Right now, numbers of infected persons are being bandied about in news media as part of the breathless coverage of Covid-19. Those numbers are meaningless without context.

Let me explain. If the media report, say, 500 Covid-19 infections, my first reaction as a science journalist is to ask how many people were tested. It’s an issue of maths and basic logic.

If, say, 100 people are tested and ten of those test positive, the *only* thing we can deduce is that ten people – or 10% – of *those tested*, at *the moment they were tested*, were positive. That’s it. We can’t extrapolate it to a broader population.

Let’s say a week later, the number of people tested positive is now 20; has the rate of infection doubled? Not necessarily. It depends on how many people were tested. If, to get more accurate data, 200 people were tested, then still only 10% of those tested were positive. The rate of infection is the same.

Now, let’s assume, (because the situation is urgent), there’s a rapid expansion in testing, and 400 people are tested. If the number reported positive is still 20, then the infection rate is actually *lower.* That’s a whole different picture.

That’s why numbers on their own are meaningless, and why mainstream media who publish numbers without context are being careless or, worse, disingenuous.

Time to dig a little deeper. Those 500 who are infected – how are they? It’s not a flippant question. If they’re all connected to ventilators then the situation is serious. If, on the other hand, they’re all self-quarantined and, outside of the aches and pains and coughing and sneezing, otherwise well, then that’s a different picture.

What about those who die? Let’s go there, with an eye on the numbers.

If the fatality rate of Covid-19 is mentioned in the media, ask how it was measured. Is the figure reported the case fatality rate (CFR) or the infection fatality rate (IFR)? There’s a big difference.

The *case fatality rate* is the number of deaths divided by the number of *known* infections. But not all infections are known. What if we include those who are infected, but aren’t tested because, say, their symptoms aren’t that serious? Then we get the *infection fatality rate*. That’s the number of deaths divided by the *true* number of infections. Obviously this statistic is more difficult to calculate – it requires estimating the number of unreported infections. If correct, it is a more accurate picture of how dangerous a pathogen may be.

Generally speaking, the CFR will be larger than the IFR, and so more compelling for news media.

‘Infected’ is a powerful word, ‘fatality’ is frightening. Better care should be exercised by new media to paint a clearer picture of Covid-19.