Earlier quoted context omitted.
There are situations where that statistic/categorization is meaningful, but not in this context of trying to rank different causative diseases and public health risks. For example, imagine that 99% of all the euthanized are people who chose it after a diagnosis of the "Insane Agony Virus." Obviously we would want all those mortal data-points to get counted as reasons to fight IAV, even if only a tiny number of people…
Because secondary effects are impossible to untangle or verify. Maybe IAV is just a made up term that people use, or far more people believe that they have IAV than is actually the case, or that they are wrongly convinced that IAV is incurable or untreatable, or the medical system has been perversely incentivized to encourage medically assisted suicide in the case of IAV diagnosis, or it's actually the case that it's…
The correct breakdown depends on goals. If someone is stabbed with a knife, the important cause(s) of death won't be "because a lot of cells couldn't synthesize ATP anymore", and probably not " saying the wrong thing to Stabby McStabberson."