Is it all false?



Medicine has a very long tradition of jumping to conclusions and explaining all kinds of fancy mechanism. 

Has anynody done the calculations of false positive covid tests ? It sounds great that a test is 99.9% accurate when it comes to specificity, but the results are terrible when very few have the so-called virus.  See the quote below from the  Norwegian authorities. 

With a prevalence of 0.01 per cent (as in Norway today), the positive predictive value would be around 7 per cent with today's PCR test (sensitivity 80 per cent and specificity 99.9 per cent). That is, 14 out of 15 who test positive are not infected with SARS-CoV-2.  
If we presume that the rate is as low as 20 positives in Australia, one in a million, then we would have 1000 false positives if we test  a million. If we are lucky we would find one of the   Infected ones.  We would then probably treat 1000 persons as if they had so-called Covid. 
All the positives in Melbourne may in principle be false positives. Maybe one or 2 have the virus. Now, if the accuracy is only 99%, we would have 10000 false positives for each hit.  
And this is the best possible scenario, not counting false negatives. 
Victoria actually did 1 million tests and got 0.3 % positives. If the accuracy of the tests used were 99.7 they could all be false positives. 

https://www.fhi.no/en/op/novel-coronavirus-facts-advice/testing-and-follow-up/test-criteria-for-coronavirus

Now  99.9 % specificity may seem a bit high, and BMJ states the following 

As current studies show marked variation and are
current estimates from systematic reviews,
numbers of 70% for sensitivity and 95% for specificity for illustrative purposes

Kommentarer

Legg inn en kommentar

Populære innlegg fra denne bloggen

Covid general thoughts

The number that we normal people can use to guide us concerning risk is not the number of deaths pr positive test. We don’t know the percentage tested, we don’t know how reliable the tests are, and we don’t know how many have had Covid 19 and are immune.  But luckily, like you indicate with your graphs, we can know risk of death by nationality. China is a bit over 2 pr million. Right now, Italy has 6820 deaths in 60 million, 113 pr million.  Annual flu deaths in Italy are around 340 pr million. It the numbers doble for Covid 19, we will still be way below normal flu deaths. And most of the people dying from Covid 19 would be the ones most likely to die from the flu, so if we hadn’t started testing for Corona, we might not have noticed anything at all, especially in China, but even in Italy.  A lot of people die every day from all causes, 1700 pr day in Italy, so many die with Corona, not from Corona. Maybe all the Corona deaths would have died anyway. We really don’t know...
Les mer

How to make a pandemic

How to create a pandemic Imagine that you want to start a pandemic, what would you need? 1. Find some vague criteria for what constitutes the symptoms that you want people to look for. Anything subjective that a lot of people can identify with,is ideal. Let us take memory problems and/or confusion + a few common ones from the Covid list. Tiredness, aches and pains are common and subjective enough. (For covid 19  the symptoms are:  fever, dry cough,tiredness. Less common symptoms:aches and pains, sore throat, diarrhea, loss of taste or smell, a rash on skin, or discolouration of fingers or toes) It would be a good idea to take something that is very common in old people  so that we can use death from old age as proof of the lethality of the new virus. 2. Then we would need something biological to test. Any RNA sequence would do, as long as it is not present in the whole population. If it were, someone might claim herd immunity very quickly. Actually it could be a R...
Les mer

Only one in 1000 positive tests should be counted as positive.

Covid tests may be 1000 times too sensitive From an article in NYT: "The PCR test amplifies genetic matter from the virus in cycles; the fewer cycles required, the greater the amount of virus, or viral load, in the sample. The greater the viral load, the more likely the patient is to be contagious. This number of amplification cycles needed to find the virus, called the cycle threshold, is never included in the results sent to doctors and coronavirus patients, although it could tell them how infectious the patients are. In three sets of testing data that include cycle thresholds, compiled by officials in Massachusetts, New York and Nevada, up to 90 percent of people testing positive carried barely any virus, a review by The Times found. On Thursday, the United States recorded 45,604 new coronavirus cases, according to a database maintained by The Times. If the rates of contagiousness in Massachusetts and New York were to apply nationwide, then perhaps only 4,500 of ...
Les mer