Talk:Artificial intelligence in healthcare

Wiki Education Foundation-supported course assignment
This article is or was the subject of a Wiki Education Foundation-supported course assignment. Further details are available on the course page. Student editor(s): Microbiomeman.

Above undated message substituted from Template:Dashboard.wikiedu.org assignment by PrimeBOT (talk) 17:34, 17 January 2022 (UTC)

Wiki Education Foundation-supported course assignment
This article was the subject of a Wiki Education Foundation-supported course assignment, between 26 May 2020 and 3 July 2020. Further details are available on the course page. Student editor(s): Junwei Sheng.

Above undated message substituted from Template:Dashboard.wikiedu.org assignment by PrimeBOT (talk) 17:34, 17 January 2022 (UTC)

Wiki Education Foundation-supported course assignment
This article was the subject of a Wiki Education Foundation-supported course assignment, between 17 August 2020 and 24 November 2020. Further details are available on the course page. Student editor(s): CJut, Brruuu.

Above undated message substituted from Template:Dashboard.wikiedu.org assignment by PrimeBOT (talk) 17:34, 17 January 2022 (UTC)

Wiki Education Foundation-supported course assignment
This article was the subject of a Wiki Education Foundation-supported course assignment, between 25 August 2020 and 10 December 2020. Further details are available on the course page. Peer reviewers: Sxl79.

Above undated message substituted from Template:Dashboard.wikiedu.org assignment by PrimeBOT (talk) 17:34, 17 January 2022 (UTC)

Wiki Education Foundation-supported course assignment
This article was the subject of a Wiki Education Foundation-supported course assignment, between 6 September 2020 and 6 December 2020. Further details are available on the course page. Student editor(s): Nectaros.

Above undated message substituted from Template:Dashboard.wikiedu.org assignment by PrimeBOT (talk) 17:34, 17 January 2022 (UTC)

class project
Hi, we are a group working on a class project to help edit an article. Our main goal is to do more research on specific topics about AI in healthcare and either add a new section/add to an already existing section.

Alex: I am interested in creating a new section called “Ethics/Controversy”. Given that AI in healthcare is relatively new, there is undoubtedly a lot of skepticism/unclear understandings about AI in healthcare. This section would aim to shine light on these controversies in an unbiased way and objectively break down the ethical aspect of using artificial intelligence in healthcare.

Sean: I am focussed on adding an “accessibility” Section, discussing how the technology may become available and to who will be able to benefit from it most. The article mentions potential uses in third world countries, but I would like to discuss the ways that individuals reading the article may either contribute to, or benefit from AI in healthcare, and where they can go to do so.

Claire: I am interested in adding more to the Telehealth section. Looking at the telemedicine article, there is not much information there about application of artificial intelligence, so it would not be redundant to add material here. I have found some sources to start researching. CJut (talk) 02:27, 22 Oct 2020 (UTC) Claire

Hello I have chosen to work on this article for my class. I will be editing and improving the grammar/sentence structure of some sentences. I will also announce any deletions I may make. Please let me know if someone may have suggestions or concerns.

Microbiomeman (talk) 02:27, 6 April 2018 (UTC) Evan

Hello, I disagree with the statement: "(2) and algorithms are black boxes; algorithms can predict extremely precise, but not the cause or the why." This is actually false, at least from the point of view of statisticians. Many algorithms used in Medicine are actually prefered because of their interpretability. They offer coefficients that quantify the importance of each feature (regression, linear, penalized or logistic). Other algorithms such as LDA or random forest allow you to create decision frontiers, which are understandable easily for any humans (ex: if some patient is over 130 kilos, then the patient is obese). Therefore you should really adapt that section.

Paciko (talk) 09:22, 6 November 2018 (UTC)

Hello, I have chosen to work on this article for a class. I will be adding details about recent (since c. 2013) developments in the use of AI in identifying drug-drug interactions. Feel free to respond with comments or concerns.

Rammor (talk) 17:27, 17 November 2019 (UTC)

Insilico
Dear Wikipedia Community, I would like to propose a few edits/additions to the 'creation of new drugs' section in this article. Following are my suggestions, and I would love to hear your thoughts on it before I edit anything directly. Sources will be hyperlinked carefully in the main edits but I have mentioned those in brackets below for reference. Thanks in advance for giving this your time

1) Include an introductory sentence:

"For small molecules (https://en.wikipedia.org/wiki/Small_molecule), computational AI approaches offer the possibility to predict properties of potential candidates by analyzing vast amounts of information and identifying suitable drug candidates for the disease variant."

2) Include additions to Insilico Medicine example: "Later in November 2021, Insilico Medicine announced that it was dosing the first healthy volunteer in a microdose phase 0 trial of ISM 001-005, Insilico's AI-generated candidate for treatment of idiopathic pulmonary fibrosis (https://insilico.com/blog/fih).

By February 2022, phase 1 clinical trials were started after phase 0 trials led to successful demonstration of clinical proof of concept, depicted by a favorable pharmacokinetic and safety profile of the drug in humans (https://insilico.com/phase1)."

3) Consider removing the Deep Genomics example (unless there is a reason we want to highlight a candidate specifically for Wilson's disease, otherwise it seems there are too many examples shared in this section?)

4) Add a paragraph on rise in use of AI applications for drug discovery during Covid-19 (marking a major world event triggering an accelerated adoption):

"Applications of AI in drug discovery have also accelerated during Covid-19, including for identification and screening of existing medicines for repurposing them for covid-19 treatment. For instance, Oxford University's Recovery trial of Covid-19 therapies confirmed the benefits of a drug originally flagged by an AI tool created by BenevolentAI. The trial found that treatment with baricitinib, a drug marketed by Eli Lilly for rheumatoid arthritis, reduced deaths by 13% in patients hospitalized with Covid-19 (https://www.ox.ac.uk/news/2022-03-03-recovery-trial-shows-baricitinib-reduces-deaths-patients-hospitalised-covid-19)."

Akhurana27 (talk) 07:58, 11 June 2022 (UTC)


 * This reads as promotional - we should not be promoting individual companies like this, certainly not based on primary sources. MrOllie (talk) 11:18, 11 June 2022 (UTC)
 * Thank you for your feedback.
 * As full disclosure, I am not associated with Insilico Medicine or other startups mentioned here in any capacity. There are some (credible) media outlets that have captured Insilico’s work in the field (see example links below), and my view is that they are a notable example in the market. Furthermore, Insilico was already called out as an example in the existing section of this wiki page. I have suggested the addition of two recent updates for Insilico since those are significant developments for the wider industry.
 * https://www.ft.com/content/704ced9a-dffd-49a1-a58f-46fc6dca0cd2
 * https://techcrunch.com/2021/06/22/a-i-drug-discovery-platform-insilico-medicine-announces-255-million-in-series-c-funding/
 * https://techcrunch.com/2022/06/06/hong-kongs-ai-drug-discovery-firm-insilico-50m-series-d/
 * https://www2.deloitte.com/content/dam/Deloitte/my/Documents/risk/my-risk-sdg3-intelligent-drug-discovery.pdf
 * https://www.fiercebiotech.com/medtech/insilico-medicine-begins-first-human-trial-its-ai-designed-drug-for-pulmonary-fibrosis
 * For minimising room for misinterpretation as promotional material, I have edited the section on Insilico as below. Thanks again for your feedback.
 * "In September 2019 Insilico Medicine reported the creation, via artificial intelligence, of six novel inhibitors of the DDR1 gene, a kinase target implicated in fibrosis and other diseases. The system, known as Generative Tensorial Reinforcement Learning (GENTRL), designed the new compounds in 21 days, with a lead candidate tested and showing positive results in mice.
 * Later in November 2021, Insilico Medicine announced that it was dosing the first healthy volunteer in a microdose phase 0 trial of ISM 001-005, Insilico's AI-generated candidate for treatment of idiopathic pulmonary fibrosis. By February 2022, phase 1 clinical trials were started."
 * 18:24, 15 June 2022 (UTC) Akhurana27 (talk) 18:24, 15 June 2022 (UTC)


 * We're not really an outlet for industry news about which company is doing one thing or another. I accept that your motivations are good, but it's just not the sort of thing that should be in a Wikipedia article. Thanks for noting the other company namedropping, I removed that as well. - MrOllie (talk) 18:59, 15 June 2022 (UTC)
 * Also, please stop altering the section title I added. Thanks. - MrOllie (talk) 11:00, 16 June 2022 (UTC)
 * It seems that you have made edits to the wikipedia page (i.e. removing an entire section) without discussion with the community on this channel. This goes against Wikipedia editing guidelines, and so I have undone the edits.
 * Regarding your concern on talking about company specific examples, I would note again that they illustrate significant industry developments. Akhurana27 (talk) 14:34, 26 June 2022 (UTC)
 * The section title for this edit discussion is being changed to "Creation of New Drugs" instead of "Insilico" since the edits suggested above are wider than just about Insilico. Akhurana27 (talk) 14:34, 26 June 2022 (UTC)
 * No, removing advertising is not 'against Wikipedia editing guidelines'. I'm not sure where you got that impression - I suggest you read those guidelines over again. Wikipedia is not a vehicle for advertising, and we're not the business page - we're not a place to regurgitate press releases. It is pretty obvious what this discussion is about - attempted promotion of Insilico and my removal of it. MrOllie (talk) 14:49, 26 June 2022 (UTC)

Addressing education of AI within healthcare
It is noted that this page does not make reference to the absence of the topic of AI education within the format curricula of healthcare provider education. This is an important topic for a variety of reasons: - There is a well acknowledged gap in current education and skills required for healthcare providers to be comfortable and knowledgeable about the use of AI-based tools in practice - There are immense patient safety, privacy and quality of care implications to this - Healthcare providers entering practice are exposed to a large number of vendors with AI tools of undetermined validity and utility, and do not have a framework for evaluating these.

I proposed that we make reference to these gaps and to work currently happening to propose curriculum changes (e.g. https://mededu.jmir.org/2023/1/e46344/). I attempted to make a suggested change though this was reverted. Happy to discuss. FrozenTroubadour (talk) 14:37, 20 July 2023 (UTC)


 * Are you connected with this paper or its authors in some way? This citation and the accompanying text strike me as redundant. MrOllie (talk) 14:42, 20 July 2023 (UTC)
 * Yes, I am affiliated with one of the institutions that was involved with the paper and know the authors (hence my awareness of it).
 * I welcome feedback on the text proposed. The main thrust of the paper is around: 1. The proposed topics and sub-topics that should be incorporated into medical education (Figure 1 in the paper); and 2. Ensuring this is done longitudinally within educational curriculum as a core competency (as opposed to a one-off or optional segment of education), given how AI can impact every facet of the care delivery process. FrozenTroubadour (talk) 14:53, 20 July 2023 (UTC)
 * Please have a look at WP:COI and WP:PAID, you have been violating Wikipedia's terms of use. MrOllie (talk) 15:03, 20 July 2023 (UTC)
 * Thanks for highlighting and my apologies in that case. This is my first time engaging with Wikipedia. If we used a different reference to which I have no organizational affiliation, my understanding is that this would then not be an issue? 99.230.171.103 (talk) 19:09, 20 July 2023 (UTC)

Wiki Education assignment: Research Process and Methodology - SU23 - Sect 200 - Thu
— Assignment last updated by FULBERT (talk) 11:36, 10 August 2023 (UTC)

Ethical concern suggestions
Olive41 (talk) 20:57, 24 September 2023 (UTC) Hello all, I am a new user here on Wikipedia. From reading the article I wanted to suggest how the article can be improved. expanding on the section discussing ethical concerns related to AI in healthcare by providing more context and discussion on these issues, including potential solutions or strategies to address them. This will be beneficial to the readers because  it provides a more comprehensive understanding of the challenges and responsibilities associated with AI in healthcare. It promotes ethical awareness, transparency, accountability, and responsible decision-making, ultimately contributing to the ethical and responsible development and use of AI in healthcare.


 * ::Hi, you make a good point about the ethical concerns of AI in healthcare. I too was wondering how AI in healthcare could expand given the limitations with data privacy and how this impacts AI training.--~ Tecasm (talk) 20:04, 29 October 2023 (UTC)

New text in Bias section.

 * Specific text to be added: As the data derived for AI-tools depends on basic research, due to limitations in existing knowledge, AI-based drug discovery based on data from nonanimal model organisms (cell culture or organoids) can have potential risks to patients in clinical trials. Clinical trials with AI‐designed drug EXS‐21546 was unsuccessful.
 * Reason for the change: Need to stress the limitations.
 * References supporting change:

https://doi.org/10.1002/ame2.12364

[Https://medicaltrend.org/2023/10/21/first-wave-of-ai-designed-drugs-faces-major-clinical-setbacks https://medicaltrend.org/2023/10/21/first-wave-of-ai-designed-drugs-faces-major-clinical-setbacks] Zebrafishgroup (talk) 12:02, 25 January 2024 (UTC)
 * While I agree that the article should cover risks and downsides, I have declined this request because: the first source is not about the risks of AI based drug development, it is about the risk of not using animal models before trying drugs on human subjects. The second source does not support the proposed content, since it focuses on commercial failures rather than risk to patients. (Also, a drug failing is not noteworthy because a large majority of drugs will never reach approval) (t &#183; c)  buidhe  20:16, 28 January 2024 (UTC)