How AI challenges the medical device regulation: patient safety, benefits, and intended uses
Open accessThe DOI matches its Crossref record; 30 of the 30 references it cites were found in a registry or an index.
Abstract
This article examines whether the EU Medical Device Regulation (MDR) adequately addresses the novel risks of AI-based medical devices (AIaMDs), focusing on AI medical imaging tools. It examines two questions: first, does the MDR effectively deal with issues of adaptability, autonomy, bias, opacity, and the need of trustworthiness of AIaMD? Second, does the manufacturer’s translation of the MDR’s requirements close a discrepancy between an AIaMDs’ expected benefit and the actual clinical utility of assessing device safety and effectiveness beyond the narrow performance of algorithms? While the first question has previously received attention in scholarly literature on regulatory and policy tensions on AIaMD generally, and work on future technical standard setting, the second has been comparatively overlooked. We argue that effective regulation of AIaMD requires framing notions of patient safety and benefit within the manufacturer’s articulation of the device’s intended use, as well as reconciling tensions. These tensions are on (i) patient safety and knowledge gaps surrounding fairness, (ii) trustworthiness and device effectiveness, (iii) the assessment of clinical performance, and (iv) performance updates. Future guidance needs to focus on the importance of translated benefits, including nuanced risk framing and looking at how the limitations of AIaMD inform the intended purpose statement in the MDR.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
The false hope of current approaches to explainable artificial intelligence in health care
Ghassemi, M., Oakden-Rayner, L., & Beam, A. L. (2021) · The Lancet Digital Health · doi:10.1016/s2589-7500(21)00208-9 · Cited by 1,724
Liu, X., Cruz Rivera, S., Moher, D., Calvert, M. J., Denniston, A. K., The SPIRIT-AI and CONSORT-AI Working Group, SPIRIT-AI and CONSORT-AI Steering Group, Chan, A.-W., Darzi, A., Holmes, C., et al. (2020) · Nature Medicine · doi:10.1038/s41591-020-1034-x · Cited by 1,421
Human-computer collaboration for skin cancer recognition
Tschandl, P., Rinner, C., Apalla, Z., Argenziano, G., Codella, N., Halpern, A., Janda, M., Lallas, A., Longo, C., Malvehy, J., et al. (2020) · Nature Medicine · doi:10.1038/s41591-020-0942-0 · Cited by 1,024
Collins, G. S., Dhiman, P., Andaur Navarro, C. L., Ma, J., Hooft, L., Reitsma, J. B., Logullo, P., Beam, A. L., Peng, L., Van Calster, B., et al. (2021) · BMJ Open · doi:10.1136/bmjopen-2020-048008 · Cited by 852
Liu, X., Rivera, S. C., Moher, D., Calvert, M. J., & Denniston, A. K. (2020) · BMJ · doi:10.1136/bmj.m3164 · Cited by 702
Beede, E., Baylor, E., Hersch, F., Iurchenko, A., Wilcox, L., Ruamviboonsuk, P., & Vardoulakis, L. M. (2020) · Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3313831.3376718 · Cited by 614
Durán, J. M., & Jongsma, K. R. (2021) · Journal of Medical Ethics · doi:10.1136/medethics-2020-106820 · Cited by 446
Daneshjou, R., Smith, M. P., Sun, M. D., Rotemberg, V., & Zou, J. (2021) · JAMA Dermatology · doi:10.1001/jamadermatol.2021.3129 · Cited by 403
Sounderajah, V., Ashrafian, H., Golub, R. M., Shetty, S., De Fauw, J., Hooft, L., Moons, K., Collins, G., Moher, D., Bossuyt, P. M., et al. (2021) · BMJ Open · doi:10.1136/bmjopen-2020-047709 · Cited by 333
Continuous Learning AI in Radiology: Implementation Principles and Early Applications
Pianykh, O. S., Langs, G., Dewey, M., Enzmann, D. R., Herold, C. J., Schoenberg, S. O., & Brink, J. A. (2020) · Radiology · doi:10.1148/radiol.2020200038 · Cited by 237
The DECIDE-AI Steering Group (2021) · Nature Medicine · doi:10.1038/s41591-021-01229-5 · Cited by 184
Heydon, P., Egan, C., Bolter, L., Chambers, R., Anderson, J., Aldington, S., Stratton, I. M., Scanlon, P. H., Webster, L., Mann, S., et al. (2021) · British Journal of Ophthalmology · doi:10.1136/bjophthalmol-2020-316594 · Cited by 183
Human Factors Challenges for the Safe Use of Artificial Intelligence in Patient Care
Sujan, M., Furniss, D., Grundy, K., Grundy, H., Nelson, D., Elliott, M., White, S., Habli, I., & Reynolds, N. (2019) · BMJ Health & Care Informatics · doi:10.1136/bmjhci-2019-100081 · Cited by 144
Felderer, M., & Ramler, R. (2021) · Lecture notes in business information processing · doi:10.1007/978-3-030-65854-0_3 · Cited by 128
Automated Classification of Skin Lesions: From Pixels to Practice
Narla, A., Kuprel, B., Sarin, K., Novoa, R., & Ko, J. (2018) · Journal of Investigative Dermatology · doi:10.1016/j.jid.2018.06.175 · Cited by 123
Algorithm Change Protocols in the Regulation of Adaptive Machine Learning–Based Medical Devices
Gilbert, S., Fenech, M., Hirsch, M., Upadhyay, S., Biasiucci, A., & Starlinger, J. (2021) · Journal of Medical Internet Research · doi:10.2196/30545 · Cited by 100
Tackling bias in AI health datasets through the STANDING TOGETHER initiative
Ganapathi, S., Palmer, J., Alderman, J. E., Calvert, M., Espinoza, C., Gath, J., Ghassemi, M., Heller, K., Mckay, F., Karthikesalingam, A., et al. (2022) · Nature Medicine · doi:10.1038/s41591-022-01987-w · Cited by 82
Not all biases are bad: equitable and inequitable biases in machine learning and radiology
Pot, M., Kieusseyan, N., & Prainsack, B. (2021) · Insights into Imaging · doi:10.1186/s13244-020-00955-7 · Cited by 73
Niemiec, E. (2022) · DIGITAL HEALTH · doi:10.1177/20552076221089079 · Cited by 55
Patient wisdom should be incorporated into health AI to avoid algorithmic paternalism
McCradden, M. D., & Kirsch, R. E. (2023) · Nature Medicine · doi:10.1038/s41591-023-02224-8 · Cited by 36
Changes in software as a medical device based on artificial intelligence technologies
Zinchenko, V. V., Четвериков, С. Ф., Akhmad, E. S., Arzamasov, K. M., Vladzymyrskyy, A. V., Андрейченко, А. Е., & Morozov, S. P. (2022) · International Journal of Computer Assisted Radiology and Surgery · doi:10.1007/s11548-022-02669-1 · Cited by 30
Vodrahalli, K., Ko, J., Chiou, A. S., Novoa, R., Abid, A., Phung, M., Yekrang, K., Petrone, P., Zou, J., & Daneshjou, R. (2023) · JAMA Dermatology · doi:10.1001/jamadermatol.2023.0091 · Cited by 30
Trustworthy medical AI systems need to know when they don’t know
Grote, T. (2021) · Journal of Medical Ethics · doi:10.1136/medethics-2021-107463 · Cited by 20
Charalambides, M., Flohr, C., Bahadoran, P., & Matin, R. (2021) · British Journal of Dermatology · doi:10.1111/bjd.19616 · Cited by 15
When Medical Devices Have a Mind of Their Own: The Challenges of Regulating Artificial Intelligence
Boubker, J. (2021) · American Journal of Law & Medicine · doi:10.1017/amj.2022.3 · Cited by 14
Distributed Governance of Medical AI
Price, W. N. (2022) · SMU Science and Technology Law Review · doi:10.25172/smustlr.25.1.2 · Cited by 14
Sources of Risk and Design Principles of Trustworthy Artificial Intelligence
Steimers, A., & Bömer, T. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-77820-0_18 · Cited by 6
Sahiner, B., Friedman, B. A., Linville, C., et al. (2018) · Exhibit - A Showcase of Scholarship, Creativity and Preservation Provided by Xavier University Library (Xavier University) · Cited by 5
Research works printed but not linked: 9
- Consider the MDR's emphasis on human-factors which includes risk management, design and specifica- tions and adverse incidence reporting; Article 2 (64) MDR; Annex I, Chapter I, Section 5 (a) MDR; Annex I, Chapter II, Section 14.6 MDR. 49
- Annex I, Chapter III, Section 23.4 (b) MDR; see also, Article 2 (12) MDR. 59 Article 61 MDR; Annex XIV, Part A MDR. 60 Article 2 (52) MDR. 132 See for instance Daneshjou et al., supra note 122;
- David Wen et al., Characteristics of Publicly Available Skin Cancer Image Datasets: A Systematic Review, 4 The Lancet Digital Health e64, 70 (2022);
- Inioluwa Deborah Raji et al., The Fallacy of AI Functionality, in 2022 ACM Conference on Fairness, Accountability, and Transparency 959, 961 (2022), https://dl.acm.org/doi/10.1145/3531146.3533158 (accessed Jun 1, 2023); · DOI · Paper page
- Federico Cabitza, Andrea Campagner & Clara Balsano, Bridging the 'Last Mile' Gap between AI Implementation and Operation: 'Data Awareness' That Matters, 8 Annals of Translational Medicine 501 (2020). 169
- See also, David Lyell & Enrico Coiera, Automation Bias and Verification Complexity: A Systematic Review, 24 Journal of the American Medical Informatics Association 423, 426 (2017);
- See Daneshjou et al. advocating for guidelines that 'address dermatology specific aspects such as image source, lack of standardization, skin tone and considerations of bias'; Roxana Daneshjou et al., CheckList for Evaluation of Image-Based AI Reports in Dermatology: CLEAR Derm Consensus Guidelines from the International Skin Imaging Collaboration Artificial Intelligence Working Group, 158 JAMA Dermatol 90, 3 (2022). 205
- Annex I, Chapter II, Section 17.1 MDR; see also, Johan Ordish, Hannah Murfet, and Alison Hall, supra note 13, at 27. 213 The European Coordination Committee of the Radiological, Electromedical and Healthcare IT Industry (COCIR), supra note 14, at 13. 214 See Annex IX, Chapter 1, Section 2.4 MDR. 215 BS EN 62304:2006 + A1:2015 Medical device software. Software life-cycle processes, supra note 37, at 8.2. 216 Berkman Sahiner et al., supra note 149, at 7. 217 See for instance discussion in BSI and AAMI, supra note 13, at 9. 218
- Mike Nix, George Onisiforou & Annabelle Painter, Understanding Healthcare Workers' Confidence in AI: Report 1 of 2, 26 (2022), https://digital-transformation.hee.nhs.uk/binaries/content/assets/digita l-transformation/dart-ed/understandingconfidenceinai-may22.pdf (accessed Mar. 28, 2024). 243 See also discussion in Sujan et al., supra note 236. 244 Article 61
Cited by
Ethical and legal considerations in healthcare AI: innovation and policy for safe and fair use
Pham, T. (2025) · Royal Society Open Science · doi:10.1098/rsos.241873 · Cited by 261
The EU Artificial Intelligence Act (2024): Implications for healthcare
van Kolfschooten, H., & van Oirschot, J. (2024) · Health Policy · doi:10.1016/j.healthpol.2024.105152 · Cited by 314
Metadata
- Authors
- Onitiu, D., Wachter, S., & Mittelstadt, B.
- Affiliations
- University of Oxford (GB)
- Corresponding author
- Onitiu, Daria
- Venue
- Journal of Law and the Biosciences
- Publisher
- Oxford University Press (OUP)
- Published
- 2024-04-09
- ISSN
- 2053-9711 (ISSN-L)
- OpenAlex
- W4394593980
- Cited by
- 61
- References
- 30
- Citations by year
- 2026: 27 · 2025: 30 · 2024: 4
- FWCI
- 157.71
- Type
- journal article
- Language
- en
- Open access
- diamond
- doi.org
- Licence
- Locations
- Journal of Law and the Biosciences (publishedVersion, cc-by)
- SSRN Electronic Journal (acceptedVersion)
- Funders
- Wellcome Trust (223765/Z/21/Z)
- Sloan Foundation (G-2021-16779)
- Department of Health and Social Care
- Luminate Group
- Trustworthiness Auditing for AI
- Governance of Emerging Technologies
- Oxford Internet Institute
- University of Oxford
- Topics
- Social Sciences › Law › Digitalization, Law, and Regulation
- Health Sciences › Medicine › Public Health, Environmental and Occupational Health › Healthcare Regulation
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Keywords
- AI-based medical devices, EU Medical Device Regulation, medical device regulation, patient safety, intended use, clinical benefit, algorithmic bias, algorithmic transparency, trustworthiness, clinical performance assessment, risk framing, fairness
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.