When IRT Configuration Decisions Disrupt Clinical Supply

A conversation with Craig Mooney, Founder & Principal, IRT Advisors Group

Clinical supply performance is often treated as a forecasting or inventory problem. But in many studies, the way the IRT system is configured can have just as much influence on whether the right product reaches the right patient at the right time.

The challenge is that IRT configuration does not operate as a collection of independent settings. Enrollment assumptions, dosing, thresholds, shipment timelines, inventory levels, and allocation algorithms can all interact, meaning a change in one part of the system can affect how the broader supply network behaves.

In this Q&A, Elizabeth Urbanek, executive editor of Clinical Supply Leader, speaks with Craig Mooney about how IRT configuration decisions affect clinical supply performance, where enrollment forecasting and supply planning can diverge, and why clinical supply teams need to be involved in IRT strategy from the beginning. He also discusses the operational warning signs that can indicate a system is no longer keeping pace with the realities of the study, and why technology still requires active supply chain oversight.

What are the most common ways IRT configuration decisions affect clinical supply performance?

The challenge is that there is no single magic configuration dial or number you can turn to get it right. It's a combination of many configurations working together, and that's where people get into trouble, treating them as groups of settings instead of a system. These include things like DNx values, algorithm schedule, thresholds, caps, shipping timelines, enrollment rates, and on and on (the list is longer than most people want to hear about at a kickoff meeting). Part science and part art, the whole exercise is knowing how these configurations work together to achieve the desired result: the right product to the right patient at the right time, every time and without exception. Get one variable wrong, and you are not just wrong about that variable, you are wrong about how the whole system behaves, because these configurations are connected. If you don't know how these work in concert, you don't know how your system will behave.

Where do sponsors typically underestimate the connection between enrollment management and supply planning?

One of my favorite phrases is, "The number of patients is not the question, but when and where they show up." My experience has shown that even under the best conditions, actual enrollment rates for individual sites, or even entire countries, can be far off from what was predicted in the pre-study phase. Well-intended sites and sponsors are often making these predictions based on assumptions that don't pan out or on an understanding of the draft protocol at the time of assessment, which is usually months before the protocol is finalized. I have watched carefully built enrollment forecasts fall apart in month two because a couple of anchor sites didn't open on schedule or the patient population was harder to match with the inclusion/exclusion criteria. Clinical research is dynamic, so the response has to be a dynamic approach, too. This is the plan for now, not forever, and a team that manages with that perspective is more likely to have success.

What study design elements create the greatest challenges for forecasting and inventory management?

Variable dosing is one of the biggest challenges. A fixed dose model is somewhat predictable. But when a design includes variable dosing, prediction gets much harder, with the degree of dose variability driving the degree of predictability. I like to remind colleagues that we are doing research. If the answer were already known, there would be no reason to run the study in the first place (obvious when you say it out loud, easy to forget in the middle of a forecasting exercise). When you don't know what dose a patient will end up on, you have to be prepared for any option, which means building in flexibility rather than betting on a single scenario. Sometimes, availability of supplies makes that flexibility near impossible. Fortunately, modern IRT systems can mitigate, though certainly not eliminate, this challenge through fractional prediction models that hedge across the range of possible doses instead of guessing at just one or overpredicting on all.

How should clinical supply teams be involved in IRT strategy discussions during study start-up?

"Involved" is too light a word. Clinical supply teams are instrumental. In many organizations they own IRT, and that is, in fact, my preferred perspective. IRT has a few primary goals, with the top two being randomization and enrollment on one side and assurance of supply for all participants on the other. Anything less than a full partnership between the responsible parties on these fronts is malpractice. Bring all the parties to the table early, not as a courtesy, but because it's impossible for one or two groups to know all the implications of decisions made in isolation. An often-cited African proverb says if you want to go fast, go alone. If you want to go far, go together.

What warning signs indicate an IRT design may create operational problems later in the trial?

The early warning signs are usually not dramatic. They show up as small operational exceptions that start to repeat. One site needs an unexpected shipment. Another site is sitting on more inventory than it can reasonably use. A depot looks healthy overall, but the wrong product is in the wrong place. Enrollment is moving differently than assumed, and the team starts making manual adjustments to keep up.

This is where I lean on my affinity for 007: "Once is happenstance. Twice is coincidence. The third time it's enemy action." Another clear signal is when the phrase "workaround" is first suggested. Workarounds are fine for one-off situations, but if you don't investigate the reason for the one-off, or too quickly dismiss that it is not systemic, you might be in trouble.

Looking across hundreds of studies, what lessons can clinical supply leaders apply to improve trial execution?

You cannot rely on a system to manage your supplies. Systems are based on math and conditions – that's it. Only supply chain professionals actively manage supplies, because they can see the things a system can't include in its algorithm, such as a site that has been quietly underperforming, a shipment stuck in customs, a protocol amendment that just changed everyone's assumptions. I have seen bad things happen when systems were treated as the only way supplies were managed. Systems can and should do rules-based heavy lifting, but they need review and insight from people who understand the study, not just the algorithm. Systems do not have context or experience. "Set it and forget it" is a recipe for disaster, in clinical supply as in most things worth doing carefully.

About The Expert:

Craig Mooney is the Founder and Principal Consultant of IRT Advisors Group, an independent consultancy focused on IRT and RTSM strategy. With more than 30 years of experience on both the sponsor and provider sides, he brings a 360-degree view to how trial technologies get built, bought, and run. Craig spent eight years as director of IRT at Bristol Myers Squibb. Recent roles include vendor leadership at Perceptive (formerly Calyx) and Taikun Pharma Services.

In 2025, Craig received the Informa IRT Lifetime Achievement Award and in 2026 completed IRT Insights Initiative, an independent capability assessment of IRT/RTSM vendors.

Craig is also an Editorial Advisory Board Member for Clinical Supply Leader.