Read on Twitter
NY updated its seroprevalence study on May 2. It's the largest (15.1k subjects) in a highly impacted area and so likely the most robust due to Ab lest limitations.
Unfortunately it revised DOWN covid prevalence for NYC.
Now 19.9 % (21.2% April 23)
Implications for IFR...
1/n
Unfortunately it revised DOWN covid prevalence for NYC.
Now 19.9 % (21.2% April 23)
Implications for IFR...
1/n
Based on only PCR positive deaths from NYC health dept upto 1 May: IFR= 0.77%
Add in deaths with COVID on death certificates from NYC health dept upto 1 May: IFR= 1.07%
(This accounts for the total excess mortality seen in NYC so most accurate).
2/n
Add in deaths with COVID on death certificates from NYC health dept upto 1 May: IFR= 1.07%
(This accounts for the total excess mortality seen in NYC so most accurate).
2/n
Updated IFR by age:
0-17: 0.003%
18-44: 0.1%
45-64: 1.0%
65-74: 4.3%
75+ : 10.3%
3/n
0-17: 0.003%
18-44: 0.1%
45-64: 1.0%
65-74: 4.3%
75+ : 10.3%
3/n
This NYC IFR is higher than recent sero-studies in Europe suggest e.g. Luxembourg 0.83%. This may be methodological due to subject selection or Ab test reliability better at higher prevalence. Could be due to demographic attack rate. Not just age but also ethnicity.
4/n
4/n
ONS multivariate analysis shows even when adjusting for age and socioeconomic grouping higher mortality risk exists for certain ethnic groups (up to 2x).
NYC is more diverse than Luxembourg or Geneva. As is the UK.
5/n
NYC is more diverse than Luxembourg or Geneva. As is the UK.
5/n
NYC is also much more unequal. Higher multi comorbidity and poor health status seen in lower socioeconomic groups in more unequal societies will nudge up IFR.
Likewise to a lesser degree for GBR compared to continental Europe. Backed by ICNARC data for UK ICU deaths.
6/n
Likewise to a lesser degree for GBR compared to continental Europe. Backed by ICNARC data for UK ICU deaths.
6/n
But nothing skews IFR like age.
Here I use NYC IFR by age and ONS data for all registered COVID deaths to April 24 for England and Wales to show attack rate.
Far from shielding we funnelled COVID at the old by allowing carehomes and hospitals to be centres of spread.
7/n
Here I use NYC IFR by age and ONS data for all registered COVID deaths to April 24 for England and Wales to show attack rate.
Far from shielding we funnelled COVID at the old by allowing carehomes and hospitals to be centres of spread.
7/n
While NYC total IFR ~1% and attack rate is more evenly distributed, in E&W due to demographic attack profile total IFR is running at 2.7%
May be underestimate. A high and increasing proportion of E&W excess deaths not registered as COVID are in care homes (?undercounted).
8/n
May be underestimate. A high and increasing proportion of E&W excess deaths not registered as COVID are in care homes (?undercounted).
8/n
However, just because we have an IFR running so high does not mean that UK lock down must go in present form.
Risk to total under 45 remains <0.1%. All it means is that the Gov really needs to sort out carehomes and hospital transmission ASAP.
END
Risk to total under 45 remains <0.1%. All it means is that the Gov really needs to sort out carehomes and hospital transmission ASAP.
END
