Tackling algorithmic bias and promoting transparency in health datasets: the STANDING Together consensus recommendationsUniv Birmingham, England; Roche Diagnost, Switzerland; Univ Cambridge, England.
MIT, MA 02139 USA.
Google Res, CA USA.
Google Res, Canada.
Google, CA USA.
Imperial Coll London, England.
Univ Hosp Birmingham NHS Fdn Trust, England; Natl Inst Hlth & Care Res NIHR, England; Univ Birmingham, England.
MIT, MA 02139 USA.
Univ Hosp Birmingham NHS Fdn Trust, England.
Independent Canc PatientsVoice, England.
NIHR Sheffield Biomed Res Ctr, England.
Univ Sheffield, England.
UCL, England.
Tethyan Consulting, Australia.
Oxford Univ Hosp NHS Fdn Trust, England; Univ Oxford, England.
UCL, England; PATH, WA USA; WellcomeTrust, England.
Newcastle Univ, England; Gateshead Council, England.
Google Res, CA USA.
Google, England.
Genom England, England; Alan Turing Inst, England.
Univ Adelaide, Australia.
Med & Healthcare Prod Regulatory Agcy, England.
Harvard Univ, MA 02138 USA.
Med & Healthcare Prod Regulatory Agcy, England.
African Biobanks & Longitudinal Epidemiol Ecosyst, Nigeria.
Hlth Res Author, England.
UCL, England.
King Hussein Canc Ctr, Jordan.
Harrison Ai, Australia; Alfred Hlth, Australia; Monash Univ, Australia.
Swansea Univ, Wales.
NIH, MD USA.
Emory Univ, GA USA.
Nat Res, England.
The Lancet, England.
Mbarara Univ Sci & Technol, Uganda.
Univ Southampton, England; Univ Hosp Southampton NHS Fdn Trust, England.
Univ Cambridge, England.
WHO, Switzerland.
Univ Cambridge, England.
Univ Hosp Birmingham NHS Fdn Trust, England; Natl Inst Hlth & Care Res NIHR, England; Univ Birmingham, England; Univ Birmingham, England; HDR UK Hlth Data Hub Acute Care, England; NIHR Midlands Appl Res Collaborat, England; NIHR Midlands Patient Safety Res Collaborat, England.
Univ Southampton, England; Natl Inst Hlth & Care Res, England.
British Stand Inst, England.
Univ Hosp Birmingham NHS Fdn Trust, England; Natl Inst Hlth & Care Res NIHR, England; Univ Birmingham, England; Univ Birmingham, England; Univ Birmingham, England; UCL, England; Moorfields Eye Hosp, England.
Univ Hosp Birmingham NHS Fdn Trust, England; Natl Inst Hlth & Care Res NIHR, England; Univ Birmingham, England; Univ Birmingham, England.
Show others and affiliations
2025 (English)In: The Lancet Digital Health, E-ISSN 2589-7500, Vol. 7, no 1, p. e64-e88Article, review/survey (Refereed) Published
Abstract [en]
Without careful dissection of the ways in which biases can be encoded into artificial intelligence (AI) health technologies, there is a risk of perpetuating existing health inequalities at scale. One major source of bias is the data that underpins such technologies. The STANDING Together recommendations aim to encourage transparency regarding limitations of health datasets and proactive evaluation of their effect across population groups. Draft recommendation items were informed by a systematic review and stakeholder survey. The recommendations were developed using a Delphi approach, supplemented by a public consultation and international interview study. Overall, more than 350 representatives from 58 countries provided input into this initiative. 194 Delphi participants from 25 countries voted and provided comments on 32 candidate items across three electronic survey rounds and one in- person consensus meeting. The 29 STANDING Together consensus recommendations are presented here in two parts. Recommendations for Documentation of Health Datasets provide guidance for dataset curators to enable transparency around data composition and limitations. Recommendations for Use of Health Datasets aim to enable identification and mitigation of algorithmic biases that might exacerbate health inequalities. These recommendations are intended to prompt proactive inquiry rather than acting as a checklist. We hope to raise awareness that no dataset is free of limitations, so transparent communication of data limitations should be perceived as valuable, and absence of this information as a limitation. We hope that adoption of the STANDING Together recommendations by stakeholders across the AI health technology lifecycle will enable everyone in society to benefit from technologies which are safe and effective.
Place, publisher, year, edition, pages
ELSEVIER , 2025. Vol. 7, no 1, p. e64-e88
National Category
Epidemiology
Identifiers
URN: urn:nbn:se:liu:diva-211315DOI: 10.1016/S2589-7500(24)00224-3ISI: 001395769600001PubMedID: 39701919Scopus ID: 2-s2.0-85212816427OAI: oai:DiVA.org:liu-211315DiVA, id: diva2:1934730
Note
Funding Agencies|NHS AI Lab; Health Foundation; National Institute for Health and Care Research (NIHR) [AI_HI200014]
2025-02-052025-02-052025-05-15