A higher proportion of the population is dying from the coronavirus disease (Covid-19) in states and union territories that test poorly despite a widespread outbreak, and thus have a higher positivity rate, according to data analysed by HT.
This would mean that adequate testing, especially when the positivity rate in a region starts rising, forms a very crucial aspect of saving lives from the virus.
States with high positivity rates, also have high deaths
Most of the country’s worst-hit areas in terms of deaths all have high positivity rates. In Maharashtra, which is the worst-hit state in the country, 20.7% of all tests done have come back positive for Covid-19 — the highest in India. In terms of people who have lost their lives, the state again ranks among the worst with 288 deaths per million residents — only Puducherry (332 deaths per million and positivity rate of 15.2%) has seen a larger proportion of its population die.
On average, 71 people have died across the country for every million residents and 8.4% of tests have come back positive.
Other badly hit states, Karnataka and Andhra Pradesh, also have overall high positivity rates – 12.3% and 12.1% respectively. Both again have seen a high number of deaths — 129 deaths per million in Karnataka and 108 in Andhra Pradesh. Tamil Nadu has seen 122 deaths per million residents, but it has fared relatively better on testing and has a positivity rate of 8.2%.
In Delhi, 9.3% of all samples tested had come back positive till Saturday — again higher than the national average. The state has seen 262 people die for every million residents —the third highest of all states and UTs in the country.
Some states do fare better. Bihar has a positivity rate of 2.6% (more on this later), in Jharkhand this is 3.9%, in Rajasthan it is 4.2% and in Assam it is 5.4%. Only seven people per million residents have died of Covid-19 in Bihar. Only 18 have died in Jharkhand and Rajasthan and 19 in Assam.
The relation between positivity rate and deaths
The statistical measure to understand the strength of the relationship between two variables is a concept called ‘correlation coefficient’. It is measured in values between -1.0 and 1.0. A correlation coefficient of 1.0 shows a perfect positive correlation, which means that every time one variable increases, there is also a proportionate increase in the other. A correlation of -1.0 shows a perfect negative correlation, where increase in one variable means a proportionate drop in the other. A zero means changes in the variables do not affect the other — they just aren’t related.
For the above analysis of deaths per million and positivity rate, the correlation coefficient comes to 0.77, which can be considered a strong positive correlation. This means that increase in positivity rate has a strong influence on the deaths in a state.
Why are testing (and positivity rate) crucial?
Positivity rate — the fraction of tests that return positive — shows how widespread the virus is in the community, and over time it gives us an idea of whether a region’s testing strategy is adequate. The World Health Organisation recommends that the positivity rate from a region that has a comprehensive testing programme should be at or below 5% for at least two weeks before it can be considered that the outbreak is under control.
This means that any region that is seeing a rising positivity rate over a period of time (say a week), needs to improve its testing in order to track and isolate people with infections. Once this number drops under 5% and remains there, we can assume testing in the area is sufficient. In India, the positivity rate on average remains fairly high (mostly due to some major states still reporting high positivity rates). Unless this number drops to under 5%, the threat of rising deaths remains high.
But can this metric be skewed?
In simple words: yes. A Reverse Transcription-Polymerase Chain Reaction (or RT-PCR) test is the gold standard in Covid-19 testing because it has a very low chance of missing cases and thus tends to result in a higher positivity rate. A rapid antigen test, meanwhile, may throw up false negatives for people who are infected with a low viral load. A greater share of rapid antigen tests in a region may lead to a lower positivity rate there.
This is why there must be a caveat when we look at the positivity rates from states such as Bihar, Uttar Pradesh and Delhi, all of which have recently boosted their testing numbers and rely largely on antigen tests – around 70% of all tests in these three states have been antigen over the past 2-3 weeks.
What experts say
Experts said it “makes sense” that if the positivity rate is high then it will mean a larger number of deaths in a state, but added that the true picture may be hidden by variables like unreported cases and deaths. “The true fraction of reported-to-unreported deaths and reported-to-unreported cases are the lurking variables here. It could also be that where you have a smaller number of cases you indeed have a smaller infection fatality rate because the healthcare system is not overstretched, it is easier to test, contact trace and isolate so it could be that the fatality rate is truly different” said Bhramar Mukherjee, the head of biostatistics at University of Michigan. – Hindustan Times