Key Covid-19 indicators surged from Feb 2021 but govts failed to act

Key epidemiological indicators such as the doubling time of Covid cases and the test positivity rate had started rising from mid-February 2021. Central and state governments ignored it until too late

Despite a growing rate of increase in Covid-19 cases and a higher proportion of samples testing positive, on average, since mid-February 2021, central and state governments did not act to impose more restrictions, upgrade health facilities and control the rise in Covid-19 cases, our analysis shows.

On May 1, 2020, the Ministry of Home Affairs (MHA) said that India would track Covid-19 hotspots on the basis of active cases in a district, the time it takes for confirmed cases to double, and testing and surveillance feedback. Starting from February 1, 2021, restrictions on public gatherings were relaxed, according to an order by the MHA, which added that guidelines to track containment zones and maintain social distancing should be enforced by local authorities.

Meanwhile, India's average doubling time of cases had been increasing since the end of July 2020, and reached a peak of 710 days on February 11, 2021. From then on, the doubling time started shrinking--from 522 days by February 28, 346 days by March 15 and 139 days by March 31, indicating that Covid-19 cases were growing fast.

All this while, states allowed large gatherings such as the Kumbh Mela and did not strictly enforce distancing, mask wearing and other guidelines. As of May 3, Covid-19 cases were doubling in India every 35.7 days. Its case numbers were increasing by over 300,000 and deaths by 3,500, on average, every day even as the media has been saturated with stories of undercounting of deaths and case numbers.

Just like the doubling time, India's average Covid-19 Test Positivity Ratio (TPR)--the proportion of positive tests in all samples tested for Covid-19 across the country--reached a low of 1.58% on February 11, 2021, after which it has been steadily increasing. The TPR reached 2% by February 28, 3% by March 16 and crossed 5% on March 27.

If a district consistently has a TPR greater than 5% or a TPR that is increasing, it is likely that the virus is spreading and the district is not testing enough people, we had reported in September 2020. The World Health Organization (WHO) says the pandemic is under control if the TPR is under 5% for a period of two weeks. India's TPR was 21.5% as of May 3, 2021.

Even with an increasing average TPR and falling doubling time, gatherings continued across the country. For instance, even though the doubling time of Covid-19 cases in Uttarakhand decreased from 1,670 days on February 27 to 1,062 days on March 8, the state hosted the Kumbh Mela, a religious gathering, where at least 3.5 million people congregated on April 12, 2021, according to estimates. As of May 3, the doubling time of Covid-19 cases in Uttarakhand was 22 days and its TPR is 17%. Similarly, restrictions in some districts were put in place in Chhattisgarh by Holi on March 28, even though the doubling time of cases had started shrinking consistently by March 4 and had reached 104 days on March 28, down from a peak of 987 on February 14, data show.

Between December 2020 and April 2021, despite the warning signs, dedicated Covid-19 facilities and oxygen-supported beds for severe Covid-19 facilities fell 6%, we reported on April 22.

We analysed the doubling time and TPR of 15 states with the most active Covid-19 as of May 3, 2021. We have also considered two states that saw mass gatherings due to elections (Assam) and government-sponsored events (Uttarakhand) in our analysis. To calculate the doubling time and TPR, we use a seven-day average to smooth daily ebbs and flows in cases and testing. We use data from COVID19.org for the calculations in this story.

To be sure, the doubling time and TPR are not the only measures to track a pandemic and experts point to a number of factors that can help understand the trajectory of cases and take corrective action. "I do not like using doubling time solely as a metric. A comprehensive evaluation of the multiple epidemiologic measures--effective reproduction number [Rt, the number of people infected by one Covid-19 patient, on average], test positivity rate, predictions of infections and deaths in two weeks and healthcare capacity estimates--should have driven policy decisions not just one metric. And in this case, all metrics were pointing to an uptick and the potential for a second surge since February," said Bhramar Mukherjee, professor of biostatistics and epidemiology at the University of Michigan in the US.

As of May 6, India has 3,573,092 active cases and has reported 230,223 deaths due to Covid-19.

Doubling time in many states fell before India's average

The doubling time helps understand how quickly the pandemic is spreading in a state. As fewer new cases are reported, the doubling time keeps increasing. As doubling time plateaus, it tells us that the number of new cases is steady, but as this indicator starts reducing, it is an alarm that each day is seeing a surge in new cases.

India's current doubling time of 35.7 days is "shockingly low" and "at a rate not seen since around the peak of the September wave", said Maxwell Salvator, who works at the Centre for Precision Health Data Science at the University of Michigan, and is part of Mukherjee's Covid-19 study group. The doubling time might fluctuate or be unreliable when total case counts are low, but becomes more robust as case counts increase in size, which makes the current reduction in doubling time even more troubling, he added.

On multiple occasions last year, the health ministry announced an increase in the Covid-19 doubling time (see for instance, June 2020 and October 2020) to show that they were successfully dealing with the pandemic. It was only on March 23, 2021 that India's average doubling time of 202 days was reported publicly, at least 40 days after this number had started dropping, our review of the health ministry's press releases since February 11 has found.

Health is a state subject. Even when the central government publicly mentioned that Covid-19 cases were increasing in some states, few additional precautions were instituted in those states by the Centre or by the state governments concerned. On February 27, a week after India's doubling time had started falling, the health ministry advised eight states to refocus on surveillance and stringent containment in districts that saw a spike in cases. This included West Bengal where elections to its legislative assembly were announced the previous day, followed by massive gatherings at rallies and road shows for the next two months.

Maharashtra was the first state where the doubling time stopped increasing and then started to drop, by the second week of February.

By mid-February, the doubling time of COVID-cases had started falling in Delhi, Haryana and Madhya Pradesh.

By the third week of February, the doubling time started dropping in states such as Telangana, West Bengal, Andhra Pradesh and Assam. In Assam, the doubling time started increasing on March 3, but dropped again within two weeks of declaration of elections, by March 9.

The doubling time has reduced the most in Assam and Andhra Pradesh between January and March. As of May 3, Andhra Pradesh has 17,222 cases daily, on average, 300 times the 55 average cases daily between February 12 and 18. In Assam, average daily cases have increased 400 times from 7.5 between February 10 and 16 to 3,254 as of May 3.

The doubling time in Uttar Pradesh, Rajasthan and Gujarat started reducing gradually towards the end of February while in Karnataka and Tamil Nadu, the doubling time fell in the first week of March.

For the rest of the states, the doubling time started reducing in the beginning of March.

Kerala is an outlier, with its doubling time not increasing in 2020, until the beginning of February 2021. The doubling time started to rise in February, but it dropped again from mid-March.

West Bengal, Karnataka, Kerala, and Odisha are some of the states with the highest effective reproduction number, and recent trends suggest that this value is no longer increasing or coming down, Maxwell said, adding that sustained elevated Rt values are "troubling". "These, combined with increasing test positive rates, and with relatively low levels of testing as in West Bengal, indicate to me that serious, immediate interventions are needed," added Maxwell.

Most states have a higher TPR than the peak of the first wave

The TPR helps estimate the undercounting of Covid-19 infection, which in turn gives an idea of how much control the state has over the spread of the pandemic, according to Professor Menon. India's current TPR of 21.5%, as of May 3, is 13 percentage points higher than India's highest TPR during the last peak in September 2020.

Because the TPR has been substantially lower for a sustained period, the current high rate indicates two things: First, that the virus is spreading at a terrifyingly high rate, and second, that not enough testing is being done, said Maxwell.

Yet, the current TPR calculations might be based on unreliable data. "At present, the data for tests conducted and confirmed cases are corrupted and questionable. There is rampant undercounting, so it is difficult to make inferences from the TPR with any level of certainty," said Gautam Menon, professor of physics and biology at Ashoka University in Haryana.

As of May 3, all 17 of the analysed states have TPR over 5%. All of these states, except Maharashtra, have current TPRs higher than the TPR of the peak of the first wave in 2020.

Most experts predict the peak of Covid-19 cases in India to occur around mid-May. "I believe this prediction is reasonable. I've seen several different models reach this conclusion, alongside some of our internal modeling. It's important to keep in mind that the peak of Covid-19 related deaths is on a lag so while cases may continue to decline, we expect daily deaths to increase for another couple weeks, peaking near the end of May," said Maxwell, in an email reply on May 3.

He added: "[N]o state is out of the woods by any means… Any state that hasn't already instituted interventions such as curfews, discouraging any event that draws large crowds and lockdowns should do so. People should be wearing masks, practicing proper hygiene, and physically distancing to the extent that it is possible--and national, state, and local governments should be doing what they can to empower their citizens to engage in these practices, such as providing masks, soaps and sanitizers and working with healthcare institutions to expand test and vaccine availability." On May 4, the health ministry reported that some states such as Maharashtra, Punjab, Delhi and Gujarat were showing early signs of improvement.