost none spend enough time thinking about when. In this episode, we unpack why study sequencing is one of the most consequential — and most consistently underestimated — decisions in IND-enabling development, and walk through the three sequencing mistakes that quietly accumulate risk while looking like efficiency.

Key takeaways:

• Sequencing is an informational problem, not a scheduling problem — and confusing the two is where programs get into trouble
• Running a GLP pivotal study before dose range is genuinely established is the most common and most costly sequencing mistake
• A clean NOAEL without any adverse effects at the high dose isn't a win — it's a failure to toxicologically characterize your margins
• Parallel execution without informational overlap isn't a compressed timeline — it's two separate bets running simultaneously with limited ability to course-correct
• Regulators aren't just evaluating what your data shows — they're evaluating whether your program was designed to answer the right questions in the right order

Links:

• Data Is Not Strategy on Amazon: https://a.co/d/0cDYM8vP

The Nonclinical is hosted by Dessi McEntee, MS, DABT — board-certified toxicologist and Fractional Head of Toxicology. Subscribe to the newsletter on LinkedIn, take the course at nonclinical.academy, or work with Dessi at toxistrategy.com.

Transcript

SPEAKER_01 0:00

Imagine burning um fifty thousand dollars a day in capital. You're completely confident your new molecule is headed straight to the clinic, right?

SPEAKER_00 0:09

Oh yeah. The board is thrilled.

SPEAKER_01 0:11

Exactly. The board is thrilled, your GATT chart is just this masterpiece of efficiency, and then boom, the FDA hands down a clinical hold.

SPEAKER_00 0:19

Which is pretty much every biotech leader's worst nightmare.

SPEAKER_01 0:22

Aaron Powell It really is. And they do it because of a delayed liver signal that you well, you could have caught three months earlier for a fraction of the cost if only your non-clinical studies had been run in the correct order. Trevor Burrus Right. So today we are talking about why moving too fast in early drug development is mathematically the slowest possible way to the clinic.

SPEAKER_00 0:40

Aaron Powell Yeah, it's the ultimate paradox of the biotech industry, honestly. Because the pressure to compress timelines, I mean, it's just baked into the financial model of early stage drug development. Trevor Burrus, Jr.

SPEAKER_01 0:51

You literally have to hit milestones to survive.

SPEAKER_00 0:53

Aaron Powell Exactly. Milestones are the only way to survive it. But there's this massive difference between achieving genuine structural efficiency and simply, you know, hiding risks under the guise of going fast.

SPEAKER_01 1:07

Aaron Powell Which is why we're dedicating this deep dive directly to you, the biotech leader who's managing these early stage programs, because you're the one in the hot seat.

SPEAKER_00 1:16

Always.

SPEAKER_01 1:17

Yeah. We're pulling from a really fascinating set of notes and insights by Desi McInty. She's a board-certified toxicologist who actually embeds with fast-moving biotech teams as a fractional head of toxicology.

SPEAKER_00 1:30

She sees all the messy behind-the-scenes stuff.

SPEAKER_01 1:32

Oh, absolutely. And her core document is simply titled Study Sequencing. And our mission today is to take her insights and fundamentally reframe how your team approaches non-clinical timelines. Because right now, uh the default mindset in most boardrooms is completely backward.

SPEAKER_00 1:48

It really is.

SPEAKER_01 1:49

Okay, let's unpack this. Because when you sit in on a project management sink, the driving questions are almost always logistical, right? It's always what can we start first? Or how do we shave two months off this critical path? Right. And it sounds like rigorous management, like you're just being efficient. It's like a contractor trying to build a house by telling all the tradespeople to work on the exact same day just because they're available. Yeah. You end up with the roofers like standing on top of the plumbers.

SPEAKER_00 2:13

Aaron Powell That's exactly what it is. And if we connect this to the bigger picture, treating non-clinical study sequencing as a logistical problem that is the core philosophical error.

SPEAKER_01 2:24

Aaron Powell How so?

SPEAKER_00 2:25

Aaron Powell Well, sequencing is not about logistics, it's an informational problem. I mean, every single study in an IND enabling program exists to generate specific data, right? Right. But that data isn't just a regulatory checkbox. It has to actively inform the design, the interpretation, and the execution of the very next study you run.

SPEAKER_01 2:45

Aaron Powell Okay. But I want to challenge that slightly. Sure. Let's say a contract research organization, a CRO, has an open slot next week, right?

SPEAKER_00 2:53

Yeah.

SPEAKER_01 2:54

And we have a batch of drug substance ready to go. Why wouldn't we just start the study? It doesn't matter. I mean, we need the data eventually anyway, right? If we just gather all the pieces simultaneously, can't we just stitch the narrative together at the end?

SPEAKER_00 3:07

Aaron Powell, you'd think so, but gathering data out of sequence destroys its context. Let's look at the mechanics of what you're actually doing there. When you prioritize initiation speed over logical progression, you break the informational chain.

SPEAKER_01 3:21

So you're just generating puzzle pieces that don't fit.

SPEAKER_00 3:23

Exactly. You might look highly productive to your investors because, hey, you kicked off three studies this quarter. But you are designing those later studies completely blind to the biological realities of your molecule.

SPEAKER_01 3:35

So it's less about project management and more about avoiding this compounding cascade of scientific assumptions.

SPEAKER_00 3:42

Yes.

SPEAKER_01 3:42

Actually, I want you, the listener, to pause for a second and visualize the gaunt charts for your current program. Look at the dependencies connecting those little blocks. Yeah, those arrows. Are those arrows based on a genuine need for scientific insight? Like meaning you literally cannot design study B without the pharmacokinetic data from study A? Or do those arrows just reflect budget cycles and CRO availability?

SPEAKER_00 4:08

That is the million-dollar question. Literally. Because if your schedule is dictated by CRO availability, you are setting yourself up for a brutal reality check during your regulatory review.

SPEAKER_01 4:18

Aaron Powell Because the FDA doesn't care about your CRO schedule.

SPEAKER_00 4:20

Aaron Powell Not at all. Data without context is virtually useless when you're trying to prove a novel molecule is safe for human exposure.

SPEAKER_01 4:29

Aaron Powell Right. And this brings us to the most literal way teams break that informational chain. It usually happens when the timeline panic sets in, you know, capital is dwindling, and the executive team decides to just rush headfirst into the pivotal studies. Yeah. So I'll play the role of the stressed-out CEO here, looking at our burn rate, and I push back on the toxicology team and say, look, we have massive amounts of literature on this mechanism of action.

SPEAKER_00 4:58

I've heard this exact speech.

SPEAKER_01 4:59

We know the target inside and out. We've even run structurally similar molecules. So let's just use a dosing range that seems biologically reasonable, skip the dose range findings, study the DRF, and get straight into the GLP pivotal study. We save a month and hit our milestone.

SPEAKER_00 5:15

Yeah, I hear that exact executive argument constantly.

SPEAKER_01 5:18

Aaron Powell It sounds so logical when you're panicked.

SPEAKER_00 5:20

It does. But the fatal flaw in that logic is the reliance on literature and structural similarity. Those are assumptions. They're not data. McInty's notes highlight this as the single point where IND programs frequently die. It's the gap between assumption and actual data on your specific molecule.

SPEAKER_01 5:39

So let's get into the biological wheeze on why guessing those doses is so catastrophic, because it's a huge mistake. If we skip the DRF and guess the doses for a fully regulated GLP pivotal study, what are the actual mechanistic outcomes? Like what happens?

SPEAKER_00 5:56

Well, let's run through the scenarios. Say you guess high, you select the top dose based on literature, but your specific molecule has a slightly different metabolic profile.

SPEAKER_01 6:05

Which happens all the time.

SPEAKER_00 6:06

All the time. So it causes unexpected letsality or severe acute toxicity in your animal models.

SPEAKER_01 6:11

Oh wow.

SPEAKER_00 6:12

Yeah. And your GLP study, which costs hundreds of thousands of dollars and takes months to run, is immediately invalidated. You just have to start over.

SPEAKER_01 6:19

That's brutal. So I would assume then that the natural executive response is to play it incredibly safe.

SPEAKER_00 6:25

Exactly.

SPEAKER_01 6:25

We guess low, we ensure all the animals survive, we hit our NOAEL or no observed adverse effect level, and we pop the champagne rate. No toxicity, clean safety profile, we move to the clinic.

SPEAKER_00 6:35

And that assumption is exactly what gets programs flagged by regulators.

SPEAKER_01 6:40

Wait, really? Being too safe gets you flagged.

SPEAKER_00 6:43

Yes. Achieving a clean NOAEL without demonstrating any adverse effects at the highest dose means you completely failed to achieve a toxicological characterization of the molecule. Oh you haven't found the margin to toxicity.

SPEAKER_01 6:57

Because regulators need to know where the edge of the cliff is.

SPEAKER_00 7:00

Exactly.

SPEAKER_01 7:00

So they know how far away from it the human patients will be. If we just wander around safely in the middle of the field, the FDA knows we didn't actually stress test the system.

SPEAKER_00 7:09

That's a perfect way to put it. Furthermore, consider your toxicokinetics, the TK. If you guess low, the toxicokinetics at that assumed high dose might not support adequate exposure multiples over your anticipated clinical dose.

SPEAKER_01 7:25

So you aren't pushing it hard enough?

SPEAKER_00 7:26

Biologically, yeah. If you aren't pushing the exposure high enough in the animal model, you cannot definitively prove that the human dose won't trigger an accumulation that overwhelms their metabolic pathways.

SPEAKER_01 7:37

Which means your safety margins suddenly become this highly stressful, drawn-out conversation with regulators.

SPEAKER_00 7:43

Aaron Powell A conversation you really don't want to have. Instead of a definitive data-backed conclusion, it's just a debate.

SPEAKER_01 7:50

Right. And a debate you absolutely do not want to be having when millions of dollars are hanging in the balance. So skipping the DRF entirely is a non-starter.

SPEAKER_00 7:58

Completely.

SPEAKER_01 7:58

But here's where it gets really interesting. Because the natural compromise I see teams make is to just shorten the non-GLP studies.

SPEAKER_00 8:08

Uh yes. The half measure.

SPEAKER_01 8:10

Right. The project manager says, fine, we won't skip it. We'll run a quick seven-day DRF just to check the box and establish tolerability, even though their GLP design requires a 28-day study.

SPEAKER_00 8:21

Right. We get some data, we save three weeks.

SPEAKER_01 8:23

Exactly. They think it's a win.

SPEAKER_00 8:25

But the problem there lies in the biology of how molecules accumulate and how organs respond to prolonged exposure. A seven-day DRF only tells you about acute tolerability.

SPEAKER_01 8:36

Meaning just the initial shock to the system.

SPEAKER_00 8:38

Right. It answers the question of what happens when the body first encounters the compound. But a 28-day GLP study, which is what you need to support a 28-day human trial, that requires an understanding of target organ progression over time.

SPEAKER_01 8:52

Let's dive deeper into the mechanisms of that, because why wouldn't a seven-day test predict a 28-day outcome? Doesn't it just scale up?

SPEAKER_00 9:00

Not necessarily. Think about metabolic pathway exhaustion or enzyme induction. For the first week, the liver might be producing enough enzymes to clear the drug efficiently.

SPEAKER_01 9:10

Okay, so it looks fine on day seven.

SPEAKER_00 9:12

Exactly. But by day 14, those metabolic pathways could become completely saturated, or the drug might actually downregulate the very enzymes needed to clear it.

SPEAKER_01 9:21

Oh wow. Yeah.

SPEAKER_00 9:22

So suddenly the systemic exposure of the drug spikes exponentially and toxicity emerges out of nowhere. Or consider bioaccumulation in specific tissues.

SPEAKER_01 9:31

Like it building up in the kidneys or something.

SPEAKER_00 9:32

Right. A seven-day study tells you absolutely nothing about whether early, subtle warning signals in the kidneys will reverse themselves, or if they will compound into catastrophic organ failure by day 20.

SPEAKER_01 9:44

Okay, so it's like testing a marathon runner's endurance by watching them sprint for a single block.

SPEAKER_00 9:48

Yes.

SPEAKER_01 9:49

Or testing a bridge's structural integrity by driving a single heavy truck over it. Just because the bridge held up to the immediate weight of that one truck doesn't mean it won't completely collapse under the microfractures caused by a month of continuous grinding rush hour traffic.

SPEAKER_00 10:06

That is an excellent analogy. You are testing the wrong kind of stress. And a seven-day study also fails to give you dose-dependent toxicokinetics over a sustained period.

SPEAKER_01 10:18

So if you use acute data to design a 28-day pivotal trial. Yikes. Okay, so if skipping steps or shortening them is the first symptom of timeline panic, running them simultaneously is the second.

SPEAKER_00 10:33

Oh absolutely.

SPEAKER_01 10:34

MATE calls this the parallel processing gamble. Now, we just talked about the bridge. Starting parallel studies that should be sequential, like running a four-week study and a 13-week study at the same time, that feels to me like trying to build the structural supports for a skyscraper while simultaneously constructing the penthouse?

SPEAKER_00 10:53

It's exactly like that. But sponsors demand that kind of compression constantly. And contract research organizations are businesses. You know, they will happily initiate multiple studies simultaneously if you sign the contract.

SPEAKER_01 11:05

They're not going to turn down the money.

SPEAKER_00 11:06

No, they're not. But MATE explicitly warns that parallel execution without informational overlap isn't timeline compression. It is two separate bets running simultaneously.

SPEAKER_01 11:17

Let's break down the actual mechanics of that double bet. Because on a spreadsheet, stacking a four-week study and a 13-week study on top of each other looks brilliant, right? The bars overlap, the completion date pulls way in. You look like a genius.

SPEAKER_00 11:30

On paper, yes. Scientifically, it is incredibly precarious. The entire purpose of that four-week study is to reveal specific liabilities.

SPEAKER_01 11:38

Like what?

SPEAKER_00 11:39

Target organ effects, dose-dependent changes, and long-term toxicokinetics. You uncover that information specifically to inform how you design and crucially how you monitor the 13-week study.

SPEAKER_01 11:52

Okay, so if week three of your four-week study reveals a highly specific, subtle signal in, say, liver enzymes, normally you would use that data to write the protocol for your 13-week study.

SPEAKER_00 12:05

Exactly. You'd add specific biomarker panels, maybe specialized ophthalmic exams if there's a vision signal.

SPEAKER_01 12:12

But if the 13-week study is already running.

SPEAKER_00 12:14

You are trapped.

SPEAKER_01 12:15

Oh no.

SPEAKER_00 12:15

Yeah, you now have a live 13-week study underway. You are forced to either rapidly amend a live protocol to add that liver monitoring, which is a logistical nightmare, introduces variables, and means you miss the baseline measurements for those vital markers, or you just let it run blind.

SPEAKER_01 12:30

And if you let it run blind, you end up sitting on this massive, expensive 13-week data set that just generates far more questions than answers.

SPEAKER_00 12:37

Exactly. The regulators will look at the four-week liver signal, then they'll look at your 13-week data and they'll ask why you didn't monitor the liver long term.

SPEAKER_01 12:44

Because obviously you knew about it from the four-week study.

SPEAKER_00 12:47

Right. You essentially paid for a pivotal study that is scientifically compromised from day one, with zero ability to course correct, because the bet was already placed before the cards were turned over.

SPEAKER_01 12:57

That is terrifying. Okay, so we've diagnosed the illness here. We've explored the biological trap of guessing doses, the mechanical failure of using acute data for long-term predictions, and the gamble of parallel processing. So what does this all mean? How does a biotech leader actually implement good sequencing?

SPEAKER_00 13:18

The blueprint requires institutionalizing one unyielding rule before any contract is signed, which is you must constantly ask, what do I need to know before I start this study? And which preceding study gives me that specific information?

SPEAKER_01 13:31

Aaron Powell That sounds so simple, but obviously it's rarely done. Walk us through what that progression actually looks like in a properly sequenced program.

SPEAKER_00 13:38

Sure. It builds through biological logic. You start with a non-GLP single dose or a three to five day maximum tolerated dose study.

SPEAKER_01 13:45

An MTD.

SPEAKER_00 13:46

Right. And the objective is narrow here. You establish rough tolerability, find the starting dose range, and identify any immediate acute signals that fundamentally change your understanding of the molecule.

SPEAKER_01 13:57

Okay, so that initial data forms the foundation.

SPEAKER_00 14:00

Exactly.

SPEAKER_01 14:00

Then we move to the repeat dose.

SPEAKER_00 14:02

Yes. You use that foundational data to properly design a non-GLP repeat dose study, typically 14 or 28 days. Because you have the rough range, you use the second phase to establish dose-dependent tolerability.

SPEAKER_01 14:15

So this is where you confirm your toxicokinetics over time.

SPEAKER_00 14:17

Right. You confirm TK, identify the specific target organs, and observe how those organs respond to sustained exposure. This study gives you the definitive data needed to select doses for your GLP study without guessing.

SPEAKER_01 14:30

No relying on structural similarity to other drugs. You have hard localized data on your specific molecules' behavior over time.

SPEAKER_00 14:36

Which culminates in the final step. The GLP pivotal study. Because you sequenced properly, this study is optimized. It generates a definitive, defensible, non-clinical safety data set.

SPEAKER_01 14:48

You secure a robust NOAL.

SPEAKER_00 14:50

You clearly characterize the safety margins to toxicity. And most importantly, you produce a clean, scientifically logical interpretation narrative.

SPEAKER_01 14:58

A clean narrative. I think that is the ultimate secret weapon here. And it brings us to how this data actually performs in the real world. Because when all this data is compiled and lands on a reviewer's desk at the FDA, what is the psychology at play in that room?

SPEAKER_00 15:14

What's fascinating here is the regulatory mindset. Regulators reviewing your non-clinical section are not merely validating the math.

SPEAKER_01 15:23

They're not just checking boxes.

SPEAKER_00 15:24

No, they are actively evaluating whether you, the sponsor, actually comprehend the biological realities of your own drug. They are scrutinizing the architecture of your program to see if it was designed to answer the right scientific questions in the correct order.

SPEAKER_01 15:37

So they are judging your strategic competence, not just your raw data.

SPEAKER_00 15:40

100%. When a program is properly sequenced, it tells a coherent biological story. The regulator reads the dose range finding study, sees exactly why you selected the GLP doses, and sees how your toxicokinetic monitoring directly addressed early warning signals.

SPEAKER_01 15:57

That makes total sense.

SPEAKER_00 15:58

The interpretation builds trust, right?

SPEAKER_01 15:59

Yeah.

SPEAKER_00 16:00

Because it's grounded in a clear progression from early signal detection to definitive safety characterization.

SPEAKER_01 16:07

Conversely, let's look at the Frankenstein package produced by the rushed timeline obsessed team. Oh boy. The CRO might have technically executed each individual study perfectly, right? The data itself is accurate, but the pieces don't connect.

SPEAKER_00 16:21

The regulator immediately sees the disjointed architecture. They see doses in the pivotal study that seem totally arbitrary compared to the early data. They notice toxicokinetic data that wasn't available in time to actually inform critical decisions.

SPEAKER_01 16:36

So the whole submission feels reactive.

SPEAKER_00 16:38

Highly reactive. It looks like a team desperately trying to survive adverse findings rather than a team systematically anticipating and characterizing them.

SPEAKER_01 16:46

And regulators are highly allergic to reactive data.

SPEAKER_00 16:49

Very. As McInty notes, poor sequencing might not trigger an automatic clinical hold every single time. But it absolutely generates deep systemic questions.

SPEAKER_01 16:59

And receiving a massive list of foundational regulatory questions when you are, what, weeks away from wanting to dose human patients.

SPEAKER_00 17:10

That is the exact nightmare scenario those teams were trying to avoid in the first place.

SPEAKER_01 17:14

It is the ultimate irony of drug development. Trying to move fast biologically guarantees you will move slow regulatorily.

SPEAKER_00 17:22

Exactly.

SPEAKER_01 17:23

You tried to save four weeks in phase one and you lost four months in the final review, burning $50,000 a day while you argue with the FDA about safety margins you should have characterized a year ago.

SPEAKER_00 17:34

Strategy over logistics. That is the fundamental shift biotech leaders must make. The sequence in which you run studies actively dictates the strategic value of the data you hold at every critical juncture.

SPEAKER_01 17:47

To synthesize this for everyone listening, study sequencing is not a scheduling exercise for your project managers. It is the architectural blueprint of your entire non-clinical defensibility. When one study actively biologically informs the next step, you don't just have a pile of data points, you have a coherent, impenetrable strategy. Now, if you're realizing that your current GONT charts are leaning heavily on logistical convenience rather than scientific progression, I highly recommend diving into the source material we use today. Desse McInty has a book available on Amazon that covers exactly how to build these defensible programs. It is titled very fittingly, Data is Not Strategy.

SPEAKER_00 18:28

It's a great read.

SPEAKER_01 18:29

It really is. Grab a copy, share it with your toxicology team, and sit down with your project managers to rethink those dependencies.

SPEAKER_00 18:36

You know, this raises an important question that extends far beyond non-clinical toxicology. I'd like to leave you with one final thought to mull over. Go for it. If the core lesson here is that gathering data out of logical sequence actively destroys its strategic value, take a hard look at the rest of your biotech's operations. Think about your hiring sequence. Think about your fundraising strategy or how you were designing your eventual clinical trials. How much of your broader corporate timeline right now is being dictated by the logistical convenience of what you can do today rather than the strategic logic of what you actually need to build tomorrow?

SPEAKER_01 19:11

Wow, that is the kind of question that forces a complete paradigm shift. Thank you so much for joining us on this deep dive. Keep scrutinizing those timelines, keep demanding context for your data, and above all, keep asking the right questions in the right order.