6 ms·

The ONS regularly tests a random selection of people. Their results typically trail the daily infection numbers by two weeks. Notably, the most recent ONS samp

The ONS regularly tests a random selection of people. Their results typically trail the daily infection numbers by two weeks. Notably, the most recent ONS samples demonstrated a sharp rise in cases. The suggestion is that cases are actually surging, but most are now either asymptomatic or at least very mild, and thus not reported. I guess we'll find out shortly whether that's the case.

ddek 5y ago
Which ONS data are you referring to? The currently published data indicates the wave has at least plateaued, as I read it, in all but the youngest age groups.
- a_humean 5y ago
  They are talking about the survey data, which doesn't rely upon the normal testing regime. They randomly contact and test individuals irrespective of their symptoms status, which is different from the normal testing regime that gets reported daily. I believe its a 20k sample size. The survey takes a lot longer time compile Most recent summary: https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/bulletins/coronaviruscovid19infectionsurveypilot/latest https://www.ons.gov.uk/peoplepopulationandcommunity/healthan... Most recent data: https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/datasets/coronaviruscovid19infectionsurveydata https://www.ons.gov.uk/peoplepopulationandcommunity/healthan... Methods: https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/methodologies/covid19infectionsurveypilotmethodsandfurtherinformation https://www.ons.gov.uk/peoplepopulationandcommunity/healthan...
  - monooso 5y ago
    Thanks for adding the relevant links.
  - tomatocracy 5y ago
    In particular, it's worth noting that the ONS survey tries to estimate an answer to the question "in the given time period, what proportion of the population would have tested positive?" Because people tend to continue to test positive for a reasonable period of time after they are first infected, this should be a lagging indicator of any peak in new infections having occurred.
  - timr 5y ago
    It's important to note that this data sample ends on July 24, which is almost exactly the day that cases started rapidly declining across the UK. OP was not accurately representing the facts.
  - makomk 5y ago
    That summary seems to show infections plateauing in every age group except school kids. (And of course, with schools breaking up since then they're probably both less likely to get infected and less likely to get tested.)
Animats 5y ago
The US CDC does this by re-using random anonymous blood samples drawn for other purposes, and by using test data from blood banks. The current data collected is for the presence of antibodies indicating people who've been infected. Not those vaccinated; they can distinguish that. This is cheap to measure, but runs about 6 weeks behind. Data here.[1] US average from late June is about 22% of the population. That's roughly how many people actually had the disease. That data should show the Delta variant surge as the data comes in. I still can't find a useful number for people with antibodies prevalent because either infected or vaccinated, although that is apparently tested. [1] https://covid.cdc.gov/covid-data-tracker/#national-lab https://covid.cdc.gov/covid-data-tracker/#national-lab
IshKebab 5y ago
The ONS data doesn't show a sharp rise. If anything it is levelling off. I think tomatocracy is right - because the ONS survey measures current infections rather than newly discovered infections it is likely to lag a bit and will probably start to drop soon.

ddek 5y ago

Which ONS data are you referring to? The currently published data indicates the wave has at least plateaued, as I read it, in all but the youngest age groups.

a_humean 5y ago
They are talking about the survey data, which doesn't rely upon the normal testing regime. They randomly contact and test individuals irrespective of their symptoms status, which is different from the normal testing regime that gets reported daily. I believe its a 20k sample size. The survey takes a lot longer time compile Most recent summary: https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/bulletins/coronaviruscovid19infectionsurveypilot/latest https://www.ons.gov.uk/peoplepopulationandcommunity/healthan... Most recent data: https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/datasets/coronaviruscovid19infectionsurveydata https://www.ons.gov.uk/peoplepopulationandcommunity/healthan... Methods: https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/methodologies/covid19infectionsurveypilotmethodsandfurtherinformation https://www.ons.gov.uk/peoplepopulationandcommunity/healthan...
- monooso 5y ago
  Thanks for adding the relevant links.
- tomatocracy 5y ago
  In particular, it's worth noting that the ONS survey tries to estimate an answer to the question "in the given time period, what proportion of the population would have tested positive?" Because people tend to continue to test positive for a reasonable period of time after they are first infected, this should be a lagging indicator of any peak in new infections having occurred.
- timr 5y ago
  It's important to note that this data sample ends on July 24, which is almost exactly the day that cases started rapidly declining across the UK. OP was not accurately representing the facts.
- makomk 5y ago
  That summary seems to show infections plateauing in every age group except school kids. (And of course, with schools breaking up since then they're probably both less likely to get infected and less likely to get tested.)

Animats 5y ago

The US CDC does this by re-using random anonymous blood samples drawn for other purposes, and by using test data from blood banks. The current data collected is for the presence of antibodies indicating people who've been infected. Not those vaccinated; they can distinguish that. This is cheap to measure, but runs about 6 weeks behind. Data here.[1] US average from late June is about 22% of the population. That's roughly how many people actually had the disease. That data should show the Delta variant surge as the data comes in. I still can't find a useful number for people with antibodies prevalent because either infected or vaccinated, although that is apparently tested. [1] https://covid.cdc.gov/covid-data-tracker/#national-lab https://covid.cdc.gov/covid-data-tracker/#national-lab

IshKebab 5y ago

The ONS data doesn't show a sharp rise. If anything it is levelling off. I think tomatocracy is right - because the ONS survey measures current infections rather than newly discovered infections it is likely to lag a bit and will probably start to drop soon.