New research has used a mathematical formula to clarify how lateral flow devices are more effective than sometimes portrayed. Professor Iain Buchan, Executive Dean of Liverpool’s Institute of Population Health, explains the findings.
“Public health works really well when there is a very close partnership between service and science,” says Professor Iain Buchan, Chair in Public Health and Clinical Informatics and Executive Dean at the University of Liverpool. “The most important thing is to have great science networked at national and international levels and great partnerships to embed and provoke the science at local levels.
“We’ve been able to move very fast in Liverpool because of the particularly close relationship between science and [the public health] service,” he says, explaining that a long tradition of action-research for dealing with public health challenges in a deprived area prepared Liverpool to deliver world-first COVID-19 insights, such as how mass rapid-antigen testing works “on the ground”.
Leading the university’s Institute of Population Health, Iain’s work focuses on data-intensive public health research for “major societal challenges”, most recently responses to COVID-19. Moving from clinical medicine to public health in 1996, Iain has a long history of “turning people’s numbers into actionable information, really trying to close that data–action gap”. He says: “I saw a lot of underused data – I was one of the first people to describe the child obesity epidemic, using routinely collected health visitor data.”
The latest research he has co-authored has found that lateral flow devices (LFDs) are more accurate than previously reported, demonstrating their reliability as a public health tool in reducing the spread of the virus. Working with researchers from University College London, Harvard University and the University of Bath, the research investigated validation studies of LFDs and showed the problem with reporting sensitivity values (the proportion of people with the infection who test positive) of LFDs relative to polymerase chain reaction (PCR) tests.
The study used a new formula for calibrating the sensitivity of LFDs, which showed they are likely more than 80% sensitive at picking up people who are infectious with the COVID-19 virus – i.e. a public health test for infectiousness and not a clinical test of for those who have been infected, but who may no longer be infectious, as is the case with PCR tests.
Different properties
Previous research into the effectiveness of LFDs has compared them with PCR, suggesting the latter is the “gold standard” to show if someone is positive or negative for COVID-19. But this research highlights that LFDs work in a very different way to PCRs – they are not equivalent.
LFDs detect material from the surface proteins of the virus and are very likely to give a positive result when someone is infectious. PCR tests detect the virus’ genetic material, which can be present for weeks after a person has stopped being infectious. In previous research, LFD sensitivity was being evaluated by ability to identify the same cases that PCRs picked up. If someone’s LFD is negative but their PCR is positive this is because they are not at peak transmissible stage.
“It’s important to… recognise that the two tests reflect different properties of the infection and have different testing utilities – clinical versus public health,” the study reads. Iain adds that the results of the study weren’t a surprise because it used a mathematical way of synthesising the differences between a public health test of infectiousness and a clinical test of having been infected.
“We had done a lot of initial research last November [2020], looking at the real-world performance of rapid antigen testing,” Iain says, adding that the latest research findings were an “unsurprising projection of those results onto the virus’ usual time course in a person’s body”. The research also makes an important distinction between viral load and viral shedding, where more research needs to be done, he says.
Using the data
The study’s authors say that it’s important health professionals and the public have clear information about the operating characteristics of the tests. “Criticisms of LFD for apparent low sensitivity have failed to take the viral biology and epidemiology into account,” the study reads.
“This has confused policymaking and damaged trust in LFDs, despite the need for better tools to control transmission of SARS-CoV-2. It is our hope that recalibrated absolute sensitivity statistics will assist policymaking and help build public confidence in LFD as a tool to aid COVID-19 resilience and recovery.”
The University of Liverpool has collaborated with the NHS and local government to put together a data legacy system for ongoing COVID-19 responses and recovery, called CIPHA (www.cipha.nhs.uk), that can be used quickly to understand the real-world, system-wide impacts of technologies such as diagnostic tests, symptom monitoring devices, new care pathways and artificial intelligence.
“This is a computational way of gathering a bigger, system-wide picture of what happens when you put a technology into a real-world setting, and it helps to target those resources for the best outcomes,” Iain says. “For the first time, we’ve got updated data every 30 minutes from COVID public health systems combined with NHS and local authority social care. That allows different parts of a service to coordinate their actions in near real-time.”
LFD Fact file
- Lateral flow devices were first rolled out in Liverpool in November 2020, as part of a pilot to test everyone living or working in the city regardless of whether they had COVID-19 symptoms.
- Around a third of COVID-19 transmissions are from people before they start displaying classic symptoms.
- A pilot study of asymptomatic testing in Liverpool found the sensitivity of the LFD was estimated to be 40% and the specificity relative to PCR (the proportion of people without the infection who have a negative test) thought to be about 99.9%.
- NHS Test and Trace figures show that over 6 million LFDs were conducted on 2 to 8 September 2021 – a 52.2% increase from the previous week. Over 1.4 million
- PCR tests were conducted, an 18.9% increase from the previous week.
- For self-reported LFD positives that were matched to a confirmatory PCR, 90% were confirmed positive by a subsequent PCR.
- For assisted LFD positives that were matched to a confirmatory PCR, 79% were confirmed positive by a subsequent PCR.
Future studies hope to improve the calibration further using series of daily repeated PCR and LFDs among substantial cohorts drawn from the general population. All of this strengthens the links between public health services responsible for communication, how people can access a test without digital means and many other factors, and clinical research. “Reverse-engineering practical actions [of research] helps close the data–action gap, and that provokes really good science,” Iain concludes.
Image credit | Alamy