1/10 Sweden week 49 winter season update. In week 40-45 (as far as we have reliable data) the seasonal increase in mortality remains the slowest on record. However, preliminary data suggest w46 will break the pattern by becoming 4th least deadly on record.

https://twitter.com/EffectsFacts/status/1332616970336952320?s=20

2/10 Introducing a new diagram to provide the full picture of cumulative deaths. Now that all individual weeks in 2020 no longer rank as the least deadly, it would otherwise become restrictively difficult to compare overall mortality for each season as we add more weeks.

The thread on mortality methodology – a vaccine against the common pitfalls of bad analysis
(~30 tweets in three subsections for each of these statements):
1. Why we start counting mortality in week 40
2. Why we correct for demography
3. Why we measure all-cause mortality 1.a Week40+asymmetric seasonality: The severity of respiratory illness follows a distinct seasonal pattern with peak infection rates and mortality in winter that is so notorious that the colloquial term “flu season” is used by the public and CDCs worldwide.cdc.gov/flu/about/seas…

0/17 Sweden. New series of weekly updated tables to track whether the increase in mortality this winter is in line with respiratory illness seasonality, or something extraordinary. So far (week 40-44) mortality has increased unusually slowly this winter. First some background: 1/17 When cases began to climb a few weeks ago it was a textbook exhibit of respiratory illness seasonality. I was surprised that many acted as if this was surprising. Others seemed a bit too eager at the prospects of redeeming their outsized mortality forecasts from spring.

1/21 Sweden. This respiratory illness season's age adjusted all-cause mortality was significantly lower than both 2016/17 and 2017/18. Will our myopic focus on one cause of death optimize public health outcomes? Why isn’t all-cause mortality the metric we strive to minimize? 2/21 Crude pop. adj. deaths this season never caught up with 2016/17 levels. It did so only briefly compared to 2017/18 (week 21) but returned to a deficiency by week 36. By week 41 cumulative deaths 2019/20 were approximately 500 and 1,500 below the previous seasons.

1/10 Sweden - cases, deaths & testing. A 2nd wave of PCR testing has created a 2nd wave of bad analysis of Swedish data. Influential commentators have eagerly picked up the increase in cases that has followed a ~3x increase in testing. It is time to revive some old diagrams. 2/10 Sweden. Firstly, 20 weeks have passed since the 1st surge in cases with no subsequent increase in deaths. There was a smaller bump in cases 11 weeks ago, no increase in deaths. Cases have increases for 6 weeks straight with no noticeable effect on deaths – how can this be?

(1/19) #covid19 Sweden. Respiratory illness IFR varies greatly with age. This season the demography-adjusted all cause mortality has been historically low in all available age brackets. A reminder that demography is a crucial factor in susceptibility that shouldn’t be overlooked. (2/19) We will explore two types of diagrams for each age bracket. The first type is straightforward. In the second one you would expect the yellow lines to follow the blue lines if mortality in that age bracket remained constant (and the grey columns to remain close to 0).

(1/12) Big thread with 16 diagrams to investigate the “dry tinder” hypothesis (mild preceding seasons affecting future mortality). We begin with a deep dive into the Nordics – in short: the data does not seem to refute the hypothesis (preview in pictures). (2/12) I provide comments in the pictures to keep the thread manageable. The diagrams are organized by country. Although I generally advise against inter-country comparison, in this case I would suggest looking at the same diagrams for different countries side by side too.

(1/9) New thread and diagrams on cases/deaths in Sweden. I will keep this updated weekly to alert us if cases suddenly transform into a leading indicator of deaths. So far, the Swedish experience strongly suggest that they do not (once there is community spread). #Covid19Sweden (2/9) A casual observer might be worried that the case count in Sweden has more than doubled over the last 5 weeks. But rather than a surge in mortality, deaths over the same 5-week period are down by 80%.

(1/9) The decision to close schools is one of the most disruptive and potentially inefficient measure that has been attempted to counter Covid19. @NahasNewman and many others have worked tirelessly to protect the rights of children against fact-resistant calls to isolate them. (2/9) Sweden didn’t close schools. If this policy was detrimental to their health, this should manifest in Swedish Covid19 statistics. What does the facts tell us? So far, out of 2.4 million people, one person below the age of 20 has died with Covid19 (0,000042% mortality).