As far as I can see, the UK is running out of time to avoid reaching a situation like Italy's or Spain's. Some data and analyses below. Apologies for any mistakes and please challenge or correct any claims. Data source: github.com/CSSEGISandData… #Covid19uk [1/11]
UK cases continue to rise at an approx. exponential rate, with a 25-30% daily increase in cases, doubling every ~2.9 days (faster rise of deaths). This is nearly identical (3.0 days) to the Italian growth at the same stage of the epidemic. [2/11] 
Like Italy, Spain or France, the UK will need to go into lockdown in a matter of days. But a few days will make all the difference. As shown below, the UK epidemic appears to be only 14-15 days behind Italy (not 3 or 4 weeks, as it has been unfortunately claimed). [3/11]
Italy closed schools on March 4th, 15 days ago. It had 3,090 cases and 107 deaths. The UK is already lagging one day behind Italy on this measure. Italy also implemented local lockdowns early on, which the UK has not used. [4/11]
Italy went into lockdown on March 9th, 10 days ago. Given the similar growth rate and the lag in the two epidemics, delaying the UK lockdown until Monday will probably push the UK into Italy’s dramatic situation. [5/11]
Surely we can learn from Italy’s and Spain’s suffering and avoid something entirely predictable:
[6/11]
[6/11]
According to the references in the table below, the UK also has fewer ICU beds per capita (6.6/100k) than Italy (12.5) or Spain (9.7). If these figures are correct, a lockdown may be needed even earlier to be close to Italy’s situation. [7/11]
en.wikipedia.org/wiki/List_of_c…
en.wikipedia.org/wiki/List_of_c…
A lockdown is an extraordinary measure with great costs, but it will need to happen within a few days. The choice, however, between doing it today or in 5 days has enormous implications. In 5 days the epidemic may be 3-4 times larger than today (1.29^5). [8/11]
The impact of a delay of a few days appears so vast in terms of lives and the economy that the government should share the modelling behind their decision making. I hope a delay is not simply motivated by models like the one below, which is conceptually very interesting… [9/11]
…but hardly reflective of the extraordinary time we are in. We can learn about other countries’ suffering, we do not need to experience it! [10/11]
bmcpublichealth.biomedcentral.com/articles/10.11…
bmcpublichealth.biomedcentral.com/articles/10.11…
The window for the UK appears to be closing very fast. Since a lockdown will happen, the best time is NOW. Stay safe everyone... and again, apologies for any mistakes. #UKlockdown #StaySafeStayHome [end]
Apologies for the wrong label. The figure on the right represents deaths...
Updated figures from yesterday here. You can see how predictable the numbers are: (Today's update coming soon)
As promised, here is the updated figure with data from yesterday. You can see how the UK epidemic is closely following Italy's, with a 14-15 day lag and the same doubling rate. We need a national lockdown by Monday just to be as badly hit as Italy. Time is running out.#UKlockdown
The same plot with a log axis. You can now see the drama that will unfold in the UK the next two weeks. I have no words. We need a lockdown by Monday just to match Italy's situation.
[Sunday update] Here is today's plot. The trend continues unabated. The UK epidemic is growing at the same rate as Italy’s 14-15 days ago, doubling in size every 2.9 days. We need a UK lockdown by tomorrow just to be as badly hit as Italy is now.
[Sunday update] And here is today's plot (same data), represented in log scale, with the latest data point being Saturday's. Two days before Italy went into national lockdown, they registered 233 deaths. That is exactly the same number of deaths that the UK registered yesterday.
[Monday update] Spain went into strict quarantine on March 14, the day they hit 196 deaths. France did it on March 17, with 175 deaths. YESTERDAY the UK registered 281 deaths. In my opinion, the UK has now lost its opportunity to avoid Italy’s fate.
[Monday update] Deaths in Italy and Spain today reflect infections that occurred over two weeks ago, before their lockdowns. Yesterday Italy registered 651 deaths in one day, Spain 391, and growing. This is Spain yesterday… rte.ie/news/2020/0322…
[Monday update] The UK epidemic is doubling in size every ~3 days, the same rate as Italy at the same stage of their epidemic, lagging just 14-15 days behind. The UK epidemic is now likely ~4-5 times larger than it was last Monday. ~4-5 times larger than it ever needed to be.
[Monday update] This will likely translate into many thousands of preventable deaths. A much larger epidemic will also be harder to control and will require a longer and economically more disruptive lockdown. In weeks to come, we will look back at last week with immense regret.
[Monday update] Every few days of delay the epidemic is doubling in size. We need a UK lockdown now. #UKLockdownNow
Well, the UK is now officially in full lockdown. A very sad day in many ways, but an unavoidable step at this stage. Other countries should learn a lesson for which we will pay a very high price (thread above). Please stay safe everyone and good luck.
An accessible account of what the next few weeks and months may hold: medium.com/@tomaspueyo/co…
[10 days on] Since my original tweet (see thread above), UK deaths have increased 10-fold and diagnosed cases (a vast underestimate of true cases) by >6-fold. We are now slightly below Italy 15 days ago, but the growth has been mostly predictable.
[10 days on] Same plot in logarithmic scale. Latest data suggests that deaths in the UK are doubling every ~3 days (similar to Italy at the same stage). They reflect the growth of the epidemic in the UK ~2-4 weeks ago.
[10 days on] Diagnosed cases are doubling every ~4 days, more slowly than deaths. Difficult to interpret due to low testing, but possibly revealing some effect of social distancing ~1-2 weeks ago. Still, largely following Italy’s growth.
[10 days on] A lockdown a week earlier could have made the UK epidemic ~4 times smaller than it will end up being, likely preventing thousands of deaths and reducing economic disruption. A lesson for countries still on time to react.
[10 days on] In the UK, a lot of damage has already been done, but much of the future is still in our hands. We need stricter measures, better PPE, more compliance (please avoid leaving the house if not essential) and building capacity for large-scale testing and contact tracing.
[14 days later] 2 weeks after the original post (thread above), we can now study the accuracy of the original predictions. Sadly, the UK has largely continued tracking Italy’s epidemic with a 14-15 day lag and similar doubling rates.
[14 days later] Since the UK went into lockdown 14 days after Italy, we may see a roughly similar evolution in the next few weeks. The strictness of the lockdowns and our compliance with them will probably determine the longer-term differences.
[14 days later] Cases offer a distorted picture due to limited testing, but have the advantage over death statistics that they inform about the state of the epidemic ~1-2 weeks ago, rather than 3-4 weeks ago.
[14 days later] With caution, given the uncertainty in diagnosed cases, the epidemic appears to have grown at ~17% per day in the UK (similar to Italy 14 days before) the week before the lockdown, despite the calls for social distancing.
[14 days later] This suggests that a lockdown was very important. It also suggests that a lockdown one week earlier could have led to an epidemic ~3 times smaller (likely translating into many thousands fewer deaths and the need for a shorter and less disruptive lockdown).
Given the attention since @carolecadwalla retweet, I should say that this was a simplistic analysis of the (imperfect and insufficient) data available. Expert epidemiologists should quantify this more accurately. Some modelling from true experts below.
