Your experience matters.

Reading the scientific-support dial

This is a source-checked editorial interpretation for the stated purpose—not a formal clinical grade, treatment ranking or probability of success. Strong support means replicated clinical benefit with support in specialist guidance. It does not mean every study is high quality. Limited support means fewer or less direct studies, or supportive-care guidance without an isolated treatment effect.

Mixed evidence means relevant findings conflict. Insufficient evidence means the checked sources do not establish the proposed benefit. Neither is plotted as zero effectiveness. These categories are not equally spaced measurements. Each rating links to its rationale, study limitations and sources; the date records our check, not a new study.

Older systematic reviews are read alongside current specialist guidance; this is not an exhaustive living review. Ratings change through versioned editorial releases, never through community votes or runtime AI. The underlying clinical grades remain unassigned. Method: scientific-support-v1.

How community ranking works

Scientific explanations and community experience are separate. Community ordering compares reported helpfulness within each group; it does not tell you which treatment to use.

Ingredient experience signals

Ingredient reports start with the complete product and its verified formulation version, not a direct vote about an isolated ingredient. Reports use fixed choices with no free text. We then derive ingredient-level aggregates through the recorded formula. At least 20 distinct eligible contributors are required before a result appears.

A product contains many ingredients, so these signals cannot identify which ingredient caused an experience. They never change an editorial suitability rating automatically.

We require at least 20 distinct eligible contributors before showing percentages or ranking. Each contributor counts once per approach version in the pooled result, using their most recent eligible report. Reports require a completed period, body area, baseline severity, adherence, other-treatment context, personal-use confirmation and separate contribution consent. Moderators check eligibility and conflicts before approval.

The ranking uses the lower bound of a 95% Wilson interval for “Helped” plus “Helped a little.” This reduces the advantage of a tiny sample. We also show all five outcomes and median reported duration, including negative experiences. A high position is not clinical proof, a safety score or a personalized recommendation.

Different body areas, formulations, schedules and concurrent treatments are pooled. Self-selection, recall and other changes can affect results. Statistical intervals do not remove those biases. Professional or commercial conflicts exclude reports from this initial ranking. Reports are moderated, not individually published.

Reporting privacy notice

Participation is for adults. Your email identifies your account but is not shown with community results. Dates, affected area, treatment context, and structured product experiences are sensitive health information. We store them privately in Supabase; authorized moderators may access them to check eligibility. No advertising, research reuse, public notes, free-text ingredient reviews, or photos are part of this feature.

Health-storage consent and aggregate-contribution consent are separate choices. You may store a report privately without contributing it. You can correct a treatment report by submitting the same approach and body area again. You can correct a product experience by submitting the same verified formulation again; this updates its one current report. You can also download your records, withdraw community consent, or delete all treatment and product-experience reports above. Withdrawal removes reports from current aggregates; it cannot recall copies of previously viewed aggregate results.

Reports are retained until you delete them. Deleting reports also withdraws health-storage consent. Idempotency receipts retain only request identifiers and a protected request fingerprint, not report text. Minimal moderation audit entries may remain after deletion without the report content. Do not include another person’s information. Policy version: community-2026-09-12.