Silencing science?

This is either a story about bad Covid-19 science deservedly side-lined or it is a tale of important research silenced by the peculiar power dynamics within an authoritarian state.

While it is possible that the first assumption is correct, it is far more likely that the second is true. Yet, it is difficult to know the answer because none of the parties involved in producing the research is willing to even talk about it — neither the US government agency that funded the study, the international public health body that undertook the work, nor the government agency that also took part.

On October 12th, three organisations  —  Bangladesh government’s Institute of Epidemiology, Disease Control and Research (IEDCR), the Bangladesh-based international public health body icddr,b and the US government’s donor agency USAID  —  organised a press conference to release the preliminary finding of their research. The study had been undertaken between April 18th and July 5th “on the prevalence, seroprevalence, and genomic epidemiology of Covid-19 in Dhaka.”

“Prevalence” here means how many people tested positive for the Covid-19 virus —  that is to say at the time of testing, how many people were infected with the virus. “Seroprevalence” shows the level of antibodies in the blood against SARS-CoV-2, the virus that causes Covid-19 which helps determine the proportion of people who had already been infected or exposed to the infection and produced antibodies in response.

A detailed joint press release from all three organisations was issued after the seminar. In relation to the research methodology for the testing of these two, this stated:

“[A] total of 25 wards were randomly selected out of the 129 wards and one mahalla was randomly selected from each ward and from each mahalla randomly 120 households were selected. To assess the situation in the slum, additional eight slums were also included in the survey. The households were categorized as symptomatic and asymptomatic based on the initial screening. … All members of symptomatic households and approximately every tenth asymptomatic household were selected for (i) COVID test using Real Time Polymerase Chain Reaction (RT-PCR) and (ii) blood test to detect antibodies.”

This methodology —  known as multi-stage cluster sampling  —  allowed the researchers to identify a randomised sample of individuals within the population from which they could test and obtain specimens for testing. The data obtained from the randomised samples then allowed the researchers to infer the situation in a wider population — both Dhaka City and Dhaka’s slums.

This is standard scientific practice. The purpose of doing this research is to be able to make deductions about both the wider Dhaka City population as well as the subset of Dhaka slums of which the randomised individuals who were tested were a part.

In relation to prevalence, the press release stated that 9.8% of those tested in Dhaka City were positive. “This implies that one in ten individuals were tested Covid positive in Dhaka city,” the press release stated. In relation to those living in the slums, the figure was 5.7%.

However, it is the seroprevalence findings  —  the level of antibodies found in the blood  —  that are particularly interesting. The press release stated:

“In Dhaka city, the seroprevalence of IgG and/or IgM was 45% while in the selected slums it was 74%. …. This indicates that by July 2020, 45% of Dhaka city’s population had already been exposed to the infection while the exposure was 74% in the slums.”

The press release quotes Firdausi Qadri, a senior scientist at icddr,b as saying that:

“[T]he exposure to SARS-CoV-2 has occurred in the study areas. Encouraging rates of seropositivity among the study population of Dhaka city indicates that we have started developing herd immunity against SARS-CoV-2.”

Note how the results from the randomised sample are extrapolated to the whole city or to slums in the city. The results are highly significant  —  suggesting that in Dhaka city generally, and in the slums in particular, a very high proportion of people had already been infected, or exposed to the infection, providing a real possibility of a high level of immunity. In many other parts of the world, the levels of seroprevalence have been found to be much lower – though they  are comparable to some other findings from South Asia.

This would also be very good news for Dhaka city as it would suggest that the rate of deaths per hundred thousand persons infected would be very low indeed. Unsurprisingly, the media reported the results widely. 

But something strange happened on the day following the press conference. IEDCR issued a short press release, completely dismissing this implication of the research. It stated (in translation):

“The interim result of a joint study conducted by the IEDCR and icddr,b on the Covid-19 pandemic situation in Dhaka metropolitan was published yesterday at a Gulshan hotel. We would like to thank concerned media outlets sincerely for the widespread coverage of the study. It was not claimed that the study depicted a representational scenario of the metropolitan area. However, since in some media outlets it was claimed that the study illustrated the whole/overall scenario of Dhaka city, its scientific basis is weakened and it caused confusion.

[We] should not make such a conclusion that results of so few samples/specimens are illustrative of the whole Dhaka city. In order to paint the actual situation of Dhaka city, [we] should conduct a study on a larger scale after collecting sufficient representational samples. We are requesting concerned media outlets to help clear the confusion by publishing our statement with due importance.” (bold emphasis added)

This is a very surprising statement as it contradicts what was set out in the original press statement, jointly released by the three organisations, which includes IEDCR. 

IEDCR says that the study had not “[D]epicted a representational scenario of the metropolitan area” and the “results” were not “illustrative of the whole of Dhaka City”. But this is exactly what the original press statement said. There would have been no point in the study unless it allowed one to infer the situation about Dhaka city.

So, why would the government body issue this statement?

Well, it is possible of course that there is something completely wrong with how the study was done —  that there is some representational problem and the results can not be used for inferring information about the wider city.

However, it would be surprising, to say the least, that icddr’b, an international public health body of global standing, would produce such a substandard research. If this was the case, one would also expect it to make a statement in response  —  which it has not done. In fact, the original press statement remains published on the icddr,b website. 

A more likely explanation is that the IEDCR statement reflects both the sensitivity of Bangladesh’s government to independent scientific research outside its control and its ability to contradict and silence research that is uncomfortable for one reason or the other. 

At first sight, it would seem odd that the government would seek to mute these results, which arguably present a positive picture for Dhaka city. However, it could well be the case that the government felt embarrassed by results which indicated a high level of infection in the capital, which it simply did not want to admit.

It is no secret that the government is very sensitive to this. When the Economist magazine in June of this year quoted John Clemens, the executive director of icddr,b as stating that the number of actual Covid-19 cases at that time was likely to be over ten times the official figure, he was reportedly put under significant pressure by government authorities to explain this and was required to issue a press release. The government has also reduced levels of covid-19 testing, presumably to allow it to portray low levels of infection.

In the context of Bangladesh, where currently few institutions can operate without government coercion, icddr,b — a supposedly independent international organisation — also has to toe the line.

To try and clarify this, the three organisations involved in the research were contacted. USAID responded by saying “speak to icddr,b”, which responded by saying “contact IEDCR”, from which there was no response. In a second approach, icddr,b repeated its inability to help, but this time adding “[P]lease note that icddr,b respects IEDCR’s views and has no further comment on this.”

This is as far as icddr,b is willing to go in distancing itself from the IEDCR statement. While one can not be sure what was actually behind the IEDCR statement, perhaps the main take-away from all this is the supineness of icddr,b, a supposedly international public health body, within the context of Bangladesh’s current authoritarian political structures. It can not even stand up for its own research.

Next year, icddr’b’s US director is standing down, and for the first time the organisation will be led by a Bangladeshi. One can not help but fear that the institution’s current limited autonomy will reduce even further in the future. Not a good sign for independent science in Bangladesh.●

David Bergman (@TheDavidBergman) — a journalist based in Britain — is Editor, English of Netra News.

🔗 icddr,b – The IEDCR and partners share insights on the prevalence, seroprevalence and genomic epidemiology of COVID-19 in Dhaka city (Oct, 2020)

🔗 icddr,b – Clarification about icddr,b’s licensing requirement (June 2020)

🔗 icddr,b – Appointment of the next icddr,b Executive Director (Sept, 2020)

🔗 The Lancet – Serotracker: a global SARA-CoV-2 seroprevalence dashboard