In the last three weeks it has been impossible to miss the slow building of tension over the Omicron variant of the SARS-CoV-2 virus. The news from Europe is not good with regard to the rise in cases and also the severity of the disease. Every day, the national news has reports of case numbers that are slowly rising: 100 last week, 200 this week. So, is there any indication of the actual number of cases in India? Should one be worried about a virus which has infected about 200 people in a land of over 1.3 billion, and when many are vaccinated?
Cause of error
Let me attempt to put the numbers in perspective. The reported cases of identified Omicron infections come from a genomic surveillance which is mounted by the Indian SARS-CoV-2 Genomics Consortium (INSACOG) . Genome sequencing is complex, and only 38 laboratories across the country have the ability to sequence a virus which is as infectious as this. As a result, only a small fraction of infected individuals contribute virus samples for sequencing. The numbers reported by the media are the number out of this small sample which show infection by the Omicron variant. So, the media makes an error when it reports this as the number of cases in the country.
A calculation
How can we do better? Let us think in terms of fractions or percentages. The Union Ministry of Health and Family Welfare (MoHFW) reports that in December, India performed about 12 lakh RT-PCR tests a day, and about 0.5% of the test results were positive. So, in the first two weeks of December, about 1.7 crore tests were performed nationally, and about 80,000 people tested positive during this time. The number of genomes sequenced by INSACOG is perhaps 1% of these.
This means that of the 800 or so samples taken, about 200 tested positive by the end of the second week of December. If the sample of the virus genomes to be sequenced was drawn randomly from the newly infected, then one would be forced to conclude that there are about 20,000 Omicron positive cases in India.
However, all public health agencies around the world have proceeded on the assumption that the Omicron variant arose recently and is still being transported around the world. So, the sequencing effort has been biased towards international travellers.
This means that the incidence of Omicron infections would be somewhat smaller. Could it be that only about 2,000 people are infected, i.e. , about 2.5% of cases? If the numbers were really that low, then about 80% of the virus samples would have been taken from travellers. The remaining 20% of the samples is then likely to give no Omicron positive results at all. But we know from news reports that at least about 10 cases are from people without a history of recent international travel.
Just based on the numbers that we know from the media and from other public sources, we realise that more than 5% and definitely less than 25% of the cases seen in the first two weeks of December are due to the Omicron variant. The number of cases is then closer to being somewhere between 5,000 and 25,000 in this period. Now that the total number of cases per day is beginning to rise, the lower number has become less likely.
Improving the estimates
Of course, these are very rough numbers. The agencies which handle the data and the scientists who run statistical models would be able to refine these estimates immensely and narrow the range of uncertainty. If the number of genomes sequenced from infected travellers and others are separately given by INSACOG, and tagged by the date on which the sample was collected, it would be much easier for you and me to make these estimates. However, there might be concerns about medical privacy which prevent the Government and its agencies from making public such details about the data.
Editorial | Warning bells: On Omicron cases in India
One should also be wary of other mistakes that the intentional bias in sampling virus genomes could lead to. If international travellers arrive more often in Delhi and Mumbai, then could the bias in sampling wrongly lead us to underestimate the speed of the spread of Omicron in the rest of India?
The numbers will change every week. Is Omicron spreading faster than Delta, the variant which gave India its second wave? If yes, then week by week, the fraction of Omicron cases would increase, as it out-competes the Delta variant in infecting people. This has happened in other parts of the world, and it could happen here too.
I have indicated here the kind of logic that an interested mediaperson or a layman can use. If you make informed judgments about whether to invest your savings in fixed deposits or in shares, then you make more sophisticated numerical estimates quite regularly. Given the numbers made available by the MoHFW and INSACOG, it is possible for you to estimate your personal health risks from COVID-19, whether you stay at home or travel on work or on vacation.
Sourendu Gupta is a theoretical physicist at the Tata Institute of Fundamental Research Mumbai. The views expressed are personal
