Everybody's favourite brutally hard interviewing British journalist Andrew Neil made this interesting tweet in the midst of the COVID-19 pandemic.

I don’t pretend to know what point Andrew Neil was making. I suspect he was making the frequently made point that COVID-19 is harmful mainly to people who have pre-existing conditions, and that criticisms of the UK government’s response to COVID-19 overlook that deaths from COVID-19 are of patients who die of something else, rather than from COVID-19. He has though stumbled upon an ancient dilemma in clinical coding. That of how did someone die.

Outside violence, people don’t tend to die of one thing. Take my favourite go to cancer, multiple myeloma for example. Two common causes of death are:

1. The patient’s immune system fails, they become susceptible to pneumonia, which they can’t fight and they die of pneumonia.
2. The burden on their kidneys of excreting Bence-Jones proteins irreparably damages their kidneys.

So, if someone who has multiple myeloma develops one of these, and they die, what do they die of? Do they die of multiple myeloma, or do they die or pneumonia/kidney failure? And since multiple myeloma is a progressive disease, if their disease progresses to plasma cell leukaemia do they die of that instead of myeloma? And if they were previously diagnosed with plasmacytoma do they die of that?

You then get other problems. As mentioned previously, COVID-19 appears to be particularly harmful to people who have pre-existing conditions. So if someone with diabetes catches COVID-19, did the diabetes kill them, as that was their underlying condition that made the COVID-19 harmful, or did the COVID-19 kill them because their diabetes was well managed at the time they caught it? As Andrew Neil notes, it’s not straightforward.

How people die is a lot more complicated than they died from being old/cancer/COVID-19/AIDS. People generally develop lots of conditions as they age, and lots of conditions have interplay with one another to contribute to a patient’s ill health. Does it really matter though?

To medical practice, yes it does. If we know that people with multiple myeloma often die of kidney failure, we can monitor a patient’s kidney function, and intervene early to reduce the risk of mortality from this cause. We can also thinking about how likely patient populations are to require particular clinical services based of complex health data. In the context of a new disease like COVID-19, we can also develop strategies to assign resources appropriately. We may be able to stratify patients into low risk patients who need limited intervention, high risk patients who need aggressive intervention, and those patients whose pre-existing conditions and infection with COVID-19 significantly reduce their chances of survival such that palliative care may be more appropriate.

But does it matter to everyone else? I don’t think it does, infection with COVID-19 is still associated with mortality in some people, so we shouldn’t be spreading it. Whether COVID-19 was responsible entirely for a person’s death or it was a final disease that killed an already ill patient seems like a fairly mute point. Not having COVID-19 is better than having it and we should avoid spreading it.

Though, perhaps don’t panic buy toilet paper, or indeed panic.