Written and maintained by CASRAI Editorial Board
Last updated
Epistemonikos is a multilingual database of health evidence built to support clinical and health-policy decisions. Run by the non-profit Epistemonikos Foundation in Santiago, Chile, it links systematic reviews to the primary studies they include. It finds reviews by screening ten bibliographic databases, using machine-learning classifiers whose output is checked by human screeners.
What It Is
Traditional bibliographic engines return citations as isolated records. Epistemonikos instead links broad syntheses, systematic reviews and the primary studies included in those reviews, so that a reader can move from a question to the reviews that address it and on to the underlying studies. The 2020 methods paper by Rada and colleagues describes it as a comprehensive database of systematic reviews for health decision-making. When weighing a narrative review vs systematic review, this review-first design is the reason the database is useful for locating the systematic kind.
A central feature is the Matrix of Evidence. The Epistemonikos site presents it as a tool that identifies the systematic reviews addressing a question and the studies they include, shown as a cross-tabulation. Laid out this way, a review team can see which trials several reviews share and which studies a given review omits.
How the database is built
According to Rada et al. (2020), the database is populated by searching ten bibliographic sources on a daily or weekly basis. A supervised random forest classifier screens records that have abstracts, and a custom heuristic classifier screens records without abstracts. The output is then checked by human screeners drawn from a network of more than 1,000 collaborators. As of January 2020 the authors reported more than 1.4 million records screened and more than 300,000 systematic reviews identified.
Who Runs It and What It Covers
Epistemonikos is operated by the Epistemonikos Foundation (Fundación Epistemonikos), a non-profit organization based in Santiago, Chile. Authors of the 2020 paper are affiliated with the foundation and with Centro Evidencia UC at the Pontificia Universidad Católica de Chile.
The ten sources searched, as listed by Rada et al. (2020), are:
- Cochrane Database of Systematic Reviews
- PubMed/MEDLINE
- EMBASE
- CINAHL
- PsycINFO
- LILACS
- DARE
- Campbell Library
- JBI Database of Systematic Reviews and Implementation Reports
- EPPI-Centre Evidence Library
The source list reflects the 2020 paper; check the Epistemonikos site for the current list.
How to Search or Use It
Epistemonikos is a web database at epistemonikos.org. The home page offers a search box and a link to Advanced search. The language menu lists nine languages, and the site states that a search can be run in any of them. The About menu links to a Multilinguality page describing this.
For review teams, a practical way to use the database is to start from a defined question. Writing that question in PICO form helps when the later step is rating certainty with the GRADE evidence certainty rating. Guidance on evidence synthesis methods and review types helps decide whether an overview of reviews, a scoping review or a standard intervention review fits the question before the Matrix of Evidence is consulted.
Strengths and Limits
Strengths
- High systematic review retrieval: A 2024 validation study in BMC Medical Research Methodology that examined 100 overviews of reviews (2,276 systematic reviews) found that Epistemonikos retrieved 94.4% of the systematic reviews, on par with Embase (94.4%) and MEDLINE (94.3%).
- Strong two-database pairing: The same study found that MEDLINE plus Epistemonikos plus reference checking gave the best retrieval of systematic reviews, at 98.1%.
- Study overlap at a glance: The Matrix of Evidence shows which studies are shared across reviews on one question.
- Multilingual search: The site offers nine languages, and the 2020 paper reports that its searches are not restricted by language.
- Free relational database: The 2020 paper’s reference list describes the database as free, relational, collaborative and multilingual. The site is licensed under Creative Commons Attribution 3.0 Unported and subject to its Terms and Conditions.
Limits
- Systematic review focus: The database is built around systematic reviews and the studies they include, so it is not a replacement for a full set of bibliographic databases in a primary search.
- Classifier error: Rada et al. (2020) estimated that the screening process missed about 3.0% of eligible reviews and included about 14.4% false positives. Their heuristic classifier for records without abstracts had a precision of 43.9%.
- Snapshot figures: Record counts in the 2020 paper date from January 2020 and have since grown.
Alternatives and Comparative Platforms
Rada et al. (2020) compared the number of systematic reviews held by Epistemonikos with several other resources as of January 2020. The table reports those published figures.
| Resource | Systematic reviews held (approximate, January 2020) |
|---|---|
| Epistemonikos | 307,119 included reviews |
| PubMed (systematic review MeSH) | about 110,000 |
| TripDatabase | about 65,000 |
| DARE | about 45,000 |
| PubMed Health | more than 40,000 |
Teams planning new trials can use what an evidence matrix shows about existing trials when writing a clinical trial protocol design. Conduct standards such as Good Clinical Practice then apply to the new study.
Frequently Asked Questions
What is the Matrix of Evidence in Epistemonikos?
The Epistemonikos site describes it as a tool that identifies the systematic reviews addressing a question and the studies they include, presented as a cross-tabulation. It lets readers see which studies are shared across reviews.
How does Epistemonikos identify systematic reviews?
Per Rada et al. (2020), it searches ten bibliographic sources daily or weekly. A random forest classifier screens records with abstracts and a heuristic classifier screens those without. Human screeners from a network of more than 1,000 collaborators then check the output.
Can Epistemonikos replace MEDLINE or Embase in a systematic search?
The 2024 validation study found Epistemonikos retrieved systematic reviews at a rate comparable to Embase and MEDLINE, and found the best retrieval (98.1%) from MEDLINE plus Epistemonikos plus reference checking. That finding concerns retrieving systematic reviews specifically, not primary studies.
Is Epistemonikos free?
The 2020 paper’s reference list describes it as a free database, and the site states its licence and Terms and Conditions. Check those terms on the site for current conditions.
What languages does Epistemonikos support?
The site’s language menu lists nine languages, and the site states that a search can be conducted in any of them.
Sources
Pages checked on 2026-11-07.
- Rada, G., Pérez, D., Araya-Quintanilla, F., Ávila, C., Bravo-Soto, G., Bravo-Jeria, R., et al. (2020). Epistemonikos: a comprehensive database of systematic reviews for health decision-making. BMC Medical Research Methodology, 20(1), 286. https://doi.org/10.1186/s12874-020-01157-x (open copy: https://pmc.ncbi.nlm.nih.gov/articles/PMC7708132/)
- Epistemonikos Foundation. Epistemonikos home page. https://www.epistemonikos.org/en/
- Methodological validation study of systematic review retrieval in MEDLINE, Embase and Epistemonikos (2024). BMC Medical Research Methodology, 24, 257. PMID 39522026. https://pubmed.ncbi.nlm.nih.gov/39522026/
