Claimscape

1. On topic (yes or no)

“Does the paper investigate the relationship described in the claim?”

• yesThe paper studies or reviews the same kind of exposure or intervention and the same kind of outcome as the claim, in any population, species or model
• noThe paper is about a different question, or mentions the claim's topic only in passing

2 and 3. Verdict, asked two ways (one option)

“What do the paper's own results say about the claim?” and “Taken as a whole, which way does the paper's evidence point on the claim?”

• supportsThe paper's results or conclusions agree with the claim For example: An effect in the direction the claim states; No effect, when the claim is that there is none.
• contradictsThe paper's results or conclusions disagree with the claim For example: No effect or no association where the claim asserts one; An effect in the opposite direction to the claim; Harm where the claim asserts benefit.
• mixedSome results agree with the claim and others disagree, for example across outcomes, doses, subgroups or time
• no_resultThe paper reports no result that bears on the claim For example: A study protocol or method; A review that describes the topic without reaching a conclusion.

4. Direct evidence (yes or no)

“Does the paper test the claim directly?”

• yesThe paper studies the same population, exposure or intervention, and outcome that the claim names
• noThe paper uses a different population, species or laboratory model, or a stand-in exposure or outcome

5. Study type (one option)

“What kind of study is this paper?”

• meta_analysisSystematic review or meta-analysis that pools results from several studies
• rctRandomized controlled trial in people
• observationalStudy of people without random assignment: cohort, case-control, cross-sectional, registry, database or retrospective analysis, or a non-randomized trial
• case_reportCase report or case series describing individual patients
• animalExperiments in live animals
• in_vitroLaboratory experiments on cells, tissues or molecules without live animals
• reviewNarrative review, commentary, editorial, guideline or opinion without its own pooled analysis
• otherModeling, simulation, methods, protocol or anything else

6. Finding (one of the abstract's own sentences)

“Which sentence of the abstract reports this paper's own result that bears most directly on the claim?”

The options are the abstract's sentence ids, so the highlighted finding is always text the paper contains.