[dekhn]

I wouldn't trust a stat like that without digging into the underlying causation. The whole twitter presentation is a nice narrative but I see many places where it could just be outright wrong, or saying misleading things (I work in medical biology and have plenty of familiarity with how hospitals run their ERs).

[notahacker]

I agree that this sort of thing needs a study.

Fortunately the previous tweet linked to the original source, a published study of 5 million ER visits which appears to have a full set of controls and specifically measures time to inpatient transfer in excess of 5 hours as a mortality risk factor https://emj.bmj.com/content/emermed/39/3/168.full.pdf

[dekhn]

A retrospective, observational (IE, after the fact data analysis from a population without controls) study found a risk factor (association). That's not enough to make any real conclusions as to whether this is causal.

[notahacker]

It may not be conclusive on the subject of whether the 16% additional deaths were specifically influenced by the additional wait, but in light of that data point, you're going to need a whole lot more evidence to justify your original claim that the cohort that dies 16% more than the baseline population in the days that follow were, without exception, not in any urgent need of medical care...

[dekhn]

All I said was: if you can wait 12 hours, it's not an emergency. The emergency room is for things where a person couldn't wait 1 hour without dying. People dying 30 days later: not an ER problem.

[notahacker]

Nope, the emergency department of an NHS hospital is absolutely not reserved for things where people couldn't wait one hour without dying (or people who are in danger of death period).

And since we're talking specifically about the subset of NHS emergency department attendees who are more likely to be dead within 30 days if it takes longer than the NHS' target of four hours for them to receive an inpatient bed, I'm pretty unconvinced that the alternative route of booking a doctor's appointment in the next few days to see if he'll refer in a week or two is the more appropriate one...

[dekhn]

I know NHS works differently from the US, but the health care providers in the US that are used by major corps (IE, benefits) all have urgent care lines. I've done this repeatedly (to avoid my wife going to the ER with a sick kid). Typically, while it takes a month to get a routine appointment, the kid's doctor always can see them the next day for something urgent.

If you think about it, this makes a lot of sense because ERs are tuned for low latency, low throughput, high cost for emergencies, and if lots of people who just need urgent care go, they make it hard for the emergencies to be seen with low latency.