Story on BBC news site today. There's an experiment running in Leeds where AI is used to check medical records to diagnose AF.I can see problems in that a) only poorly people go to the drs so their records will be out of date, and b) in the UK our health records are computerized by a mosaic of different companies, all incompatible. But these are solvable. It's good news
Is Mr Pengelly on this forum I wonder?
Interesting, thanks for sharing.
Very interesting. When I saw my nhs cons privately as my nhs appt kept being deferred, the admin at the private hosp, said I would need a GP referral as they hadn't the time to access my nhs notes with same consultant as it was a separate system. This is why I wonder how much info will be available.There is an old film involving a robot who constantly said "more input"
The interoperability of systems and records across the NHS is a huge problem which may trip up this idea and causes problems today as you describe. This is resolved elsewhere by having common standards, which is something that is missing here (too many different self-interested parties I suspect.
Hi Omniscient.
This is an interesting post for me. As you may or not know one of my business projects is AI. I'm going to do a blog post on my website about afib and AI. I can put an affilate link at the end for Karda and donate half (if not all) of any profit made to the forum. I'll run it past admin first. I'm taking a well deserved break until January 2nd but will be back in the swing of things that day and get back to work. 5am starts here we go! TBH I've recharged my batteries and can't wait.
Paul
PS. Some days it's a 5am start and a 10pm finish. The joys of working for yourself (not).
Interesting. Drop me a DM Paul
DM sent Omniscent 👍
That sounds interesting, Paul. I do hope you’re allowed to share the link to your blog post when you’ve written it - I’d be very interested in reading it.
Hi Outsidethelines.
Self promotion is not allowed on the forum - that's why I'll run it by admin first. The HUC will gain anyway as they should make a few bucks from the AI / Afib blog post.
I've already put an article on about AI use in medicine on the site. Give it a few months for Google to send the 'spiders' to check it out and you'll find it easily enough anyway.
Cheers Paul
Please DM me Paul. I co own and run a small biz with my husband so I know exactly what you mean about long days- not ideal for us AFibbers is it? But we're very interested in AI too and are experimenting with it to help with social media posts and formatting or training course content.
Hi Sing.
DM sent. Use Chatty (free) to help you (use the right promotes - I cover that on my site).
Paul
Yes, we've been using Chatty. My other half is good at giving prompts and refining the responses. We're also using Napkin for images and another App that creates ppt from text.
Hahahaha. This must be one of the funniest headlines yet. One of the redtops (Daily Star) has said that AI fridges are going to take over the planet. Allegedly they can talk to each other and are plotting to take over the world!
My neighbour has an AI washing machine. He has to take a picture of his washing and show it to the washing machine. The AI then sets the temperature and how long the wash lasts. Steve said he's fed up with it as it keeps talking to him when he goes into the kitchen and reminds him to do his washing. He says he doesn't want to take a photo of his washing and show it to the washing machine - he just wants to put the clothes in and set to wash.
Needless to say, he's turned the thing off. A little off topic I know but it gave me a laugh.
Paul
AI fridges set to take over the world
dailystar.co.uk/tech/rise-k...
This has made me snort with laughter 😃
I heard of a different use of AI (machine learning to be more accurate) which examined ECGs and could spot future Afib quite well. Probably more practical at the current state of play than trying to get someone's health records AI-accessible.
Hi I use an app called Qaly which uses AI to analyse for 30 different arrhythmias/ palpitations.
Absolutely brilliant and great use of AI technology.
Best wishes
Jezza
I've heard some nonesense about AI but this sounds like a load of rubbish.
Surely the issue is not, have I had have had AF in the past, but do I have it now which is easily tested for (Kardia et al). This sounds like an attempt by health companies to get access to records using the "god of AI" to con the gullable health authorities.
Well, there has been a (1 that I know of and several others I have not read) study using "AI" for the even more difficult prediction of future aortic dissection, using around 60 standard clinical indicators, and it was over 85% (iirc) correct. I suspect these are actually large multi-factorial regressions in the main, but such studies do demonstrate there is promise to be realised. Yes, there are issues with de-identification, and with corporate capture. We need regulators with half a brain, not like the current ones: 70+% owned by drug companies. Worth striving for, as long as we don't forget basic medicine.
I can accept that a computer program is able to spot issues, but it will have nothing to do with AI. It's just a large data base being analised for commonalities.
As in
Selet x where smoking and lung cancerand blood type 'o' and age > 30
or whatever the SQL instruction is.
However, since this data is de-identified, it's not going to help any specific patient. It might detect commonalities (eg, all aortic issues relate to over indulgence in exercise after age 75 or whatever might cause it- I dont know), but is this actually an issue worth exposing all our medical histories.
I had my aortic ultrasound at age 60 (I think it was 60, - it's lost in the mists of time - or worse 🙂). I suspect that is a better way of finding issues of that type than a program that might rule me out, when I have a condition that should rule me in.
My AF started without warning (Oddly a fortnight after starting a course of Bisodol; unrelated to AF). As it happened, I subsequntly suffered a heart attack and became a CABG in the space of six weeks. The heart attack was minor and the CABG sucessful, so in a way the AF was a pre indicator of possible issues to follow. The question is, do all people starting AF have susequent heart attacks and is all AF related to blocked arteries. I suspect not, but surely that would be a better use for a 'quick looksee study' about AF than how it starts.
Hey Ho - still breathing
I worked for the NHS for 17 years as a ward clerk. It used to drive me demented that programmes and data would not talk to each other. You had to rely on the staff to upload details separately on numerous systems. So much information was lost if that wasn't done properly. The computers and software were all ancient and not fit for modern purpose. Tony Blair spent millions on the Cerner programme from the US. As their medical process is very different to ours it wasn't any use. More money was spent getting Fuji to resolve the problems. It was a slight improvement but never what the NHS needed and wasn't compatible to anything else. I hope I haven't put any ex-NHS staff who used Cerner into AF with these recollections!
I've come back from new year to a million replies to this as my daily reminder isn't working. I have 2 thoughts.
First AI will spot patterns that a human would not, and that a simple Google search or SQL command would not
Second, the system can't be used to trawl our records identifying AF patients to call them in for appointments, but to build up a cross-population picture, and then apply that individually (with patients permission). This is allowable.
But whether it works or not is still debatable.
AF Regular check ups.
Vagus nerve stimulation for reducing AF
New to AF
Being shocked out of AF?
