Dear Sir David
Monitored Lateral Flow Tests (trice during 8-19 March 2021) for asymptomatic secondary pupils on their return to school
In respect of the above policy, England’s secondary pupils, parents, head-teachers, statisticians, parliamentarians, journalists and public have a right to expect:
- openness about relevant prior data held by NHS Test & Trace & disclosure of DfE/PHE’s prior planning assumptions;
- explanation for denying PCR-adjudication of asymptomatic pupils’ LFT-positives as part of properly informed consent for and from head-teachers, parents and secondary pupils;
- bespoke data-collection to enable rapid reporting of uptake & results of 1st LFTs (separately from 2nd LFTs and 3rd LFTs) & statistical thinking in action.
The Royal Statistical Society COVID-19 Taskforce issued its statement on schools on 5 March 2021, please see https://rss.org.uk/RSS/media/File-library/News/2021/RSS-statement-on-surveillance-in-schools-5-March-2021.pdf
With asymptomatic infection prevalence low in mid-March 2021 (around 2 per 1000) and a likely sensitivity of INNOVA Lateral Flow Test (LFT) in asymptomatic secondary pupils of around 40% (based on mainly adults in Liverpool) but specificity of 998 per 1000, the Taskforce did not anticipate that DfE/PHE would deny pupils PCR-adjudication of their LFT-positive results. Hence, when interviewed by Nick Robinson on TODAY programme at 07.21hrs on Saturday, 6 March, I was shocked: to learn on-air that a pupil’s negative PCR-adjudication was being ignored, the child had to continue in self-isolation for a further 8-days as did his parents and sibling.
The RSS Taskforce statement asked for disclosure of DfE/PHE’s prior planning assumptions and includes an illustration for use in visualizing their impact in the context of asymptomatic INNOVA LFT-screening of secondary pupils on return to school. In addition to planning assumptions for uptake of 1st, 2nd and 3rd LFT-test, other still undisclosed prior planning assumptions relate to
i) Prevalence of SARS-CoV-2 infection in asymptomatic secondary pupils
ii) Sensitivity of INNOVA LFT for asymptomatic infected secondary pupils
iii) Specificity of INNOVA LFT for asymptomatic uninfected secondary pupils.
All three cannot sensibly be collapsed as “99.6% accuracy”.
Please seek to bring these three prior planning assumptions immediately into the public domain, together with the date of their agreement – presumably before 8 March 2021.
The RSS Taskforce also asked for openness about relevant prior data held by NHS Test & Trace, such as the ct-value distribution for asymptomatic secondary pupils whose LFT-positive was PCR-confirmed. NHS Test & Trace also holds information on the percentage that was PCR-negative of PCR-confirmations in respect of asymptomatic university students whose INNOVA screening test was LFT-positive prior to their return home for Christmas. Thirdly, NHS Test & Trace holds information on the percentage that was PCR-negative of PCR-confirmations in respect of secondary pupils whose INNOVA screening test was LFT-positive in January 2021.
Please seek to bring these relevant prior data held by NHS Test & Trace into the public domain.
Robust explanation for denying PCR-adjudication of asymptomatic secondary pupils’ LFT-positives in England (not in Scotland) is needed since asymptomatic hauliers’ LFT-positives are subject to PCR-confirmation. Difference appear ageist and inconsistent unless informed by the relevant prior data held by NHS Test & Trace. The statistical thinking behind any explanation needs to be scrutinized against DfE/PHE’s prior planning assumptions. Professor Chris Whitty avoided being drawn on the issue when questioned on 9 March by Greg Clark, chairman of the House of Commons Science and Technology Select Committee, who – along with other parliamentarians – had been receiving questions from constituents, please see transcript online. The Committee is following up on the matter.
Please seek to bring the explanation for denying PCR-adjudication for secondary pupils’ LFT-positive screening-result immediately into the public domain.
Properly informed consent from head-teachers, parents and pupils requires frankness about all three of the above missing parts from the performance monitoring jigsaw.
The data-systems that NHS Test & Trace relies upon, and struggles with, are oft ill-designed to answer pertinent questions about infection-control; and ill-support performance monitoring for testing initiatives. Consider, for example, how many months it has taken for sub-optimal adoption of the July 2020 recommendations by the Royal Statistical Society COVID-19 Taskforce on how to glean greater intelligence from Test, Trace and Isolate about asymptomatic infections in those self-isolating and their adherence to quarantine, please see https://rss.org.uk/RSS/media/File-library/Policy/RSS-COVID-19-Task-Force-Statement-on-TTI-final.pdf. Too often, the need for quality assurance is used as an excuse – including to House of Commons Select Committee on Science and Technology – when experimental statistics could, and should, be put into the public domain in a timely fashion.
Please seek to end the refrain of quality assurance as a bar to bringing timely experimental statistics into the public domain. Frankness about unresolved problems posed by the need for fuzzy matching across data-systems is essential.
The pity is that NHS Test & Trace cannot identify 1st LFT-positives by secondary pupils on their return to school (8-19 March); and has not yet de-duplicated secondary pupils’ serial LFTs prior to March 2021. Statistical thinking ahead should have anticipated the difficulty now faced because analytical teams have faced similar LFT-related problems before. Statistical thinking in action would have designed a minimally sufficient system for data-collection – each evening from secondary schools – to enable rapid reporting across England of uptake & results of 1st LFTs (separately from 2nd LFTs and 3rd LFTs).
Please encourage official statisticians to consider that efficient data-acquisition remains a key part of statistical science and to make their voice heard in how performance monitoring of infection control and other policies is designed1. Late evidence is wasted evidence.
Let me end by congratulating colleagues: at NHS Test & Trace and at DHSC on the earnest efforts they do make to work with the data-systems that confront them. Hence, effort is being made to follow up my suggestion that – as a matter of urgency (if not already done . . . ) – analysts at NHS T&T do their best to link timely PCR-confirmations for LFT-positives by secondary-age children during 8-12 March [using alternative definitions for timely]; and to publish the number and percent PCR-negative on Monday 22 March to mark the end of the 3-monitored LFTs; with a corresponding analysis presented on 29 March for secondary pupils’ LFT-positives in 15-19 March.
Due to design-failures in the March monitoring of England’s secondary-school-LFT policy, analysts may have to resort to reporting PCR-adjudications for secondary-age children (eg 12-16 years) rather than all secondary-pupils on return. But their unbiased best is underway to “recover” at least some information – despite DfE/PHE’s avowed intent to ignore PCR-negatives – on the performance of monitored INNOVA-LFTs for asymptomatic screening of secondary pupils on their return to school in mid-March.
The issues raised in this letter may recur. For example, secondary schools could be asked to re-instigate monitored LFTs (thrice) when secondary pupils return after Easter. Hence, sorting both PCR-adjudications and data-acquisition may be pressing.
One of your predecessors as UKSA-chair famously described a knife-crime press-release as egregious. The current failure to monitor properly DfE’s policy of LFTs thrice on secondary pupils’ return to school is a statistical cause celebre in its own right.
By contrast, the 2021 Census was a pleasure as well as an obligation to complete.
Kind regards.
Yours sincerely,
Sheila M. Bird
Cc: Ed Humpherson, Director of the Office for Statistics Regulation
1. Royal Statistical Society Working Party on Performance Monitoring in the Public Services (chair: Bird SM). Performance indicators: good, bad, and ugly. Journal of the Royal Statistical Society, Series A 2005; 168: 1 – 27.
Related Links:
Ed Humpherson response to Sheila Bird: Transparency around asymptomatic testing in secondary schools