AP Photo/Rajanish Kakade
Science’s COVID-19 reporting is supported by the Pulitzer Center and the Heising-Simons Foundation.
MUMBAI—Last week, a panel of leading scientists appointed by the Indian government delivered a startlingly optimistic message: The world’s second largest COVID-19 epidemic has rounded a corner. India’s daily number of daily new cases has almost halved the past six weeks, and a new mathematical model suggests “we may have reached herd immunity,” some members of the panel wrote in a paper published online by The Indian Journal of Medical Research. Assuming measures such as social distancing, wearing masks, and hand washing remain in place, the group said the pandemic could be “controlled by early next year.”
But other scientists say the model overestimates the number of people already infected and warn that with colder temperatures and several religious holidays approaching, India may well see a second wave. The positive national trends hide a more complex picture, suggests Giridhar Babu, an epidemiologist with the Public Health Foundation of India. He believes the virus may have burned through large, densely packed populations but will continue to spread in rural areas, at a lower rate, for many months: “We still have large numbers of people for the virus to go through.”
Daily new infections in India have fallen from a high of 90,000 a day in mid-September to fewer than 50,000 this week. Deaths have also gone down, from a peak of 1275 per day in mid- September to around 500 now.
The encouraging projections come from the National Supermodel Committee, which modeled the past and future of India’s epidemic at the government’s request. Its work suggests 380 million Indians had already been infected by mid–September and that there might be “minimal active symptomatic infections” by late February if control measures continue. (The study also concluded that by flattening the curve, India’s lockdown last spring saved up to 2.6 million lives.)
But the model did not take regional differences in viral spread into account, and most cases are still concentrated in less than half of India’s 28 states, including those that have large cities where the virus first entered the country, such as Mumbai. Babu notes that serological surveys—which test for antibodies in a population to gauge the fraction already infected—have found much higher infection rates in cities, and in particular in slum areas. Studies in August, for example, found antibodies in 41% of residents of Mumbai’s slums, compared with 18% elsewhere in the city. “Many dense pockets that can be easily infected have [likely] already been infected,” Babu says. That could help explain the slowdown in new infections but suggests many more people remain vulnerable.
Gautam Menon of Ashoka University, a co-author on several COVID-19 modeling studies, adds that the model suffers from “a lack of epidemiological realism” because it assumes an unusually large fraction of infected people remains asymptomatic. Other models for the Indian pandemic are “better rooted in reality than this one,” says Menon, who believes 200 million to 300 million is a better estimate for the number of infected people.
Experts agree that spread in rural areas, home to more than half of the population, is a challenge to both fight and monitor. The health infrastructure in these areas is weaker, making it harder to treat patients. And testing isn’t easily available in many small towns and villages. Serosurveys suggest official testing, now at 1 million per day, vastly undercounts actual cases. Some states rely heavily on so-called rapid antigen tests, which range widely in sensitivity. Such tests comprise about 90% of testing in the most populous and rural state of Bihar, for instance, which has reported relatively few cases; in the state with the most reported cases, Maharashtra, two thirds of tests are still done using the more reliable PCR assay.
More granular data might help scientists better understand the pandemic’s trajectory. The national epidemic is a “figment of statistical imagination,” says T. Jacob John, former head of the department of virology at Christian Medical College; instead, “There are 100 or more small epidemics in different states and cities, rising and falling at different times.”
There are 100 or more small epidemics in different states and cities, rising and falling at different times.
More detailed data could also shed more light on India’s perplexingly low mortality rates, often touted by the government. The total death toll stands at a little over 100,000, less than half that of the United States, which has roughly a quarter of India’s population. India’s fatality rate has decreased the past few months and is currently about 1.5%, compared to 2.8% in the United States.
Explanations for the low death rates have ranged from the country’s young population to unproven factors such as cross-immunity from other viruses. But if age is the driving factor, India’s mortality rate would be lower than it actually is, according to a study by researchers at the U.S. National Bureau of Economic Research who used a model to predict mortality in India based on age-specific mortality rates in other countries. For instance, if India’s age-specific fatality rates were similar to South Korea’s, its overall fatality rate would have been only 0.74% in July, when it was still 2.7%.
In reality, the number of COVID-19 deaths is almost certainly higher. Death registration rates are not 100% in many states to begin with. And investigations have shown that some states are not including suspected or probable COVID-19 deaths.
Ramanan Laxminarayan, director of the Center for Disease Dynamics, Economics & Policy, and others have found surprising patterns in the mortality data. A study of confirmed cases in two Indian states showed that deaths among people over 85 were lower than in the United States, perhaps due what to Laxminarayan calls “survivorship bias:” Life expectancy is lower in India, and those who live beyond 75 in the first place are likely to be in better health and of better socioeconomic status, he says.
On the other hand, the study found surprisingly high death rates in younger cohorts. For instance, 9% of COVID-19 patients aged between 40-50 died, compared to around 2% in the United States. Underlying conditions could be one reason, says Laxminarayan: “There is a lot of hidden diabetes and hypertension in the country.” Air pollution may also be driving up mortality, he says. A study published this week estimated that long-term exposure to air pollution is linked to 15% of COVID-19 deaths globally.
The effect of air pollution will be more clearly seen in the next few months, as north India enters its traditionally smoggy winter. Meanwhile, further loosening of travel and economic restrictions may trigger an uptick in cases. Religious festivals have already spurred fresh surges in Kerala and other states that were successful in curbing infections early on. Even the National Supermodel Committee warns that cases could surge in December if precautions aren’t taken.
Babu warns against complacency. “The decline [in cases] is real and valid, but no one should rejoice yet,” he says. “It only means that the first set of formidable challenges is over, and the next set is beginning.”
Source: Science Mag