CASRAI Dictionary

Tag: reporting guidelines

  • PRISMA: The 2020 Reporting Standard for Systematic Reviews and Meta-Analyses

    PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) is an evidence-based reporting standard that tells authors what to disclose when they publish a systematic review. The current version, PRISMA 2020 (Page et al.), pairs a 27-item checklist with a four-phase flow diagram so that readers can judge a review’s methods, completeness and reproducibility.

    What PRISMA is — and what it is not

    PRISMA is a reporting guideline, not a method for conducting a review. It does not tell you how to design a search strategy, appraise risk of bias or pool effect sizes; it tells you how to report the work you did so another team could appraise or repeat it. That distinction matters. A methodologically weak review can still be PRISMA-compliant if it reports its weaknesses transparently, and a strong review can fail PRISMA if it omits required items. PRISMA sits alongside conduct frameworks such as the Cochrane Handbook, which governs how the review is performed.

    The 2020 update

    The original PRISMA Statement appeared in 2009. The 2020 revision modernised it to reflect advances in synthesis methods and terminology over the intervening decade: clearer guidance on reporting search strategies for every database, automation tools, study selection, and methods for assessing certainty of evidence such as GRADE. The flow diagram was redrawn to accommodate reviews that update an earlier search and to record records identified through methods other than database searching, such as citation chasing and contact with experts.

    The checklist and the flow diagram

    The checklist’s 27 items map onto the standard structure of a review — title, abstract, introduction, methods, results, discussion and other information. Each item names a specific thing to report: the registration and protocol, eligibility criteria, the full electronic search strategy, the synthesis methods, and any deviations from the protocol. PRISMA 2020 also provides an expanded abstract checklist so that even readers who never reach the full text can assess the review’s scope.

    The flow diagram is PRISMA’s most recognisable artefact. It tracks records through four phases — identification, screening, eligibility and inclusion — and reports, at each stage, how many records were found, how many were removed as duplicates, how many were excluded and why. A complete diagram makes the path from thousands of search hits to a handful of included studies auditable.

    What PRISMA standardises

    Section Representative reporting items
    Methods Eligibility criteria, information sources, full search strategy, selection process, data items, risk-of-bias assessment, synthesis methods
    Results Study selection (flow diagram), study characteristics, risk-of-bias results, results of syntheses, certainty of evidence
    Other Registration and protocol, support and funding, competing interests, availability of data and code

    Who maintains PRISMA

    PRISMA is one of the flagship guidelines curated by the EQUATOR Network, the international initiative that catalogues reporting guidelines across study designs. EQUATOR hosts the checklist, the explanation-and-elaboration document and the various PRISMA extensions — for scoping reviews, network meta-analyses, abstracts, individual-patient-data reviews and search reporting. Because EQUATOR maintains the canonical record, authors should download the current materials there rather than copying older versions from secondary sources.

    For a definition-first reference to PRISMA and related terms, see the CASRAI dictionary, and read our overview of reporting guidelines and the EQUATOR Network. Authors preparing a manuscript can also consult our guidance for authors and our look at next-generation systematic reviews.

    Frequently asked questions

    Is PRISMA only for meta-analyses?

    No. PRISMA applies to systematic reviews whether or not they include a meta-analysis. The statistical pooling step is optional; the structured, transparent reporting of the review process is not. A qualitative or narrative synthesis still follows PRISMA. See our explainer on the difference between a systematic review and a meta-analysis.

    Does following PRISMA guarantee a high-quality review?

    No. PRISMA improves transparency, which makes quality assessable, but it cannot fix flawed methods. Critical-appraisal and risk-of-bias tools serve that purpose. Complete reporting and sound conduct are complementary, not substitutes.

    Where should a review be registered?

    PRISMA asks authors to report registration. Many reviewers register a protocol in a public registry before screening begins, and PRISMA 2020 expects the registration number and the protocol’s availability to be stated so reviewers and editors can check for deviations.

    What is the difference between PRISMA and CONSORT?

    PRISMA governs the reporting of systematic reviews; CONSORT governs the reporting of individual randomised controlled trials. Both are EQUATOR guidelines, but they apply at different points of the evidence pipeline.

  • The STROBE Statement for Observational Epidemiology

    The STROBE Statement, short for Strengthening the Reporting of Observational Studies in Epidemiology, is a reporting guideline that specifies the information an observational study should include so that readers can judge its validity and reuse its findings. STROBE is a checklist for authors, reviewers and editors; it does not dictate how a study should be designed or analysed, only what must be reported transparently once the work is done.

    Observational studies make up a large share of epidemiological evidence, yet they cannot randomise who is exposed to what. Their credibility therefore rests unusually heavily on the clarity of their reporting, and STROBE exists to make that clarity a shared, citable expectation rather than a matter of individual habit.

    What STROBE covers

    The STROBE checklist enumerates items spanning the standard sections of a research paper: title and abstract, introduction, methods, results and discussion. The methods items are central, asking authors to describe the study setting, eligibility criteria, the variables and how they were measured, the data sources, efforts to address potential sources of bias, how the study size was determined, how confounding was handled, and the statistical methods used. The results items ask for the flow of participants through the study, descriptive data on the participants, and the main estimates reported with measures of uncertainty such as confidence intervals. Several items deal specifically with how the study handled missing data, how continuous variables were grouped or modelled, and whether any sensitivity analyses were performed, because each of these choices can materially change a result and each is easy to leave undescribed. A short, structured abstract is also expected, so that readers scanning the literature can grasp the design, population and main findings before reading the full text.

    The aim is completeness rather than a particular conclusion. When every relevant item is reported, a reader can assess whether the conclusions are supported by the design and data, and another team can attempt to reproduce the work or pool it in a synthesis. This emphasis on transparent, reusable reporting aligns directly with the wider goals of reproducibility in research, where undocumented methods are a primary obstacle to replication.

    Three observational study designs

    STROBE is written to cover the three principal observational designs, with a common core checklist plus design-specific guidance where the designs diverge.

    Design How it observes Typical question
    Cohort Follows groups over time by exposure status What outcomes follow an exposure?
    Case-control Compares exposure histories of cases and controls What exposures preceded an outcome?
    Cross-sectional Measures exposure and outcome at one point What is associated at a given time?

    Because these designs differ in how data are gathered and in the biases they are prone to, certain checklist items are tailored to each. A case-control study, for example, must report carefully how cases and controls were selected, while a cohort study must report how participants were followed and how losses to follow-up were handled. Reporting which design was used, and reporting it accurately, is itself a STROBE requirement and a prerequisite for sound interpretation, since the same association means different things under different designs.

    Extensions and related reporting tools

    The core STROBE checklist has been supplemented by extensions that address particular fields and data types while keeping the same philosophy of transparent reporting. These extensions adapt the checklist for areas such as molecular and genetic epidemiology, nutritional epidemiology, infectious-disease studies and research that reuses routinely collected health data, where additional reporting items are needed to let readers judge validity. The proliferation of extensions reflects a general principle: the more specialised or complex the data source, the more there is to report before a study can be appraised or reproduced. Authors should check whether an extension exists for their study type, because using the most specific applicable guideline captures reporting items that the generic checklist would miss. This mirrors the broader move in research toward documenting not just results but the full provenance of the data and analysis behind them.

    STROBE and the EQUATOR Network

    STROBE is one of the most widely used guidelines hosted by the EQUATOR Network, an international initiative that curates reporting guidelines to improve the reliability and value of the health research literature. EQUATOR organises guidelines by study type, so STROBE sits alongside guidelines such as CONSORT for randomised trials and PRISMA for systematic reviews. Locating a guideline through EQUATOR helps authors choose the correct checklist for their study type rather than reaching for a familiar but inappropriate one.

    Within an evidence ecosystem, having a named, citable reporting standard makes expectations explicit for everyone involved. It also connects observational studies to the population measures they rely on, such as incidence and prevalence and the denominators drawn from a census. Consistent terminology drawn from the CASRAI dictionary further helps keep the language of reporting stable across studies and journals.

    How STROBE is used in practice

    In practice, STROBE is most useful when it is consulted at the writing stage and again at peer review. Authors typically complete the checklist and indicate, for each item, the page or section where it is addressed, submitting this alongside the manuscript so that editors and reviewers can verify coverage quickly. Many journals reference STROBE in their instructions to authors for observational research, which gives the guideline practical force rather than leaving it as an optional ideal. Importantly, STROBE is not a quality score: a study can be reported completely yet still have design limitations, and conversely a strong study reported poorly is hard to trust. The checklist’s role is to ensure that whatever the study did, the reader can see it clearly. Used this way, it improves the appraisal, synthesis and reuse of observational evidence without constraining how researchers choose to investigate their questions.

    Why transparent reporting matters

    Observational studies cannot randomise exposure, so their credibility rests heavily on how clearly the methods, data sources and analytical choices are reported. Incomplete reporting makes it impossible to judge whether bias or confounding could explain the findings, and it makes the work difficult or impossible to reproduce, which weakens the cumulative evidence base. Transparent, STROBE-compliant reporting supports critical appraisal by readers, enables evidence synthesis by reviewers, and allows secondary analysts to reuse the work with confidence. Authors preparing observational manuscripts can consult the guidance for authors to align their reporting with these expectations from the outset rather than retrofitting it at submission.

    Frequently asked questions

    Is STROBE a way to design a study?

    No. STROBE is a reporting guideline, not a design or analysis protocol. It tells authors what to report so readers can evaluate the work; the design and statistical choices remain the researchers’ responsibility and are made before STROBE applies.

    Which studies should use STROBE?

    STROBE applies to observational designs, specifically cohort, case-control and cross-sectional studies. Randomised trials, systematic reviews and other study types have their own guidelines, which can be located through the EQUATOR Network’s catalogue.

    How does STROBE relate to reproducibility?

    Complete, transparent reporting is a precondition for reproducibility. By prompting authors to describe data sources, variables, bias and methods fully, STROBE makes it possible for others to appraise, synthesise and attempt to replicate observational findings.

  • The CONSORT Statement: Reporting Standards for Randomised Controlled Trials

    CONSORT (Consolidated Standards of Reporting Trials) is the international reporting guideline for randomised controlled trials. It comprises a checklist of essential items and a participant flow diagram so that a trial report contains everything a reader needs to assess its conduct, validity and applicability — and so that another team could, in principle, reproduce the study.

    What CONSORT is for

    Randomised controlled trials sit near the top of the hierarchy of evidence because randomisation balances confounders. But that strength is only realisable if the trial is reported in full: an unreported deviation, a hidden change of primary outcome or an unexplained loss of participants can quietly undermine the randomisation. CONSORT exists to make the trial transparent. Like PRISMA, it is a reporting standard, not a method for designing or running a trial.

    The checklist

    The CONSORT checklist covers the trial report end to end — title and abstract, introduction, methods, results, discussion and other information. Its items target precisely the places where trials are most often under-reported:

    Section Representative items
    Methods Trial design, eligibility, interventions, primary and secondary outcomes, sample-size calculation, randomisation sequence generation, allocation concealment, blinding
    Results Participant flow diagram, recruitment dates, baseline data, numbers analysed, outcomes and estimation with effect size and precision, harms
    Other Registration, protocol availability, funding

    Several items deserve emphasis. Reporting how the random sequence was generated, how allocation was concealed and who was blinded lets readers judge whether the randomisation was protected from subversion. Pre-specifying the primary outcome and reporting it — alongside trial registration — guards against outcome-switching. Stating the sample-size calculation shows whether the trial was adequately powered.

    The flow diagram

    CONSORT’s flow diagram tracks participants through four stages — enrolment, allocation, follow-up and analysis. It records how many were assessed for eligibility, how many were randomised to each arm, how many received the intended intervention, how many were lost to follow-up and how many were included in the analysis. Because attrition and exclusions can bias results, a complete diagram is one of the most informative parts of any trial report.

    CONSORT in the EQUATOR family

    CONSORT is one of the foundational guidelines maintained through the EQUATOR Network, the same initiative that curates PRISMA for systematic reviews and STROBE for observational studies. The EQUATOR philosophy is design-matched reporting: use CONSORT for randomised trials, STROBE for cohort and case-control studies, PRISMA for reviews. CONSORT also has a family of extensions for specific trial types — cluster trials, pilot and feasibility trials, non-inferiority designs, abstracts and harms reporting — so authors should select the relevant extension alongside the core checklist.

    Why complete reporting matters

    Incomplete trial reporting is not a cosmetic problem. It wastes research investment, frustrates the systematic reviews that depend on extractable data, and can mislead decisions. CONSORT-adherent reporting feeds directly into the meta-analyses that synthesise trials, which is why journals across many fields endorse it. For the wider context, see our overview of reporting guidelines and EQUATOR, the CASRAI dictionary and our reproducibility coverage.

    Frequently asked questions

    Is CONSORT mandatory?

    CONSORT is endorsed by a large number of journals and editorial bodies, which often require adherence as a condition of submission. It is a community standard rather than a law, but for randomised trials it is the expected reporting framework.

    How does CONSORT differ from PRISMA?

    CONSORT governs the reporting of an individual randomised controlled trial; PRISMA governs the reporting of a systematic review or meta-analysis that may synthesise many trials. Both are EQUATOR guidelines but apply at different stages of the evidence pipeline.

    Does CONSORT cover trial registration?

    Yes. CONSORT asks authors to report the trial registration number and where the full protocol can be accessed, which lets reviewers check the reported outcomes against the pre-specified ones.

    Which CONSORT extension should I use?

    Select the extension that matches your design — for example the cluster-trial or pilot-and-feasibility extension — in addition to the core checklist. The EQUATOR library lists the current extensions; our author guidance points to them.

  • Reporting guidelines and the EQUATOR Network: CONSORT, PRISMA and beyond

    A study can be impeccably designed and still be impossible to trust, for a simple reason: if the paper does not say what was actually done, no reader can judge whether to believe it, and no one can reproduce it. Under-reporting is one of the quietest threats to reproducibility — not fraud, not bad method, just the omission of the details that would let someone evaluate the work. Reporting guidelines exist to fix this by specifying, item by item, what a paper of a given type must contain. They are a central concern of the reproducibility domain and connect directly to the research-integrity domain, because honest, complete reporting is the precondition for both.

    The EQUATOR Network: a library of guidelines

    The coordinating body for reporting guidelines is the EQUATOR Network — Enhancing the QUAlity and Transparency Of health Research. EQUATOR does not publish a single guideline; it curates and promotes the whole family of them, maintaining a comprehensive online library of reporting guidelines for different study types and providing resources to help authors, editors, and reviewers use them. Its premise is straightforward: there is no point in conducting rigorous research if the publication that reports it omits the information readers need to assess and use it. EQUATOR’s role is to make the right guideline easy to find and to raise the baseline quality of reporting across health research.

    The key thing to understand is that different study designs need different guidelines, because each design has its own ways of being under-reported. A randomised trial can hide how participants were allocated; a systematic review can hide how studies were selected; an animal study can hide how many animals were used and why. A guideline is a checklist tuned to the failure modes of a particular design.

    The major guidelines

    A handful of guidelines account for most everyday use, and it is worth knowing which applies to what.

    • CONSORT (Consolidated Standards of Reporting Trials) is for randomised controlled trials. Its checklist and accompanying flow diagram cover how participants moved through the trial — enrolment, allocation, follow-up, analysis — along with the method of randomisation, blinding, and the pre-specified outcomes. CONSORT is what makes it possible to see whether a trial’s reported result matches the trial it set out to run.
    • PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) is for systematic reviews and meta-analyses. Its checklist and flow diagram document the search strategy, the inclusion and exclusion criteria, and how records were screened and selected. PRISMA is what lets a reader judge whether a review’s conclusions rest on a fair and reproducible survey of the evidence rather than a convenient selection.
    • ARRIVE (Animal Research: Reporting of In Vivo Experiments) is for research involving animals. It sets out the details — species, number and characteristics of animals, housing, the experimental procedures, the statistical methods — needed both to evaluate the work and to honour the ethical principle that animal studies should be reportable and not needlessly repeated.
    • STROBE (Strengthening the Reporting of Observational studies in Epidemiology) is for observational studies — cohort, case-control, and cross-sectional designs — where the reporting challenges differ again from trials.

    These are only the most prominent. EQUATOR’s library extends to qualitative research, diagnostic accuracy studies, case reports, economic evaluations, and many more, with specialised extensions for particular contexts. The skill is in matching the guideline to the study.

    What a guideline actually is — and is not

    It is important to be clear about what reporting guidelines do. A reporting guideline governs how a study is reported, not how it is designed or conducted. CONSORT does not tell you how to run a trial; it tells you what to disclose about the trial you ran. This distinction matters because a guideline cannot rescue a bad study — it can only ensure that a study, good or bad, is described completely enough for readers to tell the difference. Used as a checklist while writing, a guideline prompts authors to include the items that are easy to forget and tempting to omit; used by reviewers and editors, it provides an objective basis for asking “where is the allocation method?” rather than relying on impression.

    Reporting guidelines also pair naturally with the wider machinery of trustworthy research. A CONSORT-compliant trial report is far more valuable when the trial was prospectively registered, so that the pre-specified outcomes in the registration can be checked against those in the paper. A PRISMA-compliant review is stronger when its protocol was registered in advance. The guidelines define completeness of reporting; registration defines commitment in advance; together they close much of the gap between what was planned, what was done, and what was published.

    Practical guidance for authors

    1. Identify the right guideline before you write, not after. Match it to your study design — CONSORT for a trial, PRISMA for a review, ARRIVE for animal work, STROBE for an observational study — using the EQUATOR library to find it.
    2. Use the checklist as you draft, treating each item as a question your paper must answer, and complete the flow diagram where the guideline provides one.
    3. Submit the completed checklist with the manuscript; many journals now require it, and it signals to editors and reviewers that the reporting is complete.
    4. Combine the guideline with registration — register trials and review protocols in advance so that the reported outcomes can be checked against the planned ones.

    Where shared vocabulary fits

    “Reporting guideline”, “checklist”, “flow diagram”, “extension”, and the individual acronyms are used loosely, and confusing a reporting guideline with a study-design standard is a common and consequential error. A shared, federated vocabulary that defines these terms precisely — and points back to the EQUATOR Network and the individual guideline stewards — is what lets a reporting standard cited in one journal be understood the same way in another. Supplying that definitional layer is the role the CASRAI dictionary is designed to play; the relevant terms sit in the reproducibility domain, with adjacent entries in the research-integrity domain.

    Related reading

  • Reducing research waste as sustainability: better methods, less waste

    When research and sustainability are mentioned together, the conversation usually turns to carbon: the energy a laboratory consumes, the emissions of computing, the footprint of travel and equipment. These matter and deserve the attention they receive. But there is another kind of waste in research that is just as real and far less visible — the waste of research that is poorly designed, badly conducted, inadequately reported or never published. Effort, money, participants’ time and goodwill are all expended, yet the work yields little usable. Seen clearly, avoidable research waste is a sustainability problem: it squanders finite resources just as surely as energy waste does, and reducing it makes the research enterprise more responsible. This article looks at that argument and the frameworks behind it, drawing on the sustainable-research domain of the CASRAI Dictionary.

    The waste in the research process

    The case that a great deal of research effort is avoidably wasted was made influentially by Iain Chalmers and Paul Glasziou, who identified several stages at which waste creeps in. Their analysis pointed to waste arising when the wrong questions are asked — research that does not address what matters to those who would use it, or needlessly repeats what is known; when studies are poorly designed or conducted — methods that cannot give a reliable answer; when research is not published — results, especially negative ones, that disappear so others cannot learn from them; and when reporting is inadequate — studies described so incompletely they cannot be understood or replicated. The striking feature is that the waste is not in any single dramatic failure but distributed across the ordinary process of research, much of it avoidable with better practice. That is what makes it a sustainability concern: it is systemic, large, and within our power to reduce.

    The REWARD campaign

    This analysis gave rise to a sustained effort to address the problem, prominently the REWARD campaign (Reduce research Waste And Reward Diligence), launched through a series of papers in The Lancet. REWARD turned the diagnosis of waste into a programme for change, calling on the many actors in the research system — funders, institutions, regulators, journals and researchers — to take responsibility for reducing waste where each has influence. Its central message is that reducing waste is a shared responsibility distributed across the whole system, because the causes of waste are distributed too. A funder can insist on questions that matter; an institution can support good design; a journal can require complete reporting and publish sound studies regardless of their results. REWARD’s framing — reduce waste, reward diligence — reframes good methodological practice as the sustainable use of research resources.

    Better reporting: the EQUATOR Network

    One of the most tractable forms of waste is inadequate reporting, and here a concrete remedy exists. The EQUATOR Network (Enhancing the QUAlity and Transparency Of health Research) promotes the use of reporting guidelines — structured checklists of what a study of a given type should report so others can understand, appraise and use it. Different designs have their own guidelines, but the principle is constant: a study that is completely and transparently reported can be built upon, while one that omits crucial detail is, in effect, lost, however good the underlying work. Reporting guidelines combat waste by ensuring the value of a study is not thrown away at the final hurdle. They are a low-cost, high-return intervention: they ask not for more research but for research already done to be communicated well enough that it counts.

    Better questions and designs: registered reports

    Waste that arises earlier — from poor design, from selective publication, from questions chosen for their likely results rather than their importance — calls for interventions earlier in the process. Registered reports are one of the most promising. In this model, a study’s rationale and methods are peer-reviewed and accepted before the research is carried out and before the results are known. This attacks several sources of waste at once:

    • It improves design. Reviewing the methods before data collection catches weaknesses while they can still be fixed, not after the effort has been spent.
    • It combats publication bias. Because acceptance is based on the question and the method, the study is published whatever the results, so negative findings enter the record instead of vanishing.
    • It rewards the right things. It values asking a good question and answering it rigorously over producing a striking result.

    Registered reports embody the REWARD philosophy: get the question and the method right, and reward the diligence rather than the outcome.

    Reproducibility as the unifying aim

    Running through all of this is the concern with reproducibility. Research that cannot be reproduced — because it was poorly designed, incompletely reported, or never published — cannot be reliably built upon, and effort that cannot be built upon is, in the deepest sense, wasted. Better questions, sound methods, complete reporting and the publication of all results are the conditions under which research becomes reproducible and therefore cumulative. Framing this as sustainability is more than rhetoric: the resources poured into research — money, time, the participation of patients and volunteers — are precious, and wasting them through avoidable methodological failure is as much a sustainability failure as wasting energy. Reducing research waste makes the whole enterprise go further on what it is given.

    A consistent vocabulary for sound research

    For these practices to be tracked across funders, institutions and journals, the elements involved — study types, reporting standards, registration status, contribution roles — need to be described consistently. That consistency is what the CASRAI Dictionary provides: a shared vocabulary so the information describing how research was designed, reported and shared is understood the same way wherever it is recorded. And because doing research diligently is a collective effort, the contributions behind sound work can be described in the same shared framework — the CRediT taxonomy and its full set of contribution roles, with further context at our learning hub. Sustainability in research is not only about the carbon it emits; it is also about not wasting the effort it embodies — and better methods are how that waste is reduced.