They are not accurate at all - case undercounting ranges from 1: 2 (in the case of Iceland) to 1: 500 (Iran). The tests are applied to different groups and some countries (like Peru or Ecuador) even include the so-called rapid tests in their statistics. A good idea you can get on this page that collects all prevalence studies up to date (there are already more than 50): https://docs.google.com/spreadsheets/d/1zC3kW1sMu0sjnT_vP1sh4zL0tF6fIHbA6fcG5RQdqSc/edit#gid=0

That will give you at least a good idea, how bad statistics are. Many greetings from Peru!

Personally thinking, no. This disease is overadvertised. Does it kill more than malaria or hunger?. These two kill millions of people over the years but they don't receive the same attention as COVID-19. Maybe because the western countries are not so much affected as they are with COVID-19.

You can't get accurate results from such heterogeneous data. Some countries perform tests as screening for most of its populations. Some other countries screen only high-risk groups while other groups are tested when suspicious only. Other countries test just to confirm the already- obvious diagnosis of clinical picture and CT.

So you can't take these data for accurate estimation of the prevalence, It will just give you a clue. This is not bad anyhow, we need every possible piece of information.

Later, after the pandemic is over , data of testing protocols in each country could be collected and included in the statistical analysis process.

@Iman Hassan Ibrahim your are right

Misinformation and disinformation may accompany outbreaks which make the statistics looks irrelevant.

follow

Statistics on COVID-19 are manipulated, hidden, showing higher number of tests, and less number of deaths so as to avoid political upheaval.

Even the projections of numbers of people affected and expected deaths by COViD-19 statistical pundits have been proven wrong.

Statistics can be easily manipulated, that is why there is famous quotation:

"There are three types of lies: Lies, Damned lies, and Statistics" (1)

https://cdn.ymaws.com/www.ismte.org/resource/resmgr/EON_Shared_Articles/Lies,_Damned_Lies,_&_Statist.pdf

That is why the question was asked. I do not agree Muhammad Yousuf with the fact that you include statistics in a lie. It is a science, a scientific methodology based on the processing of "real and valid" data, rather than serviced false data.

In this era of COVID-19, my personal opinion, I think the statistics presented by world leaders are grossly inaccurate. Testing is not generalised and in some parts of Africa, patients contest their test results. Some individuals claim, the figures released are distorted by officials for political interest. Personally, I am tempted to believe them.

I don't think the initial data was correct, because at the beginning of the outbreak, no one thought that the disease would spread like this, but because of its importance, the data gradually became more accurate, although the data cannot be trusted.

The validity of the phenomena and processes in this case cannot be revealed without valid data. Thus, the penetration of statistics in the process of monitoring, analysis and statistical conclusion decreases.