As I've been diagnosed with MS for 31 years, I've seen many ideas on treatments come and go. I've seen theories on the cause of MS do the same thing. Any AI tool based on iterations of whatever is in fashion at the moment might look very strange a few years later. I would never trust those answers and would prefer that high end research articles were more accessible instead of being hidden behind paywalls. You make an effort to stand at the top of cliff and yell out what you think is right. Most don't even try. It would terrify me to think any provider got his information from machine learning. It's the modern equivalent of the self-help paperback. I would rather trawl through a variety of well researched but contradictory opinions than be stuck with answer based on plagiarizing the most common opinion.
I completely agree - I have a phd myself and I prefer reading a variety of research papers with all their contradicting results than blindly trust a simplistic answer that rolls out of AI.
If you're happy not understanding the underlying (scientific) reasons why you get a particular answer, you will never be able to tell when the answer you get is rubbish.
Exactly my thoughts. It’s sometimes interesting to see what stuff it spits out but you certainly need to invest substantial time verifying the information given.
In my old life I had a phd in computing science. Ask ChatGPT a specialist question about your own field of expertise and be shocked at the random uselessness of the answer. A system trained with good scientific data could be interesting to explore, but I’d much prefer a human expert’s opinion even if AI had given me some ideas.
Google put up llm or is it cht gpt answers - i found them WORSE than random useless - in fact dangerously wrong - this wasnt for ms - ive got multiple longterm condition like, i guess, many people whove had ms for many years.
- it s a BIG issue for me that neurologists think ive only got ms. and know nothing about anything else - why not simply ask me - im an expert by experience.
- its an even bigger issue that you seem to think rrms, spms, ppms are useful classifications. ive had rrms for 30 yers - first decades i was running marathons. now im definitely smouldering. so chat gpt on 'rrms' is not helpful?
Not sure if you are addressing me, but I would rather read verified literature than generic ChatGPT generated info. You can get the basic info on RRMS from many verified sources, as I’m sure you know. Maybe I’m misunderstanding your question or you are addressing Prof G, as I’ve never mentioned SPMS or PPMS in thus thread, so you lost me there.
i was agreeing strongly with you - but wanted to also expand out to comment on prof g - he opened with the categories - though he's esrlier spoken on how eg rrms was fabricated for fda
I agree. Also, I too am frustrated by the paywalls and gatekeeping of information. To your point about machine learning, I have 2 lawyer friends who as an experiment, used AI to research case law etc., on a lark, and were in fits of laughter at the responses. I know it's not the same, but is another example of the quality of the information.
Interestingly only a couple of days ago I pinged my son, who’s a solicitor, a link to an article on the BBC news website.
It describes the conclusions of research undertaken by a leading uk law firm. Chatbots were put to the test using 50 questions about English law. This was a repeat of earlier research and although they were much improved: ‘they made mistakes, left out important information and invented citations.’
Conclusion: ‘the tools are starting to perform at a level where they could assist in legal research, but there were dangers in using them if lawyers don’t already have a good idea of the answer’
This has definitely reinforced for me the notion that AI is useful but, dare I say, with a degree of hesitation, meaning information being provided needs to be double-checked for accuracy and reliability.
Wow, I replied to Andrew Scott above before I read your comment. I was speaking about my 2 lawyer friends (1 in the US and 1 in Canada) who relayed to me the same thing you wrote about your son's experience and the BBC article.
An MS specific LLM may help bridge the gap where health issues affect MS but are usually managed by a Primary Care HCP, who have little experience with MS. Here's my example: My GP instructed me to come off oestrogen only HRT (due to more than 7 years use), which then caused severe cognitive and fatigue symptoms. ChatGPT suggested increasing oestrogen and the GP said to ask my neurologist before making changes. With neurology being so inaccessible I did as ChatGPT suggested. I'm summarising. It was more complicated in reality. ChatGPT helped me with direction, which I then followed up with checking the clinical evidence online.
On a linked, but slightly separate point, I'm all for self-management of MS. However, in the detail I've missed in my example above, there were times when I was incredibly unwell with a respiratory infection, which was being 'managed' by the GP. I was too unwell to do my own research and what I needed was an HCP who understood MS and was also accessible. I wonder if an MS specific LLM might be used by the patient as well as a Primary Care HCP,
I don't trust AI models yet. I probably never will, as they are still based on scouring information that is generally available to anyone or any thing. Even though The large language models are capable of something akin to metacognition they are still incapable of true novelty, and the complexity and subtlety involved in it. They also don't seem to have the ability for intuition.
What is the endgame here? Are you suggesting LLMs can take over the role of consultants? I'm not saying there's no role for AI, but I'm very sceptical about its application in our current environment. Research has shown how AI reinforces existing prejudices and I feel like this tech-positive approach risks isolating and excluding pwMS. I want my consultant to treat me like a human, not a robot.
Gavin, I seem to deviate in opinion a bit but I must say I am very ai-in-healthcare positive. My HCP (ophthalmologist, to be fair) and chatgpt disagreed about if I had MS.
If it weren’t for me describing my optic neuritis to ai, I would not have had the courage to return THREE TIMES to him and demand an MRI (optic nerve enhanced). When I got to neurology, they quickly found t2 lesions in all 3 regions, an enhancing lesion, and OCBs.
AI diagnosed my MS. I would be progressing along without ocrevus were it not for this incredible tool that gave me the courage to advocate for myself.
In the US, here, and I concur with your timeline. If the hospital systems can use an AI and eliminate high-priced staff….they will do it in a heartbeat. We only talk to the doctor for 30 minutes, every 6 months, and when we try to ask a question in between we get the support staff running interference; and they are not Doctors, and not highly trained in MS. AI would probably be an upgrade, lol. I use Gregory AI to check on clinical trials already, and would use AI for an opinion, then follow up with more research including the crowdsourcing (social media) stuff. Sort thru it all, make my own mind up. But doctors have biases, even good ones, so I do this already. I’m 67, was a practicing professional in another field, dealt with professionals in other disciplines, and can safely say that “experts” have many levels of ability- many are barely competent, many are arrogant because they had some training…. As you hear already. Heck, the experts hear are pushing to get seniors off of DMT’s as the latest thing, because they wrongly evaluate the wrong risk/reward ratio- our symptoms are (usually) disability worsening, not new lesions, but my friends that are taken off DMT’s usually spiral down much quicker on their disability scale. Which shows that is the wrong approach- yet the experts keep pointing to no lesions to justify what they did. I have no problem with an AI as a tool, but I also have no problem with a doctor as an expert- I still question and check what they tell me.
Hear hear in the states. If older (60’s) once neid status is reached, you are discouraged from your DMT and you are “cured”. And then the real downhill begins. I do not think for a second that anything here would be replaced by AI if it helps the bottom line. I think one would need to check and recheck anything created by AI. Jmo.
Fascinating. My worst MS symptom is very severe RLS. I now have zero symptoms after doing my own in depth research & fighting tooth &.nail for the treatment I'm now taking.
My MS neurologist admits she knows nothing about RLS and UK treatment is about 10 years behind best treatment by the top RLS experts in the USA.
I asked ChatGPT about how to treat RLS correctly and it was absolutely brilliant. It had all the latest, best research and treatment protocols.imse
Sadly, most UK neurologists are still prescribing dangerous dopamine agonists and refusing iron infusions or Buprenorphine.
I wonder why UK neurologists kick back against AI or patient experts?Is it insecurity or arrogance?
What treatments are available for ms in the uk and what’s the most effective?
ChatGPT said:
In the UK, treatment for multiple sclerosis (MS) typically focuses on modifying the disease's course, managing symptoms, and improving overall quality of life. The choice of treatment depends on the type of MS (e.g., relapsing-remitting MS, secondary-progressive MS, or primary-progressive MS) and individual circumstances. Below are the primary treatments available, followed by insights into their effectiveness.
1. Disease-Modifying Therapies (DMTs)
These are used to slow the progression of MS, reduce the
You likely don’t need to create a new platform as chat gtp etc gives so much information on ms. Just like as you would know AI has now moved to writing drs’s and neuros reports etc for them. I have personal surgeon friends who it on their own reports and dr’s who current just record the appointment and it transcribes it all into full detailed report, creating better logs etc. As long as information and papers are put on the web then we can find it all through chat gtp and we can also ask for the answer to provide references of where the info is from. Basically I don’t think an ms platform one would give any different info to what chat gtp already gives up. I wonder if you asked it about your own work and whether it would pop up and reference your own papers… it could be interesting for you to try. ☺️
I do honestly think it would be a great idea for those less tech savvy but it still uses medical evidence base. Just be specific about your questions and ask it to reference the response with academic papers
We are slowly getting there with the Deep Research models (in fairness, so far I could not quite convince myself to pay for OpenAI Deep Research). The paywalls remain as big hindrance for uptodate research though.
FWIW, even the regular models hallucinate far less (and Gemini can further lower that by cross checking google itself) than they used too, probably less than most humans...
AI fills the gaps in care from HCP's. When I was first diagnosed with MS the MS nurse team said they had too many patients to respond to my questions. So that left me with MS charities who did answer my questions and were so supportive. AI can now fill that gap in care; of the HCPs too busy to actually answer your questions about your MS. I use AI eveveryday in all aspects of my life, including work, personal life, parenting, supporting school work, health and all things MS
As I've been diagnosed with MS for 31 years, I've seen many ideas on treatments come and go. I've seen theories on the cause of MS do the same thing. Any AI tool based on iterations of whatever is in fashion at the moment might look very strange a few years later. I would never trust those answers and would prefer that high end research articles were more accessible instead of being hidden behind paywalls. You make an effort to stand at the top of cliff and yell out what you think is right. Most don't even try. It would terrify me to think any provider got his information from machine learning. It's the modern equivalent of the self-help paperback. I would rather trawl through a variety of well researched but contradictory opinions than be stuck with answer based on plagiarizing the most common opinion.
I completely agree - I have a phd myself and I prefer reading a variety of research papers with all their contradicting results than blindly trust a simplistic answer that rolls out of AI.
If you're happy not understanding the underlying (scientific) reasons why you get a particular answer, you will never be able to tell when the answer you get is rubbish.
Exactly my thoughts. It’s sometimes interesting to see what stuff it spits out but you certainly need to invest substantial time verifying the information given.
In my old life I had a phd in computing science. Ask ChatGPT a specialist question about your own field of expertise and be shocked at the random uselessness of the answer. A system trained with good scientific data could be interesting to explore, but I’d much prefer a human expert’s opinion even if AI had given me some ideas.
Google put up llm or is it cht gpt answers - i found them WORSE than random useless - in fact dangerously wrong - this wasnt for ms - ive got multiple longterm condition like, i guess, many people whove had ms for many years.
- it s a BIG issue for me that neurologists think ive only got ms. and know nothing about anything else - why not simply ask me - im an expert by experience.
- its an even bigger issue that you seem to think rrms, spms, ppms are useful classifications. ive had rrms for 30 yers - first decades i was running marathons. now im definitely smouldering. so chat gpt on 'rrms' is not helpful?
Not sure if you are addressing me, but I would rather read verified literature than generic ChatGPT generated info. You can get the basic info on RRMS from many verified sources, as I’m sure you know. Maybe I’m misunderstanding your question or you are addressing Prof G, as I’ve never mentioned SPMS or PPMS in thus thread, so you lost me there.
i was agreeing strongly with you - but wanted to also expand out to comment on prof g - he opened with the categories - though he's esrlier spoken on how eg rrms was fabricated for fda
Cool, just got confused :)
I agree. Also, I too am frustrated by the paywalls and gatekeeping of information. To your point about machine learning, I have 2 lawyer friends who as an experiment, used AI to research case law etc., on a lark, and were in fits of laughter at the responses. I know it's not the same, but is another example of the quality of the information.
Interestingly only a couple of days ago I pinged my son, who’s a solicitor, a link to an article on the BBC news website.
It describes the conclusions of research undertaken by a leading uk law firm. Chatbots were put to the test using 50 questions about English law. This was a repeat of earlier research and although they were much improved: ‘they made mistakes, left out important information and invented citations.’
Conclusion: ‘the tools are starting to perform at a level where they could assist in legal research, but there were dangers in using them if lawyers don’t already have a good idea of the answer’
This has definitely reinforced for me the notion that AI is useful but, dare I say, with a degree of hesitation, meaning information being provided needs to be double-checked for accuracy and reliability.
Wow, I replied to Andrew Scott above before I read your comment. I was speaking about my 2 lawyer friends (1 in the US and 1 in Canada) who relayed to me the same thing you wrote about your son's experience and the BBC article.
Yes an online tutorial would be helpful. My problem is getting to see a neurologist at all!
An MS specific LLM may help bridge the gap where health issues affect MS but are usually managed by a Primary Care HCP, who have little experience with MS. Here's my example: My GP instructed me to come off oestrogen only HRT (due to more than 7 years use), which then caused severe cognitive and fatigue symptoms. ChatGPT suggested increasing oestrogen and the GP said to ask my neurologist before making changes. With neurology being so inaccessible I did as ChatGPT suggested. I'm summarising. It was more complicated in reality. ChatGPT helped me with direction, which I then followed up with checking the clinical evidence online.
On a linked, but slightly separate point, I'm all for self-management of MS. However, in the detail I've missed in my example above, there were times when I was incredibly unwell with a respiratory infection, which was being 'managed' by the GP. I was too unwell to do my own research and what I needed was an HCP who understood MS and was also accessible. I wonder if an MS specific LLM might be used by the patient as well as a Primary Care HCP,
I don't trust AI models yet. I probably never will, as they are still based on scouring information that is generally available to anyone or any thing. Even though The large language models are capable of something akin to metacognition they are still incapable of true novelty, and the complexity and subtlety involved in it. They also don't seem to have the ability for intuition.
What is the endgame here? Are you suggesting LLMs can take over the role of consultants? I'm not saying there's no role for AI, but I'm very sceptical about its application in our current environment. Research has shown how AI reinforces existing prejudices and I feel like this tech-positive approach risks isolating and excluding pwMS. I want my consultant to treat me like a human, not a robot.
Gavin, I seem to deviate in opinion a bit but I must say I am very ai-in-healthcare positive. My HCP (ophthalmologist, to be fair) and chatgpt disagreed about if I had MS.
If it weren’t for me describing my optic neuritis to ai, I would not have had the courage to return THREE TIMES to him and demand an MRI (optic nerve enhanced). When I got to neurology, they quickly found t2 lesions in all 3 regions, an enhancing lesion, and OCBs.
AI diagnosed my MS. I would be progressing along without ocrevus were it not for this incredible tool that gave me the courage to advocate for myself.
I consider myself very lucky.
In the US, here, and I concur with your timeline. If the hospital systems can use an AI and eliminate high-priced staff….they will do it in a heartbeat. We only talk to the doctor for 30 minutes, every 6 months, and when we try to ask a question in between we get the support staff running interference; and they are not Doctors, and not highly trained in MS. AI would probably be an upgrade, lol. I use Gregory AI to check on clinical trials already, and would use AI for an opinion, then follow up with more research including the crowdsourcing (social media) stuff. Sort thru it all, make my own mind up. But doctors have biases, even good ones, so I do this already. I’m 67, was a practicing professional in another field, dealt with professionals in other disciplines, and can safely say that “experts” have many levels of ability- many are barely competent, many are arrogant because they had some training…. As you hear already. Heck, the experts hear are pushing to get seniors off of DMT’s as the latest thing, because they wrongly evaluate the wrong risk/reward ratio- our symptoms are (usually) disability worsening, not new lesions, but my friends that are taken off DMT’s usually spiral down much quicker on their disability scale. Which shows that is the wrong approach- yet the experts keep pointing to no lesions to justify what they did. I have no problem with an AI as a tool, but I also have no problem with a doctor as an expert- I still question and check what they tell me.
Hear hear in the states. If older (60’s) once neid status is reached, you are discouraged from your DMT and you are “cured”. And then the real downhill begins. I do not think for a second that anything here would be replaced by AI if it helps the bottom line. I think one would need to check and recheck anything created by AI. Jmo.
Fascinating. My worst MS symptom is very severe RLS. I now have zero symptoms after doing my own in depth research & fighting tooth &.nail for the treatment I'm now taking.
My MS neurologist admits she knows nothing about RLS and UK treatment is about 10 years behind best treatment by the top RLS experts in the USA.
I asked ChatGPT about how to treat RLS correctly and it was absolutely brilliant. It had all the latest, best research and treatment protocols.imse
Sadly, most UK neurologists are still prescribing dangerous dopamine agonists and refusing iron infusions or Buprenorphine.
I wonder why UK neurologists kick back against AI or patient experts?Is it insecurity or arrogance?
Prof Gavin Giovannoni what’s he known for?
See what answer you get 😜
Just part of the answer….
What treatments are available for ms in the uk and what’s the most effective?
ChatGPT said:
In the UK, treatment for multiple sclerosis (MS) typically focuses on modifying the disease's course, managing symptoms, and improving overall quality of life. The choice of treatment depends on the type of MS (e.g., relapsing-remitting MS, secondary-progressive MS, or primary-progressive MS) and individual circumstances. Below are the primary treatments available, followed by insights into their effectiveness.
1. Disease-Modifying Therapies (DMTs)
These are used to slow the progression of MS, reduce the
I am all for it, but I think the ms world has already missed the boat….
Have never used AI for support for my MS. Am open to try it but wouldn't know where to start. An on-line tutorial would be great
You are in the wrong part of the world for this kind of stuff. https://www.cnbc.com/2025/02/21/google-meta-execs-blast-europe-over-strict-ai-regulation.html
You likely don’t need to create a new platform as chat gtp etc gives so much information on ms. Just like as you would know AI has now moved to writing drs’s and neuros reports etc for them. I have personal surgeon friends who it on their own reports and dr’s who current just record the appointment and it transcribes it all into full detailed report, creating better logs etc. As long as information and papers are put on the web then we can find it all through chat gtp and we can also ask for the answer to provide references of where the info is from. Basically I don’t think an ms platform one would give any different info to what chat gtp already gives up. I wonder if you asked it about your own work and whether it would pop up and reference your own papers… it could be interesting for you to try. ☺️
This was the vision:
https://gavingiovannoni.substack.com/p/ms-gpt-your-ai-enabled-ms-guide
It does need to be MS-specific or healthcare-specific, i.e. trained on the medical canon.
I do honestly think it would be a great idea for those less tech savvy but it still uses medical evidence base. Just be specific about your questions and ask it to reference the response with academic papers
We are slowly getting there with the Deep Research models (in fairness, so far I could not quite convince myself to pay for OpenAI Deep Research). The paywalls remain as big hindrance for uptodate research though.
FWIW, even the regular models hallucinate far less (and Gemini can further lower that by cross checking google itself) than they used too, probably less than most humans...
AI fills the gaps in care from HCP's. When I was first diagnosed with MS the MS nurse team said they had too many patients to respond to my questions. So that left me with MS charities who did answer my questions and were so supportive. AI can now fill that gap in care; of the HCPs too busy to actually answer your questions about your MS. I use AI eveveryday in all aspects of my life, including work, personal life, parenting, supporting school work, health and all things MS