Death and data: Are Delhi's numbers for Covid-19 outbreak believable?

National capital is reporting hospitalisations like other cities, but the proportion of people requiring in ICU bed is much higher.

Delhi’s numbers for Covid-19 deaths do not match records from cremation grounds handled by the municipal corporations. Media reports suggest that Delhi may be under-reporting coronavirus deaths by a factor of two.

A closer look at the numbers reveals many more data anomalies in the government’s statements. A Business Standard analysis finds that Delhi may be under-reporting the number of cases as well.

On Thursday, the national capital became the second city, after Bengaluru, to record over one lakh active infections. The city has been recording 23,000-25,000 cases daily, and recoveries have not kept pace.

Hospitalisations have been rising in Delhi. A fortnight ago, 7,021 patients here required hospitalisation. That number has now jumped to 19,563, putting stress on the infrastructure. The number of beds between April 12 and April 29 increased almost 74 per cent, but the demand has shot up by 178 per cent.

The Delhi administration has added 159 per cent more ICU beds, but demand has risen by 244 per cent. In this period, the number of active cases in the city has jumped 201 per cent.

The proportion of people requiring an ICU bed and a hospital bed have both declined. Puzzlingly, hospitalisations have fallen faster than ICU bed requirement. While 27 per cent of active cases required a hospital bed on April 12, on April 29, this ratio had declined to 20%, whereas the number of active cases requiring an ICU bed during this period fell from 5.5 per cent to 5.1 per cent.

Ideally, when a city reports more infections, these ratios tend to move lower. Rising hospitalisation and a higher proportion of people requiring a ventilator usually indicate that the actual number of infections are higher than the reported infections.

Even at the height of the first wave, the proportion of people requiring an ICU bed in Delhi averaged 2.7 per cent. It increased to 5.6 per cent in December, as the city ramped up testing, the numbers declined again.

Instead, in Delhi, there has been a scale back in testing over the last week. Not only has testing declined by 30%, but the proportion of RT-PCR tests—these also have a higher accuracy—has declined from 70 per cent to 60 per cent.

Besides, a comparison with other cities shows that Delhi may have the highest proportion of people requiring an ICU bed, even though hospitalisation rates are more or less equal to other cities.

The number of people requiring a hospital bed in Mumbai is higher at 22 per cent, but people needing an ICU bed is much lower at 3.8 per cent In Pune, while hospitalisations have been similar to Delhi at 20 per cent, people requiring an ICU and a ventilator is a lower 2.9 per cent.

Another instance that points to a lower-case detection is the new death ratio in the capital. Although the country averages 13 deaths per 1,000 new cases, Delhi has a much higher per capita ratio of 21. This ratio averaged 11 during the first wave of infections. A higher proportion, again, indicates that some infections may be going undetected.

India’s National Centre for Disease Control (NCDC) data released last week had indicated that the UK and double mutant variant was behind the surge in cases in the national capital. Of the 3,206 samples tested in the national capital, 15 per cent belonged to these two variants. For the country as a whole 11 per cent of cases were attributed to variants of concern. In Maharashtra, which has been worst affected by the Covid-19 wave, 28 per cent of samples belonged to the UK and double mutant variant. Punjab recorded the highest proportion for “variants of concern” with over 70 per cent of samples genome sequenced by Indian SARS-CoV-2 Genomic Consortia (INSACOG) belonging to the UK and double mutant strain

In a press conference earlier this month, the ministry of health and welfare had indicated that the proportion of deaths to cases in the second wave was similar to the first wave.