Nature Journal Sets a High Bar for Clinical Trial Transparency
Clinical research submissions to npj Materials Degradation are being filtered through a strict editorial lens that places registration, protocol fidelity, and data transparency at the center of acceptance. The journal’s clinical research policy makes clear that any study prospectively assigning people or groups to an intervention must be treated as a clinical trial, whether the intervention involves a drug, device, procedure, behavior change, education program, diet, quality improvement effort, or process-of-care change. It also requires that the trial number appear plainly in both the abstract and the methods section. Primary results from unregistered trials will not be considered.
The framing is broader than many investigators may expect. Health outcomes are not limited to hard clinical endpoints; they include pharmacokinetic measures and adverse events as well. In practice, that means an early-phase study built around safety, exposure, or biologic response still falls inside the journal’s clinical trial rules if participants are assigned prospectively to an intervention. The policy places the burden on authors to prove that the study was planned, registered, and reported with enough transparency for outside readers to reconstruct the logic of the trial.
Registration Before Enrollment Is Not Optional
The registration requirement is anchored to guidance from the ICMJE. Before the first participant is enrolled, investigators must register their trial in an acceptable registry, either one of the WHO primary registries within the International Clinical Trials Registry Platform or ClinicalTrials.gov. The policy also expects compliance with the minimum 24-item trial registration dataset. If a study is placed in another registry, it must also be entered into a WHO-approved registry.
This is not administrative housekeeping. Trial registration is what limits post hoc endpoint switching and selective reporting after the results are known. By locking the study’s core design features into a public registry before enrollment begins, the journal pushes investigators to preserve the original hypothesis, the outcome hierarchy, and the intended analytic plan. For readers, that protects the interpretation of a positive signal from appearing more certain than the protocol actually supports.
A prespecified interim analysis is permitted, but only if it appears in the protocol from the start. Unplanned looks at accumulating data are treated much more cautiously and are judged case by case. Post hoc exploratory analyses can still be published, but they must be labeled as such in both the abstract and the manuscript text. In a field where early signals can rapidly influence clinical practice, that distinction can prevent premature certainty from hardening into false confidence.
Protocol Fidelity and Statistical Plans Drive Review
Every clinical trial submission must be accompanied by the most recent protocol in English, signed and dated. If amendments were made, those revisions must be supplied along with a summary and the rationale for the changes. When a separate statistical analysis plan exists, it must be included as well, and it has to be detailed enough that another analyst could replicate the work in a similar dataset. Machine-readable versions are required whenever the original documents are in another language.
The editorial expectation is blunt: the primary endpoint analysis in the manuscript should match the prespecified analysis in the final protocol and SAP, except in unusual circumstances. Any departure from the registry entry or protocol must be explained in the Methods section. Reviewers receive the protocol and SAP alongside the manuscript, which sharply limits the ability to disguise an analytic detour as routine scientific judgment.
In a tightly reported trial, this matters because the statistical plan is not just paperwork; it is part of the trial’s scientific architecture. Endpoint definitions, censoring rules, subgroup handling, and missing-data assumptions can all change the apparent strength of a treatment effect. The journal’s policy treats those details as core evidence, not background material.
Reporting Standards Extend Beyond the Primary Outcome
The submission guidance also specifies what belongs in the main manuscript, not merely in supplementary files. For trial reports, editors recommend including the registration number, trial type, primary endpoints, a summary of safety data, and a statement on whether prespecified endpoints were met. The main text should also contain a CONSORT-style flow diagram, a table of baseline characteristics, adverse event reporting, and exact dates for first and last enrollment plus the date of data lock.
Those requirements are clinically important because they let readers assess how much of the study population remained in the analysis, whether the groups were balanced at baseline, and whether safety signals were distributed evenly or concentrated in one arm. In a trial of any therapeutic class, the adverse event profile can reshape the clinical meaning of the primary outcome. A drug that moves an efficacy endpoint but generates a troublesome toxicity burden may still be usable in a narrow niche, yet the bedside trade-off cannot be evaluated if harms are hidden in abbreviated reporting.
The journal also aligns its expectations with study design. Randomized trials should follow CONSORT 2025, cluster trials the appropriate cluster extension, non-randomized trials the CONSORT framework as far as possible, and AI trials CONSORT-AI. Observational studies are expected to follow STROBE; systematic reviews and meta-analyses, PRISMA; biomarker-outcome studies, STARD or a relevant extension; oncology biomarker studies, REMARK; prediction models, TRIPOD or TRIPOD-AI; and causal observational studies emulating trials, TARGET.
Data Sharing Has Become Part of the Manuscript Itself
A Data Availability Statement is mandatory for all original research manuscripts, and the journal says the statement must explain what data are available, when and to whom, how access can be obtained, and what restrictions apply. If custom code was used, a separate Code Availability Statement is required. Supporting source data and code must be available to editors and peer reviewers on request during evaluation.
The policy goes further for clinical trials that began enrolling participants on or after 1 January 2019. Those studies must include a data sharing statement in the manuscript. The requirement reflects a simple reality: if the data cannot be audited, the confidence attached to the conclusions weakens. Transparent access to source data and code gives reviewers a way to examine whether the published estimates follow from the actual trial record or from a selective analytic pathway.
Sensitive material may be redacted when necessary, but information relevant to design, analysis, and outcomes must remain visible. Editors reserve the right to request unredacted versions of the protocol and SAP if needed. That safeguard is especially relevant when proprietary details, complex endpoints, or post hoc adjustments create room for ambiguity.
What Investigators Need to Prove Before Review Begins
The policy sets a demanding threshold for authors who want a clinical manuscript to move smoothly through editorial assessment. Registration must precede enrollment. The protocol must be current, dated, and available. The analysis plan must be reproducible. Outcome reporting must match the prespecified structure. Data access must be explained clearly enough that a reviewer can judge what was measured, how it was analyzed, and whether the conclusions follow the dataset.
For investigators, the practical consequence is clear: a clinical trial submission is no longer judged only by whether the intervention appears effective. It is judged by whether the research process itself can be trusted. Missing protocol details, vague reporting of adverse events, or a mismatch between the registry and the manuscript can stall review before the scientific question is even weighed on its merits.
The policy does not soften that standard for convenience, and it does not treat transparency as an optional supplement. The editorial bar is set at full traceability from registry entry to final analysis, with data access and prespecified methods functioning as the backbone of review.