This Is Going to Hurt Me More…

After two weeks of insulin, Oscar had a follow-up to check his fructosamine levels. The good news was that the levels had fallen several hundred points, into a range the vet considers “healthy.” He’d also lost a bit of weight, at a rate the vet also considered healthy. Barring acute events or observable change in behavior, he’ll goes back in three months to test again.

So the 7am/7pm feeding and injection regime is “the new normal.”

Cats, it turns out, are much better at knowing when they’re hungry than they are at knowing what time it is – so Ivan generally starts demanding food a couple hours before it’s time to be fed.

That’s far less irritating in the afternoon than it is at 5:30 in the morning.

Dawnise is really good at ignoring him and either sleeping through it or falling back to sleep. I, on the other hand, am not. And the cats figured out long ago that a pretty sure fire way to get my attention (and get me out of bed) is to make destructive sounding noises in a room just out of sight.

So, yeah… Three weeks in and it’s already getting old.

In other news, the UK’s third COVID wave seems to have peaked at 60k daily reported cases (a ~12% positivity rate in England) in mid July. So far much lower than the estimates – some as high as 100k per day.

The latest eyebrow-raising news is that the NHS has only recently started gathering and reporting data that differentiates people being admitted to hospital with COVID from those being admitted to hospital due to COVID.

That data shows that over the past month about 1 in 4 admissions are of people who test positive, but for whom COVID symptoms are not the cause of admission.

Hospitals were measuring and reporting what they were asked to measure and report. And from the perspective of a hospital, counting patients with COVID is reasonable and important. COVID infected patients demand more resources, in the form of increased isolation, PPE, etc., than non-COVID infected patients. So whether a patient arrives with a broken leg and tests positive, or comes in with low blood oxygen clearly due to COVID, that patient places more demands on the hospital. Knowing how many patients are admitted with COVID is a key data point for understanding the capacity of the health care system.

But health care system capacity wasn’t the only way the data was being interpreted by the media, the public, and seemingly by the government. As the vaccination program suppressed the link between infection and acute illness, public attention turned from case counts to more focus on hospitalization and mortality rates. And this reveals that recently, the hospitalization numbers haven’t quite represented what people thought they did.

The data being gathered told how many people were in hospital with COVID – and using that data to reason about the risk of being hospitalized due to COVID isn’t straight forward, indeed may be impossible.

But it was happening all the same.

Why is this data only being gathered now? I have seen no good answers. The Telegraph quotes Professor Carl Heneghan, director of the Centre for Evidence-Based Medicine at the University of Oxford, as saying: “This data is incredibly important, and this is information we should have had a very long time ago. We have been crying out for it for nearly 18 months.” (emphasis added) Going on to say “the Government might have made very different decisions about restrictions if it had access to data which actually measured the situation accurately.”

Were the decisions wrong? Hard to say. But making decisions that affect many millions of lives based on data that doesn’t mean what you think it means certainly seems … not great.

It’s also hard to say if the difference between what we were measuring and what people thought we were measuring has been consistent over time. It’s tempting to naively retroactively apply the 25% “over estimation” to hospitalization numbers from January – indeed that same article quotes Heneghan as saying “at the peak of the pandemic in January, we were talking about close to 40,000 patients in hospital – this new data suggests that back then around 10,000 of them were primarily there for other reasons” – but that’s almost certainly wrong. Nothing else about this virus has been stable over time, this seems unlikely to be the exception.

To be clear, I am not a statistician – I don’t even play one on television – clearly neither is Boris Johnson, and based on the quote above I have my doubts about Professor Heneghan, too. Having said that, I’m pretty sure a better past estimate could be had using long term hospital admissions data. Fortunately, that data exists, hopefully someone’s working on that as I type.

It takes context and understanding to turn “data” into “information,” and again to turn “information” into “knowledge.” Decision making, certainly on a national scale, demands rigor at each step.

If there’s a moral here, I think it lies somewhere between “lies, damn lies, and statistics” and “never attribute to malice that which is adequately explained by stupidity.”

And there’s certainly been no shortage of that.

Leave a Reply