Editorial standards

Every page is produced in two passes: a research pass that drafts claims only from primary sources, and a verification pass in which every citation is fetched directly from PubMed or PubMed Central, FDA (openFDA, Drugs@FDA, the Federal Register, the eCFR), ClinicalTrials.gov or PubChem, and the claim is checked against the fetched text before publication. Trial results are reported with their prespecified primary endpoint and the result on that endpoint, not with whichever secondary outcome reached significance. Positive findings that cut against our overall framing are included on purpose. This site is produced by a research and editorial team using automated retrieval tools under human direction. No physician or clinician has reviewed this content, and we do not claim medical review: content that names no reviewer has had none.

Claims about oxytocin may cite only: peer-reviewed journals indexed in PubMed or PubMed Central; FDA sources including openFDA, Drugs@FDA, fda.gov documents, the Federal Register and the Code of Federal Regulations; ClinicalTrials.gov records; and PubChem. Compounding-pharmacy pages, supplement vendors, forums, press releases and news coverage are banned as sources. Where evidence does not exist we say so explicitly rather than borrowing a nearby claim, and absence statements are themselves cited to the query or document that establishes the absence. Where a number is arithmetic derived from two cited facts, we show the derivation instead of presenting it as something a source said.

Every page displays its published and last-updated dates. This reference is re-reviewed when any of the following happens: a new randomized trial or replication of intranasal oxytocin is published; a systematic review or meta-analysis in one of the covered areas is published or updated; FDA approves, withdraws or takes compounding action on any oxytocin product; or the Pitocin label is revised. Independent of triggers, the entire citation set is re-verified on a quarterly cycle and dead links are repaired or replaced.

If you find an error, use the contact link in the site footer. Corrections are assessed against primary sources, and substantive fixes are noted on the affected page with the date of the change. Where sources disagree we surface the disagreement in the text rather than silently picking a winner: the nose-to-brain delivery question and the conflict between the positive autism meta-analysis and the null randomized trials are both presented as live disputes, because that is what they are.

Site statistics are counts of this site's own content, computed at build time from the evidence pack and recomputable from the rendered pages. The study total is the number of rows in the study-results table. Species counts are a census of that table's species column. The null-primary count is a census of its primary-endpoint column, restricted to rows graded human RCT or preregistered replication that state a prespecified efficacy endpoint; reviews, meta-analyses, pharmacokinetic studies, case reports and observational designs are excluded from both the numerator and the denominator, and each excluded row shows why in its own grade chip. The citation total is the number of entries in the numbered reference list and the peer-reviewed and governmental percentage is a source-type census of that same list. Counts of external literature, such as how many women a Cochrane review covered, are always attributed to their source and never presented as our own measurements.