Press Release: IQuity Launches Healthcare Analytics Platform (predicts MS with 90% accuracy)

[Webdog]

IQuity has been covered in the forums before, but this looks like new news. IQuity claims they can predict Multiple Sclerosis with 90% accuracy at least 8 months before traditional methods. They also claim their approach "can be applied to any disease".

IQuity Launches Healthcare Analytics Platform
Machine learning platform predicts, detects and monitors chronic disease across patient populations
https://www.prnewswire.com/news-rel...-healthcare-analytics-platform-300685630.html

NASHVILLE, Tenn., July 24, 2018 /PRNewswire/ -- IQuity, a Nashville-based data analytics company that specializes in predicting, detecting and monitoring chronic disease, today announced the launch of a revolutionary analytics platform to apply this approach within large populations.

This new platform represents IQuity's evolution from a developer of genomic diagnostic technologies to an integrated data science company that analyzes many types of public and private data sets to provide new insights into the health of populations.

The approach was demonstrated in a pilot study that analyzed healthcare claims for 20 million people in New York – comprising four billion data points. Focusing on multiple sclerosis (MS), the approach predicted with over 90 percent accuracy the onset of MS within that population at least eight months before traditional methods would typically yield a diagnosis. Early diagnosis can lead to better patient outcomes and deliver substantial savings, as spending on healthcare tends to accelerate prior to a definite diagnosis. The approach can be applied to any disease.

Multiple Sclerosis News Today also covers the story from an MS perspective:
https://multiplesclerosisnewstoday....tform-can-accurately-predict-ms-company-says/

 

[Wonko]

This concerns me.

deliver substantial savings, as spending on healthcare tends to accelerate prior to a definite diagnosis.

Presumably the reason for increased spending prior to diagnosis is that people have something wrong and are being tested, are they suggesting that with machine learning models money will be saved at this point due to no/reduced testing?

 

[mariovitali]

OTOH Machine Learning is going to be disruptive for being able to suggest Biological Targets that are more likely to be relevant to the actual origin of a Disease or Syndrome.

 

[Inara]

This concerns me.

Presumably the reason for increased spending prior to diagnosis is that people have something wrong and are being tested, are they suggesting that with machine learning models money will be saved at this point due to no/reduced testing?

Or prior to correct diagnosis people are misdiagnosed and go to several doctors in order to get it right?

Maybe the money-saving-argument has to be mentioned; it's standard nowadays.

Personally I think machine supported diagnosis has chances. There are so many diseases, so many publications etc. Nobody can be up to date everyday. The doctors I met didn't read new publications or medicine news (with one exception). A computer can check huge databases in a short time. So as a support in diagnosis it could be a great help amd maybe it will avoid some misdiagnoses.

But it depends on the algorithm and database. If only psychological illnesses are checked or are given higher priority, for instance, it's useless.