Sources & Evidence

Minds Vary uses evidence to improve the accuracy, context, and safety of editorial work. “Evidence-informed” does not mean that every question has a settled answer, that every source has equal weight, or that an article can replace individual clinical judgement.

The aim is to connect material claims to appropriate sources, understand how those sources reached their conclusions, compare them with the wider body of knowledge, and describe remaining uncertainty in language a reader can use.

There is no single hierarchy that fits every question. For treatment, diagnosis, risk, and public-health guidance, broader syntheses and current professional guidance will usually carry more weight than one small study. For lived experience, access barriers, language, or context, qualitative research and first-person accounts may answer something a clinical trial cannot.

• Current guidelines and official guidanceHealth agencies, public-health organisations, and relevant professional bodies with a transparent method and scope.
• Systematic reviews and meta-analysesEvidence syntheses that assess multiple studies and make selection and quality methods visible.
• Peer-reviewed primary researchOriginal studies used with attention to design, population, sample, measures, effect size, and limitations.
• Official data and primary documentsStatistics, policy, diagnostic manuals, regulatory information, and original reports rather than summaries of them.
• Expert consensusUsed when stronger evidence is unavailable and labelled as consensus rather than settled fact.
• Qualitative research and lived experienceUsed to illuminate experience and context, not to prove universal clinical effects.

A source should directly support the nearby statement. A study of adults should not silently become a claim about children. An association should not be rewritten as cause. A result in a controlled setting should not be promised as an outcome for every person in daily life.

When a source supports only part of a sentence, the wording should be narrowed. When several good sources disagree, the disagreement should be described. When no good evidence answers the question, saying so is more useful than filling the gap with certainty.

Source appraisal should be proportionate to the importance of the claim. Claims that may affect safety, treatment, medication, diagnosis, or access to care need a stronger check than a low-risk description of an everyday strategy.

• RelevanceDoes the source examine the population, setting, outcome, and question being discussed?
• MethodWas the design capable of answering the question, and were comparison, measurement, and analysis appropriate?
• Scale and effectHow large was the sample, how meaningful was the difference, and how precise was the estimate?
• Bias and limitationsWhat could distort the result, what did the authors acknowledge, and what did the study not test?
• Recency and consistencyIs the source current enough for the claim, and does it agree with or depart from the wider evidence?
• Funding and conflictsWho funded the work, what interests were declared, and do they change how cautiously the result should be interpreted?

A preprint has not completed peer review and should be labelled as such. It may help identify an emerging question, but it should not carry a strong health recommendation on its own. Conference abstracts often provide too little method and detail for a firm conclusion.

News articles, press releases, search snippets, product pages, and social posts may point toward a source, but they are not substitutes for the underlying research or guidance. Commercial material can describe a product's own features; it should not be the main evidence for its health benefit.

Lived experience is valuable evidence of what a person experienced. It can reveal barriers, language, priorities, and practical realities that quantitative research may miss. It does not establish prevalence, diagnosis, cause, or a guaranteed response for somebody else.

First-person material should be presented with consent, context, and respect for the speaker's own language. It should not be edited into a more dramatic story or treated as a representative account of an entire diagnostic or cultural group.

Material health claims should have a citation close enough for a reader to understand what it supports. A reference list should identify the original source, publication, date, and a stable link or digital object identifier where available. Links should point to the primary document rather than an unattributed copy.

An article should show its publication date, meaningful update date, responsible author or editorial team, and any specific review status. A long list of loosely related papers is not a substitute for claim-level accuracy.

Health guidance and research change. A meaningful update may be needed when a source is retracted, a guideline changes, stronger evidence becomes available, a cited link breaks, or a reader identifies a material error or missing context.

A major change should be explained on the article. A corrected page should not erase the fact that an important correction occurred, and an unchanged page should not receive a new date solely to appear current.

A source cannot diagnose a reader, remove the need for individual assessment, guarantee an outcome, or decide which values and trade-offs matter most in somebody's life. Evidence can inform a decision; it cannot make every decision universal.

Minds Vary applies the greatest caution when evidence is weak, possible harm is high, or a reader is likely to act on the information. Clear limits are part of useful health content, not a weakness to hide.