Institutions & Clusters

Two measures on two axes, so no single order exists. Horizontal position is the number of papers with the measure, on a logarithmic scale; vertical position is mean field-weighted citation impact, where 1.00 is the world average for the same fields and years. The vertical line through each point is its 95% stability interval: where two intervals overlap, the two institutions cannot be told apart on this measure. 2021–2025, OpenAlex years with a citation window.

An institution with fewer than the minimum number of papers carrying the measure is left out of this chart altogether, and the table says so in its place. Fractional counting: each paper weighs by the institution's share of its authors' affiliations.

Three counts, never one total: a paper, a clinical trial and an observational registration are different kinds of record. There is no rank column. The list opens with the most publications in the view, ties A–Z; the switch above it orders it by clinical trials, observational studies or A–Z. A parent institution's row also counts its units' papers and studies, so rows overlap and are never added together. Every count above zero opens its record list.

Publications count research-related papers, which are selected by the AI study type. Clinical trials and observational studies count registrations listing the institution as a site; a listed site does not prove that participants were recruited there.

Institutions are named as OpenAlex records them. Citation influence: "—" where the institution has fewer than the minimum number of papers carrying the measure in the view.

The twenty health clusters of the Health Holding Company. Industry-led trials are shown as a count beside a band: the band says how firm the count is, because figures this small move by a record or two between releases. Observational studies are not shown by cluster. Every count above zero opens its record list.

A study is industry-led when its lead sponsor is a health-product company, or a contract research organisation with one as collaborator. Studies are assigned to clusters by the hospital listed, or by the cluster when the registration names the cluster itself.

None identified: No industry-led trial recorded a hospital here.

1–4: A count this small may move between releases.

5 or more: Firmer, though still a small number.

“None identified” does not mean none happened. It means no registration in this period recorded a site in that cluster with an industry lead sponsor. A study may have run there without the registry recording it.

Historical outputs are grouped by today's HHC cluster structure; this does not imply that the structure existed in the year of publication.

Applies: year (recorded study start year). A study listing hospitals in three clusters counts once in each, and a study at a hospital outside the clusters appears in none, so this column neither sums to the national figure nor accounts for all of it. The clusters keep their own order.