Home / news / Researchers in the US and China reported Monday they have developed an artificial intelligence tool that is able to accurately predict which newly infected patients with the novel coronavirus go on to develop severe lung disease

Researchers in the US and China reported Monday they have developed an artificial intelligence tool that is able to accurately predict which newly infected patients with the novel coronavirus go on to develop severe lung disease

Researchers in the US and China reported Monday they have developed an artificial intelligence tool that is able to accurately predict which newly infected patients with the novel coronavirus go on to develop severe lung disease

View Reddit by DoremusJessupView Source

12 comments

  1. One of the predictive markers, elevated levels of the liver enzyme ALT, is frequently caused by alcohol consumption. This is really bad news for my quarantine-era hobby of drinking alone.

  2. Here is a better article which mentions some of the actual indicators

    [https://www.sciencedaily.com/releases/2020/03/200330152135.htm](https://www.sciencedaily.com/releases/2020/03/200330152135.htm)

    >The researchers were surprised to find that characteristics considered to be hallmarks of COVID-19, like certain patterns seen in lung images (e.g. ground glass opacities), fever, and strong immune responses, were not useful in predicting which of the many patients with initial, mild symptoms would go to develop severe lung disease. Neither were age and gender helpful in predicting serious disease, although past studies had found men over 60 to be at higher risk.
    >
    >Instead, the new AI tool found that changes in three features — levels of the liver enzyme alanine aminotransferase (ALT), reported myalgia, and hemoglobin levels — were most accurately predictive of subsequent, severe disease. Together with other factors, the team reported being able to predict risk of ARDS with up to 80 percent accuracy.

  3. 53? That isn’t exactly a plethora of input data for machine learning with (I assume) a neural network with many nodes and input variables in the first layer. Hopefully the New York data will ramp that up a couple orders of magnitude.

  4. Heh, glad I barely drink then

  5. Well, SARS caused liver damage in many patients. It’s not surprising that this does, too. The enzymes may be present due to this.

    >According to a 2004 report, up to 60% of patients with SARS had liver impairment, with liver biopsy specimens demonstrating viral nucleic acids and injury. These authors noted that this may have been the result of drug-induced liver injury, given that most of these patients were treated with high doses of potentially hepatotoxic antivirals, antibiotics, and steroids.

    https://www.medicinenet.com/script/main/art.asp?articlekey=229150#

    >Patients with digestive symptoms had higher mean liver enzyme levels, lower monocyte count, longer prothrombin time, and received more antimicrobial treatment than those without digestive symptoms.

    >The top-line result that digestive symptoms occur in roughly half of COVID-19 patients presenting to hospital remains unchanged.

    https://journals.lww.com/ajg/Documents/COVID_Digestive_Symptoms_AJG_Preproof.pdf?utm_source=yahoo&utm_medium=referral&utm_campaign=in-text-link

  6. 53 examples seems to me to be a small dataset for an ai learning program. 80 percent accurate based on the small sample. Not all that good if for example, one in five is sent home rather than icu.

  7. 53 inputs is so ridiculously small that they should never have done this. It’s literally worse than not having done it at all. The feature space here is so complex compared to their data set that no meaningful validation could really be done. Plus, with this size of data, 80% isn’t really meaningful at all..

  8. Vincent will never make it off this rock.

  9. > finding that changes in three features — levels of the liver enzyme alanine aminotransferase (ALT), reported body aches, and hemoglobin levels –- were most accurately predictive of subsequent, severe disease.

    I wonder if the aches are simply dehydration from so much water moving to the lungs. I’m in my mid 30s and only recently realized I get foot cramps and leg aches if I’m dehydrated.

  10. Is it really AI or does it just run through a database of warning signs?

  11. We can’t even tell who is newly infected since our tests are shit and take a week to come back. By then they are fine or on the vent, so this is not useful unless everyone is tested and the test is rapid.

Leave a Reply