ISO has published the fourth edition of ISO 14155, the international standard for good clinical practice in the clinical investigation of medical devices for human subjects. The new edition, ISO 14155:2026, was published on 23 March 2026 and replaces ISO 14155:2020. It has been adopted at European level as EN ISO 14155:2026, with the previous European version withdrawn.
ISO 14155 is the backbone standard for designing, conducting and reporting clinical investigations. Under the MDR it underpins the clinical evidence that manufacturers generate to support conformity, so a new edition is one of the more consequential standards updates the sector will see this year. This revision is more than housekeeping. It shifts the standard towards risk-based thinking and statistical rigour, and in several places it becomes more prescriptive.
Risk Management, Now Integrated Throughout
The clearest change is the deeper integration of risk management across the investigation lifecycle. The new edition expects an investigation-specific risk analysis with a dual categorisation, separating risks that arise from the investigation procedure from risks that arise from the device itself, and it expects that analysis to be documented and maintained throughout the study rather than treated as a one-off exercise at the design stage.
Clinical Events Committees
For the first time, the standard formally addresses Clinical Events Committees (CECs). A CEC is an independent group of clinical experts, established by the sponsor, that classifies and adjudicates clinical events consistently across all investigation sites. The aim is to reduce the variability that creeps in when different investigators interpret and categorise events differently, which in turn improves the reliability of the data.
Data Monitoring Committees
The treatment of Data Monitoring Committees (DMCs) is tightened. Where a DMC is in place, the sponsor must now pre-define and document, in the Clinical Investigation Plan, the specific conditions under which the investigation would be suspended or stopped. That removes ambiguity about when and why a study should halt, and it puts those decisions on the record before the data starts to arrive.
An Estimand Framework
The new edition introduces an estimand framework, borrowing from pharmaceutical clinical trial practice. An estimand is a precise, pre-specified description of the treatment effect the investigation is designed to measure, stated clearly enough that everyone agrees on what is being measured, in whom, under what conditions, and how complications and missing data will be handled. In practice this pushes sponsors to define the precise clinical question up front, with statistical discipline, rather than settling it after the fact.
What Manufacturers Should Do
If you have a clinical investigation in design or about to start, this is the edition to build it on. Investigations planned against the 2020 version will need to be reviewed against the new requirements, particularly the risk analysis structure, the committee arrangements, and the estimand definition, since these shape the protocol and the Clinical Investigation Plan rather than just the paperwork around them. Do not assume a long grace period. ISO editions take effect on publication, and the value of aligning early is that your protocol and statistical plan are built to the current standard from the outset rather than retrofitted.
If you would like help reviewing a Clinical Investigation Plan against EN ISO 14155:2026, or working out how the new risk and estimand requirements apply to your study, that is the kind of work Specculo handles with manufacturers. Get in touch if a review would help.