How this is calculated
Inspect the versioned inputs.
Input defaults that are not sourced measurements are illustrative planning assumptions. Exact unit definitions, the 52-week convention and algorithm details are in the shared methodology.
Worked example
An example you can inspect.
A year-from filter of 2018 excludes a 2015 record. Selecting open access uses the stored bibliographic access flag, not a promise that every external full-text link remains available.
Know which records the table contains.
The explorer exposes the project's entire discovered metadata set, including included, candidate and excluded records. Those categories represent different editorial states. The default included view helps distinguish the read-source selection from the broader discovery pool, but an included label does not imply full-text review or clinical certainty. The review-scope field tells you what material was actually assessed. A candidate title can be useful for discovery while remaining unavailable as support for a website claim.
Combine filters deliberately.
Text, outcome, population, design, sauna type, year, evidence level, access and inclusion filters work together. A year value means that year or later. Text matching searches the listed metadata rather than a hidden full-text research archive. If the result becomes empty, clear one condition at a time to see which restriction removed the records. An empty filtered set means no record in this bounded catalog matched those inputs. It does not establish that the wider literature contains no relevant research.
Follow the original record and its scope.
The table links to the original source and to an internal research summary only where that summary has actually been written. Planned summaries are labeled as planned instead of linking to nonexistent pages. When assessing a record, keep its design, population, exposure and outcome together. A study of a supervised clinical protocol or a particular group cannot automatically answer a question about every home sauna user. Publication metadata identifies a paper; the appraisal must still establish what claims that paper can support.
Avoid counting records as independent proof.
A catalog can include several publications related to one research population or program. The number of displayed rows is therefore a publication count, not necessarily a count of independent experiments or participants. Open access describes availability rather than methodological quality, and an evidence label does not erase study limitations. Use the table to find the relevant records, then read the linked interpretation and primary material. More matching titles alone do not justify a stronger causal conclusion or a more precise personal prediction.
Use a reproducible search note.
For an editorial question, record the date and the filters you used, together with the relevant record identifiers. The catalog can expand later, so a result count alone is not a durable citation. Keep the individual source and the reviewed claim connected in your notes. The generated share link deliberately omits free-text research queries and population or outcome text; another reader may need to recreate those filters manually.
Questions, considered
Questions about this tool
How should I interpret the study explorer result?
The catalog is a fixed November 8, 2026 research snapshot. Most included appraisals are abstract-only. Publication counts can include overlapping cohorts; not every candidate research-summary page is written. Search and filter the entire research metadata catalog by outcome, design, population, year and access, with inclusion and review scope visible.
Which assumptions matter in the study explorer?
Search title, outcome or population text; combine exact design, sauna type, evidence-level, year-from, access and inclusion filters. The default view is the included pool. Candidate/excluded records remain accessible as metadata, never promoted into citable evidence by a filter. The public dataset contains metadata only. Link to an existing written summary when available and otherwise label the summary planned while linking the original paper.
Primary sources
Follow the source.
- Association between sauna bathing and fatal cardiovascular and all-cause mortality events ↗
Abstract, Results; original cohort of 2,315 middle-aged Finnish men; DOI 10.1001/jamainternmed.2014.8187
The baseline frequency categories were one, two to three, and four to seven visits per week. They are observational exposure groups, not assigned weekly doses or goals. Keep sessions, total minutes and temperature distinct; the tracker cannot convert these to a validated therapeutic heat dose.
Checked 2026-11-08 - Published tool assumptions and preference scoring ↗
Editorial assumptions, input defaults, uncertainty and validation
Explicit site-authored conventions, not external scientific evidence.
Checked 2026-11-08