0234 GMT September 29, 2020
At the start of August, around half a million tests were being carried out each day across India on a week's average, according to the international comparison site, Our World in Data, BBC reported.
Daily figures released by the Indian government are slightly higher than this.
This is a large number but should be put in the context of the size of India's population.
It carries out around 36 tests each day for every 100,000 people. In comparison, the figure for South Africa is 69, for Pakistan it's eight, and for the UK it's 192.
Modi's ambition is to double this number to achieve a million tests each day for a country with a population of more than 1.3 billion.
While boosting testing is regarded as a key part of the battle against the coronavirus, it's the type of testing which experts say is causing concern.
The one that's been most commonly used globally is a PCR (polymerase chain reaction) test, which isolates genetic material from a swab sample.
Chemicals are used to remove proteins and fats from the genetic material, and the sample is put through machine analysis.
These are regarded as the gold standard of testing, but they're the most expensive in India and take up to eight hours to process the samples. To produce a result may take up to a day, depending on the time taken to transport samples to labs.
In order to increase its testing capacity, the Indian authorities have been switching over to a cheaper and quicker method called a rapid antigen test, more globally known as diagnostic or rapid tests.
These isolate proteins called antigens that are unique to the virus, and can give a result in 15 to 20 minutes.
But these tests are less reliable, with an accuracy rate in some cases as low as 50 percent, and were originally meant to be used in virus hotspots and healthcare settings.
It is worth noting that these tests only tell you if you are currently infected and are different from antibody tests that tell you if you were infected in the past.
India's top medical research body, the Indian Council of Medical Research (ICMR), has approved the use of three antigen tests developed in South Korea, India and Belgium.
But one of these was independently evaluated by the ICMR and the All Indian Institute of Medical Sciences (AIIMS), which found that their accuracy in giving a true negative result ranged between 50 percent and 84 percent.
"The antigen test will miss more than half of truly infected cases," Professor K Srinath Reddy, of the Public Health Foundation of India said.
This can be for various reasons like the swab sample wasn't good enough, the viral load in the person or even the quality of the testing kit.
The ICMR had issued guidelines saying those with negative results from an antigen test should also get a PCR test if they show symptoms, to rule out a false negative.
Rapid or diagnostic tests may or may not use antigens in detecting the virus.
In the UK, the most common type of rapid test has an error margin of 20 percent for giving false negative results.
But the rapid test kits developed by Oxford Nanopore are said to pick up 98 percent positive cases, although that needs independent checking by researchers and health experts.
Both these rapid tests use genetic material, not antigens, hence are more reliable.
The World Health Organization (WHO) and the US Food and Drugs administration have also advised getting a PCR test if you test negative in a rapid test.
The United States is vying to develop such diagnostic kits you could buy at a store, swab your nose or saliva and get the results within minutes, like pregnancy test kits.
But the FDA guidelines for approval of such kits say that their performance has to be nearly as good as lab tests.
The US is already using antigen test kits by BD and Quidel which have a sensitivity of 71 percent and 81 percent respectively, higher than those used in India.
Many Indian states, which decide their own testing protocols, have been increasingly turning to the rapid antigen test.
ICMR announced on August 4 that up to 30 percent of the total tests conducted in the country were antigen tests.
Delhi was the first state to begin antigen-based testing in June, and many other states followed suit. It began using them on June 18, although there is no data publicly available until June 29.
We've looked at data from June 29 to July 28, which shows Delhi conducted a total of 587,590 tests, of which 63 percent were antigen tests.
But the available data shows that less than one percent of those who tested negative in an antigen test went on to have a PCR test, and 18 percent of those who did tested positive.
The recorded infection rate in the capital has fallen in recent weeks, but experts suggest that could be because many cases have been missed.
The authorities have now asked testing centers to conduct more PCR tests.
But data shows that more than 50 percent of the tests conducted are still antigen tests, despite the Delhi High Court's order that it should be used only in hotspots and healthcare settings.