Written and maintained by CASRAI Editorial Board
Last updated
On this page: what a core outcome set (COS) is, what the COMET Database contains and how to search it before designing a trial’s outcomes, and the two methodology standards — COS-STAD for developing a COS and COS-STAR for reporting one — that determine whether a given COS is trustworthy enough to adopt.
What a Core Outcome Set Is (and Isn’t)
A core outcome set is, in the COMET Initiative‘s own definition, “an agreed standardised set of outcomes that should be measured and reported, as a minimum, in all clinical trials in specific areas of health or health care.” The word doing the work is minimum: a COS is not a ceiling on what a trial can measure, and it is not itself a measurement instrument. It is an agreement, reached across the people who will use the results — trialists, clinicians, patients, and often regulators or guideline developers — about which outcomes are important enough that no trial in that condition should be allowed to omit them.
That distinction matters because a COS answers a different question than a patient-reported outcome measure (PROM) or, more broadly, a clinical outcome assessment (COA) answers. A COS defines what should be measured — the outcome domain itself, such as “pain” or “return to work.” A COA or PROM defines how that domain is actually captured — a specific validated instrument, scoring rule, and administration method. Adopting a published COS for a trial in, say, chronic low back pain still leaves the separate task of choosing which validated pain instrument satisfies that domain; COMET’s sister resource for that step is the COMET PROMs database, distinct from the COS database this guide focuses on.
Why this exists at all: without a COS, trials in the same condition routinely measure different outcomes, using different instruments, at different time points — a pattern documented across dozens of clinical areas before COMET’s founding. The practical cost is that a later systematic review or meta-analysis often cannot pool results across trials because there is no common outcome to pool, and selective outcome reporting (trials quietly dropping an outcome that turned out null) becomes harder to detect. A COS is the standardization mechanism that makes cross-trial synthesis possible in the first place — the same underlying problem that publication and outcome-reporting bias exploits when no agreed outcome set exists.
The COMET Database: What It Contains and How to Search It
The COMET Database is a searchable registry of COS studies — both completed core outcome sets and studies currently in progress to develop one — maintained by the COMET Initiative, a University of Liverpool–based collaboration of methodologists, journal editors, and patient representatives. Each entry is searchable by health condition, study title, abstract text, or author, and the database offers a free alert subscription that notifies a registered user when a new COS is registered or published in a chosen area.
The single highest-value use of the database is negative: checking it before designing a trial’s outcome set, not after. If a relevant COS already exists for the condition and intervention type under study, using it (rather than a bespoke outcome list) is what allows that trial’s results to be pooled with others later, and several funders and journals now expect a COS search to be documented in a trial’s protocol for exactly this reason. If a search turns up an in-progress COS registration rather than a completed one, that is itself useful information — it flags that a field is actively converging on a standard, and a new trial may be able to wait for or align early with the eventual result rather than adding one more divergent outcome list.
COS-STAD: Minimum Standards for Developing a Core Outcome Set
Not every published “core outcome set” was developed with a rigorous process, which is precisely the problem COS-STAD (Core Outcome Set–STAndards for Development) exists to address. Published in PLOS Medicine in 2017 by an international group led by Jamie J. Kirkham and Paula R. Williamson, COS-STAD sets out 11 minimum standards organized into three domains, intended less as a how-to guide and more as a quality bar a reader can check a given COS against.
| Domain | What it requires, at minimum |
|---|---|
| Scope | The COS states the health condition(s), the target population, the intervention(s), and the setting it is intended to apply to — a COS scoped too broadly is hard to apply consistently. |
| Stakeholders | The consensus process includes, at minimum, researchers/trialists who will use the COS, healthcare professionals, and patients or their representatives — omitting any one of the three is a recognized quality flag. |
| Consensus process | The process defines in advance how outcomes are scored, what counts as consensus, and the criteria for including, dropping, or adding an outcome — decided before results come in, not adjusted afterward — with outcome wording kept unambiguous. |
A COS entry in the COMET Database that documents its methods against these three domains is meaningfully more trustworthy than one that lists only a final outcome set with no visible process behind it.
COS-STAR: Standards for Reporting a Completed Core Outcome Set
COS-STAR (Core Outcome Set–STAndards for Reporting) is the companion reporting guideline, published in PLOS Medicine in 2016 by an overlapping international group of COS developers, methodologists, journal editors, and patient representatives, developed through a two-round Delphi survey with nearly 200 participants followed by a consensus meeting. Where COS-STAD is a design-quality checklist, COS-STAR is a completeness checklist for the published article itself — it specifies what a COS report needs to state explicitly (scope, stakeholder groups involved, consensus methods used, the final outcome list and how it was worded) so that a reader can actually assess whether COS-STAD’s development standards were followed, rather than taking a headline “consensus was reached” claim on faith.
In practice, the two standards are meant to be used together: COS-STAD as a design and appraisal checklist, COS-STAR as the reporting template that makes that appraisal possible from the published article alone.
The COMET Handbook and Consensus Methods
The COMET Handbook, maintained by the COMET Initiative and periodically updated by Williamson and colleagues, is the detailed methods reference behind COS-STAD’s consensus-process domain — it surveys the available approaches to reaching stakeholder consensus (most commonly a multi-round Delphi survey, sometimes paired with a face-to-face or virtual nominal group technique consensus meeting) and the evidence for how each performs. A Delphi survey circulates candidate outcomes to participants across stakeholder groups, collects importance ratings anonymously over two or more rounds with feedback shown between rounds, and stops once a pre-defined agreement threshold (commonly around 70–80% rating an outcome as critically important) is reached or stability across rounds is observed. The Handbook does not mandate one specific method; it documents the options and the trade-offs, which is why COS-STAD’s own standard is that the chosen method and its consensus definition be stated in advance, not which specific method must be used.
Why This Belongs in a Trial or Review Protocol, Not an Afterthought
Checking the COMET Database is now a step several major trial registries and funders expect to see documented at the protocol stage, alongside a systematic review protocol or a trial’s own registration. For a systematic reviewer, the same check matters in reverse: if the included trials in a review used inconsistent outcomes because no COS existed for the field at the time, that heterogeneity is itself worth flagging in the review’s PRISMA-reported methodology and discussion, since it explains why formal pooling was limited even where a meta-analysis would otherwise have been the natural next step.
Frequently asked questions
Is a core outcome set the same as a set of outcome measures?
No. A COS specifies outcome domains — the concepts to be measured, such as “pain intensity” or “physical function” — not the specific instruments used to measure them. Selecting the actual measurement instrument for each domain is a separate step, covered by PROM selection and validation guidance rather than by the COS itself.
What happens if no core outcome set exists yet for my condition?
Search the COMET Database for an in-progress registration first — many fields have a COS under active development even before one is published. If nothing exists or is underway, COS-STAD’s three domains (scope, stakeholders, consensus process) are the recognized starting checklist for designing a new one, and registering the planned project with COMET makes it discoverable to other groups working in the same area.
Does using a published COS mean I don’t need my own outcome measures?
No — a COS is explicitly a minimum. A trial can and often does measure additional outcomes beyond the core set; the COS only guarantees that a defined minimum is captured consistently across studies so that later synthesis is possible.
Who developed COS-STAD and COS-STAR?
Both were produced by overlapping international working groups convened through the COMET Initiative, with Jamie J. Kirkham and Paula R. Williamson as lead/corresponding authors on the COS-STAD recommendations (PLOS Medicine, 2017) and on the earlier COS-STAR reporting statement (PLOS Medicine, 2016); both used Delphi-based consensus processes involving COS developers, methodologists, journal editors, and patient representatives.
Related CASRAI resources
- Choosing and Validating a Patient-Reported Outcome Measure (PROM) — the instrument-selection step that follows once a COS has defined which outcome domains to measure.
- Clinical Outcome Assessment (COA) — the broader category of measurement instrument that a COS domain gets operationalized into.
- Systematic Review Protocol Template — where a COMET Database search belongs in a review’s pre-registered methods.
- PRISMA and Systematic Review Methodology — how outcome heterogeneity across included studies gets reported.
- How to Detect and Assess Publication Bias — the related risk that a COS’s consistent reporting requirement helps counter.
- Meta-Analysis — the pooling step a consistent core outcome set makes possible across trials.
Last verified 2026-08-31 against the COMET Initiative’s own site (comet-initiative.org), the COS-STAD recommendations (Kirkham, Williamson, et al., PLOS Medicine, 2017), and the COS-STAR reporting statement (PLOS Medicine, 2016). Consult the COMET Database directly before finalizing a trial or review’s outcome set, since new core outcome sets are registered on an ongoing basis.
