Social Blogs

Replies

• 

  Bup July 14, 2021 at 4:16am

  Better not buy Woods then, for the COVID risk alone.
  • 

    Uncle Wizards Sleeve indigenous elder He/Him > Bup July 15, 2021 at 12:15am

    Haha , the bat probably caught it off Woods . 
• 

  Kurupt - Sofa King Annoyed July 14, 2021 at 5:51am

  Then that begs the question "should the bulldogs be allowed inside the Queensland bubble"?
• This reply was deleted.
  • This reply was deleted.
    • 

      Parramatta Tragic July 14, 2021 at 1:12pm

      Cobwebs?

       
• 

  Prof. Daz July 14, 2021 at 9:10am

  Wile, that is not what the study actually reports. We have to be careful swapping from an association claim (Men with X also have Y) to a causal claim (Men with X get Y). 
  The report noted the study authors explicitly saying it's an association not a causal finding. They don't know if pre-existing low testosterone leads to worse outcomes if the male gets Covid-19, or getting Covid-19 really badly also reduces testosterone. I'm sure the full report also probably controlled for the obvious confounding factor: age. Older you are, the worse Covid hits. 
  • 

    Monsieur Eel > Prof. Daz July 14, 2021 at 10:02am

    Daz, I was sent this recent study on Ivermectin. Posting here because someone closed the previous blog for some reason.

    https://journals.lww.com/americantherapeutics/fulltext/2021/08000/i...

    It seems legit but I'm not sure. What do you think?

    

    Ivermectin for Prevention and Treatment of COVID-19... : American Journal of Therapeutics

    mortality, in secondary outcomes, and in chemoprophylaxis, among people with, or at high risk of, COVID-19 infection. Data sources: We searched…
    • 

      My Bob > Monsieur Eel July 14, 2021 at 11:37am

      I wondered why super closed that blog down, I assume it was him. Maybe he thinks he is You Tube and needs to protect us from alternative views?
      • 

        Monsieur Eel > My Bob July 14, 2021 at 11:53am

        Most likely because of UFC 264, i.e. the Wiz and Daz extravaganza.

        McGregor breaking his leg off was nothing compared to that show.
    • 

      Prof. Daz > Monsieur Eel July 15, 2021 at 9:42am

      Monsier Eel, this Bryant et al study is relatively new and I found it listed in the data sampling we run,  but I haven't had a chance yet to go over it (busy with some other tasks). 

      I whipped through and a few remarks just off the top of my head.

      1) the authors state no conflict of interest but that is not true. Bryant heads the UK-based BIRD group which is an ivermectin advocacy group, the UK equivalent to the US FLCCC. You could try to say advocacy is distinct to money, but I would ask why hide the link?

      2) study reports 62% reduction in hospitalization risk and 80% reduction in death risk. This is an outsized conclusion given their own data. At best they could say ivermectin might have benefit and is worth a full trial. Why exaggerate the effect?

      3) the most technical so bear with me! Look at Fig 7, the funnel plot. Their own diagram contradicts how they render it. The funnel plot measures publication bias (tendency to publish positive not negative findings). If no pub bias, the results should cluster around 1.0. Bryant et al claim they do but their plot shows the opposite. Small and low quality studies cluster lower than 1.0, meaning reduced risk of death if taking ivermectin, but larger and higher quality studies cluster around 1.0 ( o effect). Note the authors identified lower and higher quality studies so I am using their assessment. Their own funnel plot clearly shows publication bias. Statistically, I bet it you removed the small and low quality studies, the beneficial effect Bryant et al report would disappear. I say "I bet" but seriously it's obvious from their own data, as the studies they rank higher cluster to reporting no effect. 

      4) thus, the big issue with meta analysis comes home to roost here. Just because you pool a whole lot of small and low quality studies, you don't magically produce a large high quality study. They have not produced a big randomized and controlled clinical trial just by pooling data from small low quality trials. But they pretend they have by ignoring their own funnel plot. 

      5) PS: I just plugged "Kory" and "Bryant" into the search field for our algorithmic data scraper and limited results to past 2 months and I see Kory has tweeted that this Bryant et al study is a "slam dunk" proving ivermectin is efficacious. No no no. A meta analysis even if it is the greatest meta analysis ever is just a pointer to the need for a large, high quality, randomized clinical trial. The meta analysis itself flags associations. If you want evidence of conflict of interest, it's Pierre Kory completely misrepresenting what the study does. 

      I could say more but to do so would involve following the trail of the 24 studies they use. In a meta analysis, your statistical regressions are only as good as your inputs. Garbage in, garbage out. Health food in, health outcome out!  But as I noted, their own funnel plot shows their conclusion relies entirely on the small and low quality studies that they themselves identified as showing signs of publication bias. How did that pub bias ramify through to their conclusion, you should ask. It's garbage in garbage out! They should have performed regression on the two sides of the funnel plot (lower than 1.0 and cluster around 1.0) but they chose to retain the low quality studies. Why? Because it dilutes the "no effect" of the higher quality studies and drives the conclusion toward a positive result. It's publication bias pure and simple. I would not trust any conclusion based on this meta analysis. 
      • 

        Mapik > Prof. Daz July 15, 2021 at 4:42pm

        https://c19ivermectin.com/

        

        

        Ivermectin for COVID-19: real-time analysis of all 101 studies

        Ivermectin for COVID-19: real-time analysis of all 101 studies