The hospitalization curve has been the most useful and reliable curve for spotting trends – and is far more accurate than the daily new positive test cases curve or death curves.
Positive test cases is dependent on how many tests were run – thus at the start of college semesters, for example, the daily positive test case numbers shoot up because of testing all staff and students, biasing the results for a week or two. Second, WHO acknowledged this week that there is a false positive problem when the RT-PCR test is used not as a diagnostic test but to screen mass populations (WHO has revised procedures to address this). Consequently, the daily new cases trend is not a reliable indicator of the direction of the disease.
Deaths are an indicator but only as a lagging indicator and cannot be used to make near term predictions. This is because the curves are based on “daily body count reports” which are always lagging the actual dates of death by a random amount. You can see that no one collates death data on weekends – rather obvious from this chart.
But the larger problem is the daily body count is a measure of data collection efficiency – not actual deaths by date.
To illustrate, in my state, each daily report lists the actual date of death. On a peak day in December, 54 deaths were listed. However, 38 of those occurred in the prior months – in fact, that report include deaths occurring across 5 months. The most deaths to occur in a 24 hour period was 33 on December 9th. Yet 7 “daily body count” reports exceeded 33 as they caught up with past deaths, often long ago.
To illustrate – the orange line is the public “daily report” deaths.
The red line is when the deaths actually occurred. Peak date of deaths was Dec 9th. But if you go by the orange line of “daily body count reports”, peak day of deaths was in the 3rd week of January. This is what the public sees and is misleads them to think January is worse than December.
As public health fell further and further behind in death reports, their daily death report (the orange line) became further out of date.
Consequently, while the summation of dead bodies is correct, the data lags by a random amount and can not be used as a leading indicator.
Declines Began Months Ago
Declines occurred up to months ago, depending on the region. Huge population states like California, New York, Florida and Texas have experienced bad times up until now – and their numbers drive the national numbers. The national numbers, however, begin to flatten out during the past few weeks. Too early to call this a peak – but it might be the peak of the third wave, which would be great if it is.
The national curve might follow the shape of many state curves. That is, the curve will fall off over a time period similar to its rise. Within 2 months cases might be at a far lower levesl if the pattern of say, North Dakota is followed. Wouldn’t that be nice?
This pandemic may be gradually coming to and end, mostly on its own. We hope. But I do not make predictions – I make observations and ask really stupid questions.