Method & transparency

Methods and inclusion criteria

An editorially curated research resource, updated in dated editions. It is not a systematic review.

Scope

The core subject is receptive exposure to installed visual art in healthcare: murals, paintings, photographs, sculpture and related works. “Passive” describes the absence of art-making; it does not mean viewers are emotionally or intellectually inactive.

Art-making and art therapy are outside the core scope. A guided gallery visit remains receptive viewing, but differs from ambient exposure. Responsive installations, sound, movement and wider landscape design are identified as additional components.

Inclusion and scope rules

The core scope is research on receiving or viewing installed visual art in healthcare. Relevant populations include patients, staff and visitors; relevant outcomes include experience, wellbeing, stress, pain and perceptions of the environment. Eligibility does not depend on a favourable result.

  • Core evidence: studies and reviews of visual art encountered in healthcare settings.
  • Adjacent evidence: museum or gallery viewing, facilitated encounters and multi-component interventions. These are labelled because they differ from ambient exposure.
  • Context: theory, mechanisms, policy, guidance and research-gap records. These are not presented as additional health-outcome trials.
  • Outside the core scope: art-making, art therapy, and activities centred on music, performance or creative participation. Broader retained sources are identified as context rather than direct evidence for installed visual art.
  • Practice records: commissioned or donated artworks and institutional programmes are documented separately from research findings.

These rules describe the repository’s editorial scope. This edition applies them to the inherited collection; it does not represent a new prospective search or a complete census of eligible studies.

What we count

Counts on this site use the units below and are recalculated from the release data for each edition.

  • Research record: one entry in the research library. A record may be a study, a review, a framework, a policy document or an institutional account, and some records summarise the same publication. Research records are therefore not a count of unique studies.
  • Installation record: one artwork, or one ensemble presented as a single work, at one institution. The Miraculous Journey (W-09) is one installation record comprising fourteen sculptures. The same artwork at a second institution would be a second record.
  • Institutional case summary: a narrative account of practice. Some summaries cover grouped institutions, a district or a composite of environments, so 27 summaries are not 27 separate hospital art programmes.
  • Artist: an artist can appear in several records. Jaume Plensa appears in W-03, W-25 and W-32.

Filters and charts can place a record in more than one category, so category totals can exceed the number of records. No count on this site measures effectiveness.

The updated date in the data records the last change to a record. The last full editorial audit, abstract or full-text assessment, and technical website QC are separate activities. The 29 September verification ledger covers all 36 records; its snapshot date is not a new verification of every claim.

Full accessibility remediation of the legacy PDF downloads remains pending. The corresponding HTML pages are the preferred reading and form-filling route.

Source category, claim verification and current status

Each installation record keeps three separate fields.

  • Source category describes the kinds of source the record cites: institution or artist (at least one source from the hospital, its charity or collection, the commissioner or the artist); independent only (press, scholarly or heritage sources); or awaiting institutional confirmation (compiled from Wikimedia Commons and Wikidata). The date shown is the date the category was assigned, not a date on which every detail was checked.
  • Claim verification records which details have been checked against a named source, and when. Most records have not yet been checked claim by claim, and say so.
  • Current status records what is known about whether the work is still in place. Historical sources and photographs do not confirm a present location.

A source category of “institution or artist” does not mean that every detail or the current location has been verified. Where information is missing, records say “not reported”, “not assessed”, “not yet searched” or “no eligible study identified in the documented search”, not “none”. Enquiries to institutions are dated on the record and marked as sent only where the correspondence is retained.

The verification ledger (CSV) lists, for each installation record, the claims published, the source category, what has been checked, when, and what remains uncertain. The data dictionary (CSV) defines the fields and allowed values in the downloadable data.

Images

Photographs are displayed only where a licence or permission is recorded, and the basis is recorded for each image in the installation data. Permission to publish a photograph and permission to reproduce the artwork shown are checked separately. Thirteen photographs whose permission could not be verified were withdrawn on 28 September 2026; their records show a placeholder until permission is recorded.

Editorial policy and founder disclosure

Founder installations are documented as examples of practice. Their inclusion does not provide evidence of health benefits.

Records are selected against the scope rules on this page. Inclusion of research does not depend on a favourable finding, and null and unfavourable results are retained. Corrections are dated in the change log below. The founder’s practice as a sculptor is disclosed on the relevant records and on the About page.

Any future study of a founder installation should be assessed by an independent reviewer under the same criteria as other research, and the artist or donor should not control the analysis or the reporting of unfavourable findings. The installation record and any research record would be kept separate.

Understanding the filters

Study / record type describes the research design or kind of background material. Source format separately identifies preprints and conference abstracts; “Other source formats” is not a peer-review rating. Methods availability flags records with limited reporting; “See record verification” does not imply a full appraisal.

Mixed findings means outcomes or groups differed. Includes null findings means at least one reported comparison found no detectable difference. These labels overlap and apply to the record, not every outcome. Full findings and limitations take precedence over the labels.

Classification

Each record identifies its relevance, evidence type, design, sample or scope, finding and limitation. “Healthcare art” describes relevance to the topic, not a high confidence rating. Reviews, trials, observational studies, theory and institutional accounts answer different questions.

The earlier A–E scale mixed descriptions of the literature, causal evidence and practice. It has been retired from the public display. Original grades remain in the downloadable data as historical metadata, not current ratings. This is not the formal GRADE system.

Counts and overlap

The register contains 78 research records. Original IDs are preserved; PAH-020 was absent in the supplied register. PAH-057 to PAH-076 were added in Edition 4.4 and PAH-077 to PAH-079 in Edition 4.5. Several records refer to the same publication, and reviews may include trials listed separately. The 36 artwork/project records (35 documenting placement and one installation in progress) and 27 case summaries also overlap. Neither category is a count of health-outcome studies. What we count.

Verification, dates and limits

The repository is a curated update, not a systematic review. The current edition’s changes are listed first in the change log below; earlier edition notes are kept as historical records. Each research record displays its editorial review date and verification status; the publication year is shown separately.

The audit reviewed all existing pages, register records and supplied report sections. It checked key disputed claims against primary or author-institution sources and checked the accessibility of cited URLs. An accessible page does not verify every assertion made about it.

Some sources could not be fully accessed or independently re-extracted. Each research record identifies whether its result received a fresh source check or its original citation was retained. Source links may change, and historical installation records do not guarantee current access.

The material is weighted towards English-language and UK/US sources. No comprehensive global search, duplicate independent screening, formal risk-of-bias assessment or exhaustive census of hospital sculpture is claimed.

Change log · newest first. Earlier entries are historical and describe the repository as it was at the time.

Edition 4.11: tools and record relationships · 29 September 2026

Added a whole-life cost calculator, permanent pages for 78 records in both languages, two shared-publication groups, an ELSA shared-cohort link, verification checklists for 36 artworks, and a form that prepares contributions for email. This release does not constitute a new full research appraisal or contact with institutions. A unique-study total remains unestablished. Updated the Greek translation note.

Publication reconciliation · Artwork verification · Budget calculator

Edition 4.10: technical QC and usability · 29 September 2026

Corrected form printing and labels; added draft download and restore; reconciled the verification ledger with all 36 records; distinguished installations in progress; and improved record links, filters and the evidence map on mobile. This release is technical QC, not a new scientific appraisal of the sources.

Edition 4.9: verification of all research records, founder records and fuller case studies · 29 September 2026

  • The 49 research records that had no verification level were checked. 40 now carry a level (abstract checked 27, full text assessed 4, abstract or primary source checked 9); 9 remain unassigned because the source could not be accessed or only a secondary account was available, and each states why.
  • Corrections found in the check: the PubMed links for PAH-029 and PAH-037 pointed to unrelated articles and now point to the correct records; PAH-029’s full title restored; authors and volume details added to PAH-035, PAH-041 and PAH-053; PAH-043 sample (77) and PAH-041 scope (62 studies, 4,439 participants) added.
  • Structured summaries added to 12 further empirical records: PAH-001, 023, 024, 036, 037, 039, 040, 043, 045, 046, 049 and 052.
  • Three founder records added: W-34 and W-35, Egymonousa at Tzaneio Hospital and Elena Venizelou Hospital, and W-36 at Alexandra Hospital, where installation is in process. Their details come from the artist and are labelled as such.
  • The six in-depth case studies extended with further sources, including an independent evaluation of the Great Ormond Street collection and reporting on the covering and re-unveiling of the Sidra sculptures.
  • About page: the repository is funded entirely by its founder, who holds editorial control subject to the declared interest.

Edition 4.8: hospital route, case studies and verification levels · 29 September 2026

  • New home page and main navigation, with separate routes for hospital leaders and researchers. The “For hospitals” page now leads with the board brief, the commissioning checklist and the budget worksheet, and separates institutional purposes from what research has measured.
  • New project decision sheet for a specific commission or donation, in English and Greek, as a web page and a fillable PDF. The board brief now sets out what the evidence supports before its limits.
  • Research library: compact view now lets each record be expanded; a shorter opening with a “Start with these records” editorial introduction; filter help behind a help control; ordering control always visible.
  • Citation corrections: PAH-016 author name corrected to “de Boer C” and DOI added; PAH-038 DOI added; PAH-042 volume and pages added.
  • Every research record now shows a verification level (abstract checked; abstract or primary source checked; full text assessed; or not yet assigned). No record has been independently reviewed. Structured summaries (population, sample, intervention, comparator, outcome measures, timing, result, uncertainty) added for PAH-016, PAH-031 and PAH-038.
  • Six in-depth institutional case studies: Cleveland Clinic, Chelsea and Westminster Hospital, Great Ormond Street Hospital, Sidra Medicine, Mayo Clinic and Liverpool Hospital. Source labels in the case summaries now show readable titles.
  • Case summary CASE-26 (Woodhull Hospital): an earlier edition attached the rediscovery of a different mural in Amsterdam to this work; that account had been removed, and the note describing its removal has now moved from the case text to this change log.
  • Installation cards reordered (photograph, title, artist, hospital, description, source label, record); captions now give the photo credit only; unfinished status wording replaced by a single verification statement.
  • W-05 and W-06 (founder works): material and dimensions aligned with the artist’s catalogue, sources linked to the work pages, photographs credited to Petrides Public Sculpture. W-20: Tate’s catalogue entry added; source category changed to institution or artist, and the loan date settled as 1980.
  • Source library opens with the register’s own citations; the retained original bibliography remains available through the Source set filter.
  • Data: PAH-078 CSV row repaired (an unquoted comma had shifted two columns); appraisal_status updated; structured_summary added to the JSON export.
  • About page: an editorial responsibility section. Roles other than the publisher are listed as not yet appointed; external review remains pending.

Edition 4.7 update: Greek text and design · 28 September 2026

  • A language review of the Greek site corrected page descriptions, terminology, correction notes and Greek downloads; the Greek board brief PDF was regenerated.
  • The budget worksheet now records the non-cash valuation of a donated work on its own line, so it is not added to expenditure.
  • One typeface system for both languages, a sculpture on the home page, photographed works listed first in the catalogue (the founder’s works after the others), fewer visible filters in the research library, a compact view and a marker for the kind of each record.
  • Installation records W-07 and W-19 now give their latest documentary source separately from the date of the photograph; W-07 records the November 2023 removal from the main entrance, with its present location unconfirmed. The Art form filter now lists art forms only: reviews, programmes, guidance and policy that do not study one art form are grouped as “Several or unspecified”, and older links using the previous values still work.

Edition 4.7: corrections and planning tools · 28 September 2026

  • Added planning tools for hospital leaders: documented purposes with examples, catalogue records grouped by location and service, a commissioning checklist and a whole-life budget worksheet without benchmark prices.
  • Installation records now state the last documented location and the date of that source where it is recorded, and whether current placement has been rechecked. Enquiries are marked as sent only where the correspondence is retained; the W-28 enquiry is marked as prepared but not sent.
  • PAH-078: installation permanence is recorded as unclear; the one-year programme is an observation period, not evidence of permanence.
  • Photographs are optional when proposing an installation, on the About and contribution pages and in the record template.
  • Dates now distinguish the latest research update (28 September 2026) from the last full audit of all records (9 September 2026).
  • The installations page title no longer describes the records as illustrated; records without photographs are shown compactly.
  • The Greek text was checked against the English and corrected: terminology, question punctuation (including in the Greek board brief PDF), remaining English glosses, and one link that appeared only on the Greek home page.

Edition 4.6: corrections and research infrastructure · 28 September 2026

  • Installation records now keep source category, claim verification and current status as separate fields. The home page, installations page and this page use the same definitions, and the category date is labelled as the date it was assigned.
  • Withdrew 13 photographs whose publication permission could not be verified (W-01, W-03, W-04, W-08, W-09, W-11 to W-18). The records show a placeholder, and the credits page lists them as not displayed. Every displayed image now has a recorded basis.
  • Aligned the founder policy on the installations, About, sculpture and methods pages: founder installations are disclosed examples of practice, and any independent research concerning them is assessed separately under the same criteria.
  • Relabelled the six sculpture records as sculpture-related records, including one research-gap record.
  • Published a verification ledger for all 33 installation records and a data dictionary. Research records gain an evidence-family field derived from their scope and type, and structured coding of sculpture attribution, exposure, encounter and appraisal status for the sculpture-related and newest records; other records are marked “not yet coded”.
  • Recorded institutional enquiries sent on 28 September 2026 for 14 installation records.
  • Added a proposed evaluation planning template, and a visible edition and date on the board brief PDFs.
  • Not yet done: the six substantial institutional cases, full-text appraisal of PAH-077 to PAH-079, the multi-database search update, publication and study reconciliation, outcome-level data, linked practice and research maps, and an appointed independent reviewer.

Edition 4.5: sculpture as a research question · 28 September 2026

  • Added a sculpture page setting out research questions about sculpture in healthcare, with routes to installations, sculpture research and the evaluation guide.
  • Added definitions of what the site counts, and a verification status with a filter on every installation record.
  • Added PAH-077 (Hitokuwada 2026, student maquettes exhibited at Hiroshima University Hospital), PAH-078 (Monsuez et al. 2019, Le Louvre à l’Hôpital) and PAH-079 (Trupp et al. 2026, guided viewing at the National Gallery). The library now holds 78 records, and the evidence map is updated.
  • Added W-32 Blau (Jaume Plensa, Hospital Clínic Barcelona) and W-33 Mor og barn (Ivar Johnsson, Haukeland University Hospital, Bergen), both from institutional sources and published without photographs until permission is recorded.
  • W-19 confirmed against Guy’s and St Thomas’ sources. W-20 and W-25 given independent sources; W-20 records conflicting loan dates. W-16 marked: placement since the reported 2022 theft and recovery is not confirmed.
  • Corrected the home page, which described the installation records as sculptures by 31 different artists; Jaume Plensa appears more than once. Corrected the Greek installations page, which still gave 18 works.
  • Replaced the home page introduction and counters; each counter now states what it counts.
  • Added an editorial policy and founder disclosure to this page, and placed the current method description ahead of this change log.
  • Reviewed and held: a donor-recognition installation of crystal fish at Sengkang General Hospital, Singapore, pending a scope decision (no named artist).

Edition 4.4: new research, installations and evidence map · 28 September 2026

Adds 15 research records (PAH-057 to PAH-071), bringing the library to 70; 13 installation records from nine countries (W-19 to W-31), bringing the register to 31, each with a Creative Commons photograph; an evidence map; a downloadable board brief in both languages; and a sculpture tag on PAH-024, whose collection includes sculpture. New installation records are compiled from Wikimedia Commons and Wikidata and are marked for confirmation with each institution.

Adds five further records on the Gulf and Middle East (PAH-072 to PAH-076), a regional overview page with an Arabic summary, and an expanded Sidra Medicine case summary. The library now holds 75 records.

Search update · 28 September 2026

This edition adds the first documented search. Three PubMed searches were run on 28 September 2026:

  • S1 (152 results): (artwork*[tiab] OR "visual art"[tiab] OR "works of art"[tiab] OR painting*[tiab] OR mural*[tiab] OR sculpture*[tiab] OR "art installation"[tiab] OR "public art"[tiab]) AND (hospital*[tiab] OR "healthcare environment*"[tiab] OR "health care environment*"[tiab] OR "healthcare setting*"[tiab] OR ward[tiab] OR wards[tiab] OR "waiting room*"[tiab] OR clinic*[tiab]) AND (view*[tiab] OR exposure[tiab] OR displayed[tiab] OR installed[tiab] OR "built environment"[tiab]) NOT ("art therapy"[tiab]) AND 2019:2026[dp]
  • S2 (29 results): ("public art"[tiab] OR "public sculpture"[tiab] OR "street art"[tiab]) AND (health[tiab] OR wellbeing[tiab] OR "well-being"[tiab] OR stress[tiab] OR mood[tiab]) AND 2019:2026[dp]
  • S3 (18 results): ("art"[ti] OR "artwork"[ti] OR "artworks"[ti] OR "paintings"[ti]) AND (hospital[ti] OR patients[ti] OR "healthcare"[ti]) AND (anxiety[tiab] OR stress[tiab] OR pain[tiab] OR satisfaction[tiab] OR wellbeing[tiab]) AND (randomized[tiab] OR randomised[tiab] OR trial[tiab] OR controlled[tiab] OR "quasi-experimental"[tiab]) NOT ("art therapy"[tiab] OR "art-making"[tiab] OR "music"[ti]) AND 2015:2026[dp]

They returned 199 unique records, of which 198 could be retrieved. One editor screened titles and abstracts: 8 were included, 6 were already in the repository, and the rest were excluded, most because they did not concern viewing installed art in healthcare. Seven further records came from sources suggested by readers and were checked against their primary sources. Screening decisions are published in the search log (CSV).

Limits: one database, one screener, English-language search terms and no formal risk-of-bias assessment. This is an update, not a systematic review.

Quality check, 28 September 2026: added a sitemap, a 404 page, sharing images and security headers; compressed large photographs; fixed a narrow-screen overflow on the Sources page; corrected one status in the review queue and one out-of-date self-reference.

Edition 4.3: orientation, frameworks and contributions · 28 September 2026

Following comments received after the September 2026 launch, this edition adds entry points by role, theme and country on the home page; a page on frameworks that propose how art might affect people; a contribution page with minimum and optional fields and a public queue of suggested sources awaiting review; a page for the Athens roundtable; an independent-review statement; and a rights-holder contact on the image credits page. No research record, finding or grade was changed, and suggested sources are not counted in the 55 records until appraised.

Edition 4.2: navigation and usability · 13 September 2026

Added three starting routes, a credited installation photograph, simpler search, a sculpture-specific shortcut and compact installation filters. Study type, source format and methods availability are now separate fields. Mixed and null findings have distinct, overlapping labels. Research summaries, citations and source-verification dates are preserved; this update is not a new research audit.

Proposed research-development plan

The next research update should publish its databases, exact search strings, search date, eligibility rules, screening decisions and duplicate-publication checks. A starting search would combine visual art terms (artwork, mural, painting, photograph, sculpture) with healthcare settings (hospital, healthcare, inpatient, outpatient, clinic), with the full syntax adapted and reported for each database. This is a proposed approach, not a search already conducted.

A proposed quarterly review would check new studies, corrections and preprint publication status, with an annual review of the search strategy and coverage. No recurring review process is currently claimed. Independent reviewers have not yet been appointed; names, qualifications and interests should be published when their participation is confirmed.

Edition 4.1: repository revision

Replaced the promotional homepage with research search, removed founder artwork from the homepage and social metadata, made limitations visible beside findings, standardised finding labels, corrected evidence-type labels for PAH-001, PAH-003, PAH-008, PAH-009 and PAH-019 to match their recorded designs, expanded inclusion and disclosure information, and increased text contrast throughout. The 55 research/context records, 18 works, 27 institutional cases and original citation IDs are retained.

Research corrections retained from Edition 4.0

  • PAH-004: corrected the visitor count to 534, applying the journal’s April 2026 correction.
  • PAH-016: added the trial’s finding: better perceptions of the environment, with no detected psychological or clinical difference.
  • PAH-031: distinguished 525 responses from the 327 survey responses and retained the preprint and cross-sectional limitations.
  • PAH-042: aligned the summary with the review’s conclusion about eudemonic wellbeing and lack of strong support for other outcomes.
  • W-14: identified Emma Rodgers’s sculpture as New Life (2013), outside the Wirral main entrance, rather than an untitled indoor Liverpool work.
  • Woodhull: removed the restoration story belonging to a different Keith Haring mural in Amsterdam.
  • Eskenazi / Kaiser: separated Rob Ley’s May/September at Eskenazi from Double Exposure at Kaiser Permanente, Dublin.
  • Houston: separated the broader Texas Medical Center district from MD Anderson-specific claims.
  • Clinical and financial claims: removed unsupported generalisations about inflammation, dopamine, length of stay, reimbursement and return on investment.
  • Evaluation: removed the arbitrary 30-person minimum and clarified measurement timing, comparison, oversight and instrument suitability.

Latest editorial update: 29 September 2026 (Edition 4.11). Last full audit of all records: 9 September 2026. Corrections and proposed additions can be sent to research@publicart.health, with a citation and primary-source link.

What the repository holds

Research records
78
Studies, reviews and context records. Some describe the same publication. Browse
Artwork/project records
36
Sculptures in healthcare settings, each with its sources. Browse
Institutional case summaries
27
Accounts of institutional art programmes; some cover several sites. Read

Counts from the Edition 4.11 release data, 29 September 2026. Of the 36 artwork/project records, 23 cite institutional, commissioner or artist sources, 3 cite independent sources only and 10 await institutional confirmation; a source category does not mean every detail has been verified. How we count Of these, 35 document placement and one is an installation in progress; current placement is not confirmed for every work.