Decentralized Trials Are Everywhere. Nobody Agrees What They’re For.

Pick up any clinical development conference agenda from the last two years and a pattern emerges: decentralized clinical trials get their own keynote slot, their own breakout session, their own pre-conference workshop. Platform vendors have multiplied. CROs have built dedicated DCT service lines. Pfizer’s public statements have credited remote elements with accelerating COVID-19 vaccine trial logistics. The FDA issued a final guidance on conducting clinical trials with decentralized elements in September 2024, lending the whole enterprise institutional gravity. By almost every surface-level measure, decentralized trials have arrived.

The data can be made to look compelling. According to research from the Tufts Center for the Study of Drug Development’s PACT Consortium, DCT-enabled trials enrolled Asian participants at a rate of 20.9% compared to 14.2% in traditional trials, and American Indian or Alaska Native participation reached 1.9% — nearly four times the 0.5% seen in conventional designs. Those numbers get cited in investor decks, regulatory submissions, and op-eds. They are real. They are also cherry-picked from trials where DCT elements were carefully selected, well-resourced, and operationally compatible with the patient population in question.

Yet underneath the momentum, there is no shared understanding of what decentralized trials are actually supposed to accomplish — and when the stakeholders who must execute them want genuinely incompatible things, the model does not quietly underperform. It collapses.

The Stakeholders Are Not Pulling the Same Direction

Sponsors approach DCTs primarily as an enrollment acceleration tool. The pitch is intuitive: remove geography as a barrier, extend your catchment area to the entire country, reduce dropout by eliminating burdensome site visits. Under this logic, decentralization is a throughput solution. The problem with that framing surfaces when you examine what sponsors are actually willing to build. DCT infrastructure — home nursing networks, direct-to-patient drug shipment logistics, remote monitoring integrations, telehealth platforms cleared for clinical use — requires upfront capital investment that most sponsors want the CRO to absorb. The result is a structural tension where sponsors demand a DCT-enabled trial at traditional CRO margins.

CROs, for their part, have built DCT capability as a revenue diversification strategy, not because they have resolved the operational complexity. Every major CRO now has a DCT division with a logo and a slide deck. What they have not uniformly built is the trained home nursing bench, the site-agnostic data reconciliation workflows, or the regulatory affairs depth to manage the protocol amendment load that comes when decentralized elements require mid-trial adjustment. The FDA’s September 2024 final guidance is explicit that incorporating decentralized elements does not diminish sponsor responsibility for data integrity and participant safety oversight — language that should be read as a warning, not a permission slip.

Sites occupy the most uncomfortable position. A fully decentralized trial, in theory, routes participants directly to their homes and bypasses the traditional site entirely. A hybrid DCT still requires the site to manage consent, coordinate with home nursing vendors, reconcile remotely collected data, and troubleshoot technology failures — without the visit-based revenue that made that administrative burden worthwhile. The investigator who built a practice around seeing trial participants in-clinic every four weeks now has a logistical coordination role and a reduced fee schedule. Site enthusiasm for DCTs, accordingly, is stratified by financial model, not by ideology.

Then there is the FDA’s position, which is the most consequential and the most misread. The May 2023 draft guidance — later finalized in September 2024 — carefully avoided mandating any specific DCT architecture. Regulators recommended; sponsors decided. That flexibility, which was intended to accommodate diverse therapeutic areas and patient populations, has functioned instead as an ambiguity that every stakeholder interprets in their own favor. Sponsors read it as endorsement. CROs read it as a sales opportunity. Sites read it as an unfunded mandate. And the agency, holding the guidance document, watches all three groups describe incompatible implementation strategies as if they read the same text.

The principle underneath this chaos is straightforward: all four stakeholders are optimizing for different outcomes and calling it the same thing. Sponsors want faster enrollment. CROs want margin. Sites want preserved revenue. The FDA wants data integrity and participant safety. Those four objectives can coexist in a well-designed hybrid trial. They frequently do not, because nobody sat in the same room and agreed on which objective was primary before the protocol was written.

When the Model Fails, It Fails Quietly

Consider the TELEPIK trial — a fully decentralized study in patients with advanced estrogen-receptor positive, HER2-negative breast cancer. The design was built on the premise that removing the site visit barrier would unlock an underserved population. The trial was terminated early after achieving an 8% enrollment rate, with only 2 patients enrolled out of 25 estimated eligible participants. The investigators noted that the rate was similar to what traditional oncology trials had already experienced — meaning the decentralized format provided no enrollment advantage whatsoever in that population.

That finding deserves more attention than it has received. The standard counterintuitive assumption in DCT advocacy is that geography and site burden are the primary enrollment barriers. TELEPIK suggests that for complex oncology populations, the dominant barriers are disease severity, treatment eligibility, and patient risk tolerance — none of which a telehealth platform resolves. Removing the site visit does not fix the underlying reasons a terminally ill patient declines trial participation.

And yet TELEPIK is rarely cited in DCT marketing materials. The Tufts diversity data is cited constantly. Both findings are real. The selective use of evidence is how a contested methodology becomes an industry consensus — not through dishonesty, exactly, but through the gravitational pull of what each stakeholder needs to be true.

What the Guidance Left Unanswered

The FDA’s September 2024 final guidance runs to dozens of pages and addresses remote consent, local laboratories, digital health technologies, and sponsor oversight obligations with considerable specificity. What it does not address is the threshold question: for which trials, in which indications, with which patient populations, do decentralized elements actually improve trial outcomes rather than simply redistribute operational complexity?

That absence is not an oversight. Regulators are not in the business of telling sponsors which design to choose for efficiency purposes. But the practical consequence is that the industry is running DCT adoption without a validated selection framework. A cardiovascular outcomes trial enrolling ambulatory patients with well-controlled disease faces an entirely different calculus than an oncology trial enrolling patients on active systemic therapy. The Tufts diversity data came from trials where DCT elements were suited to the patient population. TELEPIK’s failure came from a population where they were not. Without a published framework for distinguishing these contexts, sponsors are making DCT design decisions based on commercial pressure and conference consensus rather than evidence.

The FDA has the data to build that framework. Across hundreds of IND submissions over the past three years, the agency has seen which decentralized elements were incorporated, in which indications, and with what downstream effects on protocol deviation rates, dropout, and data completeness. Whether that knowledge ever becomes public guidance — a tiered indication framework, a DCT feasibility checklist with regulatory weight — remains the open question that the September 2024 document conspicuously declined to answer.

Until it does, every stakeholder will keep reading the guidance as confirmation of what they already wanted to do. And somewhere, another TELEPIK is being designed with complete confidence.

References

1. FDA — “Conducting Clinical Trials With Decentralized Elements: Guidance for Industry, Investigators, and Other Interested Parties” (September 2024)
2. Ropes & Gray — “FDA Guidance Clarifies Approach to Decentralized Clinical Trials” (May 2023)
3. Applied Clinical Trials — “New Data Links Decentralized Clinical Trial Approaches to Improved Diversity in Patient Populations” (Tufts CSDD PACT Consortium)
4. PMC / National Library of Medicine — “TELEPIK Trial: Early Termination Due to Low Accrual in a Decentralized Oncology Trial”