Sorry old graph above for NW England. Replaced here.
Regarding the '31 March issue', looking at it regionally, see the attached graph for E of England, there's a steep rise 23 March to 28 March, which seems plausible enough, and then the data for 29/30 March seem heavily inflated, then 31 March is down to 25 March levels, 1 April still that low, and then from 2 April something of a plateau.
For London it's more straightforward as a huge dip on 31 March (only), likewise the Midlands, for the NE & Yorkshire a ludicrous dip again only on 1 day. For the NW the pattern is 33,35,60,60,30,82,80,81, so again a big dip but also odd patterns around that day.
For the SE 34,25,62,63,36,59,53, so similarly to the NW and E of England there's a big boost in deaths for the 29th and 30th, a big fall on the 31st and then resumption in service after that.
SW England is a couple of weeks behind the rest of England and the number of deaths at this point were very low, but still a drop on the 31st of March.
Spreadsheet here:
gofile.io/?c=tShowr