If you’ve been following the coronavirus outbreak in the US, you have likely heard that case numbers in the United States are reaching record highs. This is true: the US averaged nearly 70,000 new cases per day in late July, more than twice what it was in April. As a result, some in the media are proclaiming that the outbreak is currently the worst it’s ever been. Meanwhile, Trump and some on the right have bragged about how the US’s case fatality rate – deaths as a percentage of cases – has dropped significantly since April. What often isn’t mentioned, however, is that the number of cases being recorded is heavily dependent on the number of tests being conducted, and the US is currently conducting about 5 times more tests than we were back in early April.
This isn’t to say that we are now identifying every coronavirus infection: unless you are testing the entire US population daily, it is impossible to know the exact number of people currently infected in the US. Still, the larger percentage of the population you test, the more cases you will identify. This testing differential is why the US is currently experiencing far fewer deaths than April, despite having more official new cases now.
Positive test percentage – the share of total tests coming back positive – is a much better indicator of the scale of the coronavirus outbreak than total cases. This can be seen from a statistical analysis as well. In fact, there is no statistically significant relationship between total coronavirus cases and coronavirus deaths in the US:
Positive test%, meanwhile, explains 96% of the variance in US coronavirus deaths, assuming a 19 day lag between positive test% and deaths:
Using positive test%, we can see the true severity of the outbreak over time. And a clear pattern emerges: while the current positive test rate is not even close to the peak of 20% back in April, there has been a significant increase in the past couple months, doubling from a low of 4.3% in June to a high of 8.6% in late July:
Coronavirus Positive Test% Over Time, United States
As of this writing, the positive test rate stands at 7.5%, slightly lower than where we were a few weeks ago, but higher than we were for much of May and June. This is likely why daily deaths have roughly doubled from early July to today:
The lesson here is simple: Confirmed case numbers are practically useless without considering the number of tests being conducted. Monitoring the positive test rate nationally as well as within individual states is a good way to identify the severity of the outbreak within the US, and I will be doing this analysis in the future.
