A fact-check of Simon Thornley's appearance on @NZQandA yesterday. Thornley begins by citing the "Ioannidis study" showing a #Covid19 death rate "only marginally higher than the standard flu viruses". I believe this is the study he's referring to. (1/16) medrxiv.org/content/10.110…
In the paper, Ioannidis conducts a meta-analysis of antibody surveys to determine how widespread Covid-19 is in different places and then compares the deaths associated with those places to arrive at an estimate of the IFR (see explainer) of 0.24%. (2/16) who.int/news-room/comm…
Part of the problem is many of these antibody surveys - including Ioannidis' own - have been criticised for using inaccurate tests and applying estimates to entire populations from too-small sample sizes. More on these surveys here, from May. (3/16) newsroom.co.nz/is-asymptomati…
It's worth noting that the 0.24% figure is well below the CDC's IFR estimate (0.65%) and the WHO's (0.5-1%). But is Thornley right that 0.24% is "only marginally higher" than the IFR of the flu? No - seasonal flu IFR is about 0.04% per below. (4/16)
That means, even with Ioannidis' watered down IFR, #Covid19 still kills six times as many infected people as the flu. If the CDC is right, then it's 11 times more deadly. If the WHO's upper bound projection is right, then it's 25 times worse. Hardly marginal. (5/16)
Next fact-check - Thornley compares the WHO's early case fatality rate projection of 3.4% with current IFR projections. This is apples and oranges as the below thread demonstrates - it simply makes no sense. By the way, CFR today is still 3.3%. (6/16)
.@JackTame rightly points out that Thornley is wrong to point to empty hospitals during Level 4 as evidence of our ability to scale up to beat the virus. NZ still hasn't had a massive outbreak that threatened our health system capacity - 40,000 cases, per @PunahaMatatini. (7/16)
Thornley cites an analysis of the impact of lockdowns on #Covid19 spread. While he's right that it didn't show an impact on death or case numbers, it did show an impact on recovery rates. This below is from the conclusion of the paper. (8/16) 
Of course, there are more factors at play than just "did country x lockdown, yes or no?" The timing is important -NZ has always locked down with relatively few cases compared to the UK or US locking down after weeks of community transmission. (9/16)
Plus, as Tame points out, it's weird to cite an analysis of how lockdowns went in other countries when we *know* it worked in New Zealand as evidenced by spending two months at Level 1 with more freedoms than almost anywhere else in the world. (10/16)
Thornley argues against a strategy of "ongoing lockdowns to control the virus". He'll be glad to know that's not New Zealand's strategy! More below. (11/16) newsroom.co.nz/eliminating-co…
Not a fact-check, but it's worth noting from a journo perspective that Thornley never answered how he would protect vulnerable people. Here's who he'd have to protect. (12/16)
How are we going to have a functioning health system and economy if we have to lock away for months 1.1 million over 60s, 250,000 people with type 2 diabetes, 700,000 asthmatics, 1.5 million current and former smokers, etc.? Probably an important q to answer.... (13/16)
Thornley says we should follow Iceland's strategy for restrictions. But since May 14, a majority of New Zealanders have been living freer lives than people in Iceland, including through the present outbreak where two thirds of us have remained at L2. (14/16)
Thornley is wrong in claiming European mortality data shows #Covid19 deaths are those who we'd have expected to die. Excess mortality data shows 200,000 more Europeans have died so far than would be expected to, including 187,000 people over 65. (15/16) euromomo.eu/graphs-and-map…
Finally, Thornley says there's been little debate over #Covid19 strategy in NZ. That's an odd claim to be making on live television, after writing an op-ed for Newsroom the week prior and amidst a concerted effort by columnists to back the Plan B path, but 🤷. (16/16)
