The ONS data

I recently had a really interesting conversation with a colleague who spent years in public health, including assembling death statistics. What an eye opener.

The speed and fluency with which he was able to disentangle the statistics was truly impressive. Despite the best efforts of an increasingly prescriptive training system, it appears that doctors still just go off the traditional training pathways and explore the world as it interests them. We are all the richer for it. I hope to do even a small amount of justice to a couple of their arguments. Also interesting just how dispassionately folks who talk about such a large number of deaths can be.

The ONS data is a couple of weeks behind due to the time taken to report deaths, but it well worth a look as it compares this year’s deaths against the 5 year average for a similar time of the year.

The first point made was that there are about 10-12,000 deaths each week in the UK, with seasonal variation due to winter cold and viruses. Death is a good endpoint because reporting is effectively universal and does not require any subjective opinion in itself, although the cause of death is up to the interpretation of the attending doctor. After initially being trepidatious in reporting COVID as the cause, many doctors are apparently more comfortable to cite COVID even where testing is negative.

The second point made was that the data show a clear increase in the death rate this year over the 5 year average. There is no denying from this data when the epidemic hit the UK. The ONS helpfully shows the official COVID-19 deaths as well, so it is easy to see that there are excess deaths over and above those attributed to the virus. The view of the doctor I spoke with is that the timing of this is such that these apparently non-COVID deaths are in fact COVID and not folks dying of other diseases because they couldn’t get to the hospital.

The third point made to me — not shown on this data but well appreciated by those who work from these datasets — was that there is often a spike in excess deaths about ten days or so after the first cold snap of winter. This is then followed by a reduction in deaths that basically matches the excess. In other words, the first cold snap of the winter tends to bring forward deaths by a few months but these are frail people who were likely to die anyway.

There is one school of thought that at least some of the initial deaths represent very frail patients from care homes – some of whom may never have made it into the hospital. It is possible that the combination of evolving herd immunity plus the tragic early deaths of the most vulnerable citizens may tend to attenuate any second or third waves.

In the meantime, the number of non-COVID patients continues to rise. The lockdown is increasingly in name only (LINO?).