Background: Little is known about the impact of changes in mobility at the sub-city level on subsequent COVID-19 incidence or the contribution of mobility to socioeconomic disparities in COVID-19 incidence .
Methods: We compiled aggregated mobile phone location data, COVID-19 confirmed cases, and features of the urban and social environments to analyze linkages between population mobility, COVID-19 incidence, and educational attainment at the sub-city level among cities with> 100,000 inhabitants in Argentina, Brazil, Colombia, Guatemala, and Mexico from March to August 2020 . We used mixed effects negative binomial regression to examine longitudinal associations between changes in weekly mobility (lags 1-6 weeks) and subsequent COVID-19 incidence at the sub-city level, adjusting for urban environmental factors . Findings: Among 1,031 sub-cities representing 314 cities in five Latin American countries , 10% higher weekly mobility was associated with 8.5% (95% CI 7.4% to 9.5 %) higher weekly COVID-19 incidence the following week . This association gradually declined as the lag between mobility and COVID-19 incidence increased and was not different from the null at a six-week lag . We found evidence that suggests differences in mobility reductions are a driver of socioeconomic disparities in COVID-19 incidence . Interpretation: Lower population movement within a sub-city is associated with lower risk of subsequent COVID-19 incidence among residents of that sub-city . Implementing policies that reduce population mobility at the sub-city level may be an impactful COVID-19 mitigation strategy that takes equity into consideration and reduces economic and social disruption at the city or regional level.