Welcome to Atheist Discussion, a new community created by former members of The Thinking Atheist forum.

Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
AI equal with human experts in medical diagnosis, study finds
#1

AI equal with human experts in medical diagnosis, study finds
https://www.theguardian.com/technology/2...tudy-finds

Quote:Artificial intelligence is on a par with human experts when it comes to making medical diagnoses based on images, a review has found.

The potential for artificial intelligence in healthcare has caused excitement, with advocates saying it will ease the strain on resources, free up time for doctor-patient interactions and even aid the development of tailored treatment. Last month the government announced £250m of funding for a new NHS artificial intelligence laboratory.

However, experts have warned the latest findings are based on a small number of studies, since the field is littered with poor-quality research.

One burgeoning application is the use of AI in interpreting medical images – a field that relies on deep learning, a sophisticated form of machine learning in which a series of labelled images are fed into algorithms that pick out features within them and learn how to classify similar images. This approach has shown promise in diagnosis of diseases from cancers to eye conditions.

However questions remain about how such deep learning systems measure up to human skills. Now researchers say they have conducted the first comprehensive review of published studies on the issue, and found humans and machines are on a par.
The following 2 users Like GenesisNemesis's post:
  • Tres Leches, Alan V
Reply
#2

AI equal with human experts in medical diagnosis, study finds
So WebMD was correct?
The following 1 user Likes no one's post:
  • GenesisNemesis
Reply
#3

AI equal with human experts in medical diagnosis, study finds
This doesn't surprise me. In my experience, doctors vary by a couple orders of magnitude in their ability to observe and correctly diagnose. The average is insufficient. In this regard they are like software developers ... many of whom could not program (and certainly not architect) their way out of a paper bag.

I'm in the process of changing internists for this very reason. I noted him cutting corners on my annual physical. For example it's a best practice to do a digital prostate exam via the rectum. I certainly do not look forward to this (a doctor I had years ago would say, as he pulled on his latex gloves, "and now I'm going to ruin your day AND mine.") But that he hasn't done it in 4 years is a little concerning. I am sure it's because he doesn't like doing it.

The last straw was he misdiagnosed my stepson with a toe fungus which miraculously cleared up on its own, and he misdiagnosed me with a planter's wart which ended up being a self-treatable plugged sweat gland. He also has let me go for 2 years without a physical. Basically he's not engaged, and may not be all that competent.

In the face of all this, at least an AI observes meticulously and will not miss any observed patterns, so it's bound to be better than the average, mediocre chutes-and-ladders diagnostic skills of most doctors. Its biggest problem is probably that it's still following a mostly blood test-based diagnostic (if it's not in a blood test the symptom doesn't matter and possibly doesn't exist) and mostly a pharmaceutical-oriented treatment regime (it would not be good at eliciting or noticing that you're depressed or stressed or anxious for example).
The following 1 user Likes mordant's post:
  • GenesisNemesis
Reply
#4

AI equal with human experts in medical diagnosis, study finds
I'd imagine that visuals are on the simpler side of diagnostic skills.
The following 2 users Like Thumpalumpacus's post:
  • GenesisNemesis, brewerb
Reply
#5

AI equal with human experts in medical diagnosis, study finds
Personally, I would like to receive an AI diagnosis, but I would like a session with the doc to ask pertinent questions. It's not just about looks, it's also about feels and interactions.

This reminds me of computerized translations - it finds words to match the words it sees, but it cannot see context, inflections of the voice, and it doesn't know the myriad of idioms and sayings that exist in all languages. Some of the results are hilariously wrong. 

So, while I think AI can be a part of establishing diagnosis, I wouldn't want to rely on it just yet.
[Image: color%5D%5Bcolor=#333333%5D%5Bsize=small%5D%5Bfont=T...ans-Serif%5D]
The following 3 users Like Dom's post:
  • brewerb, GenesisNemesis, Thumpalumpacus
Reply
#6

AI equal with human experts in medical diagnosis, study finds
(09-28-2019, 10:28 PM)mordant Wrote: ...I noted him cutting corners on my annual physical. For example it's a best practice to do a digital prostate exam via the rectum. I certainly do not look forward to this (a doctor I had years ago would say, as he pulled on his latex gloves, "and now I'm going to ruin your day AND mine.") But that he hasn't done it in 4 years is a little concerning. I am sure it's because he doesn't like doing it.

Considered clinical opinion in Australia now says that digital rectal examinations are a waste of time due both to the poor
skills of most GPs, plus the fact that our prostates vary in size and density from one bloke to the next.  The accepted test
is now the PSA (prostate specific antigen) blood test, although the Cancer Council Australia and the Royal Australian College
of General Practitioners both recommend against routine screening in men without symptoms.  I personally disagree with
this view, as a gradual increase in PSA levels over a period of years can definitely point to prostate cancer.

My late father had a subtotal prostatectomy, so I'm aware of the possible hereditary cancer syndrome, and prefer to have the
PSA test every couple of years, even if just for peace of mind.  As far as I know, no computer program can diagnose prostate
cancer without human intervention.
I'm a creationist;   I believe that man created God.
The following 2 users Like SYZ's post:
  • GenesisNemesis, mordant
Reply
#7

AI equal with human experts in medical diagnosis, study finds
(09-29-2019, 04:29 PM)SYZ Wrote:
(09-28-2019, 10:28 PM)mordant Wrote: ...I noted him cutting corners on my annual physical. For example it's a best practice to do a digital prostate exam via the rectum. I certainly do not look forward to this (a doctor I had years ago would say, as he pulled on his latex gloves, "and now I'm going to ruin your day AND mine.") But that he hasn't done it in 4 years is a little concerning. I am sure it's because he doesn't like doing it.

Considered clinical opinion in Australia now says that digital rectal examinations are a waste of time due both to the poor
skills of most GPs, plus the fact that our prostates vary in size and density from one bloke to the next.  The accepted test
is now the PSA (prostate specific antigen) blood test, although the Cancer Council Australia and the Royal Australian College
of General Practitioners both recommend against routine screening in men without symptoms.  I personally disagree with
this view, as a gradual increase in PSA levels over a period of years can definitely point to prostate cancer.

My late father had a subtotal prostatectomy, so I'm aware of the possible hereditary cancer syndrome, and prefer to have the
PSA test every couple of years, even if just for peace of mind.  As far as I know, no computer program can diagnose prostate
cancer without human intervention.

Well there's a situation where if the doc would simply explain his thinking it might not bother me so much. I've had annual DREs from three different doctors over the past twenty years and suddenly this guy does one and that's it for four years ... seems like a little engagement would go a long way here. I'm the last to wax nostalgic for a DRE, so I'd be happy to find that PSA tests supercede them.

Now that you mention this I did a bit of reading and it does seem like US practitioners are starting to question the utility of the DRE also.

At least I HAVE been administered PSA tests more or less annually.
The following 1 user Likes mordant's post:
  • SYZ
Reply
#8

AI equal with human experts in medical diagnosis, study finds
The medicos (in Australia) are now saying that for men over age 70 years (like me) any/all tests for
prostate cancer are waste of time and money.  Apparently prostate cancers are so slow growing that
older blokes will "outlive" the cancer, and that such invasive surgery—a prostatectomy—is a needlessly
invasive procedure that could throw up more problems than it solves.

In fact, that the prostate gland surrounds the urethra is one of the proofs(?) that some omniscient god
didn't design the human body.  The gland should've been off to one side, so its enlargement, with
ageing, wouldn't cause urinary problems.

And as I said earlier, prostate cancer is diagnosed without the need for any robotic involvement, due
to the fact it relies solely on human assessment.  I know of course there are several surgical procedures
that can now be performed purely robotically, but even those rely on critical human input during the
actual procedure. The robot doesn't possess any "intelligence" per se, and I'm guessing I won't see
true clinical AI in my lifetime.

I also note (from the link) that Prof David Spiegelhalter, the chair of the Winton centre for risk and
evidence communication at the University of Cambridge, said the field was awash with poor research.

"This excellent review demonstrates that the massive hype over AI in medicine obscures the
lamentable quality of almost all evaluation studies" he said.
I'm a creationist;   I believe that man created God.
The following 1 user Likes SYZ's post:
  • GenesisNemesis
Reply




Users browsing this thread: 1 Guest(s)