A Closer Look At Respiratory Death (part 3)

Lessons from an undisclosed NHS Trust

In the first part of this mini-series I mentioned the rather awkward fact that an adult in-hospital death whose diagnostic record included both an acute respiratory diagnosis and a positive COVID-19 test result may not necessarily have died directly from respiratory failure attributed to SARS-COV-2. There’s sepsis to consider, plus organ failure and cardiac troubles as well as neurological issues including stroke, not to mention iatrogenic causes. Whilst COVID-19 may well contribute to some or all of these - especially through induction of a cytokine storm - or even be the genuine underlying cause of death I am hoping members are appreciating just how muddy the water is when it comes to assigning death to a single underlying cause.

A clue to the difficulty is given in the acronym MUSE (multicausal and unicausal selection engine), this being the software that ONS use to automatically process the thousands of death certificates registered each week. Only in the case of inquests and post-mortems are the MUSE algorithms over-ridden. Multicausal death is most certainly a ‘thing’ yet the authorities present data as if death is unicausal when it comes to COVID. If we are being honest about this we should be counting deaths where COVID was indeed a genuine underlying and sole cause, where COVID was a contributory cause, where COVID was a co-morbidity to other primary causes, and where SARS-COV-2 was falsely detected.

By way of getting the ball rolling on this I am going to present a slide for all adult deaths identified as positive COVID cases with at least one acute respiratory diagnosis (symptomatic COVID death) that reveals the distribution of additional and therefore potentially causal conditions other than SARS-COV-2 infection. In cases where SARS-COV-2 infection led to development of full blown COVID-19 symptoms ending with an acute respiratory condition such as respiratory arrest which was the one and only primary cause of death then there would be no other diagnoses made in the patient record. If other diagnoses were indeed made at the time of death then we’d have the insane problem of ascribing COVID as sole cause in a potentially multicausal situation.

Just to underline just how tricky all this really is let us imagine a situation where an emergency COVID admission with a deteriorating acute respiratory condition is wheeled through to ICU where they suffer respiratory arrest. The clinical team pull the patient through this episode but they die on a general ward three days later from liver failure owing to the pharmacological load. Is this a COVID death? In a way it is because COVID put them in a risky situation, but in a way it is not – their liver actually killed them, and that liver might have been functioning poorly because of alcoholism, disease or even prescription drugs. Let’s get the crayons out and see just how awkward the situation on the ground really is…

Now this is a sizzling surprise! We discover that none of the 1,574 symptomatic COVID deaths can be attributed to SARS-COV-2 infection alone, which dunks us right into the murk tank of multicausal death. Announcing that all 1,574 symptomatic COVID deaths are multicausal at this point is a bit strong since we need to pick over the fine detail of the patient record – we have no idea whether any of these additional diagnoses are potentially causal (e.g. stroke) or a comorbidity (e.g. high blood pressure). Nevertheless, I hope this slide reveals just how murky the COVID waters are, for 67.4% of symptomatic COVID deaths exhibited patient records with three or more additional diagnoses and a staggering 91.2% of symptomatic COVID deaths exhibited patient records with two or more additional diagnoses.

Just how reliable are the COVID death stats that are produced by the Office for National Statistics on a weekly basis after they have mauled death certification using an automated processing system like MUSE that is going to run a bulldozer through these medical intricacies? Let us remind ourselves once again of what the WHO desire…

"With reference to section 4.2.3 of volume 2 of ICD-10, the purpose of mortality classification (coding) is to produce the most useful cause of death statistics possible. Thus, whether a sequence is listed as ‘rejected’ or ‘accepted’ may reflect interests of importance for public health rather than what is acceptable from a purely medical point of view. Therefore, always apply these instructions, whether they can be considered medically correct or not."