Ethical considerations in patient-directed artificial intelligence platforms

Abstract

Dear Dr Dermatoethicist: An anxious 45-year-old presents for evaluation of an arm lesion after using an at-home machine learning photography tool which raised suspicion for a malignancy. Upon clinical evaluation, the lesion is a seborrheic keratosis. How should I ethically advise this patient about their future use of this application?—Concerned Dermatologist Dear Dr Dermatoethicist: An anxious 45-year-old presents for evaluation of an arm lesion after using an at-home machine learning photography tool which raised suspicion for a malignancy. Upon clinical evaluation, the lesion is a seborrheic keratosis. How should I ethically advise this patient about their future use of this application? —Concerned Dermatologist Dear Concerned Dermatologist: Dermatology practices have extended waiting times, which can be challenging for patients and delay evaluation of lesions. Patients may seek alternative diagnostic methods, including artificial intelligence (AI) applications, due to ease of accessibility and availability. These applications may provide reassurance in the case of a benign diagnosis or encourage the patient to seek medical attention sooner. Many applications exist and continue to emerge in the market to evaluate and monitor skin lesions. For example, several popular applications on the iphone operating system and Android App stores incorporate machine learning for diagnostics based on photos uploaded by the patient. They also enable users to take serial photos of skin lesions, allowing them to track changes over time. The ability to evaluate and monitor skin lesions on these platforms increases patient autonomy. Additionally, when they make correct diagnoses, they provide reassurance for the patient, or in the case of a concerning diagnosis can prompt sooner medical attention, meeting the ethical principle of beneficence and reducing maleficence from a delayed diagnosis. These tools can also be useful to clinicians by visually demonstrating interval changes in skin lesions. Some limitations of these platforms include an inferior ability to evaluate uncommon diseases and lesions occurring in darker skin tones.1Beltrami E.J. Grant-Kels J.M. Consulting ChatGPT: ethical dilemmas in language model artificial intelligence.J Am Acad Dermatol. 2023; (S0190-9622(23)00364-X. https://doi.org/10.1016/j.jaad.2023.02.052)Abstract Full Text Full Text PDF Scopus (11) Google Scholar, 2Daneshjou R. Vodrahalli K. Novoa R.A. et al.Disparities in dermatology AI performance on a diverse, curated clinical image set.Sci Adv. 2022; 8eabq6147https://doi.org/10.1126/sciadv.abq6147Crossref PubMed Scopus (39) Google Scholar, 3Liu Y. Primiero C.A. Kulkarni V. Soyer H.P. Betz-Stablein B. Artificial intelligence for the classification of pigmented skin lesions in populations with skin of color: a systematic review.Dermatology. 2023; 239: 499-513https://doi.org/10.1159/000530225Crossref PubMed Scopus (0) Google Scholar As demonstrated by our patient's experience, inaccurate diagnoses may lead to undue anxiety and inadvertently compromise the principle of nonmaleficence. Patients receiving incorrect diagnoses from these applications may request biopsies which would increase unnecessary biopsies, overall cost of care, and patient maleficence. In addition, many of these applications are not verified by an outside expert, although this may be changing. In 2021, Google announced the development of an AI-powered dermatology tool that was comparable to board-certified dermatologists in accuracy for select conditions, though this is not yet commercially available or fully verified.4Liu Y. Jain A. Eng C. et al.A deep learning system for differential diagnosis of skin diseases.Nat Med. 2020; 26: 900-908Crossref PubMed Scopus (306) Google Scholar The worst scenario would occur if the device misdiagnosed a malignancy as benign resulting in further delay in diagnosis and maleficence. To best balance the risks and benefits of using these tools, I suggest you educate your patient on the appropriate use of these applications as a screening device that does not replace evaluation by a dermatologist. Your patient also needs to be educated regarding the need to formally consult a dermatologist for bleeding, nonhealing, changing, and/or pruritic or painful lesion as they have a higher risk of malignancy. AI companies should continue to engage dermatologists who could provide expanded data sets to help develop and optimize the AI algorithms and improve their future diagnostic capabilities.5Zakhem G.A. Fakhoury J.W. Motosko C.C. Ho R.S. Characterizing the role of dermatologists in developing artificial intelligence for assessment of skin cancer.J Am Acad Dermatol. 2021; 85: 1544-1556https://doi.org/10.1016/j.jaad.2020.01.028Abstract Full Text Full Text PDF PubMed Scopus (23) Google Scholar While AI based tools have made significant progress, these technologies are still in their nascency, and dermatologists need to be informed so they can guide patients on their appropriate and ethical use. —Dr Dermatoethicist None disclosed.

Ähnliche Arbeiten