Effects of city design on transport mode choice and exposure to health risks during and after a crisis: a retrospective observational analysis

Background

Rapid declines in city mobility during the early stages of the COVID-19 pandemic in 2020 resulted in reductions in citizens’ exposure to transport-related air pollution and associated health risks as many cities introduced non-pharmaceutical interventions designed to curb the spread of COVID-19. However, these benefits soon reversed during the pandemic's recovery phase (ie, from September, 2020, onwards), especially in cities with designs that afforded mode shifts away from public and active transport in favour of private motor vehicles. The aim of this study was to understand the association between global city designs, transport mode choices, and population-level risk exposure during 2020.

Methods

In this retrospective observational analysis, we assembled and analysed spatial datasets (including historical and predicted pollution levels, mobility indicators, and measures of individual disease transmission) and clustered 507 global cities using a graph neural network approach based on measures of the structural dimensions of each individual city's design and network structures of urban transportation systems. We compared city types on the basis of transportation mode shifts, air pollution levels, and associated health outcomes (ie, cardiovascular disease, ischaemic heart disease, respiratory disease, asthma, and reported COVID-19 cases) throughout 2020. We estimated risk reductions for these health outcomes across four phases of the pandemic, which we defined as the pre-pandemic, entry, mid-crisis, and recovery phases. We also identified city designs showing sustained reductions at the end of 2020 in transport-related air pollution (fine particulate matter [PM2·5] and nitrogen dioxide [NO2]) associated with reduced estimated risk of acute and chronic disease outcomes (ie, all-cause mortality, ischaemic heart disease mortality, cardiovascular disease, respiratory disease, and asthma).

Findings

The mean estimated reduction of global NO2 concentrations across the observed cities from the beginning of the entry phase until the mid-crisis phase was 3·76 parts per billion (ppb), calculated as the difference between observed 2020 mean levels of 12·63 ppb and predicted mean levels (if the pandemic and mobility restrictions had not occurred) of 16·39 ppb. The mean estimated reduction of global PM2·5 concentrations across the observed cities was 9·76 μg/m3 (the difference between observed 2020 mean levels [29·03 μg/m3] and predicted mean levels [38·79 μg/m3]). If maintained over the long term, the estimated NO2 reduction could have a substantial effect on reducing health risks for both acute and chronic disease, equating to an estimated overall reduction in all-cause mortality risk of 1·5% (95% CI 2·2–3·0), a reduction in cardiovascular mortality risk of 4·1% (2·6–6·0), and a reduction in respiratory disease mortality risk of 1·9% (0·8–3·0). If the reduction in PM2·5 concentration estimated in this period was maintained over the long term, all-cause mortality risk reductions of 18·9% (95% CI 13·2–25·0), asthma risk reductions of 46·8% (18·7–65·5), and ischaemic heart disease morbidity risk reductions of 0·25% (0·2–0·3) could be achieved. In the later stages of 2020, city designs (primarily in the Americas and Oceania) that afforded a mode shift away from public transit to private motor vehicles during the pandemic's recovery phase tended to show the poorest outcomes across all air pollution and health measures, even increasing risk levels above pre-pandemic baselines in some cases. By contrast, cities located in Japan and South Korea showed little change in pre-crisis and post-crisis transport mode choice, maintaining comparatively low levels of air pollution and associated disease risk, and reduced rates of infectious disease transmission throughout the 2020 observation period. Contrasting experiences of road injury in the post-pandemic phase (ie, post 2020) were also observed between these locations.

Interpretation

Our results highlight the transient environmental and health benefits observed during the early stages of the COVID-19 pandemic, driven by substantial reductions in transport-related air pollution and associated health risks due to imposed non-pharmaceutical public health interventions. City design appears to have played a crucial role in observed pollution and health risk differences between cities, with those that afforded a shift away from public and active transport towards private vehicles witnessing a rapid erosion of pollution-related health benefits gained in the entry to mid-crisis phases of the pandemic. These negative effects appear to have also transferred through to increased rates of road trauma in these cities, with a resurgence in road injury above pre-pandemic levels, particularly within countries reliant on private motorised transport. Conversely, cities in Japan, South Korea, and some European regions, which did not experience modal shifts towards cars, sustained their reductions in air pollution and have continued along a trend of declining road transport injuries. These findings underscore city design as a key factor in navigating pandemic-related challenges and suggest that city designs with higher levels of public and mass transit show greater levels of resilience when confronted with infectious disease threats.