For a sponsor running a real-world study, patient-led digital enrollment lets participants find a study online, check their own eligibility, and consent without a site coordinator. Adoption has moved fast. What had not been measured at scale is how that route performs once a study is running, next to traditional site-based enrollment. Castor and Thermo Fisher Scientific set out to close that gap and presented the results at the 42nd ISPE Annual Meeting in Milan.
Patient-led digital enrollment reached half of a study’s participants in a median of 1 month, against 6 months for site-based enrollment, about 6 times faster, with no evidence of disengaging over time once the recruitment funnel is separated from enrolled participants.
At a glance
In a Castor and Thermo Fisher Scientific collaboration at platform scale, we asked whether the way people enroll shapes the data quality that follows. A matched retrospective analysis of 707 patient-led studies against 8,423 site-based studies found that patient-led enrollment filled studies about six times faster, with no long-run engagement penalty.
The gap this set out to close
Patient-led digital enrollment is growing in real-world evidence research, but how it performs once a study is running, next to site-based enrollment, had not been measured at platform scale. The team compared the two routes across enrollment speed, long-run survey engagement, enrolled follow-through, and retention, using de-identified operational data from 15,912 Castor studies matched like with like on therapeutic area, design, size, time period, and survey schedule.
What the analysis found
Enrollment fills half a study in one month, not six
Patient-led studies reached half of their final enrollment in a median of one month, against six months for site-based. Enrollment front-loads under patient-led designs.
Long-run engagement converges
Survey completion by position converges over the study. Site coordination gives a small early bump that fades within a few months, and patient-led shows no sign of disengaging over time. This is a descriptive convergence, not a claim that one route is better.
A measurement point for patient-led data
Open enrollment links create a record for every anonymous click. Any metric that pools those clicks understates the true performance of the patient-led route. Separating anonymous records from enrolled participants is necessary for a fair comparison, and is the methodological contribution the poster sets out.
The honest cost
Enrolled follow-through is close between the two routes. Patient-led protocols here ran shorter, so observed follow-up was shorter. The poster reports this openly rather than around it.
Abstract, as presented at ISPE 2026
Authors. Derk Arts, MD, PhD (Castor). Mariah Baltezegar (Castor). Susan A. Oliveria, ScD, MPH, FISPE (Prospective Real-World Studies, Thermo Fisher Scientific, New York, USA). Christine Varner (Prospective Real-World Studies, Thermo Fisher Scientific, North Carolina, USA). Sebastiaan Knijnenburg, PhD (Castor).
Background. Decentralized designs increasingly use patient-led digital enrollment, where participants self-identify online, prescreen, and eConsent without investigator mediation. Whether enrollment route affects downstream data completeness and retention has not been measured at platform scale.
Objective. Compare enrollment velocity, long-run ePRO engagement, enrolled follow-through, and retention between patient-led and site-based studies on a single platform (Castor).
Methods. Matched retrospective study of de-identified operational data from 15,912 production studies, compared like with like on therapeutic area, design, size, time period, and survey schedule. A tiered evidence model identified 707 patient-led studies (462,812 participants) against 8,423 site-based, giving a matched set of 472 patient-led and 1,078 site-based. Patient-led classification is predominantly behavioral inference, with a confirmed 19-study subset (gold 8, silver 11) retained as a sensitivity check. Timing outcomes are reported as medians with interquartile range, other outcomes as proportions with bootstrap 95 percent confidence intervals.
Results. Patient-led studies reached half of final enrollment in a median of 1 month versus 6 for site-based, about 6 times faster. Long-run ePRO completion converged, with no evidence of disengagement over time. Enrolled follow-through was close (0.886 versus 0.931), and 180-day retention favored site-based (0.65 versus 0.12), confounded by protocol length.
Conclusion. Patient-led digital enrollment fills studies about 6 times faster with no long-run engagement penalty, at the cost of a shorter observed follow-up. It suits front-loaded, short-cycle real-world evidence programs.
Read the full poster, including the cohort construction, the enrollment and engagement figures, and the limitations.
Design a real-world evidence study around patient-led enrollment
From patient-led enrollment and eConsent to real-world evidence at scale, with AI-driven data extraction and human review.
Explore Castor for RWE