Cost savings
Real
Devices, shipping, returns
Risk
Conditional
Instrument-dependent
FDA position
Accept
With equivalence evidence
BYOD Decision Model
Instrument · Population · Endpoint · Support
BYOD — Bring Your Own Device — lets patients complete eCOA/ePRO assessments, eConsent, and some sensor capture on their own smartphone instead of a sponsor-provisioned phone or tablet. The economics look obvious: no devices to ship, no inventory to track, no returns. In short studies with high-acuity populations and unburdensome instruments, BYOD often works well.
The problem is that "BYOD" is treated as a single decision when it's actually four decisions: about the instrument, the population, the endpoint, and the support model. Sponsors who collapse all four into a procurement conversation usually save money on devices and lose it twice over on protocol deviations, helpdesk volume, and remediation visits.
The FDA has accepted BYOD-collected ePRO data in regulatory submissions. The ISPOR ePRO Good Research Practices Task Force (Coons et al., 2009; updated 2023) describes the evidence required when an instrument is migrated to electronic — including across patient-owned devices. None of this is a green light. It's a checklist.
See also: eCOA & ePRO · Device & Provisioning · Concierge-as-a-Service
Plain Likert items, simple numeric ratings, and text-light instruments tend to migrate cleanly to BYOD. The ISPOR mixed-modes work classifies these as "minor modifications" — usability testing and cognitive interviewing are typically enough; a quantitative equivalence study is not always required.
Adult oncology, cardiometabolic, and many chronic disease populations now have near-ubiquitous smartphone ownership in most regions. When access is uniform across the protocol's geography, BYOD reduces friction at screening rather than introducing a new selection bias.
Self-reported symptom diaries, eConsent, daily wellness check-ins, medication adherence prompts, and qualitative quality-of-life instruments don't depend on a specific sensor. BYOD is a reasonable default for these.
BYOD studies that pair a clean digital design with multilingual concierge support — onboarding calls, OS-specific troubleshooting, password resets, lost-phone recovery — keep compliance curves where provisioned studies do. BYOD without operational support is where the data gaps appear.
Visual analog scales (VAS), pain body maps, complex matrices, fixed pagination, and instruments validated on a specific screen size are mode-sensitive. Muehlhausen et al. (2018) and the 2023 ISPOR update both flag VAS items as the canonical failure case for unscreened BYOD migration. If the instrument is in this group, equivalence work is not optional.
Older patients, rare disease cohorts with pediatric subgroups, and trials in regions with low smartphone penetration can introduce selection bias when BYOD is the only option. A hybrid model — BYOD where it works, provisioned where it doesn't — typically protects enrollment without breaking the data model.
Many digital health technology (DHT) endpoints — six-minute walk tests, sleep architecture, validated activity counts, PPG-derived HRV — depend on a known sensor, sampling rate, and firmware. BYOD wearable choice fragments those inputs. Use sponsor-provisioned, validated devices when the endpoint is sensor-driven; let BYOD handle the questionnaire layer.
The most common BYOD failure pattern is invisible until interim analysis: the cost savings on devices are reabsorbed by helpdesk tickets, app reinstalls, OS-update failures, and missed assessments. Without a concierge layer, BYOD shifts cost from procurement to clinical operations — and from a fixed line item to an unpredictable one.
Is the instrument validated on paper or a specific device size? Does it use VAS, body maps, complex matrices, or strict pagination? If yes, plan equivalence work before BYOD goes anywhere near the protocol.
What is realistic smartphone penetration in the target population and geography? What proportion of expected enrollees will need a provisioned device to participate? A 5–15% provisioned tail in an otherwise BYOD study is usually the right operational mix.
Does the primary or key secondary endpoint depend on standardized sensing? If yes, BYOD should be limited to the eCOA/ePRO layer. Sensors and digital biomarkers should run on validated, provisioned hardware.
Who answers the phone at 8pm on a Sunday when a patient can't log in? In the absence of a defined human escalation path — ideally multilingual — BYOD is a cost-deferral plan, not a cost-reduction plan.
Standard symptom diary, Likert-style PRO-CTCAE-style items, weekly cadence, adult population with high smartphone access. BYOD is usually appropriate. Equivalence evidence: usability testing + cognitive interviewing; quantitative equivalence may not be required if modifications are minor.
Visual analog scale as a primary endpoint instrument is the textbook BYOD failure scenario. Either provision a device with a calibrated VAS implementation, or run a formal quantitative equivalence study before approving BYOD.
Hybrid model: BYOD for the daily eCOA, sponsor-provisioned validated wearable for activity, HRV, and sleep endpoints. The instrument layer and the sensor layer are separated. This is the most common defensible design for DCT/hybrid studies. See: decentralized clinical trial execution.
Caregiver-mediated reporting, mixed device ownership, small cohort. Default to provisioned devices — the cost of a few protocol deviations in a small cohort is larger than the device savings. See rare disease compliance.
No. The ISPOR ePRO Task Force defines three modification levels — minor, moderate, substantial. Minor changes typically require usability and cognitive interviewing only. Moderate or substantial changes — including most VAS, complex matrix, or strict pagination items — require quantitative equivalence.
Cross-OS rendering is a design and QA problem, not an evidence problem. Address it at the eCOA build stage with platform-specific testing and a locked rendering specification. Document the rendering control in the system validation package.
Yes. Mixed-modes studies are explicitly addressed by the ISPOR mixed-modes guidance and have been accepted in submissions. The study must pre-specify how mixed-mode data are analyzed and that the analysis plan accounts for any mode effects.
The most common failure is not technical — it's the absence of a defined patient support escalation path. Patients abandon their assessment after one or two friction events. A concierge layer with same-day response prevents the abandonment curve from forming.
BYOD does not exempt the data system from Part 11. Audit trails, electronic signatures, and access controls apply equally — the data system, not the device, is the controlled environment.
Delve Health combines a configurable eCOA/ePRO platform, validated wearables, and human concierge support — so the BYOD decision is operational, not improvised. We help sponsors design the right mix per protocol and stand up the support layer that keeps BYOD trials on plan.
Talk to DelveeCOA vs ePRO vs DHT: clear definitions · How to choose an eCOA vendor · Why eCOA data fails in real-world trials · Multilingual patient support and data quality · Reducing site burden in hybrid trials
Sources referenced: Coons SJ et al., Value in Health 2009 (ISPOR ePRO Task Force); Eremenco S et al., Value in Health 2014 (PRO Mixed Modes Task Force); Muehlhausen W et al., Therapeutic Innovation & Regulatory Science 2018 (BYOD measurement equivalence review); ISPOR 2023 updated recommendations on measurement comparability across modes; FDA PFDD Guidance 3 (October 2025); Pugliese L et al., Therapeutic Innovation & Regulatory Science 2022 (regulatory acceptance of BYOD).