The 4 nations update their stats on COVID-19 patients in hospital at a slightly different pace.

If you take the most recent available data:
England: 1,048 (19 Sep)
NI: 21 (17 Sep)
Scotland: 61 (18 Sep)
Wales: 68 (18 Sep)

That's a total of 1,198.
Summary figure is just 1,081.

In other words, if you're looking only at the summary page, you're already 10% too low by virtue of peering into the rear-view mirror.

Same problem for new patients admitted...

Detailed data:
England: 199 (17 Sep)
NI: 0 (17 Sep)
Scotland: 3 (7 Sep)
Wales: 42 (17 Sep)

Total of detail: 244
On the summary page: 134

That's a *much* bigger gap.

Why does this happen?

Well, the summary page takes its data from the last date for which full 4-nation data was available, whenever that happened to be.

So for hospital patients, 17 September. For admissions, 7 September.

This means that, in a rapidly rising second spike, the view that most people (and most journalists - sadly) rely on can be quite far behind the best available data, just because one or other of the 4 nations hasn't contributed a recent update.

Result: a false sense of security.

NOTE: underlying data comes from...
https://coronavirus.data.gov.uk/healthcare