Stunning estimates released by Statistics Canada earlier this year showing that hundreds of people more than normal died in New Brunswick during the summer and fall of 2021 have been discarded by the agency.
It is now reporting the opposite — that in New Brunswick in 2021 hundreds fewer people than normal died.
It's a major swing in a critical pandemic measurement that has surprised health and data professionals and generated calls for a detailed explanation from all government agencies involved.
Tara Moriarty, an associate professor and infectious disease researcher at the University of Toronto, has been tracking the issue of "excess mortality" across Canada during the COVID-19 pandemic. She is baffled by what's happened with New Brunswick's numbers.
"I've never seen this before. I've never seen it," Moriarty said of a sudden drop in New Brunswick death counts.
"The province and Statistics Canada need to explain."
In July Statistics Canada released estimates showing 8,138 people had passed away in New Brunswick during the first 50 weeks of 2021.
The number shattered previous records and was calculated to be 567 more deaths than would be expected in a normal, non-pandemic year.
That appeared to show the COVID-19 pandemic was taking a much larger human toll on New Brunswick residents than official government counts of COVID deaths were recording, and it triggered sharp questioning in the legislature and expressions of concern from public health officials.
However, on Thursday, Statistics Canada released new estimates lowering the death total in New Brunswick for those 50 weeks to 7,272.
That's 866 fewer deaths than it reported in July and now puts fatalities in New Brunswick in 2021 lower than normal, not higher.
There have been no similar revisions in other provinces.
Explanations from Statistics Canada have been limited but appear to focus on ongoing problems it has been having with the quality of mortality information being provided to it by New Brunswick.
"Recently, New Brunswick has been submitting preliminary information on deaths to Statistics Canada," read a statement issued by Statistics Canada on Thursday about the significant change in its New Brunswick death counts.
"While this information does provide more timely results, this incomplete data has increased the uncertainty associated with provisional estimates for recent reference periods.
"As a result … this information is no longer being included in the provisional estimates for New Brunswick."
Statistics Canada has been tracking deaths in each province monthly during the COVID-19 pandemic and comparing them to what would have been expected in a normal year in an effort to detect "excess mortality" caused by the virus both directly and indirectly
"To understand the direct and indirect consequences of the pandemic, it is important to measure excess mortality, which occurs when there are more deaths than expected in a given period," Statistics Canada notes in an explanation of the project.
The agency had been showing New Brunswick with one of the highest death rates among provinces during the second half of 2021, but that has flipped to one of the lowest.
And although the numbers were always presented by Statistics Canada as being "estimates" and "provisional" no other province has experienced a change as large.
Some of the changes are so large, not everyone is prepared to accept the new estimates as being any more accurate than previous ones.
For example during the first two weeks of December 2021, Statistics Canada had been estimating 439 deaths in New Brunswick as recently as two months ago but now puts the number at 326, a 26 per cent reduction.
And it went from estimating 61 deaths above normal in New Brunswick in the last two weeks of October 2021 (359) to 20 deaths below normal (278) in its current estimate.
Moriarty said the involvement of multiple provincial and federal bodies in the construction of death estimates in New Brunswick makes identifying where the problem in the numbers lies difficult, but she doubted Statistics Canada alone is to blame.
"Professionally, they are exceedingly careful about being very conservative and cautious about their estimates and about trying to be as right as they can be," said Moriarty.
"It's incredibly important for trust in the institution, so they can't be too happy at the moment."
Ray Harris, a Fredericton-based data professional, has worked with New Brunswick COVID numbers throughout the pandemic, including the excess mortality numbers.
He noted the province has not always presented other COVID data in a straightforward way and said such large revisions in how many people died in New Brunswick in 2021 will need to be fully explained to be believed.
"This is a hot topic. People are looking at these numbers," Harris said. "I'm of the belief that you should always try to build credibility, especially with data."
In a statement, the Department of Health said it is working co-operatively with the federal government on the mortality issue but no one was made available for an interview.
"The department continues to collaborate with Statistics Canada to ensure the data and methodology are appropriately understood by both parties," read the statement.
"The Department of Health will also complete an analysis of death certificates of its own to better understand mortality trends observed during the pandemic, as well as the contribution of COVID-19 to excess deaths."