Choosing between BYOD and provisioned devices in trials

The choice between letting participants use their own devices (BYOD) or shipping pre-configured ones touches more than logistics. It shapes the participant experience, the consistency of data collection, and the operational overhead your team takes on for the duration of the study.

Neither approach is universally right. The right answer depends on who your participants are, what the protocol requires, and what your team can actually support. It's tempting to assume smartphone ownership is now close to universal and that BYOD is therefore the safe default. A community survey of 104 adults in a low-income New York neighbourhood found overall smartphone ownership was genuinely high (around 72% of respondents), but that headline figure hid meaningful gaps: ownership was notably lower among older participants, those identifying as Latino, those insured by Medicare, and those with household incomes below $30,000 a year. Participants with at least a high school education were roughly seven times more likely to actually use health apps than those without one. The digital divide has narrowed. It hasn't closed, and it doesn't close evenly across every population a study might recruit from.

When BYOD makes sense

Bring-your-own-device tends to work well when:

• The study involves simple, frequent inputs like daily symptom scores or quick surveys
• Participants are comfortable with apps and typically use modern smartphones
• The protocol is observational or low-risk, where strict device standardisation is less critical
• Shipping and tracking physical devices would add more complexity than it solves

The main advantages are familiarity and convenience. Participants are already using their device. There is no second gadget to carry around or worry about losing.

When BYOD creates problems

The risks of BYOD grow when:

• The app requires a specific OS version that older devices may not run
• Participant populations include people with less reliable internet access or older hardware
• The study collects passive or background sensor data that needs consistent hardware to be comparable
• Intervention timing is precise enough that OS-level variability could affect the experience

Inconsistent app performance across devices can also create data quality differences that are difficult to account for in analysis. And if your target population overlaps with any of the groups the East Harlem study found had lower ownership rates, a BYOD-only design risks quietly excluding exactly the participants whose data the study most needs to be representative.

When provisioned devices are worth it

Pre-configured devices make more sense when:

• Standardisation is genuinely important for data comparability
• The participant population may be unfamiliar with apps or smartphones
• The software requires specific hardware or performance capabilities
• You need to control exactly what is installed and what is not

The upsides are control and consistency. Technical support is simpler when all devices are identical. Setup errors at the participant end are much less common.

The downsides are real too: shipping, tracking, and returns add administrative overhead, and some participants are reluctant to carry a second device.

Three questions to guide the decision

1. What is the minimum data you need, and how often do you need it?
2. Who are your participants, and what do you know about their digital access and comfort, not just in aggregate but broken down by the demographic factors most likely to predict a gap?
3. What level of device support is your team actually equipped to provide?

A hybrid model is also an option

Some trials offer BYOD as the default and ship provisioned devices to participants who request them or who do not have compatible hardware. This adds a layer of operational complexity but improves equity and flexibility without forcing one model on everyone, and it directly addresses the specific disparity the research above documents rather than assuming it away.

There is no universally correct answer here. The decision should follow the study, not precede it, and it's worth revisiting the decision partway through recruitment too. A gap between who signed up and who's actually completing tasks reliably is often the first visible sign that a BYOD-only design has quietly excluded a segment of the intended population, and it's far easier to add a provisioned-device option mid-study than to explain the resulting gap in the data at the end of it.