Interactive response technology (IRT) is at the heart of clinical trials: the software system is used to randomise patients and manage the study supply chain. IRT should be used by default for any trial from the Phase II stage, says Calyx scientific eTech enabled services vice president Craig Mooney. Even Phase I trials would benefit from an IRT to track patients, he adds.
Trial sponsors and CROs have used IRT in earnest since the mid-1990s, so most kinks have been worked out, Mooney says. Consequently, there is a risk of overlooking its importance or viewing it as an afterthought.
But, if the randomisation or drug assignment is erroneous, it can make a trial invalid, says Almac head of strategy and commercialization Cheryl Kole. If the drug supply is not sent to a site at the right time, the site could end up having too much or too little inventory. If the IRT is not performing well, there is the risk of regulatory rejection of a post-trial study report, adds Sharp IRT services manager Melissa Peirsel.
With such high stakes, CROs and sponsors need to have a thoughtful shopping list when choosing an IRT.
IRT randomises patients and manages supply
There are two key elements to IRT. First, it is a patient management system to bring patients into the trial, which covers screening and randomisation, says Suvoda cofounder and CEO Jagath Wanninayake. An IRT is critical in blinded trials as it assigns patients to a trial arm, and keeps the studies blinded, Kole explains.
Second, the software is pivotal in managing the clinical trial supply chain. “Trials need the right drug at the right location at the right time every single time,” Mooney notes.
IRT systems handle the digital supply chain and the physical supply. Once a drug or device is manufactured, it is often assigned a serial number that enables tracking, Kole notes. Serial numbers, along with batch IDs and expiry dates, are then loaded onto the IRT, says Natalie Townsend, vice president, randomisation, and trial supply management (RTSM) strategy, at Veeva. IRT systems take into consideration labeling specifications of different regulatory bodies in different countries, including language requirements, Kole adds.
Once a trial is activated and starts screening or recruiting participants, then the supply can be sent. IRT allows supply traceability and warns supply managers if there are any shortages, Townsend adds. It allows for supply accountability in terms of what inventory is unused, notes CTI Clinical Trial and Consulting information technology executive director Louis Minham.
“Essentially, IRT systems automate what otherwise would have been based on paper or email chains,” Kole says.
IRT forecasting critical
While developing an IRT for a specific trial, forecasting when drug supplies should be sent to the site is a critical step.
There are predictive and nonpredictive methodologies to forecast supplies, says Sharp IRT project manager Richard Coxon. Predictive methodologies forecast how much supply a site needs for future patient visits, he explains. The patient has a schedule, and the IRT makes sure there is supply for those visits, he adds.
To enable a nonpredictive method, sites must be within “floor” and “ceiling” stock levels, Coxon says. For example, if the supply reaches the floor number, the IRT can be triggered to ensure the site goes back up to the ceiling level, he explains. A floor number is necessary as drugs can be damaged during transit or fall out of the temperature range.
Forecasting errors can expose trials to supply issues. For instance, there can be overstock at sites or the site may not have room to accommodate extra supply, which can lead to waste, Townsend says. The just-in-time (JIT) inventory management approach labels and dispatches drug supply when needed, but it does not randomise patients and its uptake is yet to peak.
Forecasting not straightforward
Studies often throw curveballs that test the forecasting methodology. It can be tough to predict the recruitment pace, which varies among sites and countries, Wanninayake says. Forecasting is more complex if a trial requires drug titration or weight-based dosing, Townsend adds. “There needs to be enough stock sent for any eventuality,” she notes.
There are more physical locations and accompanying logistics to consider with large trials, Kole says. For a US-based trial, one depot may be supplying the investigational drug to sites, and associated labelling is US-specific, she adds. For multicountry trials, there can be different depots for different countries. In some cases, the sponsor provides the comparator drug for one country, while other sites must source the drug themselves.
But Mooney underscores it is not about trial size. “It is [about] when and where patients decide to show up,” he adds. “IRT planning should consider patients you don’t know about.” For example, drug supply could be planned for five patients, but a site could recruit 10 in the first week, Mooney notes. Hence, there is a need to have a floor stock level, he adds.
An effective IRT can marry the predictable and unpredictable sides of managing supply, Wanninayake says. “An effective IRT can be configured on the fly to accommodate how the trial is going.”
Configured versus customised
There are two ways to develop an IRT: “out of the box” IRT with trial-specific configurations to match the study design, and completely customised IRT systems.
The setup for a configurable IRT system is straightforward and can be as easy as ticking a box so the number of site visits in the software adjusts according to the trial design, Coxon says. But some are not designed to be configurable. A bespoke IRT is ideal for complex trials, or if the sponsor has specific processes or integration needs, Kole notes. The IRT can be tailored to the needs of an individual sponsor or protocol and be adjusted if there is a protocol change, she adds.
A customised IRT comes with a higher price tag. Minham notes that it is easier to justify the IRT cost in a trial with more sites. “The cost of implementing an IRT is the same at eight sites or 58,” he adds. But there might be some additional cost in maintaining an IRT in a larger trial with a larger inventory.
A bespoke IRT can also take a longer time to develop, risking initiation delays. Complex, customised IRT can take six-to-ten weeks, Kole says. In contrast, a standard IRT can be finalised in one-to-two weeks for a trial with a simple design, and a system with some degree of customisation can take around four weeks, she adds.
It is critical for sponsors to land a vendor as early as possible to have enough time for adjustments, Kole notes. Also, striking a balance between being configurable and customisable is ideal to maximise the best of both approaches, Wanninayake adds.
Due to the time and resource investment required to build an IRT, there can be a degree of loyalty between CROs or sponsors and their chosen IRT vendors. Suvoda’s Wanninayake says it had worked with some sponsors for five years who, after changing CROs, dictated to the new CRO to use its IRT. CTI’s Minham adds the CRO only works with a small number of providers as their IRT have been integrated into its systems.
Large pharma aligns with one-to-three IRT providers regularly as there is likely an established team at the vendor to meet their needs, Mooney says. “They would rather fix a problem than start from scratch.” While mid-size firms prefer certain vendors, they are open to working with others who might be a better fit, he adds. Small pharma companies without many studies are not aligned to a specific provider because repeatability is not an issue, he notes.
The IRT shopping guide
In the future, IRT is likely to have an even more prominent role, such as in the decentralised approach of direct-to-patient drug deliveries. Therefore, it is imperative for CROs and sponsors to be prepared when shopping for an IRT.
For starters, even before a demo or a request for information, companies should review their trial design needs, which can allow sponsors to fully vet which IRT works best for them, Peirsel says. “It is not just covering your current needs but also potential needs in the future,” Townsend adds.
The sponsor or CRO should ensure the IRT works with its existing systems for all technologies to work seamlessly, Townsend says. The IRT that can be easily integrated into a CRO’s systems has an edge, Minham adds.
The IRT should be easy to use, as large sites with many trials will have to juggle different IRT systems, Mooney says. The time taken to run an IRT report can be different for a 30-patient trial and a 5,000-patient trial, Coxon notes, adding many people accessing the same report simultaneously can make it even harder. The software must be able to withstand such traffic.
Moreover, experience is also crucial, Mooney notes. Oncology trials can require multiple doses with different escalation schedules, and vaccine trials require supply to arrive on schedule, he says.
Vendors providing effective customer support is also vital, Wanninayake says. Typical software company support centres on how to use a certain function but within this space, questions relate to why the IRT is not reflecting the protocol, he explains.
With evolving clinical trial designs, these interactive technologies are also changing. And not all IRT systems are created equal, which makes it critical to choose the right provider. A model IRT should have a rigid foundation, with additional development reflecting the needs of the trial design. It should also be flexible to deal with the inherent unpredictability of running a study.
As Wanninayake puts it: “We shouldn’t forget that clinical trials are experiments—IRT systems should also have the flexibility to meet the unexpected.”
- Forecasting the supply of drugs to sites does not only depend on the dosing schedule, but also unpredictable factors such as recruitment pace and dropout rates.
- CROs and sponsors need to find an IRT that strikes a balance between being configurable and customisable for quicker software development and reflects the needs of the trial design.
- Companies shopping for an IRT need to be prepared to elucidate their trial needs even before looking for a vendor to make the most of early conversations.