CCTMB #005 What are we talking about when we talk about decentralised clinical trials (DCTs)?

By Annie Wright, NIHR BRC PhD Candidate based at Imperial Clinical Trials Unit (ICTU)

Decentralisation is the concept of moving trial procedures, e.g. recruitment or consent, away from a central location and towards participants. Different ways of decentralising can occur across the lifecycle of a trial:

Stage of the Trial	Traditional	Decentralised
Recruitment	Recruitment at the trial site, with participants speaking to study staff	Online recruitment, remote screening
Consent	Consent discussed and documented during an in-person site visit	e-Consent, completed remotely
Intervention	Participants collect medicines from the trial site	Medicines delivered to participants' homes
Follow-up/outcome collection	In-person visits and assessments at the trial site	Remote visits, electronic PROMs, wearables, home measurements (biological and physiological)
Monitoring	Source data verification and monitoring visits conducted at the trial site	Remote source data verification, remote monitoring visits.
Participant engagement	Communication primarily through site visits, telephone calls or paper materials	Apps, messaging, digital reminders

You might be reading this list and thinking, “we already do this in our trials, but we don’t call them decentralised?”

While the use of Decentralised Clinical Trial (DCT) terminology has increased, many of the individual approaches we now describe as “decentralised elements” aren’t particularly new. Telephone follow-up, postal questionnaires, and centralised trial procedures have been used for some time. Therefore, we might not use the label DCT to describe what we are doing.

Where do we draw the line?

I have previously thought about decentralisation as a spectrum from traditional → hybrid (some elements remote) → fully decentralised (completely remote). I wrote a blog post at the start of my PhD on the decentralised trials and what I thought about the concept then. Throughout my PhD though, I have noticed that the understanding and definitions of DCTs vary between trialists and there is not an agreed definition.

Overwhelmingly, trials we run would probably fall into the hybrid category. Is it useful to say that they are hybrid DCTs, or is this just the changing landscape of trial conduct?

There are benefits to having a label for decentralised trials. It gives researchers a way to find examples of innovative delivery, makes case studies easier to share, and provides a common language for a growing area of methodology. But there are also risks. If every trial with e-consent is called “decentralised”, we lose the ability to spot more substantial changes to trial delivery. And “The Multicentre Platform Bayesian Adaptive Decentralised Cluster Randomised Controlled Trial” is a bit of a mouthful…

The FDA uses the term “trials with decentralised elements” in their most recent guidance, which may be more helpful. Rather than asking “Is this a DCT?”, we could describe which elements of a trial have been decentralised. One way to make these elements more visible could be at the clinical trial registration phase, for example on ClinicalTrials.gov: add a short tick-box section for decentralised elements. Trial teams could simply indicate which parts of the trial were delivered in a decentralised manner.

Mock up of selecting decentralised elements on ClinicalTrials.gov (not available in practice)

The key idea is to make it easy to search for trials using similar decentralised elements, so we can find and learn from real examples. A possible counterpoint is that researchers could search for individual terms, such as “e-Consent” or “remote monitoring”, and that reporting guidelines such as SPIRIT and CONSORT already encourage trial teams to report these details. However, a registration-based approach would complement, rather than replace, existing reporting guidance. It would allow users to filter directly for specific decentralised elements or combinations of elements, including in trials where these details are not clearly reported in the final publication.

This may be particularly important for elements such as centralised monitoring. In my systematic review, some elements were poorly reported in publications, meaning that they could easily be missed when relying on publication searches alone. A small number of structured tick boxes at registration could therefore improve the visibility and discoverability of decentralised trial practices.

Perhaps the more useful question then for the future isn’t “Is this a decentralised trial?” but “Which parts of this trial could be decentralised, and why?”

When designing a trial, we might want to think about the following questions to decide which elements could be decentralised:

1. What are we trying to achieve with decentralisation?

   • Reduce participant burden? Improve recruitment? Make follow-up easier?

2. Which activities could be decentralised?

   • Consent? Outcome collection? Intervention delivery? Monitoring?

3. Who benefits, and who might be disadvantaged from decentralisation?

   • Participants, sites, researchers? And does it work for everyone?

4. What are the trade-offs?

   • Technology requirements, data quality, workload, digital exclusion, cost, etc.

These questions matter in practice. As part of my PhD, I conducted two surveys: one exploring trialists’ experience of decentralised trials, and another exploring the public’s views on using technology to collect outcome data.

For trialists, 93% had already used at least one decentralised trial element. However, cost, operational complexity and a lack of suitable data-management systems were among the main barriers they reported. So, even when a decentralised approach seems appropriate, there can be practical challenges to putting it into practice.

The public survey also highlighted the importance of thinking about who a particular approach works for. Many people were open to using digital methods in clinical trials, with 75% preferring to complete patient-reported outcomes online and 71% preferring an app. Comfort with the devices included in the survey was also high, with 82–93% reporting that they were comfortable using them, and 80% were willing to share routine medical records as part of a trial. However, participants also raised concerns about how comfortable and easy devices were to use, as well as privacy.

Together, these findings reinforce the idea that there is no single answer to whether an approach should be decentralised. Instead, we need to consider what it is trying to achieve, who it works for, and what trade-offs it brings.