Abstract
As 3D total‐body photography for the early detection of melanoma is not yet widely used in clinical practice, we do not have a full understanding of patient's concerns about use, privacy and confidentiality, and if their concerns differ depending on the use‐case. We conducted a virtual consumer forum to assess patients concerns about privacy and confidentiality in dermatology imaging for research, artificial intelligence development and for their own clinical care.
Keywords: confidentiality, diagnostic imaging, melanoma, patient acceptance of health care, privacy, telemedicine
INTRODUCTION
Imaging technologies such as 3D total‐body photography involves photographing the whole body and is showing promising results to be able to accurately monitor lesions over time. 1 3D total‐body photography presents unique problems in terms of privacy and confidentiality. Images may be visually identifiable, sensitive and potentially unflattering (nude and semi‐nude). They can include the patient's face and other identifiable features (e.g. tattoos, body piercings, birthmark and body habitus) which are easily recognisable. Addressing privacy is imperative for the clinical adoption and secondary use of digital total‐body imaging.
We aimed to investigate the factors that determine consumers' comfort and willingness to share 3D total‐body images for research, artificial intelligence (AI) development, clinical and teaching scenarios.
METHODS
This study was approved by the University of Queensland (UQ) Human Research Ethics Committee (approval number 2022/HE000090). Consumers from previous and current 3D total‐body imaging studies at the UQ Dermatology Research Centre were invited to participate in a one‐hour virtual consumer forum in March 2022 conducted using Zoom video conferencing software.
The forum included 15 optional multi‐choice poll questions spread throughout. At the beginning of the forum, participants were asked to tick the polling box on the screen ‘yes’ if they consent to participate in the polls. Participants were asked their feedback on (i) non‐identifiable images (individual skin lesions that show only the mole or skin lesion, including dermoscopy) and (ii) identifiable and/or sensitive images (3D total‐body images, as well as regional images of individual body parts). Participants were shown examples to define these image categories. Participants were provided with a brief concept of how 3D images would utilise AI prior to asking the AI questions. Respondents answered questions about ‘how comfortable’ they were by responding ‘yes or ‘no’ for each scenario. Once the majority of the audience completed each poll question, the results were then displayed on the screen and read out to the audience by the moderator in de‐identified, aggregated form. Anytime during the forum participants could enter written feedback into the Zoom chat function.
RESULTS
54 attendees joined the consumer forum, and 39 (72.2%) participants consented to participating in the polls. Scenarios where participants were not comfortable sharing non‐identifiable images included: social media (n = 4/39, 10.3%); researchers in overseas research institutes (n = 3/39, 7.7%); publicly accessible databases managed by a research organisation (n = 2/39, 5.1%); and education and training in a classroom setting (n = 1/39, 2.6%) (Table 1). All participants were comfortable sharing non‐identifiable images in all other scenarios presented.
TABLE 1.
Consumer forum poll questions (n = 39)
| Hypothetical scenario | Non‐identifiable n (%) | Identifiable or sensitive n (%) |
|---|---|---|
| Would you feel comfortable having your images taken in a doctor's office and kept for your personal medical record and to assist with diagnosis? (n = 39) | ||
| Yes | 39 (100) | 33 (84.6) |
| No | 0 | 6 (15.4) |
| Did not respond | 0 | 0 |
| Would you be comfortable for your images to go to a controlled database? | ||
| Yes | 33 (84.6) | 18 (46.1) |
| No | 0 | 15 (38.5) |
| Did not respond | 6 (15.4) | 6 (15.4) |
| Would you feel comfortable having your images sent between doctors for a second opinion for diagnostic purposes? | ||
| Yes | 37 (94.9) | 31 (79.5) |
| No | 0 | 6 (15.4) |
| Did not respond | 2 (5.1) | 2 (5.1) |
| Would you feel comfortable having your images sent to a university for research purposes? | ||
| Yes | 38 (97.4) | 23 (58.9) |
| No | 0 | 15 (38.5) |
| Did not respond | 1 (2.6) | 1 (2.6) |
| Would you feel comfortable having your images used in a presentation at a conference? | ||
| Yes | 31 (79.5) | 9 (23.1) |
| No | 0 | 22 (56.4) |
| Did not respond | 8 (20.5) | 8 (20.5) |
| Would you feel comfortable having your images used in a presentation on social media? | ||
| Yes | 28 (71.8) | 2 (5.1) |
| No | 4 (10.3) | 30 (76.9) |
| Did not respond | 7 (17.9) | 7 (17.9) |
| Would you feel comfortable having your images used for medical, nursing or allied health education and training in a classroom setting? | ||
| Yes | 31 (79.5) | 14 (35.9) |
| No | 1 (2.6) | 18 (46.2) |
| Did not respond | 7 (17.9) | 7 (17.9) |
| Would you be comfortable for your images to go to a publicly accessible database managed by a research organisation? (n = 33) | ||
| Yes | 31 (79.5) | 6 (15.4) |
| No | 2 (5.1) | 27 (69.2) |
| Did not respond | 6 (15.4) | 6 (15.4) |
| Would you feel comfortable with your images being shared with researchers from multiple research institutes within Australia? | ||
| Yes | 34 (87.2) | 17 (43.6) |
| No | 0 | 17 (43.6) |
| Did not respond | 5 (12.8) | 5 (12.8) |
| Would you be comfortable for your images to go to a restricted database? | ||
| Yes | 33 (84.6) | 24 (61.5) |
| No | 0 | 9 (23.1) |
| Did not respond | 6 (15.4) | 6 (15.4) |
| Would you feel comfortable having your images shared with researchers in overseas research institutes? | ||
| Yes | 29 (74.4) | 10 (25.6) |
| No | 3 (7.7) | 22 (56.4) |
| Did not respond | 7 (17.9) | 7 (17.9) |
| Would the country influence your decision? | ||
| Yes | 18 (46.2) | |
| No | 14 (35.9) | |
| Did not respond | 7 (17.9) | |
| Artificial Intelligence | ||
| In the past 12 months, have you heard, read or seen anything about Artificial Intelligence? | ||
| Yes | 32 (82.1) | |
| No | 2 (5.1) | |
| Did not respond | 5 (12.8) | |
| To what extent do you feel you know a lot about Artificial Intelligence? | ||
| Low | 11 (28.2) | |
| Moderate | 17 (43.6) | |
| High | 6 (15.4) | |
| Did not respond | 5 (12.8) | |
|
Would you be comfortable for your images to be used to create artificial intelligence algorithms to automatically diagnose melanoma? A university developed the AI The AI algorithms may then be sold for financial gain. | ||
| Yes | 31 (79.5) | 23 (59.0) |
| No | 0 | 8 (20.5) |
| Did not respond | 8 (20.5) | 8 (20.5) |
|
Would you be comfortable for your images to be used to create artificial intelligence algorithms to automatically diagnose melanoma? A company or an industry organisation developed the AI. The AI algorithms may then be sold for financial gain. | ||
| Yes | 31 (79.5) | 13 (33.3) |
| No | 1 (2.6) | 19 (48.7) |
| Did not respond | 7 (17.9) | 7 (17.9) |
The scenarios where more than half of participants were uncomfortable sharing identifiable and/or sensitive images were social media (n = 30/39, 76.9%); publicly accessible databases managed by research organisations (n = 27/39, 69.2%), presentation at a conference (n = 22/39, 56.4%) and researchers located overseas (n = 22/39, 56.4%). Country in which the database was held influenced the latter decision in 18 (46.2%) participants. Participants were most comfortable with sharing identifiable images for their own personal medical record (n = 33/39 84.6%).
Most participants had low or moderate understanding of AI (n = 28/39, 71.8%). If a university developed the AI, 59.0% (n = 23/39) of participants would share sensitive and/or identifiable images, which dropped to 33.3% (n = 13/39) when developed by a company or industry.
Participant comments included: ‘I wouldn't want my face shared’, and ‘obscuring the facial features won't always de‐identify. Scars, tattoos could be identifying’. One participant mentioned: ‘The parts of sensitive images that are used for personal biosecurity, i.e. facial recognition to open your iPhone, etc., need to be protected from outside hacking’. If data were shared overseas participants were concerned the country ‘may not have the same regulations regarding privacy’, and ‘once it leaves Australia, it is very hard to ensure security of data’.
DISCUSSION
As the potential sensitivity of images increased, an increasing proportion of participants reported that they would have reservations about sharing such data for multiple purposes. A greater proportion reported reluctance to share images on public platforms and for business‐centric uses, such as social media, classroom teaching, public databases and AI, whereas there was higher comfort for use in clinical or research settings. Participants were less willing to share data with researchers located outside of the country they reside in.
Most participants in our forum had low or moderate understanding of AI, similar to previous research, 2 which may contribute to their concerns about their use of images in this setting. Commercialisation and profit have been previously identified as factors that limit willingness to share health data, 2 , 3 which are elements in creating AI.
An Australian prospective cohort study of 3D total‐body photography for monitoring naevi in the general population (n = 149) found high levels of comfort and satisfaction after use of the technology. 4 In their research setting, the main barriers identified were trust in the diagnosis, followed by physical privacy (the requirement to undress and be photographed in their underwear), cost and travel requirements.
Due to the concerns participants in our forum raised about control over their images and a lack of current guidelines, researchers and healthcare providers should prioritise privacy concerns to facilitate uptake, by providing clear privacy policies and safeguards, including outlining transparently any secondary uses of data. 5 Participants should be given the option to opt‐in to certain secondary uses, as their level of acceptance changes in various scenarios.
FUNDING INFORMATION
This research was conducted with the support of the following: Australian Cancer Research Foundation; NHMRC Clinical Trials and Cohort Studies Grant (2001517); NHMRC Centre of Research Excellence (2006551); NHMRC Synergy Grant (2009923). HPS holds an NHMRC MRFF Next Generation Clinical Researchers Program Practitioner Fellowship (APP1137127).
ETHICS APPROVAL
This study was approved by the University of Queensland Human Research Ethics Committee (approval number: 2022/HE000090).
ACKNOWLEDGEMENTS
We thank the participants who contributed to the consumer forum for their valuable input. Open access publishing facilitated by The University of Queensland, as part of the Wiley ‐ The University of Queensland agreement via the Council of Australian University Librarians. Open access publishing facilitated by The University of Queensland, as part of the Wiley ‐ The University of Queensland agreement via the Council of Australian University Librarians.
Horsham C, Janda M, Kerr M, Soyer HP, Caffery LJ. Consumer perceptions on privacy and confidentiality in dermatology for 3D total‐body imaging. Australas J Dermatol. 2023;64:118–121. 10.1111/ajd.13952
Prof H Peter Soyer is an Associate Editor of the Australasian Journal of Dermatology.
REFERENCES
- 1. Grochulska K, Betz‐Stablein B, Rutjes C, Chiu FP, Menzies SW, Soyer HP, et al. The additive value of 3D Total body imaging for sequential monitoring of skin lesions: a case series. Dermatology. 2022;238:12–7. 10.1159/000517900 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Aggarwal R, Farag S, Martin G, Ashrafian H, Darzi A. Patient perceptions on data sharing and applying artificial intelligence to health care data: cross‐sectional survey. J Med Internet Res. 2021;23:e26162. 10.2196/26162 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Ghafur S, Van Dael J, Leis M, Darzi A, Sheikh A. Public perceptions on data sharing: key insights from the UK and the USA. Lancet Digit Health. 2020;2:e444–6. 10.1016/s2589-7500(20)30161-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Horsham C, O'Hara M, Sanjida S, Ma S, Jayasinghe D, Green AC, et al. The experience of 3D Total‐body photography to monitor nevi: results from an Australian general population‐based cohort study. JMIR Dermatol. 2022;5:e37034. 10.2196/37034 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Abdelhamid M, Gaia J, Sanders GL. Putting the focus Back on the patient: how privacy concerns affect personal health Information sharing intentions. J Med Internet Res. 2017;19:e169. 10.2196/jmir.6877 [DOI] [PMC free article] [PubMed] [Google Scholar]