Intro
A few years ago, before doing a full-time PhD whilst working a full-time job, and side-hustling extra papers, left me with precisely ZERO time, I used to write weeknotes covering all the things I was currently working on, thinking about, and reading. Now it's 2024, I'm #PhDone and my full-time job is as a postdoctoral researcher at the Yale Digital Ethics Center so I thought it was time to bring the weeknotes back. Let's get into it.
Things I worked on
Programme of research outline. Primarily I have been working on developing the programme of research I want to complete for my postdoc over the next couple of years. I find it very helpful to have a framework in my mind of how all the things I'm working on fit together, and I also like to have a "big project" that I'm working towards. I'll write it up properly I a separate blog when it is fully fleshed out, but for now - in outline - it looks like this:
Literature searches. To get the ball rolling on this programme of research, I conducted two literature searches, that I will now systematically work my way through when I have time and/or my brain is too tired to actively think but it can passively absorb. I conducted both searches using a combination of scopus, pubmed, and Google Scholar. Specifically:
Search A: Related to the question "is competence a prerequisite for digital privacy?" This was a question largely inspired by the rise of Dementia villages and ongoing debates about the ethics of mental health surveillance of online spaces. Currently my resultant Zotero library has 232 items in it.
Search B: Related to a concept I am trying to develop that I have tentatively called "personalised unwellness." The desire to develop this concept is inspired by the fact that personalised newsfeeds on social media - most explicitly TikTok's "For You Page" very clearly blur the lines between 'diganosis', targeted advertising, and recommender systems. I think this is a seriously concerning trend that is currently receiving insufficient attention. Currently the resultant Zotero library has 1211 items in it.
App Store Audits. I'm working with a couple of other members of the DEC to update a large audit I conducted a few years ago of health, wellness, and medicine apps available on the Apple and Google 'App Stores' and the availability/quality of the evidence available to support their claims of efficacy. The updated audit is much larger - the app sample is 3x the size - and it aims to compare the US vs. the UK app stores to see if there is any significant difference given differences in the governance of the app stores and in the governance of Software as a Medical Device between the two countries.
The BMJ's Future of the NHS Commission. You may have seen that back in the summer of last year the BMJ launched a Commission on the future of the NHS. I am fortunate enough to be the digital/data/tech commissioner and as a consequence I'm involved in two of the working groups for the commission working on two papers. Both are now in the stages of wrapping up, and so there's been a few meetings and some work on drafting text - and most especially finding references to support the arguments made in the papers.
Reviewed 'pro-ethical design tools.' Recently, I have been debating whether to update my paper from a few years ago on the 'what to how' of AI ethics to see if the field has developed, if new patterns have emerged etc. To help me decide whether this would be a good idea, I spent a couple of hours (not-systematically) searching online for new tools that have been developed and/or written about since 2020.
Things I did
Recorded a lecture for an event coming up in Oxford that I will not be able to attend in 3D. The lecture was based largely around my PhD thesis "Designing an algorithmically enhanced NHS"
Had a kick-off meeting for a new project on evaluating the use of LLMs for healthcare (more below).
Had several interest meetings/discussions about what should be the next priorities for NHS digital/data policy as we head into an election year.
Met with some super cool folk from the Yale Medical School who have very similar interests to me, and will hopefully be fruitful future collaborators.
Peer-reviewed 3 different papers for 3 different journals (and accepted 4 more).
Worked with the brilliant postgrad researchers we have in the Digital Ethics Center on their projects - all of which are very cool but not mine to talk about.
Reviewed the literature alerts that I receive on the first of every month, to identify newly published papers/research that is relevant to my work and that I therefore need to read.
Things I thought about
The evidence behind the use of LLMs for healthcare, and how their use could a) be evaluated robustly and b) audited once deployed. I do not think there are clear answers to either of these questions which is concerning given the rapid rise in interest in their use. It is extremely difficult to meaningfully evaluate a model that is constantly shifting and the performance of which varies dramatically depending on how the end-user interacts with the model; the use of clinical vignettes is not sufficiently robust and yet expecting clinical trials to be conducted is probably unrealistic and not suitable (+ clinical trials involving AI/ML are already often weak). Similarly, it is clear that some form of auditing technique will be needed to monitor the performance in terms of safety and accuracy in both real-time and with regards to edge-cases. This will like to be an automated process (or at the very least a semi-automated process) but how exactly to achieve this is not yet clear to me.
How to select the most appropriate method for policy-related research depending on the specific research question and whether it involves scoping, critical analysis, evaluation, design, or more. In general, I think the methodological literature in this area is weak, and it is difficult to navigate unless you know exactly what you are looking for which makes things very hard for junior researchers/ those new to policy-related research. I think a lot more needs to be done to improve this. Why is there always so much to do and so little time.
The risks of skill-based vendor lock-in. As you may have seen late last year, NHS England announced the winner of its controversial "Federated Data Platform" contract. I wrote about why the contract is so controversial for the BMJ here. At the beginning of this week, the NHS published a heavily redacted version of this contract. Why it is quite so heavily redacted raises a number of important and pressing questions. Yet, what was most on my mind about the FDP is the risks associated with outsourcing all the knowledge, understanding, familiarity of NHS data and all the skills needed to work with it from an analytical perspective to a third party provider. I think this is creating a form of highly-risk 'skills-based' vendor lock-in, that has little to do with the actual software platform itself, that is on a scale we have not really witness before. What worries me most is that it is unclear how to mitigate this risk.
The role of information exchange in continuity of care across the whole health and care system, including community care, social care, podiatry, optometry, dentistry, pharmacy etc. and the barriers to achieving this continuity-enabling information exchange in the NHS:
continuing problems with system interoperability
lack of data standardisation
limited resources in terms of time
limited investment in core information infrastructure
lack of clarity regarding data ownership/clarity
unnecessarily complex regulatory regime
no read/write API for GP records
(A selection of) Things I read
The highlighted papers are those I particularly enjoyed.
Al Meslamani, Ahmad Z. “Beyond Implementation: The Long-Term Economic Impact of AI in Healthcare.” Journal of Medical Economics 26, no. 1 (December 31, 2023): 1566–69. https://doi.org/10.1080/13696998.2023.2285186.
Bhatt, Ami B., and Jennifer Bae. “Collaborative Intelligence to Catalyze the Digital Transformation of Healthcare.” Npj Digital Medicine 6, no. 1 (September 25, 2023): 177. https://doi.org/10.1038/s41746-023-00920-w.
Biasiotto, Roberta, Jennifer Viberg Johansson, Melaku Birhanu Alemu, Virginia Romano, Heidi Beate Bentzen, Jane Kaye, Mirko Ancillotti, et al. “Public Preferences for Digital Health Data Sharing: Discrete Choice Experiment Study in 12 European Countries.” Journal of Medical Internet Research 25 (November 23, 2023): e47066. https://doi.org/10.2196/47066.
Brookes, Gavin, and Paul Baker. “What Does Patient Feedback Reveal about the NHS? A Mixed Methods Study of Comments Posted to the NHS Choices Online Service.” BMJ Open 7, no. 4 (April 2017): e013821. https://doi.org/10.1136/bmjopen-2016-013821.
Buonora, Michele J., Sydney A. Axson, Shawn M. Cohen, and William C. Becker. “Paths Forward for Clinicians Amidst the Rise of Unregulated Clinical Decision Support Software: Our Perspective on NarxCare.” Journal of General Internal Medicine, November 14, 2023, s11606-023-08528–2. https://doi.org/10.1007/s11606-023-08528-2.
Chae, J. “A Comprehensive Profile of Those Who Have Health-Related Apps.” Health Education and Behavior 45, no. 4 (2018): 591–98. https://doi.org/10.1177/1090198117752784.
Corbucci, Luca, Anna Monreale, Cecilia Panigutti, Michela Natilli, Simona Smiraglio, and Dino Pedreschi. “Semantic Enrichment of Explanations of AI Models for Healthcare.” In Discovery Science, edited by Albert Bifet, Ana Carolina Lorena, Rita P. Ribeiro, João Gama, and Pedro H. Abreu, 14276:216–29. Lecture Notes in Computer Science. Cham: Springer Nature Switzerland, 2023. https://doi.org/10.1007/978-3-031-45275-8_15.
De Boer, Christopher, Hassan Ghomrawi, Suhail Zeineddin, Samuel Linton, Soyang Kwon, and Fizan Abdullah. “A Call to Expand the Scope of Digital Phenotyping.” Journal of Medical Internet Research 25 (2023): e39546.’’
Fazakarley, C A, Maria Breen, Paul Leeson, Ben Thompson, and Victoria Williamson. “Experiences of Using Artificial Intelligence in Healthcare: A Qualitative Study of UK Clinician and Key Stakeholder Perspectives.” BMJ Open 13, no. 12 (December 2023): e076950.https://doi.org/10.1136/bmjopen-2023-076950.
Fitzpatrick, Geraldine. “Integrated Care and the Working Record.” Health Informatics Journal 10, no. 4 (December 2004): 291–302. https://doi.org/10.1177/1460458204048507.
García-García, Julián Alberto, Manuel Carrero, María José Escalona, and David Lizcano. “Evaluation of Clinical Practice Guideline-Derived Clinical Decision Support Systems Using a Novel Quality Model.” Journal of Biomedical Informatics 149 (January 2024): 104573. https://doi.org/10.1016/j.jbi.2023.104573.
Gillner, Sandra. “We’re Implementing AI Now, so Why Not Ask Us What to Do? – How AI Providers Perceive and Navigate the Spread of Diagnostic AI in Complex Healthcare Systems.” Social Science & Medicine 340 (January 2024): 116442. https://doi.org/10.1016/j.socscimed.2023.116442.
Giuffrè, Mauro, and Dennis L. Shung. “Harnessing the Power of Synthetic Data in Healthcare: Innovation, Application, and Privacy.” Npj Digital Medicine 6, no. 1 (October 9, 2023): 186. https://doi.org/10.1038/s41746-023-00927-3.
Greenhalgh, T., K. Stramer, T. Bratan, E. Byrne, J. Russell, and H. W. W. Potts. “Adoption and Non-Adoption of a Shared Electronic Summary Record in England: A Mixed-Method Case Study.” BMJ 340, no. jun16 4 (June 16, 2010): c3111–c3111. https://doi.org/10.1136/bmj.c3111.
Habib, Anand R., and Cary P. Gross. “FDA Regulations of AI-Driven Clinical Decision Support Devices Fall Short.” JAMA Internal Medicine 183, no. 12 (December 1, 2023): 1401. https://doi.org/10.1001/jamainternmed.2023.5006.
Ham, C., J. Dixon, and C. Chantler. “Clinically Integrated Systems: The Future of NHS Reform in England?” BMJ 342, no. mar28 1 (March 28, 2011): d905–d905. https://doi.org/10.1136/bmj.d905.
Hernandez-Boussard, Tina, Shazia Mehmood Siddique, Arlene S. Bierman, Maia Hightower, and Helen Burstin. “Promoting Equity In Clinical Decision Making: Dismantling Race-Based Medicine: Commentary Examines Promoting Equity in Clinical Decision-Making.” Health Affairs 42, no. 10 (October 1, 2023): 1369–73. https://doi.org/10.1377/hlthaff.2023.00545.
Hille, Eva Maria, Patrik Hummel, and Matthias Braun. “Meaningful Human Control over AI for Health? A Review.” Journal of Medical Ethics, September 20, 2023, jme-2023-109095. https://doi.org/10.1136/jme-2023-109095.
Jensen, Iselin, and Vilde Hellevik Borgen. “# Autism: A Cross-Sectional Study of the Quality of Diagnostic Information on TikTok,” 2023.
Jones, Monica Catherine, Tony Stone, Suzanne M Mason, Andy Eames, and Matthew Franklin. “Navigating Data Governance Associated with Real-World Data for Public Benefit: An Overview in the UK and Future Considerations.” BMJ Open 13, no. 10 (October 2023): e069925. https://doi.org/10.1136/bmjopen-2022-069925.
Knott, Alistair, Dino Pedreschi, Raja Chatila, Tapabrata Chakraborti, Susan Leavy, Ricardo Baeza-Yates, David Eyers, et al. “Generative AI Models Should Include Detection Mechanisms as a Condition for Public Release.” Ethics and Information Technology 25, no. 4 (December 2023): 55. https://doi.org/10.1007/s10676-023-09728-4.
Lavis, Anna, and Rachel Winter. “#Online Harms or Benefits? An Ethnographic Analysis of the Positives and Negatives of Peer‐support around Self‐harm on Social Media.” Journal of Child Psychology and Psychiatry 61, no. 8 (August 2020): 842–54. https://doi.org/10.1111/jcpp.13245.
Lee, Jessica T., Alexander T. Moffett, George Maliha, Zahra Faraji, Genevieve P. Kanter, and Gary E. Weissman. “Analysis of Devices Authorized by the FDA for Clinical Decision Support in Critical Care.” JAMA Internal Medicine 183, no. 12 (December 1, 2023): 1399. https://doi.org/10.1001/jamainternmed.2023.5002.
Li, Linda T., Lauren C. Haley, Alexandra K. Boyd, and Elmer V. Bernstam. “Technical/Algorithm, Stakeholder, and Society (TASS) Barriers to the Application of Artificial Intelligence in Medicine: A Systematic Review.” Journal of Biomedical Informatics 147 (November 2023): 104531. https://doi.org/10.1016/j.jbi.2023.104531.
Li, Edmond, Jonathan Clarke, Hutan Ashrafian, Ara Darzi, and Ana Luisa Neves. “The Impact of Electronic Health Record Interoperability on Safety and Quality of Care in High-Income Countries: Systematic Review.” Journal of Medical Internet Research 24, no. 9 (September 15, 2022): e38144. https://doi.org/10.2196/38144.
Morales, Makayla, Alexis Fahrion, and Shannon Lea Watkins. “# NicotineAddictionCheck: Puff Bar Culture, Addiction Apathy, and Promotion of e-Cigarettes on TikTok.” International Journal of Environmental Research and Public Health 19, no. 3 (2022): 1820.
Mordecai, Chandler. “# Anxiety: A Multimodal Discourse Analysis of Narrations of Anxiety on TikTok.” Computers and Composition 67 (2023): 102763.
Ndlovu, Kagiso, Maurice Mars, and Richard E. Scott. “Interoperability Frameworks Linking mHealth Applications to Electronic Record Systems.” BMC Health Services Research 21, no. 1 (December 2021): 459. https://doi.org/10.1186/s12913-021-06473-6.
Raab, René, Arne Küderle, Anastasiya Zakreuskaya, Ariel D Stern, Jochen Klucken, Georgios Kaissis, Daniel Rueckert, et al. “Federated Electronic Health Records for the European Health Data Space.” The Lancet Digital Health 5, no. 11 (November 2023): e840–47. https://doi.org/10.1016/S2589-7500(23)00156-5.
Roppelt, Julia Stefanie, Dominik K. Kanbach, and Sascha Kraus. “Artificial Intelligence in Healthcare Institutions: A Systematic Literature Review on Influencing Factors.” Technology in Society 76 (March 2024): 102443. https://doi.org/10.1016/j.techsoc.2023.102443.
Stevens, Alexander F, and Pete Stetson. “Theory of Trust and Acceptance of Artificial Intelligence Technology (TrAAIT): An Instrument to Assess Clinician Trust and Acceptance of Artificial Intelligence.” Journal of Biomedical Informatics 148 (December 2023): 104550. https://doi.org/10.1016/j.jbi.2023.104550.
Stoody, Vishvanie B, Hannah R Glick, Annie C Murphey, Julie M Sturza, and Ellen M Selkie. “A Content Analysis of Transgender Health and Wellness Themes Shared Through Social Media.” Clinical Pediatrics, 2023, 00099228231219499.
Shull, Jessica Germaine. “Digital Health and the State of Interoperable Electronic Health Records.” JMIR Medical Informatics7, no. 4 (November 1, 2019): e12712. https://doi.org/10.2196/12712.
Tal, Eran. “Target Specification Bias, Counterfactual Prediction, and Algorithmic Fairness in Healthcare.” In Proceedings of the 2023 AAAI/ACM Conference on AI, Ethics, and Society, 312–21. Montreal QC Canada: ACM, 2023. https://doi.org/10.1145/3600211.3604678.
Till, Alex, Hanish Sall, and Jonathan Wilkinson. “Safe Handover : Safe Patients - The Electronic Handover System.” BMJ Quality Improvement Reports 2, no. 2 (2014): u202926.w1359. https://doi.org/10.1136/bmjquality.u202926.w1359.
Triptow, Christina, Jason Freeman, Paige Lee, and Thomas Robinson. “# HealthyLifestyle: AQ Methodology Analysis of Why Young Adults like to Use Social Media to Access Health Information.” Journal of Health Psychology, 2023, 13591053231200690.
Vo, Vinh, Gang Chen, Yves Saint James Aquino, Stacy M. Carter, Quynh Nga Do, and Maame Esi Woode. “Multi-Stakeholder Preferences for the Use of Artificial Intelligence in Healthcare: A Systematic Review and Thematic Analysis.” Social Science & Medicine 338 (December 2023): 116357. https://doi.org/10.1016/j.socscimed.2023.116357.
Warren, L., J. Clarke, and A. Darzi. “Measuring the Scale of Hospital Health Record System Fragmentation in England.” Health Services Research 55, no. S1 (August 2020): 43–44. https://doi.org/10.1111/1475-6773.13386.
Wellman, Mariah L. “‘A Friend Who Knows What They’re Talking about’: Extending Source Credibility Theory to Analyze the Wellness Influencer Industry on Instagram.” New Media & Society, 2023, 14614448231162064.
Wharton, George A, Harpreet S Sood, Amanda Sissons, and Elias Mossialos. “Virtual Primary Care: Fragmentation or Integration?” The Lancet Digital Health 1, no. 7 (November 2019): e330–31. https://doi.org/10.1016/S2589-7500(19)30152-9.
Comments