A lot of it is the orthodox approach shining through again. In many ways this comes down to the long-term UK medical establishment approach towards testing in general, as well as testing at different stages of epidemics, pandemics and observing the seasonal Influenza-like-illness picture every year.
Theres no emphasis on, expectation of or capacity for mass testing. Instead they go for various surveillance systems, testing a small subset of people within communities and extrapolating the wider picture. Vigorous testing of every possible case is normally reserved for the very early phase only, where limited numbers apply and there is specific data they are trying to obtain. Later testing goes back to the sample-based approach, with some exceptions along certain lines, including some clinical need ones.
I believe that related orthodox thinking also shows up in some other healthcare phenomenon in this country. There are a bunch of common conditions that GPs dont seem terribly interested in testing for upon initial presentation, and from what I've experienced it seems the first approach is usually to have a look at the patient and pick the thing it is most likely to be, rather than testing to see what it actually is. A different sort of numbers game, one where you deal with the most common possibility without much fuss or testing resources. But where some of your patients will end up with the wrong treatment and some frustrating follow-ups and lost time trying to get far enough along the path of possibilities to actually stumble on the right cause, or if all else fails then very well, finally test. And I'm sure many people who have dealt with the passing of a frail relative might have experienced the cause of death being recorded in terms of a combination of their existing health conditions, and something like pneumonia. Perhaps some pathogens that actually caused the pneumonia were formally ruled in or out by testing, perhaps not. The pathogens actually responsible certainly cant be relied upon to always make it onto the official cause of death of that person.
Obviously there are areas of exception to what I've said about testing and routine healthcare. Some pathogens and diseases and patient conditions evoke a very different approach. I'm sure there are some areas that have just the sort of testing protocols we would ideally like to see universally. That they are missing from this pandemic so far in the UK is bad but sadly not surprising.
To compensate for the fact that the system wont properly identify and record every case of everything, the data from cases they did directly detect and data from sampling/surveillance schemes tends to be combined with the excess mortality statistics, in order to estimate the number of deaths that a particular epidemic has caused.
So yeah, thats my opinion of the nature of some relevant parts of the orthodoxy. Mostly the same themes I've gone on about before. And plenty of areas where we can clearly see the ugly and awkward contrast between how they are used to conducting these matters, and what this pandemic calls for.
In this particular context, given their testing capacity limits right now, they could at least publish a separate 'suspected deaths' or 'possible community deaths' figure.