Problem is, the common cold coronavirus will trip a covid test positive.
So, need to factor out the common colds.
So, need to factor out the common colds.
That depends on the type of test.
I will admit, I have not personally compared the RNA sequences of SARS-CoV2 and other coronaviruses. However, it is typical for the nucleic acid sequences to show more variability than the protein sequences between closely related organisms. So I expect rtPCR tests to be the most reliable for detecting Covid-19.
On the other hand, I don't expect antibody tests to be able to differentiate between Covid-19 and other coronaviruses because antibodies only recognize very short pieces of proteins, just 6 amino acids long. It is very common for antibodies to cross-react between related viruses. So all of those antibody (serology) tests that purport to show that vast numbers of people have had Covid-19 during cold and flu season are just plain faulty. We can't even use statistical data to filter out positive serology results, since we don't keep track of how many people catch coronavirus colds.
In any case, it is just as bad to inflate the number of Covid-19 cases with dubious serology tests or modeling as it is to dismiss cases on the basis that tests always have false positive and false negative rates. In my opinion, it's best to just stick with the confirmed numbers and keep collecting data. We can only beat this if we have an accurate understanding of how the virus transmits and who is actually sick with it.