Should you allow participants to pause their involvement?

Some participants fall behind. Work stress, illness, caregiving, travel: the reasons are endless and mostly unpredictable. In a decentralised trial, it is easier than ever for people to quietly drift rather than formally withdraw. But instead of counting that as attrition, what if the protocol made space for a pause? Most protocols were written as if every participant's life would stay stable and uninterrupted for the full duration of the study, which is a strange assumption to build into something that can run for months or years.

This post looks at that question from operational, ethical, and data quality perspectives.

The reality of partial participation

When a participant cannot keep up, the options under a traditional protocol are binary: continue fully, or withdraw. Neither reflects what most people actually want, which is usually to stay involved, just at a reduced pace for a while.

Without a pause option, the likely outcomes are either formal dropout (which loses the participant entirely) or informal noncompliance (which loses data quality without losing the participant). Neither is ideal.

What a pause could look like in practice

A structured pause option might include:

• A participant marking themselves "on hold" in the app with a reason
• Site staff entering a pause code in the eCRF
• Automated prompts and reminders pausing for a defined period
• A recontact at a set interval to check if they are ready to resume

This is not a dropout. It is a documented, time-limited gap with a clear re-entry path.

Where this approach helps

• Retention. Participants who might otherwise withdraw permanently have a way to step back without leaving.
• Data integrity. Paused status is distinguishable from missing data, which makes analysis more honest. This distinction matters more than it might seem: a comparison of multiple imputation methods on real longitudinal trial data, checked against ground truth values collected two years later, found that well-characterised intermittent gaps could be imputed with minimal bias, even across a large number of repeated measures. A structured, labelled pause gives statisticians exactly the kind of clean, well-understood gap that imputation methods handle well. Undocumented, silent noncompliance does not.
• Trust. Acknowledging that life gets complicated, and designing for it, feels respectful to participants.

What to think carefully about

A pause option is not without trade-offs:

• Re-engagement after a pause is not guaranteed, and some pauses will effectively become withdrawals
• If pause use is uneven across arms or demographics, it could introduce bias that needs to be addressed in analysis
• Statistical methods must account for these gaps, which adds complexity, even if the evidence suggests that complexity is manageable when the gap is well-documented
• Study timelines may extend if pauses are common

Where it works best

Pause options tend to be most useful in:

• Longitudinal studies with frequent data collection, especially behavioural or nutritional
• App-based trials where prompts are the main touchpoint
• Minimal-risk interventions where a gap does not create safety concerns

Implementation basics

If you build it into the protocol:

• Define the maximum pause duration (two weeks? four weeks?)
• Limit how frequently a participant can pause
• Include clear language in consent forms about what a pause means for their data
• Flag any re-onboarding tasks that need completing when they return
• Agree the imputation or analysis approach for pause-related gaps with your statistician before the study starts, not after the data is already patchy

The broader mindset shift is worth naming: designing for continuity rather than just completion. Studies that expect perfection from participants tend to see more dropouts. Studies that expect real life and plan for it, including at the level of how the resulting gaps get analysed, tend to keep more people involved for longer.

None of this requires treating a pause as a lesser form of participation to be tolerated reluctantly. Framed well, in the consent process and in how staff talk about it, a documented pause option signals that the study is designed around how people actually live, not around an idealised participant who never has a difficult month. That signal tends to matter more to retention than any single feature of the app or platform itself.