First test for long Covid gets EU approval, August 2022 - IncellDx incellKINE Long COVID test

Post copied from the Long Covid in the media thread.
See also this thread:
USA: InCellDx Dr Patterson - New lab service offering cytokine tests to Covid longhaulers
____________

https://uk.finance.yahoo.com/news/first-diagnostic-test-long-covid-070000516.html

 

Merged thread

taken from pressreader
https://www.pressreader.com/

 

Really?

This is a 'fact' is it?

I have other 'objections' to this paragraphs assertion but.....maybe I have the sniffles.

 

This seems a bit out of nowhere - I've heard of the man and the company but not in the context of anything produced being 'nearly there'. Then suddenly something suggesting a regulator is taking the 90% effectiveness claim as correct. What have I missed? does the 'Kine' bit assume they are looking for cytokines or something?

 

A blood test for people feeling 'rubbish'. Umm, they can't walk far, stand upright or process information.

 

Another scam no doubt.

 

Bet they're regretting picking that name for their company, though.

 

there are some sceptical threads on IncellDX and Dr. Patterson already, as this one

https://www.s4me.info/threads/usa-i...ng-cytokine-tests-to-covid-longhaulers.20434/

 

My concerns are generally the lack of validation. It seems like the test was developed using machine learning. When you do this, you should have two cohorts--one for training, another for validation, to guard against overfitting (classifying data based on coincidences in the particular dataset you're studying). But they used only one. I also haven't seen evidence it can distinguish between long Covid and other conditions. It's a test of immune markers, so people with autoimmune diseases unrelated to LC or ME might get false positives.

Thus, my hopes are modest. Maybe the test is half-decent after all, and will be better-validated later. Maybe its arrival will increase interest in biomarkers and lead to development of a better one. I don't know. But my gut says this is a development, even if it's an overhyped one.

 

Isn't Patterson the guy who claims to have cured something like 6000 patients yet not a single anecdote was to be found anywhere? Surely institutions are not falling for this garbage are they?

 

Yes he is. I have seen a bunch of people on the longhaulers sub talking about the results they got and the treatment, none has said it cured them so far. Patterson claims it distinguishes between LC and ME/CFS and other conditions but I haven't seen a paper to that effect yet. If he was curing so many people I would have expected it being shouted hard on the long haulers sub, but its not those that recover have done so slowly and naturally like most post viral sufferers and those that haven't are stuck like pwME. We would be hearing a lot more buzz around something that has cured so many that wasn't from the organisation itself.

 

Why wasn't Dr. Bruce Patterson who went on to a residency in Pathology (focusing on viral pathogenesis) and HIV interested in M.E years ago?

Was there not a significant amount of data for ME/CFS patients?

 

There is a deeper problem with this approach. The gold standard for machine learning is the existing clinical diagnosis. If all the test does is come up with the same population, clinical diagnosis is just as good. And if clinical diagnosis is not considered reliable then the test is not reliable either.

Diagnostic tests are useful to the extent that they provide some sort of explanation that adds confidence to the clinical diagnosis or that they can predict the development of a clinical diagnosis. The latter does not apply here. A good test is not one that correlates best but one the picks out the cases that do not fit and explains why. A machine learning based algorithm almost by definition does not provide any explanation - it is just what the computer spits out.

 

There are two primary ways to use machine learning, tagged or untagged data. With tagged data you would be taking positive results and negative ones determined manually and ensuring it correctly identified which is which through the training. I suspect this is using untagged data where the machine learns groupings based on the data and then Patterson's team assigns a tag to the groups as to whether is positive or not for the condition. The untagged approach is going to be hard to accept in most medical fields I think, places like Netflix used it quite successfully for finding new categories for movies and TV (but they don't expose them as they make no sense to viewers) but medically its backwards compared to what people are used to. The issue we have with a lot of machine learning based research at the moment is a lack of access to the code and the models to replicate results and often they aren't made available even on request. Papers based on this can't be considered high quality as a result. Hopefully the authorities at least have assessed it appropriately and have all the pieces they need to replicate the results. A diagnostic test would be wonderful and I think it will change the fate of the condition dramatically.

 

.....at least until it emerges that what was actually being picked up was people who didn't have ready brek for breakfast.....or something else entirely unrelated.

 

What does this mean?:

"Products Regulatory Agency, before it can go on sale here."

 

ME Association: New Test For Long Covid

"The Times reports on a diagnostic test from US-based diagnostics company IncellDx which claims to detect Long Covid and can distinguish between this and ME/CFS. The article is behind a paywall but can be accessed from the link below, https://www.thetimes.co.uk/article/...elldx-blood-treatment-symptoms-2023-hjz3rgdq0"

https://meassociation.org.uk/2023/02/new-test-for-long-covid/

Also reports on comments on the test from a certain Jonathan Edwards.

 

The author is Francesca Steele who has previously written about (ie, advertised) the Lightning Process for LC.

 

I'm pretty certain that my blood marker profile during PVFS in the early years is quite different from 32 years of having post-infectious M.E.

At what point does it become post-covid and no longer long-covid?

 

Copied post
(Paywall)
I tried the new test for long Covid

Francesca Steele suffered fatigue and flu symptoms for a year after she had coronavirus. Now she’s taken a new private blood test, what would her diagnosis be?
https://www.thetimes.co.uk/article/...elldx-blood-treatment-symptoms-2023-hjz3rgdq0