Thursday, April 16, 2020

New York Not Close to Exiting Lockdown

The above shows the daily count of new Covid cases in New York state (blue) and the seven day centered moving average (orange).  The black line marks when the Governor issued his "PAUSE" (shutdown) order.  It continues to appear that around 12-14 days after that order, cases peaked and are now in a slow decline.

The quality of this date is appalling - many symptomatic people are not getting tested, the criteria to get a test vary widely depending on location within the state, and have changed over time in poorly documented ways.  The seeming day-to-day noise in the data is also much too large - a random process producing a count approaching 10000 should have tiny day-to-day fluctuations, so something is very wrong with the data collection - possibly the count is terminated at different times of day or something of that kind.

One measure of the data quality is the fraction of new test results that are positive, shown here:

It peaked at around 50% (ignoring whatever caused that outlier at 75%) and has now declined into the mid thirties.  That suggests that a little more testing is happening relative to the size of the epidemic, and tends to confirm the idea that the slight decline in new cases is real.

There is no sign at all that we can exit lockdown any time soon.  To remove most of the lockdown measures, it would be good to have few cases (in the hundreds statewide), and be aggressively chasing them all down, testing all their contacts and confining anyone found positive or symptomatic.  That would imply a positive test ratio of 5% or less.

We would appear to be months away from these kinds of exit criteria on present trends.

One possible new factor is the governor's order, taking effect yesterday, that everyone needs to wear masks anywhere social distancing may be imperfect.  We don't know how well this will work in New York, but it's at least possible it will cause another significant decline in the case volume.  We'll find out in 12-14 days.


Lars-Eric Bjerke said...

In the graf below you can compare number of death of all countries. All western countries have lock downs of various kinds except Sweden. Our state epidemiologist claims the virus will spread in all western countries when they exit the lock down and that the only way to reduce the number of deaths is to protect the elderly and the sensible (where Sweden has failed especially in Stockholm). He assumes that a vaccin will be too late. The measures taken awaiting the vaccin should of course be sufficient not to overload the hospitals. It may be noted that Stockholm has a high death rate and the major part of the country a rate comparable to the other Scandinavian countries which implies that one remedy does not fit all countries or part of a country.

Lars-Eric Bjerke said...

The Stockholm region has 2,4 million inhabitants, of these 5826 have been tested positive for Covid 19 and 921 have died (April 19). Calculations by The Public Health Agency estimates that about 700 000 inhabitants (30 %) are or have been infected. Thus the estimated death rate so far is 0,13 %. Half of these cases have occurred att retirement homes, a terrible failure. To protect the elderly was a major strategy of the Agency.
Of the remaining 7.6 million Swedish inhabitants 620 have died.