Understanding Partial Compliance in Remote Data Collection

In remote trials, participant data often arrives in fragments. A few diary entries here, a skipped week there. A missed video visit. An adverse event report that never quite gets submitted. None of this is unusual. But deciding what to do with partial data is one of the harder parts of running these studies.

‍

Traditional trial models often emphasised complete datasets. Participants who missed multiple visits were excluded from analysis. But in decentralised and real-world settings, expecting perfect compliance is not only unrealistic, it risks missing valuable information from the people who are least able to follow a rigid protocol.

‍

The first step is accepting that partial compliance is part of the picture. That does not mean lowering standards. It means planning for variability and deciding early what counts as usable data. Are five out of seven days enough to calculate an average? Does a missed check-in invalidate a week? The answers depend on the study question and the analysis strategy, but the questions should be asked up front.

‍

Patterns can also help. One missed entry is often noise. A steady decline over time might reflect real disengagement, a technical issue or an early signal of a clinical event. Looking at data in context can provide meaning that goes beyond individual fields.

‍

Some missingness is random. But some follows patterns linked to participant characteristics. People with higher symptom burden, lower digital literacy or more demanding schedules may contribute less complete data. If these gaps are not acknowledged, the analysis could skew toward people who find the process easier, not necessarily those for whom the intervention matters most.

‍

Design decisions can also play a role in minimising unnecessary gaps. Consider:

• Reducing the number of required daily entries when a weekly average is sufficient
• Giving participants recovery options so they can return to the study after missing a few days
• Avoiding language that suggests a missed task is a failure
• Providing visible feedback on progress, even when data is partial

‍

Finally, how sites and coordinators respond matters. A missed entry does not always need a query. Sometimes a brief message of support or a reminder that all contributions are helpful is enough to bring someone back on track.

‍

Partial compliance does not always mean poor engagement. It often reflects real life. And studies that are built with this in mind are more likely to succeed in gathering useful, inclusive and credible data.