You asked
1. Does the ONS model, used to calculate the percentage positives that you publish, have as one of its inputs data on the case numbers diagnosed and published by PHE?
2. Has the Dynamic Bayesian multi-level regression post-stratification (MRP) used to produce your modelled estimates remained constant throughout the pandemic? If so, why has data been altered retrospectively such that later publications have different results to earlier ones?
3. Please provide, for each fortnight that a report has been produced, the number of or percentage of positive swabs that had a Ct value less than or equal to 25 and more than 2 positive genes identified.
4. How many of the positive swab results have had successful full genome sequencing (what percentage of total swabs)?
5. What methods are used to identify possible false positive results and what is done to ensure these are minimised?
6. What proportion of participants who had antibody testing and who have tested positive on PCR had a subsequent positive antibody test?
7. Of all the participants, what percentage have ever had a positive PCR test? Of these what percentage had no symptoms at the time of their positive test?
8. Of all the participants, what percentage have tested positive on at least two genes with a Ct value of 25 or less? Of these what percentage had no symptoms at the time of their positive test?
9. What proportion of the households surveyed have never reported symptoms?
10. The number of people testing positive, the number of people swabbed and their symptomatic status for England for each report.
We said
Thank you for your request.
1. The model used for the ONS COVID-19 Infection Survey only uses data collected from the sample as part of the survey and does not use PHE data. For more information on our sample and data collection, please see our methodology article.
2. The underlying MRP model has not changed over time. However, the modelling uses data from a set period of time, smoothing estimates over the most recent period; this is why the estimate for a given period can change over time, as we use the most recently available data to re-estimate values. You should not look at the modelled estimates for specific days in previous weeks using the latest model. Our official estimates should be used to understand the positivity rate for a single point in time. This estimate is our most stable estimate and remains the same for a given week over time. The official estimates can be found on tab 1a in the respective datasets for England, Wales, Northern Ireland and Scotland, accompanying our latest bulletin.
3. You can find the percentage of COVID-19 positive cases by gene pattern and Ct value in the UK on tab 1a of the technical dataset accompanying our latest bulletin. We do not hold information on positive swabs with a Ct value less than 25, but more information on Cycle threshold (Ct) values can be found in this article by our academic partners.
4. We try to read all letters of the virus's genetic material for every positive nose and throat swab with sufficient virus to do so (Ct less than 30) (whole genome sequencing). Sequencing is not successful on all these samples, or only part of the genome is sequenced. This is especially so for the higher Ct values, which are common in ONS data as we often catch people early or late in infection when viral loads tend to be lower (and hence Ct values are higher). Where we successfully sequence over half of the genome, we use the sequence data to work out which virus is which type of variant. Of the 13,808 cases with Ct<30, we have successfully sequenced 6,641 (48%) of them. There are also recent sequences still being processed and so we cannot provide a conclusive final number.
5. Test specificity and test sensitivity are used to understand the accuracy of RT-PCR testing. Test specificity measures how often the test correctly identifies those who do not have the virus, and so a test with high specificity will not have many false-positive results. A study using data from the COVID-19 Infection Survey over the period 26 April -11 October by our academic partners on cycle thresholds showed an overall false-positive rate for the period was below 0.005%, i.e. there was at least 99.995% specificity. The false-positive rate would be expected to be approximately constant over time since it is either random or driven by external factors. More information on test specificity can be found in our methodology article and this article by our academic partners.
6. To provide information on the proportion of participants who had antibody testing and who have tested positive on a PCR test, who had a subsequent positive antibody test, we would need to create bespoke analysis. Under the Freedom of Information Act 2000, Public Authorities are not obligated to create information in order to respond to requests. We therefore consider this to be information not held. Such services would be subject to legal frameworks, disclosure controls, resources and agreement of costs. If you would like to request this bespoke dataset, please contact health.data@ons.gov.uk to discuss your enquiry further. Please note, there may be a charge for this work which would be subject to our charging policy.
7. You can find the most recent swab test results (latest 2 and 6 week periods) including the number of people testing positive for COVID-19 and number of people in the sample in England, Wales, Northern Ireland and Scotland on tab 1c of the respective dataset accompanying our latest bulletin. You can find our latest symptoms analysis of the percentage of people testing positive for COVID-19 by symptom (including no symptoms) on tabs 3a, 3b, and 3c in the dataset accompanying our latest characteristics article.
8. Please see answer to question 3 for data on Ct values. To provide information on the percentage of people with at least two genes with a Ct value of 25 or less who had no symptoms at the time of their positive test, we would need to create bespoke analysis. This is therefore considered information not held. If you would like to request this bespoke dataset, please contact health.data@ons.gov.uk to discuss your enquiry further. Please note, there may be a charge for this work which would be subject to our charging policy.
9. Please see answer to question 7 for our latest symptoms analysis. This analysis is conducted at participant level rather than household level.
10. In the spreadsheet accompanying our latest bulletin you can find the official estimates of the percentage of people testing positive in England on tab 1a and the number of people in the sample for the most recent two and six weeks on tab 1c. See our weekly bulletin series for previous datasets covering earlier periods.Please see answer to question 7 for our latest symptoms analysis.