Summary:

Group screening takes samples from multiple people, mixes them together, and tests them as one. People within a group that test positive are individually retested, or an algorithm is used to identify positive individuals.

Pros: Fewer tests to process; more people tested; lower cost; detection of asymptomatic cases.

Cons: It is not efficient when the infection rate is high.

Context

As COVID-19 continues to spread and infect, the UK has experienced problems with testing capacity, whether from limited staffing, equipment or chemical reagents.

The Centre for Disease Control’s most recent estimate is that 40% of people with COVID-19 may be asymptomatic. Because the number of tests that can be processed every day are limited, current UK testing, including Northern Ireland, is focused on people with classic COVID-19 symptoms: fever, a new continuous cough and/or loss of taste or smell. Consequently, people without these symptoms are advised not to get tested, but they may still be infectious. Presymptomatic people are also infectious. 

The absence of an approved vaccine, along with reinfections and uncertainty about immunity also indicate that mass testing for COVID-19 will be needed, at least in the short term.

How does group screening work?

Currently, the main method of COVID-19 testing is a PCR test. This detects active viral infections from a swab taken from a person’s nose and throat. FactCheckNI recently addressed some inaccurate claims made about PCR tests. (While current group screening focuses on PCR tests of nasal and throat swabs, research into pooled saliva samples has been suggested by scientists in the UK and Ireland.)

Group screening—also referred to as pooled testing—takes swabs from multiple people, mixes them together and tests them as one.

This method was first proposed by Robert Dorfman during World War II to test thousands of soldiers for syphilis. It has since been used successfully in other contexts to detect chlamydia and gonorrhoea, HIV and West Nile virus.

As Smriti Mallapaty outlines in Nature journal, four methods for COVID-19 group testing are currently being practised or trialled.

Method 1:

A number of swab samples are tested together. If the test comes back negative, everyone in this group is ruled out. If the test comes back positive, each member of the group is retested individually to pinpoint the positive case(s).

If a workplace had five teams of 20 people, then a mixed group sample could be tested for each team. If tests came back negative for four teams, then 80 people would  have tested negative. If the fifth team’s group test was positive, then all 20 people in just that team would need to be retested to find the positive case(s). A total of 25 tests would be required rather than the 100 necessary without group testing.

This method works best when infection numbers are low, as most groups come back negative and do not need a retest. If too many groups come back positive, they need individual retests anyway so time is wasted and there is little saving. It works best when the prevalence of the disease in a community is around 1%. Group screening is not effective when the COVID-19 positivity rate is over 10%.

Group screening allows mass testing at a reduced cost. How much can be saved depends on the number of infections. American researchers estimated that even in the hardest hit states, with 2% prevalence of COVID-19, at least 75% of costs could be saved with group testing.

Another benefit of group screening is to identify people who have mild or no symptoms who would not otherwise have been tested. This is therefore useful in the early identification of COVID-19 hotspots.

The ideal group size depends on prevalence of COVID-19. The lower the prevalence of the virus, the larger the group can be. Real world pools have ranged from five to 50. An NHS England report in September 2020 recommended groups of 6-12.

The FDA has cautioned that samples may be diluted in group tests which could make the virus harder to detect. However, false negatives are also a problem with individual tests. The key point is that in areas of high COVID-19 prevalence, group size should be reduced to increase test sensitivity. 

Method 2:

This follows the same approach as method 1, except any groups who test positive in the first round of testing are then divided into subgroups.

Using the example above, the fifth team that tested positive could be divided into four subgroups of five people. This requires four more tests. If three subgroups test negative, they are ruled out. The other group of five will have individual retests. Overall this could result in 14 rather than 100 tests.

This method uses less tests than method 1, but the disadvantage is that it can be slow, as it takes several hours to get a result for each group test.

Method 3:

This method uses one retest and an algorithm for positive groups.

The first round of group tests are completed as per method 1. For groups that test positive, individuals are retested and their samples are combined into a number of overlapping groups. 

By testing different combinations of samples, an algorithm can identify the positive individuals, without a need for the third round of retesting used in method 2.

Method 4: 

This method uses one round of tests. Tests are conducted on many combinations of overlapping groups. Each person’s sample is tested a number of times as part of a number of different groups. This forms a matrix, from which positive cases can be isolated using an algorithm. 

This method requires more tests to be done at the beginning (to have enough overlapping groups), however it happens within one testing cycle and saves the time and expense of retests. 

One limitation is that it is labour intensive. Manoj Gopalkrishnan, a computer scientist at the Indian Institute of Technology Bombay in Mumbai, thinks it will be most effective if conducted robotically rather than by technicians.

Has group screening been used for COVID-19?

In May, method 1 of group screening was used to test 2.3 million people in two weeks in Wuhan in China. They used five samples per group, and 56 infected people were identified.

Group tests have been approved by the FDA in the USA, and a small number of programmes are now underway.
Method 1 group screening has been especially useful in low resource settings, for example in Ghana.

Method 3 has been used in Rwanda to test groups of up to 50 people, analysing second round retests using a ‘hypercube’ algorithm. The Rwandan research team estimates that their method may be cutting the cost of testing from US$9 per person to 75 cents. A similar approach has been used in Uruguay, allowing mass screening with limited resources.

In September 2020, some UK universities began to use method 1 group screening. For example, Cambridge University launched a pooled asymptomatic screening programme for students living in college accommodation.