Skip to main content
Apotheon
§ METHOD·for /compounds

Our Review Methodology

Every compound monograph on Apotheon ends with a Practical Verdict — a calm, evidence-anchored summary of what the research actually supports versus what's hyped. Here is how each monograph is built, vetted, and cited.

The short version

Every monograph on Apotheon ends with a Practical Verdict — a calm, evidence-anchored summary of what the research actually supports about a compound, distinct from the hype, the marketing, or any single positive trial. Reading the verdict alone tells you whether to pay attention to the rest of the monograph. Every PMID is live-verified against PubMed before publishing; the verdict is dated so you can see when it was last reviewed.

How each monograph is built

The pipeline is the same for all 70+ compounds:

  1. Lit review — fetch the relevant systematic reviews, RCTs, and mechanism papers from PubMed and OpenAlex. We start with the highest-quality evidence and work down only if needed.
  2. Adverse-event scan — query FAERS (FDA Adverse Event Reporting System) and ClinicalTrials.gov via BioMCP for safety signals and registered trials.
  3. Evidence classification — every major claim is graded against the Bradford-Hill criteria (causal strength on a 9-point scale) and labeled with the actual relationship: Direct Causation, Probable Causation, Unreplicated Causal Claim, Surrogate Endpoint, Mechanistic Extrapolation, or No Evidence.
  4. Bias and marketing review — we surface confounders, funding patterns, single-region clustering, single-institution clustering, and the manipulation patterns that show up in marketing copy (proprietary blends, MLM funnels, cult-adjacent lineages, testimonial farms).
  5. Practical synthesis — what the evidence justifies for whom, at what dose, with what monitoring, at what cost.
  6. Citation lock — every PMID written into the monograph is fetched live from PubMed and confirmed to match the cited topic. The verifier runs on every build.

What's inside a Practical Verdict

Every Practical Verdict on Apotheon contains the same canonical sections, applied uniformly across all 70 compound monographs:

  1. Evidence Classification — a table grading each major claim against Bradford-Hill criteria and labeling its relationship to the underlying outcome. This is the most important table in the monograph; it tells you which marketing claims survive scrutiny and which don't.
  2. Quality Concerns — explicit flagging of claim inflation, surrogate-endpoint dominance, replication failures, and evidence-strength mismatches. Where the published narrative outruns the data.
  3. Evidence Gaps — what we don't know. Long-term safety, sex-stratified data, head-to-head formulation comparisons, pharmacogenomic modifiers, replication status. The honest list of where the science stops.
  4. Bias Flags — single-region clustering, single-institution clustering, publication bias, financial conflicts of interest. The metadata of the evidence itself.
  5. Marketing Red Flags — specific marketing patterns and industry incentives (who benefits, both pro- and anti-supplementation) that should make you pause.
  6. Practical Considerations — a short, practical answer to "should I take this?" That includes a health utility score (0-10), an opportunity-cost note, and a simple ADD / WATCH LIST / SKIP / CONDITIONAL recommendation.
  7. Bottom Line — the one-paragraph compression. If you read nothing else in the monograph, read the Bottom Line.

This structure exists so the verdict is a falsifiable assessment, not an opinion. Every claim has a citation. Every citation has been live-verified against PubMed before publishing. Every safety flag has a reason.

Why every verdict is dated

The dated (as of YYYY-MM-DD) on every Practical Verdict heading is not decoration. It is a commitment that the verdict is provisional. If a future trial reverses the verdict on a compound — and several have, even within the lifetime of this site — the verdict updates and the date moves with it.

Apotheon does not pretend to have settled science when the science is unsettled. When the evidence shifts, so does the read.

Citation integrity — what "PubMed-verified" actually means

Every PMID that appears in a monograph has been live-fetched from PubMed (via the NCBI E-utilities API) and confirmed to match the compound and topic it is cited for. The mechanism is a citation lock at src/data/citations.lock.json that records, for every PMID we cite, the title, journal, year, abstract excerpt, and the verification date. The verifier runs:

  • on every commit that touches compound data (pre-commit hook),
  • on every production build (CI gate),
  • on demand via npm run verify:citations.

The trust anchor: a citation that doesn't survive the verifier doesn't ship. The defense is concrete because the alternative — fabricated PMIDs paired with unrelated papers — is a documented failure mode in LLM-assisted health writing, and we have caught and corrected six such cases in the source vault during this site's lifetime.

How to use a verdict

  • For deciding whether to read the rest of the monograph: read the Bottom Line first. If the compound's evidence base is strong and you're already inclined to use it, the Practical Considerations section gives you the practical guardrails. If the evidence is thin or the hype is heavy, the Quality Concerns and Marketing Red Flags will tell you how to recognize it elsewhere.
  • For arguing with someone (or yourself): the Evidence Classification table is the canonical artifact. Bradford-Hill scores and causal-strength labels are the same vocabulary that careful clinicians and reviewers use; quoting them disarms most popular-press claims faster than anything else.
  • For pricing your time: the Opportunity Cost note in Practical Considerations translates "is this worth it?" into honest dollar and minute terms.

What's not in a verdict

  • A purchase recommendation. Apotheon does not tell you what to buy. The Practical Considerations section names where a compound has a justified niche; whether you fit that niche is yours.
  • A clinician relationship. Every monograph carries the same disclosure: this is research, not medical advice.
  • A guarantee. The dated heading is the explicit disclaimer that the verdict is the best honest read of the evidence on the day it was written. Evidence updates, and so does the verdict.

The verdict is a tool for thinking calmly about a compound when most of the surrounding noise is anything but calm. That is the whole point of the section.