acealgorithmic bias in AI health technologies stemming from health dataset limitations and its effect on health inequalities
Source article: Tackling algorithmic bias and promoting transparency in health datasets: the STANDING Together consensus recommendations
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 stakeh…
Contested: both sides are scored from claims and sources, not community votes.
Business ethics 1 by Jan Cardino13. CC BY-SA 4.0 · https://creativecommons.org/licenses/by-sa/4.0
On Dec 18, 2024, The Lancet Digital Health published the STANDING Together consensus recommendations, developed through a systematic review, stakeholder survey, Delphi process with 194 voters from 25 countries, public consultation, and international interviews involving over 350 representatives from 58 countries. The process produced 29 recommendations in two parts covering documentation of health datasets and use of health datasets.
The work matters because AI health technologies built on incompletely described datasets can scale existing health inequalities, and transparent reporting of composition and limitations is positioned as a way to enable proactive bias evaluation. Uncertainty remains about adoption across the AI lifecycle and whether prompting inquiry rather than checklist compliance will translate into measurable reductions in disparate performance.
- More than 350 representatives from 58 countries provided input, with 194 Delphi participants from 25 countries voting on 32 candidate items across three electronic survey rounds and one in-person consensus meeting.
- The 29 final recommendations are presented in two parts: Documentation of Health Datasets for transparency around composition and limitations, and Use of Health Datasets for bias identification and mitigation.
- Draft items were informed by a systematic review and stakeholder survey, with development supplemented by public consultation and international interview study.
STANDING Together consensus recommendations provide guidance for documenting health datasets and for using them to identify and mitigate algorithmic biases, aiming to support AI health technologies that are safe and effective across population groups.
Without careful dissection of how biases are encoded into AI health technologies, underlying health dataset limitations risk perpetuating existing health inequalities at scale.
The rundown
The initiative was developed to address bias encoded in AI health technologies through its data foundations. Draft recommendation items came from a systematic review and stakeholder survey, then refined via Delphi voting and supplemented by public consultation and an international interview study.
Input came from more than 350 representatives from 58 countries, with 194 Delphi participants from 25 countries voting on 32 candidate items over three electronic rounds and one in-person meeting, resulting in 29 recommendations split between documentation guidance for curators and use guidance for bias evaluation across population groups.
No health dataset is free of limitations, so absence of transparent communication about limitations should itself be treated as a limitation, and the recommendations are meant to prompt inquiry rather than serve as a checklist.
Sources
- Peer-reviewedThe Lancet Digital Health2024-12-18
How should this claim be treated?
The debate