An evidence summary is useful only when readers can see how it was assembled. This page explains the editorial method used for records in the Serum Research Library.
The short version: we identify the exact compound first, prefer primary and official sources, label the study model, report important limitations, separate study registration from study results, and never turn a research finding into a recommendation for personal use.
1. Resolve the identity before reviewing claims
Similar names do not always identify the same material. A base peptide, a salt, an ester, an analog, a conjugated form, and a branded finished drug may have materially different identities or regulatory histories.
For a record prepared under this method, Serum Research applies an editorial identity screen and documents, where available:
- preferred compound name and common synonyms;
- amino-acid sequence or chemical formula from an authoritative record;
- database identifiers from authoritative records, such as a UNII or a trial intervention name;
- distinctions between similarly named forms;
- any catalog record considered for linking and whether its identity can be reconciled with the source identity.
If the catalog identity cannot be reconciled with the source identity, the page must not imply that they are the same.
2. Start with primary and official records
We use literature reviews to understand a field and find citations, but important statements are checked against the underlying study or official record whenever possible.
Preferred source types are:
- FDA databases, approval documents, and labels for U.S. approval status.
- ClinicalTrials.gov records for registered study design and status.
- Primary research identified through PubMed or publisher records, followed by review of the actual study record.
- Authoritative chemical-identity databases and original sponsor or patent records when needed to resolve naming.
A search result, retailer page, press release, or unsourced article may lead to a source, but it is not enough on its own.
PubMed is a discovery and indexing resource. Inclusion in PubMed does not by itself establish peer review, primary-study design, validity, or evidence strength.
3. Label the study model
We do not combine unlike evidence under a vague phrase such as “studies show.” The model labels below are a Serum Research editorial taxonomy for organizing sources, not a validated certainty score:
| Study model | Library label | Required context |
|---|---|---|
| Isolated target, biochemical assay, or cultured cells | Biochemical or cell-based | System, concentration, endpoint, and whether the finding was replicated |
| Excised tissue or organ preparation | Ex vivo | Species or tissue source and the limits of the model |
| Whole-animal experiment | Animal | Species, model, route, comparator, and measured endpoint |
| Human study without investigator assignment | Human observational | Population, exposure definition, confounding, and association-only language |
| Prospective investigator-assigned human study | Human interventional | Registration, design, sample size, intervention, comparator, endpoints, duration, and results status |
ClinicalTrials.gov distinguishes interventional trials from observational studies. We use those terms in the same narrow sense and read protocol fields separately from posted results fields. A listed study is submitted by its sponsor or responsible party; listing does not mean that the U.S. government evaluated its safety or science, that results exist, that a finding was favorable, or that FDA approved the intervention.
4. Calibrate confidence to the exact proposition
Confidence language is calibrated to a specific proposition, not permanently assigned to an entire molecule. The review considers study model, relevance, replication, risk of bias, precision, directness, publication or selective-reporting concerns, and whether the studied material matches the identity being discussed. This is an editorial writing rule, not a validated certainty score or a completed GRADE assessment.
| Evidence situation | Appropriate interpretation |
|---|---|
| Multiple relevant, adequately designed human studies | Greater confidence for the exact material, population, comparison, outcome, and time frame studied; important limitations still remain visible |
| A single, small, preliminary, indirect, or observational human study | Limited or preliminary human evidence; no assumption of replication, causality, or broad applicability |
| Whole-animal evidence without adequate human evidence | A finding in the named species and model, not an established human outcome |
| Biochemical, cell, or ex-vivo evidence | A response in the defined laboratory system, not an established whole-organism or human effect |
| Unresolved identity, no dependable primary support, or a claim that exceeds the cited evidence | Unsupported or unresolved; the claim is omitted or explicitly limited |
For example, evidence that a compound binds a receptor can be stronger than evidence that it changes a human outcome. Those propositions must not be presented as though they have the same support.
5. Review limitations before writing the summary
The checklist below is an editorial screen informed by the official and primary methods sources listed below; it is not a completed risk-of-bias assessment. At minimum, the editor checks:
- sample size and attrition;
- randomization, masking, comparator, and prespecified endpoints where relevant;
- model-to-human translation;
- formulation, route, exposure, and duration;
- whether the reported outcome is a surrogate or direct outcome;
- statistical uncertainty and multiplicity;
- sponsor role and declared conflicts;
- retraction, correction, expression of concern, or unavailable results;
- whether later work replicates or contradicts the finding.
Absence of evidence is not proof of no effect. It is also not permission to imply an effect.
6. State regulatory status precisely
When a record makes a regulatory-status statement, that statement is dated and jurisdiction-specific.
Acceptable examples include:
- “FDA approved the named finished drug product and application for the specified context; this statement does not apply to Serum Research material.”
- “No matching approval record was identified in Drugs@FDA for the documented exact search terms as of the stated date; this database result is not a universal status determination.”
- “Investigational status is reported by the cited trial, sponsor, or regulatory source; that statement is separate from the dated Drugs@FDA search.”
- “Registered clinical studies identified; registration is not approval and does not establish a favorable result.”
- “Preclinical research record; no relevant human interventional study identified using the documented search as of the page’s stated review date.”
The record identifies the regulatory source and check date. If identity is ambiguous, the status remains unresolved until it is corrected.
7. Keep commercial claims separate from evidence summaries
Research-library pages may describe published evidence and, only when identity and product-link gates pass, catalog availability. They must not imply that a catalog product is equivalent to a studied formulation or an approved medicine.
Statements about a Serum Research product, identity, quantity, purity, method, sterility, endotoxin, origin, storage, or batch performance, require product-specific records. A scientific paper about a similarly named compound does not, by name alone, substantiate those commercial specifications.
8. Maintain an auditable record
For each page, the working record should retain:
- the responsible organization or named editor, and a reviewer only when that review occurred and is claimed;
- first-published and last-reviewed dates;
- source identifiers and access dates;
- exact regulatory database searches;
- a short change log for material corrections;
- an internal claims-check result;
- the identity record used to connect the article to any product page.
Corrections
If you find an identity error, missing primary study, incorrect status, or misleading interpretation, contact Serum Research through the contact page. A substantive correction should be dated and described on the affected record.
Primary and official references supporting the method
- About PubMed, National Library of Medicine
- ClinicalTrials.gov protocol data-element definitions
- ClinicalTrials.gov results data-element definitions
- ClinicalTrials.gov disclaimer
- About Drugs@FDA, U.S. Food and Drug Administration
- FDA approved-drug questions and answers
- FDA UNII Search
- NIH Rigor and Reproducibility
- CONSORT 2025 published statements
- FDA/ICH E9 Statistical Principles for Clinical Trials
- NIH NCATS Translational Science Spectrum
Sources checked: August 31, 2026. Questions or corrections can be sent through the Serum Research contact page.
Editorial responsibility: Serum Research. This page was prepared from the primary and official sources listed above. No independent scientific, regulatory, or clinical review is claimed.