State of decentralized clinical trials in 2026

The European Commission, the Heads of Medicines Agencies and the EMA published a recommendation paper on decentralised elements under the ACT EU initiative, dated 13 December 2022. It is explicitly non-binding. It covers elements such as home health visits, remote monitoring, direct-to-patient shipment of investigational product and electronic informed consent, and it includes an appendix summarising national provisions. The paper itself says it is a first step that would be updated as experience grows, and that it may not always match national rules, so the country-level appendix is where practical differences show up.

FDA issued a draft guidance on decentralized clinical trials in May 2023 and finalised it in September 2024 under the title Conducting Clinical Trials With Decentralized Elements. Sources differ on whether the day was 17 or 18 September, so we say September 2024. Secondary summaries note two things worth knowing: the final title shifted from a trial type to a set of elements, and investigators remain responsible for obtaining informed consent even when consent is done remotely. We did not open the FDA document itself for this post, so read the current text before you cite it in a protocol.

ICH adopted the E6(R3) principles and Annex 1 at Step 4 on 6 January 2025. Annex 2, which addresses decentralised elements, pragmatic elements and real-world data in interventional trials, reached Step 4 on 3 June 2026, according to ICH and PMDA documents and trade press. It is meant to be read with the principles and Annex 1 and does not prescribe a particular operational approach. Secondary guides report that the EU's CHMP adopted it on 25 June 2026 with effect from 15 January 2027, and that FDA had not issued Annex 2-specific guidance; we could not confirm either point on a primary page, so check the current text. Our Annex 2 guide walks through the structure, and the E6(R3) guide covers the parent text.

Read together, they ask the same things of the sponsor: decide per element whether remote conduct is fit for the protocol, keep the investigator in control of safety and consent, make sure data from remote sources is attributable and trustworthy, and document the reasoning. None of them says that more decentralization is better. That is a design decision, not a regulatory goal.

The Tufts CSDD and Science 37 analysis in May 2026 is the most recent attempt to put numbers on this. Its authors describe the evidence on direct-to-patient sites as scarce before this work. The reported results favour the site model, but the figures come from a company that was a party to the study and cover trials run through that company's own model. That does not make them wrong. It does mean you should ask what the comparator was, how the trials were chosen and whether the populations match yours.

Remote data is not automatically better or worse. A diary completed daily at home may be more timely than a recalled answer at a clinic visit, but it also adds responsibility for attribution and for what happens when a participant skips entries. The regulatory texts above care about attributable, traceable records. In practice that points at the audit trail, role-based access and a clear account of who entered what. Our post on the technology stack deals with that layer by layer.

Remote elements help when they remove a real burden for participants and do not remove oversight. They hurt when they are added because the word is fashionable. The evidence supports the first half of that sentence more than it supports a general claim about decentralization.