A compound name is a starting point, not a conclusion. To understand a finding, read what was studied, in whom, against what comparator, and with what limitations.
Start with the question
Write down the question you want the paper to answer. A study about a biochemical pathway may not answer a question about symptoms, long-term outcomes or safety. Keep the question next to your sources in My research.
Identify the study setting
| Setting | What to look for | What remains uncertain |
|---|---|---|
| Laboratory / in vitro | Cell type, assay, controls and concentrations | Whether the finding translates to a whole organism |
| Animal / preclinical | Species, model, sample size and replication | Whether the findings translate to humans |
| Human observational | Population, confounders and follow-up | Whether an association reflects causation |
| Randomized clinical trial | Randomization, comparator, blinding, endpoints and adverse events | Generalizability beyond the studied population and formulation |
| Review or meta-analysis | Search method, eligibility rules and quality of included studies | Gaps and biases inherited from the underlying evidence |
The National Cancer Institute's clinical-trials overview explains the role of clinical studies. A trial phase is a stage of investigation, not a score for the strength of every claim about a compound.
Read the endpoint, not just the headline
Check whether the endpoint was specified before the study, whether it measures an outcome that matters to the question, and whether the paper reports absolute as well as relative changes. Look for uncertainty intervals, dropouts and adverse events. Record both positive and negative findings.
For example, the ELIXA lixisenatide trial did not establish a significant reduction in major cardiovascular events. Sharing a drug class with another compound does not make their trial findings interchangeable.
Separate a compound from a product
A paper studies a particular material or formulation under particular conditions. A supplier listing with the same compound name does not establish equivalence. Check the batch identifier, testing method and identity documentation; read our COA guide and lab-verification hub.
Understand the reference labels
Linked references counts the sources attached to an entry. It is a navigation aid, not a quality score, an approval status or a complete literature review. A higher count does not establish stronger evidence.
No references located means the recorded search did not identify compound-specific indexed literature. References not reviewed means we have not recorded a literature review for that entry. Neither statement proves that no research exists.
Dates on individual entries describe the recorded editorial review. Older entries are not silently relabeled as newly reviewed when the design changes. Read our editorial policy for how the library is maintained.
A repeatable reading checklist
- Confirm the exact compound, formulation and study population.
- Identify the design, comparator and predefined endpoints.
- Read the limitations, adverse events and conflicts of interest.
- Check the primary source, publication date, corrections and retractions.
- Save the reference and write down what the study does not answer.
Browse the compound explorer, open the references on a profile, and use My research to keep your reading organized. For research and educational purposes only. This is not medical advice.