Pakistan can learn from the UK data collection
Is it worth using a precious Covid-19 testing kit on a person who is minutes away from death—only if you care about your data.
A report in The News suggested that 300 patients with “pneumonia-like symptoms” had been brought to hospitals in Karachi in the last 15 days. These patients were either pronounced dead on arrival or died in a few hours. The story led to speculation that deaths from Covid-19 might be under-reported in Pakistan.
Then Faisal Edhi appeared on SAMAA TV’s show Naya Din to reveal that Karachi had seen 70% more deaths in April compared to last year. Most of the deceased were elderly and had breathing problems. Still, Edhi stopped short of making any claims. “It cannot be conclusively said that these people had the coronavirus,” he said, citing the closure of OPDs as a probable cause.
The speculation made enough headlines to warrant a response from the Special Assistant to Prime Minister on Health. “It would be premature to say that the deaths being reported are from coronavirus,” said Dr Zafar Mirza. Claims cannot be made without lab tests, he added.
Data from provincial governments and the official dashboard seems to be recording deaths only recorded in hospitals as caused by coronavirus. Can it simply be assumed that all people who are coronavirus victims died in a hospital bed? If not, is the government prepared to factor these deaths into its data?
Here’s how the United Kingdom is doing it:
In the UK government’s daily press briefing on Covid-19 Sunday night on BBC, deputy chief medical officer Jenny Harries showed the following graph containing data of deaths in the UK compared to other countries.
“You can see there are two lines here, the UK hospital-only data, this is very robust data which is readily accessible,” Harries said. Pakistan is reporting the same type of deaths that are only recorded in hospitals.
But there is a second line on the UK graph. “[Y]ou will see a UK all settings line which is reflecting the fact that we have a lag, if you like, between our community death reporting,” said Harries. The implication of the word ‘lag’ here not only means that the death count takes a few days to be processed but is also an admission that there might be more people dying from Covid-19 than the government can accurately record.
This means that the government is already keeping track of the possibility of more deaths from Covid-19 than those that are recorded in hospitals. In Pakistan where hospitals and testing centers are concentrated mainly around urban areas, the number of deaths from Covid-19 outside of hospitals is something that warrants more careful scrutiny by authorities.
But what if a person has coronavirus but dies of something else?
The death of the Sindh chief minister’s brother-in-law was probably the most high-profile of all Covid-19 related deaths in the country. However, CM Murad Ali Shah clarified that his brother-in-law did not in fact die from coronavirus. In a video message, Shah said that his brother-in-law had recovered from Covid-19 but died because the disease had severely affected his lungs.
So how do we attribute a person’s death to coronavirus if he or she has recovered from it but has been affected to the extent that they lose their life? If the deaths recorded in Pakistan only consider patients who had no other ailments, our calculations are dangerously skewed.
A recent Guardian report clarified that the UK government is keeping track of this problem as well, at the cost of discrepancies in its data. It revealed that there are differences between the data being reported by the Office of National Statistics and the National Health Service, because the ONS records every mention of coronavirus on a person’s death certificate even if it is not the cause of death.
Note: The graph is UK government document reproduced here by a screen grab from Sky News.