The COVID-19 situation in Bihar has improved lately – but in July and August things were bad. There were reports of hospitals at breaking point and a significant number of healthcare workers being infected with COVID-19. There were also reports of dysfunctional crematoria and bodies being burnt in the open. Even politicians were reportedly succumbing to the disease.
Yet such reports seem at odds with Bihar’s relatively low total count of COVID-19 cases and deaths, especially the latter. How come?
In a previous article, this author had examined two possible explanations for Bihar’s COVID-19 data: slow spread of the disease, or poor disease-surveillance. Recent data indicating that a large number of people in Bihar had been exposed to COVID-19 by August suggests that weak disease surveillance is the main factor. The data paints a stark picture of very poor detection of infections and probably very many missed deaths, and also contradicts the notion that the spread of the novel coronavirus is necessarily slower in rural areas.
The results are from six districts – Arwal, Begusarai, Buxar, Madhubani, Muzaffarpur and Purnia – surveyed in late August this year, as part of the second national seroprevalence survey. These results have not been widely reported. A search on Google for some keywords and direct searches on news websites, on October 31, found no English language reports among the results. A few Hindi outlets, including Dainik Bhaskar, had reported the results on October 12.
The limited coverage is perplexing given the high interest in Bihar in the run up to its legislative assembly elections and the data is so inconsistent with the National Democratic Alliance’s narrative of Bihar’s successful COVID-19 response.
Bihar versus Chhattisgarh
In the survey, Bihar’s headline figures were as follows: By late August, about 16% of the population in the surveyed districts had developed antibodies to SARS-CoV-2, which is the virus that causes COVID-19. This implies that about 34 lakh people had been exposed to the virus in these districts. But not many of these infections had been detected – fewer than 1% in fact. And the detection was poorest in the most rural districts. Moreover, according to official figures, there had been only 72 deaths in these districts – that is, just about 1 out of every 47,000 people who had been infected had officially died. This is a tiny fraction of the deaths we expect from international and Indian data.
These numbers are best viewed relative to the numbers from nearby Chhattisgarh. After early success in controlling its COVID-19 epidemic, Chhattisgarh had by mid-September become a COVID-19 hotspot, recording around 3,000 new cases every day. A seroprevalence survey in 10 districts in the second half of September indicated that about 6.7% of people in these districts – roughly 10 lakh individuals – had been exposed to the coronavirus.
Also read: Bihar’s Health System Is Woefully Underequipped to Handle COVID-19
About 6.4% of these infections had been detected by tests. This drops somewhat to 4.6% if we leave out the predominantly urban districts of Durg and Raipur from the calculus. But even in the most rural districts (that had been surveyed), Chhattisgarh was detecting more infections than Bihar was.
Officially, about 1 out of every 1,700 of those infected in Chhattisgarh had died. The epidemic was still raging, and this rate goes up considerably if we assume that deaths were being reported late. But taking the number of reported deaths at face value, Chhattisgarh’s death rate was 30-times higher than that of Bihar. Better death-reporting in the urban districts of Durg and Raipur could have biased the picture, somewhat – but even if we leave these districts out, Chhattisgarh’s death rate was still 15-times higher than Bihar’s.
So we have to ask: was the same virus really much deadlier in Chhattisgarh than in Bihar? Or was Bihar failing to detect or report most of its COVID-19 deaths?
Missing fatalities
In Chhattisgarh’s surveyed districts, the naïve infection fatality rate – the number of deaths as a fraction of the estimated number of infections – is comparable to Delhi’s but lower than that of Mumbai or Chennai. It is also less than half of what we expect from international fatality data that also accounts for different differing age structures. There could be two to three missed deaths for every recorded death in Chhattisgarh, likely predominantly from the rural districts. It is also possible that the disease had spread faster in younger populations, lowering the death rate, but the data to validate this isn’t available.
This said, Bihar’s numbers are in a different league altogether – because either the people of Bihar were mysteriously protected from severe disease or almost none of Bihar’s COVID-19 deaths had been recorded. From the international data, we expect 70- to 140-times as many deaths as recorded in Bihar’s surveyed districts. From Mumbai’s data, we expect about 60-times the recorded deaths. From Delhi’s and Chhattisgarh’s data, we expect about 30-times the recorded deaths. Even with data only from Chhattisgarh’s largely rural districts, we expect about 15-times the number of deaths as have been officially recorded. So however we look at it, Bihar’s death count just doesn’t add up.
Aside from poor detection of COVID-19 infections and probable massive death-underreporting, Bihar’s seroprevalence survey data also indicates rapid and undetected rural spread.
Also read: Why Educated Biharis Are Unable to Improve Bihar’s Human Development Indices
To illustrate this, consider Madhubani, a large district whose population was, in the 2011 Census, 96% rural – making it the most rural of the districts that were surveyed and almost the most rural district in the state. And yet, Madhubani recorded the highest antibody levels of Bihar’s six surveyed districts: 23%. Over 10 lakh people had been exposed to the disease but fewer than 1 in 200 of these infections had been detected. And with only six reported COVID-19 deaths in the district, almost no one was dying – officially anyway.
Madhubani is an extreme case; the disease’s spread was in general greater in more rural districts of Bihar. On the other hand, infections were most likely to be missed in the more rural districts. These two observations are probably linked – with very poor rural detection of infections, the disease could spread in rural areas without the local people being aware of it or being able to take action to slow it down.
In Chhattisgarh, by contrast, the disease levels were generally higher in more urban districts, with the highest infection levels in Durg and Raipur. We have what seems to be the more common pattern – that COVID-19 spreads most easily in urban areas. The fraction of infections being detected through testing was higher in more urban districts, but this trend was less marked than in Bihar.
§
Comparing Bihar with Chhattisgarh highlights some striking features of Bihar’s data. Bihar’s detection of infections was weaker, while the state reported implausibly few deaths and rapid rural spread.
The numbers suggest that we should not rush to interpret the low COVID-19 case load and deaths in predominantly rural states as being the result of low levels of infection or as signs of successful control. Although more rural states do in general report fewer cases and deaths, transmission can be rapid in rural settings. Unfortunately, COVID-19 monitoring in Bihar has been so poor that we may never know how and why COVID-19 spread so rapidly in the state’s rural areas.
Also read: Meet Sanjay Sahni, the MGNREGA Worker Contesting the Bihar Polls
In fact, the data strongly supports the claim that Bihar’s ‘success story’ is founded on poor disease and death surveillance, and that the fatality data in particular should be treated with extreme caution. Without a proper survey of mortality in Bihar during the epidemic so far, we – including the people in the state – may never know its true toll.
A more detailed version of this piece is published here. Many thanks to Nikhil Rampal for pointing me to reports on Bihar’s seroprevalence survey.
Murad Banaji is a mathematician with an interest in disease modelling.