Death statistics, both the daily deaths and the cumulative, were as expected. The new daily infection cases, however, broke the model. They were far higher than expected, even considering the known, weekly data oscillations. Over 36,000 new cases were reported. That's the largest number since...ever! The next highest daily infection was 33,578 on April 10. Prior to yesterday, all the data and the model indicated that we had passed the peak infection on or about April 9. You can see the effect of this large daily increase in the cumulative data where the last point jumps above the best fit model curve.
There are really only two choices here. Either the data has some problems and inconsistencies or the infections in the US have started down a more troubling trajectory. The first choice is possible. There has been quite a bit of chatter about states adjusting their statistics: counting cases that had previously been unknown. Also, testing is ramping up, so maybe the sudden increase is just a reflection of that increased testing. The second choice is also possible. As states relax restrictions, the number of daily infections would be expected to increase.
It may be possible to disambiguate the possibilities with time. If it's simply a matter of a data adjustment and an increase in testing, then we might expect to see the data jump up for the next few days (or even over a week given the oscillation issue). After that, the slope of the curves should then return to where they were. That is, there will be a discontinuity and offset between the historical data and the new data, but the curves will be roughly parallel to each other.
If the behavior is due to a real increase in infections, then the curve will continue to steepen, and we may be able to identify the states that are most responsible. Remember that NY/NJ account for a disproportionate number of daily cases and cumulative cases. As those two states settle down, other states are rising and coming to dominate the statistics more strongly. Perhaps the most telling signature will be if the number of deaths begins to rise. Death statistics lag infections. If infections really are increasing dramatically, deaths should follow within a week or so.
No comments:
Post a Comment