NOEL KING, HOST:
The CDC's director, Rochelle Walensky, testified before a Senate committee yesterday. She told the committee that the delta variant is responsible for most new COVID cases in this country.
(SOUNDBITE OF ARCHIVED RECORDING)
ROCHELLE WALENSKY: CDC has released estimates of variants across the country and predicted the delta variant now represents 83% of sequenced cases.
KING: Eighty-three percent. Now, the delta variant was first identified in India. The Indian government says COVID has killed around 400,000 people there, but researchers from Brown, Harvard and the Center for Global Development believe the number is much higher. Arvind Subramanian took part in this research. He's at Brown University. Good morning to you, sir.
ARVIND SUBRAMANIAN: Good morning, Noel.
KING: Your research looked at excess deaths in India, and I know you wanted us to be very clear about that. What are excess deaths?
SUBRAMANIAN: Excess deaths are - you know, you take how many people died during the COVID period, and then you compare it against some reference or base year. So the distinction is between saying COVID caused so many deaths, which we will never know because, you know, when you measure deaths during COVID, there are also other deaths happening and that they may be happening indirectly because of COVID and, indeed, deaths that would be averted because of COVID, because of - you know, if you have a lockdown, you know, there are fewer accidents. So what we're measuring is, basically, how many died during the COVID period, how many died in a reference year. You compare the two; it's likely that a high proportion of those are going to be due to COVID.
KING: And so the Indian government says, we believe around 400,000 people have died of COVID. How many people do you and your research team believe have died of COVID in India?
SUBRAMANIAN: Yeah. You know, this is - because it's an inherently difficult thing to measure, especially in a developing country like India where the health systems and the health measurement systems are kind of challenged, we come at this in three different ways because, you know, no one data source is perfect. Each has its limitations. And we kind of - you know, but all three broadly tend to suggest that somewhere between, you know, 3, 3 1/2 to 5 million people...
KING: Oh.
SUBRAMANIAN: ...With a wide margin of uncertainty, died - more people died during COVID, so somewhere between 3 to 3 1/2 and, you know, 5 million.
KING: Why would there be such a large difference between the government's official death toll and what you found?
SUBRAMANIAN: Yeah, as I said, you know, if you take - even in the early part of the pandemic in this country and in many advanced countries, it was actually quite difficult to, you know, measure these excess deaths because, you know, you're in the middle of the pandemic. You can't really measure things. You know how - you don't know how many people are dying, you know, indirectly due to COVID. So there were even difficulties in advanced economies. But in a country like India, it's going to be even more difficult to actually do these measurements. And, you know, the way the Indian government measures deaths in normal times is by doing a survey, and that survey has not been done for two or three years. So the data comes kind of patchily from all parts of the country. India is a very large country. You know, like the United States, there are about 29 states, and each state does its own measurement, you know. And these complications, you know, lead to the fact that your measurement is especially poor in these pandemic times.
KING: But it doesn't sound like you think the Indian government is deliberately undercounting.
SUBRAMANIAN: Well, you know, that's a difficult kind of charge to make. I mean, it's not clear, you know, how you would say that, and so I would not exactly say that. It's just, you know, mostly due to weak systems. It is true that, you know, the Indian government has not, you know, shared the data that it has more generally, and a number of scientists have wrote to the government saying, you know, it's easier to respond to the pandemic if you make these data available. But that's slightly different from saying, you know, deaths were undercounted.
KING: OK, fair enough. Arvind Subramanian is with Brown University's Watson Institute. Thank you so much for joining us this morning.
SUBRAMANIAN: Yeah, thanks for having me. Transcript provided by NPR, Copyright NPR.
