Thread by @drStuartGilmour

Will the UK soon see a tidal wave of COVID-related hospital admissions? Is Omicron really causing a drop in hospitalizations? Let’s look at the data.

I’m spending the afternoon watching the year end mixed martial arts, #rizin33, so it’s as good a time as any to examine the new surge in cases in the UK. There’s a lot of talk that omicron is less severe and hospitalizations won’t rise, but I’m not convinced. Let’s look.

The Guardian reports 12000 people in hospital in the UK, with past peak of ~35,000. Early reports of a UK government study found a lower risk of hospitalization and suggestions “it won’t be as bad as last time” even as cases are sky-rocketing.

I’m suspicious about the idea that the hospitalization rate is going down with omicron. Let’s look at new cases and new hospitalizations. This figure shows the number of new cases, along with new hospitalizations (scaled by a factor of 10) since January 2021. Notice anything?

It looks like hospitalizations decoupled from cases in ~July, and as cases rose after “Freedom day” daily hospitalizations remained stable. I checked this by plotting the rate of admissions on day X per 1000 cases on day X-7 [assuming ~7 day delay from case report to admission]

This is the effect of vaccination and the surge in cases in children (who have lower risk of hospitalization) under the UK govt’s reckless back to school policy. You can see the clear shift in increased hospitalization rate after freedom day followed by the downward trend.

We only have a few weeks’ data on omicron, but it looks like hospitalization rates were dropping anyway. So I built a model of hospitalization as a function of past cases, with a term for omicron from 20th December, when it became dominant.

This model found an ~0.5% a day reduction in hospitalization rate since September, and an ~30% reduction in hospitalizations after omicron became established in mid-December. There was no significant effect of omicron on the downward trend.

Using this model I calculated predicted cases for the rest of the year based on the number of cases by specimen date. Here is the whole prediction since September. It’s not perfect but the trend is clear – daily admissions will rise to 2000 a day in the new year.

It looks like the UK will reach its past hospitalization peak by mid-January. If omicron hospital stays are shorter then the growth in total cases in hospital will be slower than the last peak, but if cases don’t drop rapidly the pressures on the NHS are unavoidable.

Some caveats: we don’t have much data on the omicron period (only a month); hospitalizations depend a lot on the age structure of infections, and I don’t have this info; most of my predictions have been “not even wrong” and I did this on my NYE break while watching MMA.

Nonetheless, the trend is worrying, and even more worrying is the ridiculous predisposition of UK politicians *and* many of its leading public health and medical establishment for breezy optimism in the face of repeated failures to handle a lethal infectious disease.