Supplying investigational product to decentralised participants
A site-based study solves investigational product supply once, at the site: a pharmacy receives, stores, and dispenses product under controlled conditions, with a paper or digital trail for every unit. A decentralised study has to solve the same problem for every individual participant's home, which turns one controlled logistics chain into as many separate ones as there are participants.
Work from the Trials@Home RADIAL proof-of-concept trial looked directly at what this actually requires in practice, and the findings are a useful corrective to an assumption that sometimes goes unexamined: that decentralised supply is simply "the same process, delivered by courier instead of collected at a site."
What actually gets harder
A few specific challenges consistently show up when investigational product supply moves into participants' homes:
- Temperature control without professional storage infrastructure. A pharmacy has validated refrigeration and monitoring built in. A participant's home has a domestic fridge, which is a meaningfully different environment to guarantee cold-chain integrity in.
- Confirming receipt and condition, not just delivery. A courier confirming a package was delivered isn't the same as confirming the product arrived in the condition it needs to be in, and the person best placed to notice a problem, the participant, may not know what to check for.
- Accountability without a pharmacist physically dispensing each unit. Drug accountability at a site is straightforward because a professional is present at the point of dispensing. At home, accountability depends on participant self-report and remote verification, which introduces genuinely different risks of error.
- Handling returns, unused product, and disposal. A site has established processes for this built into pharmacy operations. Coordinating the same thing across dispersed participant homes is a logistics problem the study has to design for explicitly, not assume will sort itself out.
The same four functions, solved two different ways
A pharmacy doesn't do anything mysterious that a decentralised process can't replicate. What it does is solve four specific functions with dedicated infrastructure and trained staff, functions a decentralised design has to solve deliberately instead of assuming away.
| Function | How a site pharmacy solves it | What has to replace it at home |
|---|---|---|
| Temperature control | Validated, monitored refrigeration on-site | Active-monitoring shipping, with an escalation process for a detected excursion |
| Confirming condition on arrival | A pharmacist inspects the product on receipt | A guided prompt telling the participant exactly what to check |
| Accountability | A pharmacist logs dispensing at the point of use | Remote logging built into the participant's normal routine, not a separate step |
| Returns and disposal | Established pharmacy operations procedures | A clear process communicated at delivery, not left to be figured out later |
Read this way, decentralising drug supply isn't really about removing the pharmacy function. It's about relocating four specific, previously professionally-staffed jobs into a process a participant can complete correctly without training, which is a considerably harder design problem than it sounds until it's broken down function by function like this.
What actually happens when one of these functions is skipped
It's worth being concrete about the failure mode each function prevents, because "temperature excursion" and "accountability gap" can sound abstract until they're tied to what actually happens next. A product shipped without active monitoring can arrive having spent time outside its validated temperature range with nobody aware of it, and unless the packaging itself flags the excursion, it may simply get used, silently compromising the data generated from that dose. A participant left with a generic "let us know if there's a problem" receipt process, rather than specific prompts about what to check, may notice damaged packaging but not think to mention it, assuming it's not their place to judge whether it matters.
Neither failure looks dramatic in the moment. Both show up much later, in an unexplained result or a query that can't be fully resolved because the information needed to resolve it was never captured, precisely because the site pharmacy's usual job of catching these issues at the point of dispensing simply wasn't done by anyone.
What the proof-of-concept trial suggests actually works
The RADIAL trial's approach points toward a few practical mitigations worth building into any decentralised supply chain from the start:
- Purpose-built shipping with active temperature monitoring, not just insulated packaging assumed to be sufficient, with a clear escalation process if a monitored excursion is detected.
- A simple, guided receipt-confirmation process for participants, specific prompts about what to check (packaging integrity, any visible temperature indicator, expected quantity), rather than a generic "did it arrive" question.
- Remote accountability logging built into the participant's normal workflow, so recording what was taken and when doesn't require a separate, easily-forgotten step.
- A clear, participant-friendly returns and disposal process, communicated at the point of delivery, not left for the participant to figure out or ask about later.
Why this matters for anyone considering a fully or partially decentralised design
The lesson from this work isn't that decentralised supply is impractical. Real proof-of-concept trials are demonstrating it can be done reliably. The lesson is that it requires deliberate, specific design attention, exactly the same category of attention that pharmacy operations already receive at a traditional site, rather than being treated as an implementation detail that a good courier service will handle by default.
A study team considering a decentralised or hybrid design for a product with real storage or handling requirements needs to ask, explicitly and early: who verifies condition on arrival, how is accountability logged without a pharmacist present, and what happens to product that isn't used. None of these questions have an obvious universal answer, but all of them have a wrong answer, which is not asking them until after the study is already running.