Heterogeneous Variations in Human Mobility Influenced by COVID-19 Interventions and Risk Perceptions in the UK

Human mobility exhibited significant variations as movement patterns fluctuated in response to evolving public health restrictions and health risks during the COVID-19 pandemic. Non-pharmaceutical interventions (NPIs), such as national lockdowns and Work-From-Home (WFH) recommendations, were implemented in the UK to reduce mobility and control the spread of the virus. However, existing studies have not adequately accounted for the evolving nature of public health risk perceptions or the regional and socio-demographic disparities in mobility responses. This research proposes a novel data-driven methodology that quantifies time-varying health risk perceptions and links these perceptions with selected mobility data, applying a causal machine learning approach. The methodology advances traditional parametric regression by synthesising time-varying risk perceptions and enhances policy evaluation through machine learning, enabling a more nuanced understanding of mobility variations. Case studies of three national lockdowns in the UK revealed varied effectiveness across mobility types, lockdown periods, and durations. Notably, the first lockdown showed non-monotonic mobility variations within three months, with lockdown effects peaking initially and fluctuating in response to government announcements. The case study on three UK cities demonstrated that the WFH recommendation post-lockdown had uneven effects across socio-demographic groups. These findings challenge some initial studies that overestimated the effects of NPIs, showing that mobility responses are shaped not only by NPIs but also by evolving public health risk perceptions. In summary, the impact of NPIs on human mobility varies across different periods and socio-demographic groups, providing evidence of heterogeneous policy effects in the UK. These insights and proposed analytic framework could help design adaptive, equitable, and targeted public health policies
that account for mobility variations and regional disparities, ensuring sustained policy effectiveness and minimising unintended costs during future crises.