Objective: to evaluate the effectiveness of strict social distancing measures applied in China in reducing the incidence and mortality from COVID-19 in two Chinese provinces.

Methods: We assessed incidence and mortality rates in Hubei and Guangdong before and after the lockdown period in cities in Hubei. An interrupted time series study was conducted to evaluate the effectiveness of lockdown on reducing the number of cases and deaths from COVID-19. We assumed a slope change following a lag impact model and analysed scenarios with different time-lag periods.

Findings: Daily relative risk reduction in cases was 6.43% (CI -10.25% / – 2.31%) in ascenario with a 17 day time-lag in Hubei. In Guangdong province, this reduction was 8.43% (CI -14.07% / -2.09%) in a scenario with a 7 day time-lag. Daily relative risk reduction in mortality in Hubei was 7.88% (CI -11.06% / -4.39%) in a scenario with a 10 day time-lag.

Conclusion: Strict social distancing measures were effective in reducing incidence and mortality rates. Our results suggest that the onset of reduction effects on incidence and mortality are observed after a period ranging from 7 to 17 days and 10 days, respectively. Effectiveness and the time required for changes seem to be associated with the number of undiagnosed patients and post-lockdown home transmission. Improving epidemiological surveillance may help stakeholders make decisions in a timely manner, enabling the use of more effective and specific social isolation strategies with less economic and social impact.