Association between Covid-19 Test, Test Positivity and Death

A M Zakir Hussain

18th September, 2020 11:18:04 printer

Association between Covid-19 Test, Test Positivity and Death

How far is it true that, the more the tests for Covid-19, the more will be the number of cases identified? We argue that the relationship is not linear or not ‘dose-response’ dependent. The number of positive cases logically, should be dependent on the degree of infection in the community and much more in fact, on the pathogen’s infectiousness.

Association between number of tests and number of (positive) cases

The graph below shows that the weekly average tests done (expressed in per cent against the total tests) and the per cent of test positivity (the daily rate) are not commensurate. For example, 12.79% cases were identified against 0.8% tests done on the 6th week of reporting the first case in Bangladesh. On the 8th week, 2.0% tests identified 11.51% cases. From 7th to 11th weeks the number of tests and the number of cases were relatively high and corresponded reasonably well. But on the 12th week the number of tests fell below that of the 11th week, while the number of cases rose higher than the previous weeks. Between the 13th and17th weeks the number of tests kept rising compared to previous months and so did the number of cases but not point for point. For example on the 16th and 17th weeks, the number of tests was higher than the 15th, but not the number of cases. From the 17th to 22nd weeks, while the number of tests was falling gradually, the cases were rising, to the extent that the highest numbers of cases were identified when the tests were the least, in the entire series (24.01% test positivity against 4.0% of the total tests done). Tests rose between the 22nd to the 26th weeks above those of the 19th to the 22nd weeks, but the cases started falling gradually below the level between the 12th and 24th weeks.

The table below also testifies the fact that just increasing the number of tests did not really end-up in higher number of cases straightaway. Evidence is more cogent that it was the months, rather than the number of tests which were more deterministic in spawning the positive cases, with some exceptional hiccups - higher tests and higher numbers of positive tests. By and large the fact is attested however, that lower number of tests may give higher number of positive cases.

From the 18th to 26th weeks, close to the same number of tests identified varied numbers of cases, e.g. 88,209 tests found 20,973 Covid-19 cases in the 19th week; 83,806 tests found 19,112 instead of 19,926 (if calculated based on 19th week’s result) in the 20th week, while 92,096 tests in the 23rd week found out 19,412 cases, and 93,960 tests found 16,300 in the 25th week. Should these not be treated as a function of time rather than the number of tests?

Association between number of tests and number of deaths

The figures and the graph above do not show any firm linear relationship between the per cent of weekly tests done on average and the per cent of weekly deaths on average against total deaths. From 7th to 9th weeks the tests increased while the deaths decreased. In the 12th week, the number of tests fell but not the number of deaths. In the 16th week the number of death fell but not the tests. Tests started falling from the 17th week, but deaths did not plummet, except in the 22nd week. From 23rd to 26th week, tests remained more or less equal but the deaths fell in the 26th week, which could be due to weaker virulence of SARS-CoV-2, since no change in the protective behaviour in people was noticed.

Implication of increasing number of tests

If the number of tests is to be increased at all, why not increase it to three lakhs per day (twice that of India, proportionate to the population of Bangladesh), which will give then a test positivity of 5% or lower (which will mean Covid-19 is under leash, as per WHO’s suggestion). At this rate, 30,000 tests will identify 15,000 Covid-19 cases per day. Now, these 15,000 positive people will need isolation and 10 to 30 times of their contacts- 150,000 to 450,000, will have to be quarantined, as per WHO recommendation, per day. Does the nation have the capacity to manage, finance and monitor them every day, over a period of 14 days for each day’s catch- compounded by days after days of identified new cases?

The formidable question, however, will be - will all these endeavours arrest further spread of infection? It did not in India in four months, although India started its heroic endeavour of testing more than a million people per day since May. Will the people remaining outside the remit of daily reach stop on strutting, and roaming, and spreading the infection? Is it not possible that those who tested negative earlier might get infected later and be spreaders? Will those requested to obey self-quarantine abide by? Will those who are in good health, despite being positive, confine themselves to the remits of the four walls of their homes? They are active, robust, mobile, young and fun loving people! We need to remind ourselves that contact tracing did not work in the past in Bangladesh, because people lied and they lie or they forget to who they came in contact with in so many places they frequented to. Another important point to be considered is the number of people, who despite getting infected will be found negative due to the limitation of the diagnostic processes, which when an antigen based testing is used may run up to 65% (as found in India). These false negative cases will spread the disease gleefully, without any such intention though! These ground realities have to be considered by those who urge for increasing the number of tests.

The writer is a former director, Primary Health Care and Disease Control, former director of IEDCR, DGHS, former regional adviser of SEARO, WHO and former staff consultant, Asian Development Bank, Bangladesh.