Skip to content
DraconiticResearch
Search Sources API
186,123 studies31 sourcesUpdated 53 min ago
Peer-reviewedCC BY 4.0PLOS Digital Health

When the algorithm passes the test but fails the patient

  • Alex Mirugwe1
  1. 1Makerere University, School of Public Health, Kampala, Uganda
Read full textPDFDOICite

Abstract

Tens of thousands of AI diagnostic tools have been authorised for clinical diagnosis worldwide, justified primarily by benchmark accuracy (performance on a fixed, curated test dataset), a number that says little about what the tool will do at the bedside. The field rests on three flawed foundations: datasets that do not reflect the populations served, development pipelines that exclude clinicians, and regulatory pathways that mistake on-paper accuracy for evidence of benefit. Closing the gap requires global coordination and the political will to treat clinical validation as a prerequisite rather than an afterthought.

Related

  1. When the Cytoskeleton Fails: Primary Structural Defects Underlying Disorders with Neurological Involvement
    Gerasimos Terzopoulos, Evangelos Axarloglou, Evangelie Sintou +6
    International Journal of Molecular SciencesPeer-reviewedCC BY 4.0Full textNeuroscienceCrossrefEurope PMC
  2. A Connectome Test of the Fly Hashing Algorithm
    Sebastian Senge
    arXivPreprintCC BY 4.0Full textNeurosciencearXiv
  3. To be or not to appear tired: When the « good patient » takes the stage
    Victor Poupard, Stéphane Héas, Omar Zanna
    Anthropologie et SantéPeer-reviewedCC BY-NC-ND 4.0Full textGeneral biologyOncologyHAL
  4. In vivo CAR-T cells: When the patient becomes their own factory
    François Vilcot, Carole-Anne Brugère, Jaime Fuentealba +1
    Bulletin du CancerPeer-reviewedCC BYFull textGeneral biologyOncologyHAL
  5. Patient experiences with hereditary cancer genetic test result release in the setting of the 21st Century Cures Act final rule
    Cameron Bell, Samantha Pipito, Ashika Mani +22
    Familial CancerPeer-reviewedCC BY 4.0Full textGenetics & genomicsPMC

Cite

BibTeXRIS
BibTeX
@article{mirugwe2026when,
  title = {{When the algorithm passes the test but fails the patient}},
  author = {Alex Mirugwe},
  journal = {PLOS Digital Health},
  year = {2026},
  month = oct,
  doi = {10.1371/journal.pdig.0001774},
  url = {https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0001774},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - When the algorithm passes the test but fails the patient
AU  - Alex Mirugwe
PY  - 2026
DA  - 2026/10/06
JO  - PLOS Digital Health
DO  - 10.1371/journal.pdig.0001774
UR  - https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0001774
AB  - Tens of thousands of AI diagnostic tools have been authorised for clinical diagnosis worldwide, justified primarily by benchmark accuracy (performance on a fixed, curated test dataset), a number that says little about what the tool will do at the bedside. The field rests on three flawed foundations: datasets that do not reflect the populations served, development pipelines that exclude clinicians, and regulatory pathways that mistake on-paper accuracy for evidence of benefit. Closing the gap requires global coordination and the political will to treat clinical validation as a prerequisite rather than an afterthought.
SN  - 2767-3170
ER  -