The selective deployment of AI in healthcare

Machine‐learning algorithms have the potential to revolutionise diagnostic and prognostic tasks in health care, yet algorithmic performance levels can be materially worse for subgroups that have been underrepresented in algorithmic training data. Given this epistemic deficit, the inclusion of underr...

Full description

Saved in:
Bibliographic Details
Published in:Bioethics Vol. 38; no. 5; pp. 391 - 400
Main Authors: Vandersluis, Robert, Savulescu, Julian
Format: Journal Article
Language:English
Published: 01.06.2024
Subjects:
ISSN:0269-9702, 1467-8519
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Abstract Machine‐learning algorithms have the potential to revolutionise diagnostic and prognostic tasks in health care, yet algorithmic performance levels can be materially worse for subgroups that have been underrepresented in algorithmic training data. Given this epistemic deficit, the inclusion of underrepresented groups in algorithmic processes can result in harm. Yet delaying the deployment of algorithmic systems until more equitable results can be achieved would avoidably and foreseeably lead to a significant number of unnecessary deaths in well‐represented populations. Faced with this dilemma between equity and utility, we draw on two case studies involving breast cancer and melanoma to argue for the selective deployment of diagnostic and prognostic tools for some well‐represented groups, even if this results in the temporary exclusion of underrepresented patients from algorithmic approaches. We argue that this approach is justifiable when the inclusion of underrepresented patients would cause them to be harmed. While the context of historic injustice poses a considerable challenge for the ethical acceptability of selective algorithmic deployment strategies, we argue that, at least for the case studies addressed in this article, the issue of historic injustice is better addressed through nonalgorithmic measures, including being transparent with patients about the nature of the current epistemic deficits, providing additional services to algorithmically excluded populations, and through urgent commitments to gather additional algorithmic training data from excluded populations, paving the way for universal algorithmic deployment that is accurate for all patient groups. These commitments should be supported by regulation and, where necessary, government funding to ensure that any delays for excluded groups are kept to the minimum. We offer an ethical algorithm for algorithms—showing when to ethically delay, expedite, or selectively deploy algorithmic systems in healthcare settings.
AbstractList Machine‐learning algorithms have the potential to revolutionise diagnostic and prognostic tasks in health care, yet algorithmic performance levels can be materially worse for subgroups that have been underrepresented in algorithmic training data. Given this epistemic deficit, the inclusion of underrepresented groups in algorithmic processes can result in harm. Yet delaying the deployment of algorithmic systems until more equitable results can be achieved would avoidably and foreseeably lead to a significant number of unnecessary deaths in well‐represented populations. Faced with this dilemma between equity and utility, we draw on two case studies involving breast cancer and melanoma to argue for the selective deployment of diagnostic and prognostic tools for some well‐represented groups, even if this results in the temporary exclusion of underrepresented patients from algorithmic approaches. We argue that this approach is justifiable when the inclusion of underrepresented patients would cause them to be harmed. While the context of historic injustice poses a considerable challenge for the ethical acceptability of selective algorithmic deployment strategies, we argue that, at least for the case studies addressed in this article, the issue of historic injustice is better addressed through nonalgorithmic measures, including being transparent with patients about the nature of the current epistemic deficits, providing additional services to algorithmically excluded populations, and through urgent commitments to gather additional algorithmic training data from excluded populations, paving the way for universal algorithmic deployment that is accurate for all patient groups. These commitments should be supported by regulation and, where necessary, government funding to ensure that any delays for excluded groups are kept to the minimum. We offer an ethical algorithm for algorithms—showing when to ethically delay, expedite, or selectively deploy algorithmic systems in healthcare settings.
Author Savulescu, Julian
Vandersluis, Robert
Author_xml – sequence: 1
  givenname: Robert
  orcidid: 0009-0003-7197-104X
  surname: Vandersluis
  fullname: Vandersluis, Robert
  email: bob_vandersluis@post.harvard.edu
  organization: GSK
– sequence: 2
  givenname: Julian
  orcidid: 0000-0003-1691-6403
  surname: Savulescu
  fullname: Savulescu, Julian
  organization: National University of Singapore
BookMark eNotj0FOwzAQRS1UJNLChhP4AKR44sSOl6UqEKlSN2Vt2dOxYpQmVRJAuT1t4W3-Xz3pzdms7Vpi7BHEEs48-9jREmRWwg1LIFc6LQswM5aITJnUaJHdsfkwfIozpigS9rSviQ_UEI7xm_iBTk03HakdeRf4quKx5TW5ZqzR9XTPboNrBnr43wX7eN3s1-_pdvdWrVfb9AuUhFQjeoUBlch1GYiAnENAUWqAnDJTaA-onctkIKO1UuBz7dCg9OFgykIuGPx5f2JDkz318ej6yYKwl0h7ibTXSPtS7TbXJ38BKHFImg
ContentType Journal Article
Copyright 2024 The Authors. published by John Wiley & Sons Ltd.
Copyright_xml – notice: 2024 The Authors. published by John Wiley & Sons Ltd.
DBID 24P
DOI 10.1111/bioe.13281
DatabaseName Wiley Online Library Open Access
DatabaseTitleList
Database_xml – sequence: 1
  dbid: 24P
  name: Wiley Online Library Open Access
  url: https://authorservices.wiley.com/open-science/open-access/browse-journals.html
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Medicine
Biology
Philosophy
EISSN 1467-8519
EndPage 400
ExternalDocumentID BIOE13281
Genre article
GroupedDBID ---
--Z
-ET
.3N
.GA
.GJ
.Y3
04C
05W
0R~
10A
186
1OB
1OC
23N
24P
31~
33P
36B
3O-
4.4
44B
50Y
50Z
51W
51Y
52M
52O
52Q
52R
52S
52T
52U
52V
52W
53G
5GY
5HH
5LA
5VS
66C
6J9
6PF
702
7PT
8-0
8-1
8-3
8-4
8-5
8UM
930
A01
A04
AABCJ
AABNI
AAESR
AAHHS
AAHQN
AAIPD
AAKAS
AAMNL
AANHP
AAONW
AAOUF
AASGY
AAWTL
AAXRX
AAYCA
AAZKR
ABCQN
ABCUV
ABDBF
ABEML
ABIVO
ABJNI
ABLJU
ABPPZ
ABPVW
ABQWH
ABSOO
ABTAH
ABXGK
ACAHQ
ACBKW
ACBWZ
ACCFJ
ACCZN
ACFBH
ACGFO
ACGFS
ACGOF
ACHQT
ACJZB
ACMXC
ACPOU
ACPRK
ACRPL
ACSCC
ACUHS
ACXQS
ACYXJ
ADBBV
ADBTR
ADEMA
ADEOM
ADIZJ
ADKYN
ADMGS
ADMHG
ADNMO
ADOJX
ADXAS
ADZCM
ADZMN
ADZOD
AEEZP
AEGXH
AEIGN
AEIMD
AENEX
AEQDE
AEUQT
AEUYR
AFBPY
AFEBI
AFFNX
AFFPM
AFGKR
AFKFF
AFPWT
AFWVQ
AFYRF
AFZJQ
AHBTC
AHEFC
AHMBA
AIACR
AIAGR
AIFKG
AIURR
AIWBW
AJBDE
ALAGY
ALMA_UNASSIGNED_HOLDINGS
ALUQN
ALVPJ
AMBMR
AMYDB
ASPBG
ASTYK
AVWKF
AZBYB
AZFZN
AZVAB
BAFTC
BDRZF
BFHJK
BMSDO
BMXJE
BNVMJ
BQESF
BROTX
BRXPI
BY8
CAG
COF
CS3
D-6
D-7
D-C
D-D
DC6
DCZOG
DPXWK
DR2
DRFUL
DRMAN
DRSSH
DU5
EAD
EAP
EAS
EBC
EBD
EBS
ECF
ECT
ECV
EHE
EIHBH
EJD
EMB
EMK
EMOBN
ENC
EPT
ESX
F00
F01
F5P
FEDTE
FUBAC
FZ0
G-S
G.N
G50
GODZA
GXZFM
HGLYW
HVGLF
HZI
HZ~
H~9
IHE
IX1
J0M
KBYEO
LATKE
LC2
LC4
LEEKS
LH4
LITHE
LOXES
LP6
LP7
LUTES
LW6
LYRES
MEWTI
MK4
MRFUL
MRMAN
MRSSH
MSFUL
MSMAN
MSSSH
MVM
MXFUL
MXMAN
MXSSH
N04
N06
N9A
NF~
O66
O9-
OIG
OVD
P2P
P2W
P2Y
P2Z
P4B
P4C
PALCI
PQQKQ
Q.N
Q11
QB0
Q~Q
R.K
RIWAO
RJQFR
ROL
RWL
RX1
RXW
SAMSI
SUPJJ
SV3
TAE
TEORI
TN5
TUS
UB1
UPT
V8K
W8V
W99
WBKPD
WGLLI
WH7
WIH
WII
WIJ
WOHZO
WQ9
WQZ
WRC
WSUWO
WXI
WXSBR
XG1
XJT
XSW
YCJ
YUY
ZGI
ZY4
ZZTAW
~IA
~WP
ID FETCH-LOGICAL-u1631-7ccb6cfc60478fee1eaac1c087114e2957b1c7aa23fe977661b47ac9c3bfd9853
IEDL.DBID 24P
ISICitedReferencesCount 13
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001193591300001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 0269-9702
IngestDate Wed Jan 22 17:20:44 EST 2025
IsDoiOpenAccess true
IsOpenAccess true
IsPeerReviewed true
IsScholarly true
Issue 5
Language English
License Attribution
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-u1631-7ccb6cfc60478fee1eaac1c087114e2957b1c7aa23fe977661b47ac9c3bfd9853
Notes The views expressed in the paper are the author's own, and do not necessarily reflect of views of GSK or GSK.ai.
ORCID 0000-0003-1691-6403
0009-0003-7197-104X
OpenAccessLink https://onlinelibrary.wiley.com/doi/abs/10.1111%2Fbioe.13281
PageCount 10
ParticipantIDs wiley_primary_10_1111_bioe_13281_BIOE13281
PublicationCentury 2000
PublicationDate June 2024
PublicationDateYYYYMMDD 2024-06-01
PublicationDate_xml – month: 06
  year: 2024
  text: June 2024
PublicationDecade 2020
PublicationTitle Bioethics
PublicationYear 2024
SSID ssj0000955
Score 2.420889
Snippet Machine‐learning algorithms have the potential to revolutionise diagnostic and prognostic tasks in health care, yet algorithmic performance levels can be...
SourceID wiley
SourceType Publisher
StartPage 391
SubjectTerms algorithm
artificial intelligence
bias
exclusion
machine learning
melanoma
Title The selective deployment of AI in healthcare
URI https://onlinelibrary.wiley.com/doi/abs/10.1111%2Fbioe.13281
Volume 38
WOSCitedRecordID wos001193591300001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVWIB
  databaseName: Wiley Online Library - Journals
  customDbUrl:
  eissn: 1467-8519
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0000955
  issn: 0269-9702
  databaseCode: DRFUL
  dateStart: 19970101
  isFulltext: true
  titleUrlDefault: https://onlinelibrary.wiley.com
  providerName: Wiley-Blackwell
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1LSwMxEB6KVelFbVV8lhw8iSub3W2yAS9VWyxoLWKhtyVP7GUrfQj99ybZbqtH8ZZDAmGSmfkmmZkP4CpOhBEGqyDUNAwSkwqrUloFpCU5IVozwUNPNkH7_XQ0YoMK3JW1MEV_iPWDm9MMb6-dgnMx-6HkYjzRtzaWcnXXVYzj1BE3RMlgY4eZ5zy1QQYLGHV5PPUyj2ez9jcm9U6lu_-_7RzA3gpMonZx-nWo6LwBOwW95LIBuy-rj_MG1AYlZcHyEG7s3UAzT4BjbR1S2pH-uldCNDGo3UPjHH2s08KOYNjtvD88BSvWhGBhsRUOqJSCSCOJ67tjtMaac4llaCMjnOiItajAknIexUZb8Gf9s0gol0zGwihmvfcxbOWTXJ8AojjCPObGroxdxW3KuUoiqSyqI4rw6BSuvUiyz6IzRlZGFE4YmRdGdt977fjR2V8mn0MtsuihyMm6gK35dKEvYVt-zcezadOfcxOqj2_d4fM3gJarfA
linkProvider Wiley-Blackwell
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1LTwIxEJ4YfHFRQY1ve_BkXLPtLi09ooFABOSACbdN220Dl8XwMOHf23YX0KPx1sM22Uw7M99MZ-YDeIhiaaTBaRBqFgaxqUurUjoNaE0JSrXmUoSebIL1-_XRiA-K2hzXC5PPh9gk3JxmeHvtFNwlpH9ouZxM9bMNplzj9W5s3Yy75iQebA0x96SnNsrgAWeukKeyLuTZ7v0NSr1XaR3_839O4KiAk6iRn38FdnRWhf2cYHJVhYNe8XRehfJgTVqwOoUnezvQ3FPgWGuHUu1of12eEE0NanTQJEPjTWHYGXy0msPXdlDwJgRLi65wwJSSVBlF3eQdozXWQiisQhsb4VgTXmMSKyYEiYy28M96aBkzobiKpEm59d_nUMqmmb4AxDDBIhLG7oxcz21diDQmKrW4jqZUkEt49DJJPvPZGMk6pnDCSLwwkpfOe9Ovrv7y8T0ctoe9btLt9N-uoUwslsgrtG6gtJgt9S3sqa_FZD6784f-DW1zrcc
linkToPdf http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1LT8JAEJ4YUMJFBTW-3YMnY023LV32iEIjEZEYSbg1-4xcCuFhwr93d1tAj8ZbD91kM92Z-Wb7zXwAt2HENddYer4ivhfpJjcupaQXNwSLY6UoZ74TmyD9fnM0ooOCm2N7YfL5EJsLN-sZLl5bB1dTqX94OR9P1IMppmzjdTmyKjIlKLffk2FvG4qpkz01dQb1KLFUntqayrNd_RuWurySHPxzR4ewXwBK1MpPQA12VFaHvVxiclWHymvx87wO1cFatmB1BPfmfKC5E8Ex8Q5JZYV_7U0hmmjU6qJxhj431LBjGCadj6dnr1BO8JYGX2GPCMFjoUVsZ-9opbBiTGDhm-oIRyqgDcKxIIwFoVYGAJoczSPCBBUh15KaDH4CpWySqVNABAeYhUyblaHtum0yJqNASIPsYhmz4AzunE3SaT4dI11XFdYYqTNG-th967in87-8fAOVQTtJe93-ywVUAwMmcorWJZQWs6W6gl3xtRjPZ9fFV_8GsNiu3Q
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=The+selective+deployment+of+AI+in+healthcare&rft.jtitle=Bioethics&rft.au=Vandersluis%2C+Robert&rft.au=Savulescu%2C+Julian&rft.date=2024-06-01&rft.issn=0269-9702&rft.eissn=1467-8519&rft.volume=38&rft.issue=5&rft.spage=391&rft.epage=400&rft_id=info:doi/10.1111%2Fbioe.13281&rft.externalDBID=10.1111%252Fbioe.13281&rft.externalDocID=BIOE13281
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0269-9702&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0269-9702&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0269-9702&client=summon