Supplementary Files in Medical Journals: The Peer Review Gap Authors Need to Know in 2026

Supplementary materials were supposed to solve a specific problem. A paper about a large clinical cohort has methods too detailed for a 3,000-word article. A trial has pre-specified secondary outcomes that belong in the record but would overwhelm the results section. A statistical analysis script might help a reader reproduce findings even if most readers will never open it. The supplementary appendix was designed to hold all of that without cluttering the main narrative.

That logic is sound. What has happened in practice is somewhat different. Supplementary files now appear in roughly 40 percent of articles indexed in PubMed Central as of 2023, and they vary enormously in what they contain: raw data, additional tables, sensitivity analyses, ethics approvals, consent forms, statistical code, long questionnaires, and occasionally several hundred pages of ancillary results that few readers will ever open. A 2025 article in Theory and Society raised a pointed concern about this expansion under the title “Are supplementary materials the new file drawer?” The authors argued that supplementary appendices, when misused, offer authors a way to make information technically available while effectively shielding it from scrutiny.

The peer review dimension is the part that should concern authors most. Unlike the main manuscript, supplementary files often receive only partial or no review. Understanding that gap, and designing your supplementary materials with it in mind, is now part of writing a paper that can survive serious post-publication scrutiny.

How the Peer Review Gap Developed

For most of the twentieth century, supplementary materials barely existed in print journals. There was no mechanism for them. The shift to online publication in the late 1990s and early 2000s opened the possibility of hosting additional files alongside articles, and journals responded by allowing, and then expecting, supplementary appendices for papers with extensive methods or data. What journals did not do, at least not consistently, was update their peer review processes to match.

IOP Publishing states its position plainly in its submission guidance: supplementary files “will not be included in peer review by default.” Reviewers are invited to examine them if they wish, but the editorial decision rests on the main manuscript. Nature Portfolio journals take a similar approach, asking reviewers to note if supplementary files appear inconsistent with the main text, but not treating them as a formal review obligation. This is common across many publishers, though the exact language varies and a minority of journals do ask referees to confirm they have checked supplementary materials specifically.

The mismatch matters because academic publishing has evolved on the assumption that peer review is the primary quality check before publication. When a substantial portion of a clinical paper's methodological detail, sensitivity analyses, or additional outcome data sits in a file that reviewers are encouraged but not required to read, that assumption breaks down in ways authors may not anticipate.

Reviewers are not ignoring supplementary files out of laziness. They are working within norms they did not design. When a reviewer is given two weeks to evaluate a 4,000-word paper, has four other reviews in progress, and is told explicitly that supplementary materials are optional, the outcome is predictable. The files that authors spent weeks compiling may receive only a brief scan, or none at all.

The File Drawer Concern: When Appendices Hide Rather Than Reveal

The term “file drawer” comes from publication bias research. The classical file drawer problem describes studies with null or negative results that are never submitted or accepted, disappearing into researchers' file drawers while the published record fills with positive findings. The 2025 Theory and Society paper applied the same logic to supplementary appendices, observing that authors can use them to make inconvenient information technically accessible without making it visible.

This is not necessarily intentional. The dynamics can be subtle. A sensitivity analysis that undermines the main finding can be included in the appendix as an apparent act of transparency, while the main text presents only the primary analysis with language that downplays the discrepancy. A prespecified secondary outcome that came out negative gets listed in a supplementary table of outcomes, satisfying the requirement to report it, while the results section focuses on the endpoints that worked. A statistical model with a poor fit in one subgroup is noted in a supplementary figure that appears on page 34 of a 40-page appendix.

None of these practices necessarily violates any specific journal rule. That is what makes the pattern worth understanding. The question is not whether a piece of information appears somewhere in the published record. The question is whether it appears where readers and reviewers can evaluate it in context.

Post-publication readers, statistical reviewers, and journal integrity teams are increasingly alert to these patterns. If you are hiding information in the appendix because you are worried about how it will look in the main text, that is a signal that it belongs in the main text with a clear discussion of what it means.

What Different Journals Actually Require

Journal policies on supplementary materials differ more than most authors realize, and they are not always easy to find. Some journals publish detailed supplementary file guidance on their author instructions page. Others mention supplementary materials only in the submission system, after you have already built the manuscript. Reading the target journal's instructions carefully before you structure the paper is not optional if you want to avoid a reformatting request after submission.

PLOS ONE allows supplementary materials but asks authors to confirm that the methods and results sections of the main paper can stand alone without the supplementary files. The journal explicitly states that the paper should be interpretable without reference to the appendix. This is a meaningful constraint: it rules out the common practice of putting key methodological details in the supplementary and leaving the methods section too sparse to evaluate. PLOS Biology went further in January 2026 by requiring that analysis code be deposited in a public repository rather than uploaded as a supplementary script file, on the grounds that code in supplementary files is harder to find, use, and cite.

BMJ journals vary in their approach. Some allow supplementary materials for additional tables and data. Others, particularly for shorter article types like research letters and rapid communications, prohibit supplementary files entirely on the grounds that they expand the effective length of papers that are supposed to be concise. The New England Journal of Medicine has size limits for supplementary appendices and advises authors to restrict the appendix to material that would be too detailed for print but is genuinely needed to understand the study design or statistical approach.

The Lancet family of journals requires that supplementary materials be mentioned in the main text at the point they become relevant, a policy that at least ensures readers are told the files exist. JAMA Network journals ask that tables and figures in the supplementary appendix be numbered separately from the main paper and that each be cited in the main text. These citation requirements do not guarantee that reviewers will read the supplementary files, but they do prevent the files from being structurally invisible.

What Belongs in Supplementary Materials (and What Does Not)

The most useful heuristic for deciding what goes in the appendix is the interpretability test: can a reader who has not seen the supplementary files understand the study design, evaluate the analysis, and assess whether the conclusions are supported by the results? If the answer is no, something that belongs in the main paper is being pushed to the appendix.

Materials that genuinely belong in supplementary files include: full versions of instruments or questionnaires where only a summary appears in the methods; extended data tables with subgroup breakdowns that support but do not drive the primary finding; pre-specified analysis code deposited for reproducibility; ethics committee approvals and consent form templates that are required for the record but do not affect how a reader interprets the results; and detailed derivations or technical proofs for statistical methods already summarized in the main paper.

Materials that belong in the main paper, not the appendix, include: the primary outcome definition and its operationalization; any sensitivity analysis whose result diverges meaningfully from the main analysis; the rationale for sample size; a description of missing data and how it was handled; any covariate adjustment that influenced the primary estimate; and limitations that a reviewer would need to know about in order to evaluate the study's validity. When in doubt, keep it in the main text. Moving something to the appendix does not make it less important. It makes it less likely to be read.

One area where authors frequently misjudge this is CONSORT compliance for clinical trials. CONSORT 2025 requires reporting of specific items in the main paper. Putting a CONSORT-required item only in the supplementary appendix and then pointing to it in the main text with a sentence like “full details in Supplementary Appendix A” is not the same as reporting it. Reviewers familiar with the checklist will flag this, and journals that require a completed CONSORT checklist at submission will notice it.

How to Structure Supplementary Files So Reviewers Actually Use Them

If you have materials that genuinely belong in a supplementary appendix, structure them so they can be navigated by someone who has limited time. A 60-page PDF of unsorted tables, figures, and text blocks, with no internal headers and no table of contents, functions for most reviewers as a black box. They may note that supplementary materials are “extensive” in their review and move on.

A supplementary appendix with a one-page table of contents, clearly labeled sections that mirror the structure of the main paper, and numbered items that correspond to in-text citations in the main manuscript is a different thing. It reads as organized, deliberate, and honest. A reviewer who decides to check a specific analysis can find it in two minutes rather than ten.

For analysis code, a standalone script file buried in a supplementary archive is increasingly the wrong answer. Many journals now prefer or require that code be deposited in a public repository such as GitHub, Zenodo, or Open Science Framework, with the DOI or URL cited in the data availability statement. This approach gives the code a persistent, citable location, makes it easier for readers to find, and keeps the supplementary appendix from becoming an unnavigable archive of heterogeneous files.

The Push for Better Supplementary File Standards

The research community has begun producing tools and frameworks aimed at the supplementary file problem. In October 2025, PLOS Biology published a paper introducing FAIR-SMART, a system for making supplementary files computationally accessible in a way that mirrors the FAIR data principles (findable, accessible, interoperable, reusable) applied to primary datasets. The paper documented that 40 percent of PubMed Central articles published in 2023 carry at least one supplementary file, and observed that the lack of consistent formats and metadata has limited how findable and usable those files actually are. FAIR-SMART generates structured metadata and persistent identifiers for supplementary files so they can be indexed, cited, and retrieved separately from the main paper.

Whether FAIR-SMART or a similar approach will be adopted by major medical publishers remains to be seen. The incentives are not perfectly aligned: journals benefit from supplementary materials as a way to manage word-count constraints without turning away methodologically rich submissions, and redesigning the submission and production workflow to accommodate structured supplementary metadata requires real investment. But the direction of travel is toward more structure, not less. Authors who get into the habit of treating supplementary materials as a rigorous extension of the main paper, rather than as a holding area for overflow, will be better positioned as standards tighten.

COPE, the Committee on Publication Ethics, has not yet issued formal guidance specific to supplementary materials, though its broader retraction and correction guidelines apply to errors that originate there. The ICMJE recommendations, updated in January 2026, address data availability and deposition but do not specifically define standards for what goes in supplementary appendices versus public repositories. Both organizations have signaled awareness of the issue through working group discussions, and guidance is likely to emerge over the next reporting cycle.

Post-Publication Scrutiny and Supplementary Files

The supplementary file peer review gap does not close at publication. Post-publication readers, particularly statistical reviewers, systematic reviewers, and research integrity specialists, are more likely than casual readers to work through supplementary materials in detail. A growing number of post-publication errors and integrity concerns originate in discrepancies between the main paper and the supplementary appendix: effect sizes that appear in both places but do not match, sample sizes that change between sections without explanation, or sensitivity analyses that contradict the stated primary conclusion.

This matters practically because corrections and expressions of concern based on supplementary file discrepancies carry the same reputational costs as errors in the main paper. When a post-publication commenter identifies that a sensitivity analysis buried on page 22 of the appendix shows no significant effect while the abstract claims a significant result, the correction process is the same whether the error was in the main text or the appendix. The defense that “the information was available in the supplementary files” is not a strong one.

It is worth applying the same internal consistency check to your supplementary files that you apply to the main paper before submission. Run all numbers against the main text. Confirm that table headings match the variable names used in the methods. Make sure the sample size in every supplementary table matches the sample size in the results section. These checks take time but they prevent the kind of post-publication scrutiny that is difficult to explain even when the discrepancy was genuinely unintentional.

What the Evolving Landscape Means for Authors Right Now

The supplementary materials conversation is at an early stage compared to where data sharing or AI disclosure norms are now. There is no equivalent of the ICMJE data availability statement requirement or the JAMA Network AI use prohibition that applies specifically to supplementary file conduct. What exists is a growing body of scholarship pointing out that the current arrangements are insufficient, a set of publisher policies that vary widely, and an emerging tool ecosystem aimed at making supplementary files more structured and citable.

For authors, that means the decisions about what goes in the supplementary appendix are still largely yours to make, with the journal's author guidelines as the primary constraint. But those decisions have real consequences for how thoroughly your methods are reviewed, how easily post-publication readers can evaluate your analysis, and how well your paper holds up when someone works through the appendix carefully.

The most durable approach is to write supplementary materials the way you would write anything else that carries your name: clearly, honestly, and with the reader in mind. If a sensitivity analysis matters, describe it in the main paper and put the full output in the appendix. If a methodological detail is complex, summarize it accessibly in the methods section and deposit the full specification somewhere citable. If a secondary outcome was prespecified and came out null, it belongs in the results section with appropriate framing, not only in a supplementary table where the absence of commentary does the work of omission without technically omitting anything.

Supplementary materials are not a problem to be managed. They are part of the published record. Writing them with the same care as the main manuscript costs time up front and saves considerably more later, both in revision cycles and in the kind of post-publication attention no author wants.

Further Reading