ICMR must allow testing of asymptomatics, Delhi must back up antigen test with RT-PCR to catch false negatives
The rapid antigen test (RAT) for Covid-19 has a significant time-advantage over the RT-PCR test that is the gold standard; it generates results in 30 minutes, at the point-of-care, while the latter takes a minimum of 2-5 hours and requires laboratory settings. Given its high specificity (no-to-low false positives), it makes eminent sense to roll it out to ramp up testing. Indeed, the Delhi government has told the Delhi High Court that of the 15,012 samples collected on June 18, 7,040 were tested through rapid antigen test while 7,972 were tested through RT-PCR. On June 21, 13,345 samples were collected, of which 9,356 were tested through RAT and 3,989 through RT-PCR. However, as Dr K Srinath Reddy of the Public Health Foundation of India has pointed out in these pages (bit.ly/37SPS4I), RAT has a low sensitivity (all negative results may not be true negatives), and thus a reliance on RAT alone could lull authorities and those testing negative into thinking they are true negatives. This is why, and Dr Reddy has highlighted this, ICMR recommends in its advisory dated June 14 that “suspected individuals who test negative for Covid-19 by rapid antigen test should be definitely tested sequentially by RT-PCR to rule out infection”. More so, given gold-standard RT-PCR itself has been reported to have a sensitivity of around 70%, dependent on factors such as how the sample was collected and the viral load in the sample, and RAT has a sensitivity rate that is nearly 60% of what RT-PCR delivers. If the Delhi numbers mean that the government is increasing RAT solely—the low RT-PCR numbers despite the high RAT negatives seems to indicate this—to expand testing, the national capital is making itself comfortably numb to the possibility that many cases could be escaping the radar; it isn’t very hard to imagine what that means for future infection numbers.
Dr Reddy also speaks of the flipside of having multiple tests in the arsenal, especially given one test is to be used as a confirmatory for the other in case of negatives; it would be an administrative nightmare to keep track of new tests, re-tests, confirmatory tests run on the same sample and map it to infection numbers. If that problem can be taken care of, say, with test data segregation and analysis with the help of artificial intelligence and machine-learning—ICMR had recently partnered IBM’s Watson for its testing data—it could make for a meaningful roll-out. It would also most certainly help if the ICMR were to ease its testing criteria—the ICMR has just revised its testing strategy to allow testing of all symptomatics, but asymptomatics, except for those meeting ICMR-specified criteria are still excluded. Given the sensitivity levels of RT-PCR, Covid-19 diagnosis will still need clinical observation, X-rays and CT reports, in many cases. But, with second waves of infection being reported in Singapore, South Korea and even New Zealand that had declared zero cases recently, India’s testing strategy has to be wider than the present one. At the same time, sero-surveys, like the one ICMR conducted in May, need to be conducted periodically till the world has a pharmacological intervention that works, to which the most vulnerable classes have access.