Decentralized Clinical Trials: Benefits, Limitations and When to Use Them

During the pandemic, a Phase 3 trial for a cardiovascular drug shifted from 100% site-based visits to a hybrid model with remote monitoring and home nursing visits. Enrollment continued. Retention improved. The sponsor saved an estimated $12 million in site costs.

Three years later, that same sponsor tried to run a fully decentralized Phase 2 oncology trial. It failed. Patients needed in-clinic infusions. Remote monitoring could not replace physical assessments. Site investigators felt disconnected from their patients. The trial reverted to a traditional model after six months.

Decentralized clinical trials (DCTs) work – but not for everything. The question is not whether to use DCT approaches. It is when, how much, and for which trial types. This article provides the decision framework clinical operations teams need.

The Problem: The DCT Conversation Has Become Binary

The industry discussion about DCTs often frames it as an either/or choice: traditional site-based trials versus fully decentralized trials. In practice, virtually every successful DCT implementation is a hybrid model that decentralizes specific elements while keeping others site-based.

According to ClinicalTrials.gov data, approximately 15-20% of newly registered trials in 2024 incorporated at least one decentralized element, up from less than 5% in 2019.

The problem with the binary framing is that it leads to two types of mistakes:

Over-decentralization. Trying to remove all site visits when the protocol requires in-person procedures, leading to protocol deviations and data quality issues.

Under-adoption. Rejecting all DCT elements because the trial “needs sites,” missing opportunities to decentralize specific visits that do not require physical presence.

The Insight: What Can and Cannot Be Decentralized

The real insight: The decision to decentralize should be made at the visit level, not the trial level. Every protocol has some visits that can be remote and some that require a site. The highest-value DCT approach is identifying which visits can be decentralized and designing the protocol accordingly.

Decision Intelligence: When DCT Works Best

The Solution: Building a Hybrid Trial Design

A practical hybrid approach works in three steps.

Step 1: Visit-level analysis. Map every protocol visit. For each, determine: which assessments require a site? Which can be done remotely? Which require home nursing?

Step 2: Technology selection. Match each remote element to validated technology: ePRO platform, wearable devices, home sample collection service, telemedicine platform.

Step 3: Regulatory alignment. Confirm with the FDA or relevant agency that the decentralized elements meet data quality and patient safety requirements.

The Value: What DCT Approaches Actually Deliver

Example: A Successful Hybrid DCT in Dermatology

A mid-size pharma running a Phase 3 trial for an oral psoriasis treatment redesigned the protocol using a hybrid approach. Out of 14 scheduled visits over 52 weeks, 8 were converted to remote: ePRO-based PASI assessments guided by standardized photography, home blood draws for safety labs, and telehealth check-ins. The remaining 6 visits required in-person dermatology assessment.

Results compared to their previous traditional Phase 3 in the same indication: enrollment completed 4 months faster (broader geographic recruitment), retention improved from 72% to 89%, and per-patient cost decreased by 22%. Data quality met regulatory requirements without additional queries.

Conclusion

Decentralized clinical trials are not binary. The most effective approach is hybrid – decentralizing specific visits and elements while maintaining site-based procedures where clinically necessary. The decision should be made at the visit level based on protocol requirements, not at the trial level based on trend adoption.

For clinical operations teams, the key action is mapping every protocol visit against the decentralization potential table, selecting validated technology for each remote element, and confirming regulatory acceptance.

Explore deeper clinical operations topics. Learn about clinical trial intelligence for operations teams and how site selection frameworks are evolving in 2025.

Frequently Asked Questions

DCT and Regulatory Submissions: What FDA Expects

FDA’s 2023 DCT guidance specifies that decentralised trial data is acceptable in regulatory submissions when: electronic data capture includes a complete audit trail, patient-reported outcome instruments used in the trial have been previously validated in the indication, informed consent processes meet the legal requirements of each jurisdiction, and the protocol explicitly describes all remote assessment procedures with the same specificity as in-person procedures. Sponsors should request Type B meeting feedback on DCT design elements before initiating a trial intended for NDA/BLA submission.

Hybrid DCT Design: The Most Common Real-World Implementation

Fully decentralised trials are rare in practice – most clinical development programmes implement hybrid designs that combine site visits for high-acuity assessments with decentralised technology for routine monitoring. A typical hybrid oncology trial requires: in-person site visits at baseline (full eligibility assessment, first drug administration, baseline imaging), then telehealth visits every 2-4 weeks for clinical assessment and adverse event review, wearable monitoring for continuous activity and physiological data between visits, and in-person imaging at pre-specified assessment timepoints. This hybrid design reduces site visit burden by 50-60% while maintaining regulatory-grade data at the critical assessment points where in-person evaluation is genuinely necessary.

DCT Vendor Landscape: Key Players and What They Offer

The DCT platform market includes specialist vendors and integrated CRO solutions. Medable provides end-to-end DCT technology (eConsent, ePRO, telemedicine, site coordination). Science37 operates as both a technology platform and a network of virtual site coordinators. Oracle Clinical One includes DCT modules within its established CTMS and EDC infrastructure. Veeva Vault ePRO is the most widely used ePRO platform for patient-reported outcomes collection. For wearable integration, Medidata Sensor Cloud and BioTel Research provide validated device integration with EDC systems. Sponsors evaluating DCT platforms should confirm: regulatory acceptance precedents in their specific indication, integration with their existing EDC, and patient-facing app accessibility for their target patient demographics.

DCT Implementation Checklist: Before Your First Hybrid Trial

• Engage FDA and EMA early: request Type B meeting feedback on DCT elements before protocol finalisation for trials intended for NDA/BLA submission.
• Select DCT technology compatible with your existing EDC and CTMS to avoid manual data bridging.
• Validate remote assessment procedures with the same specificity required for in-person site procedures.
• Design informed consent process to meet jurisdiction-specific legal requirements for electronic consent.
• Plan device distribution and technical support for patient-facing wearables and ePRO apps.
• Train site coordinators on hybrid DCT workflow – the coordinator role changes substantially in a hybrid design.

Measuring DCT Success: The Metrics That Matter

DCT programmes should be evaluated against three outcome metrics. Patient retention rate: DCT programmes consistently achieve 15-25% higher retention versus traditional designs; if retention is not improving, the decentralisation design is not reducing patient burden effectively. Data completeness rate: the proportion of expected data points that are actually collected per patient; DCT designs with poor patient technology support often show lower data completeness than traditional site visits. Enrollment speed: the primary DCT promise is faster enrollment through geographic expansion; if enrollment rate is not 20-40% higher than a comparable traditional design, the DCT investment is not generating its primary expected value.