Can you deny emergency admission for cardiac failure or an accident victim for want of confirmation of a coronavirus infection? Do you really need two negative tests to be discharged from an isolation ward?
The accepted way to test for the presence of the novel coronavirus is in human nasopharyngeal samples using the real-time reverse transcriptase polymerase chain reaction (rRT-PCR).
The rRT-PCR test’s sensitivity depends on several factors, including method of sample collection, site of collection, time into disease, viral load, sample storage, cross-contamination, etc. Limiting factors include availability of testing kits, reagents, PCR equipment, trained personnel, protective gear for each collection and cost (Rs 4,500 per test).
So as such, the rRT-PCR is not entirely feasible in a country like India where people confirmed to have the infection have to be tested once every 48-72 hours and twice within 24 hours for those expecting to be discharged. For the latter, a nasal sample that tests negative could test positive in the sputum. Likewise a positive PCR test doesn’t always mean viruses in the body are still replicating. For the latter, labs need to undertake cumbersome viral viability tests.
Pooled testing is a cost-effective option in settings with a case prevalence of less than 5%. In India, the states of Uttar Pradesh, Haryana and the Andaman & Nicobar Islands have recently started doing this.
Similarly, antibody tests with good quality rapid diagnostic kits (RDKs) and enzyme-linked immunosorbent assays are cheaper, require less labour to perform and are less skill-intensive than PCR tests. They can be used to determine when staff and contacts can return to work, identify infected but asymptomatic persons in the community and locate new clusters of infection. But on the flip side, their diagnostic sensitivity varies according to days after the onset of symptoms.
So for diagnosis and care, it’s important to combine each patient’s test results with their history and clinical picture.
Further, swab samples were sent only to the National Institute of Virology, Pune, for tests until March 17, when the government qualified 72 more labs. And until March 20, India had conducted only 10,000 tests, and they began to climb sharply after. Testing has since increased sevenfold, although the share of positive cases has hovered at 3.5-4.1% of samples tested.
Now, our hypothesis is that expanding the number of tests of the general population during the lockdown – and early in the disease transmission stage – will not increase the number of suspected patients we can discover relative to the current sentinel surveillance system. This is simply because people with the coronavirus are not hiding out there now; they will become infected in future as community transmission speeds up.
Data from South India already indicates this is the case. As on April 17,
– Karnataka had conducted 182 tests per million, with 2.3% testing positive
– Andhra Pradesh had conducted 217 tests per million, with 4.5% testing positive
– Tamil Nadu had conducted 285 tests per million, with 5.6% testing positive
– Kerala had conducted 483 tests per million, with 2.3% testing positive
Now, India has attracted some criticism for not testing enough, especially after the WHO publicised its mantra to “test, test, test”. When critics quote the ‘tests per million population’ indicator, the denominator for India is 1.3 billion, its total population. However, 364 out of 733 districts in the country aren’t affected, so their people should be subtracted from the denominator. India also can’t be compared to smaller countries like South Korea, Singapore and Italy. India is vast, and exposure rates in Kerala, Hyderabad and Mumbai different strongly from the entire Himalayan and Northeastern regions.
So our testing rates will go up when the unaffected population is deducted from the denominator. Indeed, India’s tests per diagnosed case – 25.9 – is better than those of Japan (11.4), Italy (7.6), the US (5.3) and the UK (3.3).
As on April 7, 136 government labs and 56 private labs were performing tests. Recently, the owner of a large chain of diagnostic labs in India said during a discussion on TV that he wanted the government to allow private labs to undertake large-scale community testing. However, he didn’t accompany his request with an epidemiological logic, and it’s hard to see what it could be.
Undertaking large-scale tests in the community is only bound to be a waste of resources. Instead, the government should go for contact-tracing and to check people living near infection clusters. As things stand, the sentinel surveillance system is adequate for containment measures and to indicate area-specific searches, as Kerala has demonstrated.
The Indian Council of Medical Research (ICMR) allocated 100,000 RDKs to Kerala for disease surveillance and to identify hotspots. Instead of testing the general population, the Kerala government identified four priority groups. Group one was those healthcare workers who had served patients with COVID-19, who were given about 10,000 kits, and who had served only other patients, about 15,000. The second group received 25,000 kits: 20,000 for government staff with public contact, like police personnel, ASHAs, anganwadi workers and panchayat and municipality staff, and 5,000 kits for workers of community kitchens, food and grocery deliverers, ration shop vendors, etc. The third group, of all people quarantined at home, also received 25,000 kits. The fourth group, of all senior citizens, received 20,000 kits.
Since this exercise however, the ICMR has rolled back the use of antibody tests due to low quality RDKs, and so rRT-PCRs remain the only accepted way to test patients for COVID-19. However, this is not an entirely desirable situation given rRT-PCR’s many limiting factors, so other states can follow Kerala’s lead to identify some risk groups and allocate tests for them.
It also bears noting that testing isn’t the only way to reduce mortality; better case management is. That is, public health officials should concentrate on containment measures in known hotspots and better manage admitted patients to reduce mortality. They should also prevent hospital-acquired infections with better supply and use of personal protective equipment; practice personal hygiene, disinfect fomites, etc.; and include tests to rule out other common infectious causes.
In all, to prevent short- and long-term collateral damage, along with our efforts to care for people with COVID-19, essential care services for people with other communicable and non communicable diseases shouldn’t be interrupted.
Dr K.R. Antony is a consultant of public health, child survival and development in Kochi. He has served with UNICEF and as the director of Chhattisgarh’s State Health Resource Centre, and is now an independent monitor for the National Health Mission.
Dr Vasundhara Rangaswamy is a clinical microbiologist with international experience and a public health activist in Baroda.
The views expressed here are the authors’ own.