I said I would post some analysis of other countries' experiences post-peak. As mentioned previously, I'm not saying for definite that we have peaked, but the main indicators are positive and, even if we haven't quite reached it, the consensus would seem to be that we will do so in the coming days.
So what does a post-peak world look like? The best examples that we have are Italy and Spain.
The first thing I should say is that I haven't had chance to update my figures for Italy and Spain today. Secondly, and more importantly, I don't measure the peak as being a specific day. As individual days can vary from one another, leading to outliers within a general trend, I use the 4-days which showed the highest average during that period.
In Italy, using that measure, new infections peaked during the period 26th-29th March when the average daily infection rate was 5,826. As there have been 16 days since that period (in my analysis), if we go back 16 days from the peak we can see there were 70,440 infections at a daily average of 4,403. The 16 days since the peak have seen a further 64,799 infections at a daily average of 4,050 (8.0% lower).
Spain is even more pronounced, where new infections peaked between 30th March-2nd April, when the average rate was 7,989. The 12 days before the peak period saw 73,187 infections at an average daily rate of 6,099. The 12 days following the peak have seen 61,955 further infections at a daily average of 5,166 (15.3% lower).
So, there is reasonable evidence (albeit limited to a sample of two countries) that we could expect to see a marked fall in our daily infection rate once the peak has been confirmed (it may have occured between 8th-11th April but too early to confirm yet). Of course, the UK testing data has been made more complicated by the inclusion of tests for NHS staff in recent days, but it is still possible to strip this out to make a like-for-like comparison.
If we now turn to new deaths, unfortunately we see an entirely different picture. In Italy, new deaths peaked between 27th-30th March when the average daily death rate was 844. The 15 days before the peak had seen 8,118 deaths at a daily average of 541. The 15 days following this peak have witnessed 9,476 deaths at a daily average of 632 (16.8% higher).
Spain shows a similar, but slightly lower, effect. New deaths peaked between 30th March-2nd April (same as new infections), when the average daily death rate was 886. The 12 days before the peak saw 7,078 deaths at a daily average of 590. The 12 days following the peak saw 7,907 deaths at a daily average of 659 (11.7% higher).
Therefore, the available evidence would suggest that the death rate falls much less steeply than the infection rate. This lends some credence to the University of Washington model (although not the actual projections) which suggested that we will see more deaths after the peak than we experienced before it. In other words, the curve for new deaths has a long tail.
Of course, the health warning that I need to add is that we are not bound to follow the experiences of Italy and Spain. Thus far, however, we have followed Italy's trajectory quite closely and have actually been tracking above them for a couple of weeks.