When Ontario implemented regionally targeted lockdowns at the height of the COVID-19 pandemic, a common presumption was that travel from highly restricted regions to those with low levels of restriction will increase, negating the intent of the lockdowns. Now, new research is showing that wasn’t necessarily the case.
Researchers, led by Western University’s Jed Long, used de-identified and aggregated network mobility data to determine the effectiveness of Ontario’s regionally targeted lockdowns to reduce movement during the pandemic. The research found the lockdowns did not significantly reduce mobility from one public health region to another.
Long and his research team measured inter-regional mobility (outflows) between public health regions. They used network mobility data from TELUS’ Data for Good program through its Insights platform, a privacy-preserving system for analyzing mass-mobility patterns within Canada. Device locations were aggregated by determining which cell towers people most commonly connected to, and were divided into aggregate disseminated areas (ADAs), which have populations of approximately 5,000 to 15,000 each.
In one paper, Do regionally targeted lockdowns alter movement to non-lockdown regions? Evidence from Ontario, Canada, Long, with Milad Malekzadeh and Ben Klar of Western University and Gina Martin of Athabasca University, focused on two specific intervention dates to determine the effectiveness of regionally targeted lockdowns.
Through 2020, the Ontario government implemented travel restrictions for citizens, based on the boundaries of regional health units. The researchers focused on the week before and after July 17, 2020, and the week before and after November 23, 2020, when significant portions of the province had different levels of travel restrictions in place. In July, most of Ontario moved into what was then termed Stage 3 restrictions, while a number of areas such as Windsor, Toronto and Peel maintained higher levels of restrictions. In November, Toronto and Peel re-entered lockdown restrictions, while the rest of Ontario was not put into lockdown until December 26.
“One hypothesis was there would be an increase of people leaving the targeted area to do shopping and other activities. Media talked about this as happening, but the data does not bear this out,” said Long, professor in the department of geography and environment. “An alternative hypothesis was that people would see a responsibility and not leave their area. In the end, we did not see much of a change.”
In a second project, Associations between mobility and socio-economic indicators vary across the timeline of the Covid-19 pandemic, Long and Chang Ren, also from Western, examined how socio-economic factors were associated with mobility patterns, and how this relationship changed through 2020. They found that those from neighbourhoods with higher economic deprivation, as determined by census data, did not change their mobility patterns as much as those from more economically well-off areas. Long said it is important to note the socio-economic consequences of the lockdown measures, and to understand that different areas were affected differently.
“Initially, early in the pandemic, people from more economically well-off areas could reduce their mobility more, as they had the option to work from home,” said Long. “In areas where people had to go to work, who could not work from home; these areas maintained their original mobility patterns. The strong association between economic deprivation and relative changes in mobility levels was one of the only consistent findings we observed throughout all of 2020.”
The researchers also found that these associations changed over the course of the pandemic and depended on how they measured mobility. “In the first wave of the pandemic in spring of 2020, we found that those in more urban/densely populated areas did not change the variety in the places they visited as much as those in less densely populated areas,” said Long.
Further research is underway using the de-identified and aggregated network mobility data to determine whether restrictions would be more effective if targeted on smaller regions or boundaries rather than using public health regions.
“With targeted interventions, you are trying to eliminate the spread in, and between, sub-populations,” said Long. “Maybe sub-populations can be determined based on where people travel as opposed to the province’s public health regions. If we ever have to implement these types of restrictions again, it could be a better approach.”