Chronic. While execution errors are inevitable in the fight against the pandemic, some mistakes seem more absurd than others. This appears to be the case with the loss of 15,841 cases of Covid-19 that tested positive in an official database in the United Kingdom, erased between September 25 and October 2, because the database used an old version of Excel which had a limit of 65,536 lines per file. What the users did not know …
These cases were, therefore, not notified to the tracing system set up by the government, and the contacts of those who tested positive were not alerted and therefore could not isolate themselves to protect others. .
Each case that was not notified to authorities would have resulted in around 25 additional positive cases in the following six weeks.
Mistakes are not always unnecessary for scientific research. In the field of the effectiveness of tracing systems, the debate is made difficult by the fact that the characteristics of countries where the tracing effort is more systematic are often different from those of countries where this effort is less. If Taiwan, South Korea and New Zealand have succeeded better than us in controlling the pandemic, is it thanks to their careful tracing systems, or thanks to many other logistical and behavioral factors that may have played a role in them. favor?
Two researchers from the University of Warwick (United Kingdom) and Harvard Business School (United States) saw the British Excel database fiasco as an opportunity to measure the utility of tracing. If in some regions the tracing was blocked by a simple technical error, which did not influence the other conditions of the implementation, can we see in the regions concerned a more rapid expansion of the pandemic?
Their analysis (“Does Contact Tracing Work? Quasi-Experimental Evidence from an Excel Error in England”, Thiemo Fetzer and Thomas Graeber, CEPR Discussion Paper, n ° 15494, November 2020), published on November 25 in a “working document” (that is to say that it has not yet been validated by the reading committee of a scientific journal), provides a fascinating sketch of an answer.
The authors first tried to identify the localities affected by the lost cases – as this information was not publicly available, they relied on the fact that the lost cases were added late to the base from October 3. . We therefore see in some localities a greater drop than elsewhere in the trend of notified cases, followed by a catch-up thereafter. This is the indication of a greater presence of lost cases for which tracing was not carried out.
You have 34.96% of this article to read. The rest is for subscribers only.
