Registered Reports in Clinical Research: A 2026 Guide for Medical Authors

Clinical medicine has known for decades that published research skews positive. Journals have historically been far more likely to accept studies showing that a treatment works than studies showing it does not, which means the literature accumulates an optimistic signal that does not always match what a drug or intervention does in broader clinical practice. Registered Reports were designed specifically to correct that skew by moving the editorial decision before data collection rather than after. The format has now been adopted by over 300 journals worldwide, and in 2026 Nature expanded its Registered Reports offering to cover all research disciplines, including clinical and translational research areas the journal had not previously considered. Despite this momentum, fewer than 1% of journals indexed in MEDLINE currently offer the format, and the gap between where Registered Reports are available and where most medical researchers actually publish remains wide.

That gap is narrowing, though slowly. A small but growing set of medical journals now accept Stage 1 submissions, and researchers who understand how the format works are better positioned to use it strategically, both to strengthen their own work and to navigate journals that ask about preregistration practices without offering a formal Registered Report pathway. This guide covers the mechanics of both stages, the journals currently accepting medical Registered Reports, the structural reasons medicine has been slow to adopt the format, and the practical alternatives available when your target journal does not offer it.

What a Registered Report Actually Is

The format was introduced in 2013 at the journal Cortex by Christopher Chambers, a neuroscientist who saw how publication bias was distorting the literature in his own field. The mechanism was straightforward: submit the introduction and methods to a journal before collecting data, have those sections reviewed on their merits, and receive a conditional acceptance (called an in-principle acceptance) before beginning the work. The final paper is then submitted after data collection and reviewed for compliance with the pre-approved protocol, not for the favorability of the outcome.

From a structural standpoint, Registered Reports differ from clinical trial registration, which has been standard practice in medicine since the ICMJE made it a condition of publication in 2004. Clinical trial registration records your protocol in a public database before enrollment begins. Registered Reports go a step further by having that protocol formally peer-reviewed and conditionally accepted by the target journal before you begin. The distinction matters because registration does not guarantee publication; in-principle acceptance does.

It also matters because registration in a clinical trial database like ClinicalTrials.gov or ISRCTN does not involve external expert scrutiny of your design and analysis plan. A reviewer at Stage 1 of a Registered Report will push back on a power calculation that seems optimistic, a primary outcome that is insufficiently specific, or an analysis plan that would not distinguish between the competing hypotheses you claim to be testing. That quality check, before data exists, is something trial registration alone does not provide.

The Publication Bias Problem That Makes This Format Worth Understanding

Publication bias is not a theoretical concern in medicine. A 2014 analysis of FDA trial data found that roughly half the studies submitted in drug approval applications went unpublished, and the unpublished studies were disproportionately those with negative or inconclusive results. A subsequent decade of work on outcome switching, selective outcome reporting, and HARKing (Hypothesizing After Results are Known) reinforced the same message: the published medical literature overstates how well treatments work and understates how often they fail to show an effect.

These problems do not stem primarily from outright fraud. Most arise from ordinary decisions made by researchers and editors under ordinary pressures. A researcher who ran six secondary analyses selects the two that reached significance, not out of dishonesty but because space is limited and reviewers expect tidy stories. An editor returns a trial showing no effect with a brief note that the journal is looking for more definitive conclusions. Registered Reports address these patterns not by detecting or punishing them after the fact, but by removing the incentive to engage in them in the first place.

A growing evidence base supports the claim that the format works. Studies comparing outcome distributions of Registered Reports against conventional articles in the same journals have consistently found that Registered Reports produce null or mixed results at much higher rates than standard publications. This does not mean Registered Reports are less rigorous than standard papers. It means they give a less distorted picture of what the research actually found.

How Stage 1 Peer Review Works

Stage 1 is the distinctive element of the format. You submit a manuscript that includes the introduction, the theoretical or clinical rationale for the study, a precise formulation of the research questions and hypotheses, a detailed description of the methods and materials, the sampling plan and power calculation, and a fully pre-specified statistical analysis plan. No results, because no data has yet been collected.

Reviewers at Stage 1 are evaluating whether the research question is important, whether the design is capable of answering it, whether the analysis plan is appropriate to the data that will be collected, and whether the power calculation is realistic. They are not evaluating the novelty of findings, because there are none. This shift in reviewer focus is one of the format's genuine strengths: it catches design problems at the stage when they can still be corrected, before you have committed resources to a methodologically flawed study.

Stage 1 can involve multiple rounds of revision. Reviewers may ask for a more precisely stated primary outcome, a more detailed description of blinding procedures, or a clearer explanation of how missing data will be handled. These are legitimate requests, and the process typically takes two to four months, comparable to the timeline for standard peer review of a completed study. Authors who have gone through Stage 1 frequently report that it improved their research design in ways that benefited the final study, independent of any effect on publication outcome.

In-Principle Acceptance and What It Commits You To

If Stage 1 review is successful, the journal issues an in-principle acceptance (IPA). This is a conditional commitment. The journal is agreeing to publish the final paper provided you conduct the study as described, report the pre-specified analyses, and document any deviations from the original protocol with clear explanations. The IPA is not a blank check. It does not mean you can alter the primary outcome after seeing the data, switch from a superiority to a non-inferiority framework once results come in, or quietly drop a study arm that produced inconvenient findings.

What the IPA does provide is a genuine assurance that a well-designed study with a null result will be published. This changes the research environment in important ways. Researchers can pursue questions where null results are scientifically valuable, without fearing that negative findings will consign months of work to a file drawer. For clinicians studying treatments that may not work, or researchers trying to replicate prior findings before building on them, the IPA represents a meaningful shift in the economics of research effort.

Once the IPA is issued, most journals agree a submission deadline for Stage 2 with the authors. For short laboratory or survey studies, Stage 2 might be expected within six to twelve months. For clinical trials with longer follow-up periods, timelines of several years are generally accommodated, provided the authors communicate progress and notify the journal of any protocol amendments that arise from ethical review.

Stage 2: From Data Collection to Publication

The Stage 2 submission is the full manuscript: introduction, complete methods, results, and discussion. Reviewers at Stage 2 are checking primarily that the study was conducted as described in the approved protocol and that the pre-specified analyses are reported fully and accurately. If the study deviated from the Stage 1 protocol, because of participant dropout, an amended ethics approval, a recruitment challenge, or an unforeseen methodological issue, those deviations must be transparently described and explained.

Exploratory analyses that were not pre-specified are allowed at Stage 2, but they must be clearly distinguished from the confirmatory analyses that were pre-specified at Stage 1. Standard practice is to label them explicitly in the results section, typically as post-hoc or exploratory, and to interpret them with appropriate caution rather than treating them as hypothesis tests. This separation is one of the format's intellectual contributions: it makes the distinction between confirmatory and exploratory reasoning visible in the paper itself, rather than leaving readers to guess which analyses were planned and which emerged from the data.

Stage 2 review is generally faster than standard peer review, because reviewers are checking protocol compliance rather than evaluating the research question from scratch. If the study was conducted as described and the results are reported faithfully, Stage 2 rejection is uncommon. The main grounds for failure are serious undisclosed deviations from the approved protocol, concerns about data quality, or problems with how post-hoc analyses have been framed as though they were pre-specified.

Which Medical Journals Currently Offer Registered Reports

The global list of journals offering Registered Reports is maintained by the Center for Open Science. Of the more than 300 journals on that list, only a small fraction are primary medical or clinical research venues. The British Journal of General Practice (BJGP) actively solicits Registered Reports of trials and observational studies through its normal editorial system. BMC Medicine introduced a Registered Report format specifically for randomized controlled trials in 2017, making it one of the earliest medical journals to formalize the pathway. The Journal of Medical Internet Research (JMIR) and several of its specialist journals assign an International Registered Report Identifier (IRRID) number to accepted Stage 1 submissions, allowing the pre-approved protocol to be tracked in the literature independently of the Stage 2 paper.

Nature, which began publishing Registered Reports in 2023 and expanded the format to all disciplines including health and clinical research in 2026, is technically now an option for translational and clinical work, though the journal's selection criteria remain as competitive as for any other Nature submission. Nature Human Behaviour, PLOS ONE, and BMJ Open Science have all published health-related Registered Reports. PLOS Medicine has published RCTs in a registered report format on occasion, though it does not have as formalized a Stage 1 pathway as BJGP or BMC Medicine.

The honest picture is that the highest-profile clinical journals, including The Lancet, The New England Journal of Medicine, JAMA, and BMJ, do not currently offer a formal Registered Report pathway. This does not mean they are indifferent to preregistration. All four require clinical trial registration as a condition of publication, consistent with ICMJE recommendations. But formal Stage 1 review with a binding in-principle acceptance is not part of their current editorial process. Researchers aiming for those outlets should be aware of this and plan alternative strategies for demonstrating preregistration transparency.

Why Clinical Medicine Has Been Slow to Adopt

The reasons are structural as much as cultural. Clinical trials are governed by Good Clinical Practice (GCP) requirements, ethics committee review, and regulatory approval processes that typically precede data collection by months or years. IRB approval in many institutions happens well before any journal submission would be contemplated, which creates a timing mismatch: by the time researchers submit a Stage 1 protocol to a journal, ethics approval may already exist, study materials may be finalized, and recruitment may be days from beginning. The Stage 1 review process, which can take two to four months, does not fit easily into that sequence.

There are also editorial capacity concerns on the journal side. Stage 1 review requires reviewers willing to evaluate a study that has not yet produced data, which means the reward of encountering novel findings is absent. Reviewer motivation for a Stage 1 submission depends on genuine enthusiasm for the research question and design, rather than the usual draw of being among the first to assess new results. Medical journals with already-strained reviewer pools have been understandably cautious about adding this different kind of commitment.

The format works best for research where data collection has not yet begun and the timeline is flexible. Many clinical researchers are working with existing datasets, conducting secondary analyses of ongoing trials, or running studies that have already passed ethics review. For these situations, Registered Reports at journals requiring pre-collection submission are inaccessible in practice, though alternatives exist that capture some of the same transparency benefits.

What to Do When Your Target Journal Does Not Offer Registered Reports

For researchers working toward journals that have not adopted the format, the most accessible alternative is preregistration on a public registry before data collection begins. For clinical trials, registration on ClinicalTrials.gov, ISRCTN, or another WHO ICTRP-approved registry is already a standard requirement. The additional step is to specify the primary and secondary outcomes, the statistical analysis plan, and the power calculation in the registration record before enrollment begins, rather than providing only the minimum required fields. A detailed registration is verifiable, publicly dated, and carries weight with reviewers even at journals that do not offer formal Registered Reports.

The Open Science Framework (OSF) offers a preregistration pathway for non-trial research. Researchers can upload a time-stamped preregistration document specifying hypotheses and analysis plans before collecting data. The OSF registration is publicly visible and can be cited in the methods section of the final paper. This is not equivalent to in-principle acceptance, which only a formal Registered Report provides, but it is a credible and verifiable record that the analysis plan preceded the results.

PCI Registered Reports (Peer Community In Registered Reports) offers a third option worth understanding. This is an overlay journal model in which a community of volunteer researchers conducts Stage 1 peer review independent of any specific journal. A Stage 1 recommendation from PCI RR can then be taken to affiliated journals, some of which will accept it without additional review. JMIR is among the journals that have expressed interest in PCI RR-reviewed protocols. For researchers whose work does not fit the journals with native Registered Report pathways, PCI RR widens the submission landscape considerably.

Whichever route you take, the practical implication is the same: specify your analysis plan in writing and date it before running any analyses. A time-stamped protocol does not carry the formal guarantee of an IPA, but it demonstrates to reviewers and readers that your confirmatory tests were planned rather than identified post-hoc. As open science norms tighten across medical publishing, that demonstration is increasingly expected even at journals that do not yet offer formal Registered Reports.

What Registered Reports Do Not Fix

The format is not a universal remedy for problems in medical research. Registered Reports do not fix flaws in study design that exist at Stage 1, which means a poorly conceived research question can receive in-principle acceptance and proceed to publication of a rigorous but uninformative study. Stage 1 review should catch obvious design weaknesses, but reviewers make judgments under uncertainty and errors occur in both directions. A Registered Report is not immune from methodological critique after publication.

The format also does not protect against poor execution. A study that was designed well at Stage 1 but carried out poorly, with high dropout rates, compromised blinding, or protocol violations that go undisclosed at Stage 2, can still enter the literature in a misleading form. Registered Reports shift where integrity risks are most likely to surface, but they do not eliminate those risks. That is worth bearing in mind when reading Registered Report papers, which carry a justified presumption of reduced selective reporting but not a presumption of perfect conduct throughout the study.

For medical authors evaluating whether the Registered Report format fits their current project, the most useful question is whether the work involves a genuinely prospective design where the analysis plan can be fixed before the data exists. If yes, and if a suitable journal is available, the format offers real structural advantages that standard publication cannot replicate. If the work is retrospective, involves a dataset where analyses have already begun, or is heading to a journal that does not offer the format, a detailed public preregistration on OSF or a clinical trial registry is the more realistic path to the same transparency goals.

Further Reading