Hi Everyone,
On balance, trends we’ve seen before are continuing.
US Cases
· The number of new cases outside of NY continues to decline slowly (Fig. 10), with the decline being statistically significant after day 50 (19 April, P < 0.0066).
· The number of tests outside of NY continues to increase approximately linearly (Fig 2).
· Correcting new cases per day for number of tests performed, we saw last time that while this index of number of actual cases has declined from a peak around day 40 (9 April), that decline had plateaued around day 71 (10 May). This plateauing appears to be continuing: after day 71 there is no statistically significant decline (P = 0.2053). This suggests that the actual number of new cases per day has stopped decreasing, which could be an effect of relaxing social-distancing practices.
· The US data for cumulative number of cases per test still fits a modified logistic growth curve well (Fig. 4).
· Projections from the fitted curve indicate that most infections from the first wave of the virus will have occurred around day 100. Specifically, for the US outside of NY, on days 96, 124, and 163 (4 June, 2 July, and 10 Aug, respectively—these dates are about 3 days later than projected last time), 95%, 99%, and 99.9% of the total infections during the first wave of the virus will have occurred. On these dates, the minimum number of infections yet to come are 95,041, 22,030, and 2,475. On days 124 and 163, the expected minimum number of new cases will be 1,294 and147.
Fig 1
Fig 2
Fig 3 Trend after day 71 not significant P = 0.2053
Fig 4 Fig 5
US Deaths
· The number of new deaths in the US outside of NY continues its downward trend (Fig. 5). This trend is significant after day 45 (14 April, P < 0.0001). A linear projection of this trend indicates that there will be 0 new deaths by day 142 (20 July). As always, there is a large error associated with this projection, with a zero death day of 232 (18 September) being with the margin of error. This remains a very good trend.
Fig 5 Significant downward trend after day 45 P < 0.0001 0 death day = 142 max 232
NC Cases
· The number of new reported cases in NC continues to rise in an essentially linear fashion (Fig. 6).
· Last time we saw that the number of new cases per test, an index of actual infections, had risen over the most recent 10 days or so. With the new data, this rise seems to have plateaued (Fig. 7; there is no trend after day 80 (P = 0.5228)), meaning that the number of actual new cases has likely leveled off. This is in part good news, but it also means the number of new cases is not declining, not good news.
· As has been the case for a while, the number of new cases per test deviates systematically from a modified logistic growth curve (Fig. 8). The trend over the three weeks has been more linear than the fitted curve, probably due to the plateauing of new cases per test, which has been happening for a similar amount of time. Because of this deviation, I am not making any projections.
Fig 6
Fig 7
Fig 8
NC Deaths
· Last time we saw that what had appeared to be a trend toward decreasing numbers of new deaths had been erased by spikes in new cases. That erasure continues in the present data (Fig. 9): There is no significant trend after day 45 (P = 0l.8966), indicating that the number of deaths has plateaued. This is consistent with the plateauing of number of new cases per test, and indicates that spread of the virus is not being controlled well in NC.
Fig 9
Conclusions
The trends we’ve been seeing recently are continuing. While the number of new reported cases in the US continues to decline slowly, the number of new cases per test actually seems to have stopped declining about three weeks ago. This likely indicates that spread of the virus isn’t being controlled as well as previously. The number of new deaths in the US continues to decline, with projections indicating that if the trend continues, we should see essentially 0 new deaths by mid July. This assumes, of course, that infections won’t pick up again and will actually begin declining.
In NC the news is still somber. The decline in new cases per test has clearly come to a halt, which again means the spread isn’t being checked as much as it was three and more weeks ago. In addition, death rates continue to occur at a more-or less constant rate, which is consistent with the rate of infections. These patterns suggest that we may need to return to more stringent social-distancing measures locally to get declines in new infections and in death rates as are seen in the exemplary patterns from NY.
Stay safe!