Evaluation of the impact of tiered restrictions introduced in England during October and December, 2020, on COVID-19 cases: a synthetic control study

Background In 2020, a second wave of COVID-19 cases unevenly affected places in England, and different levels of tiered restrictions were introduced in different parts of the country. Previous research has examined the impact of national lockdowns on transmission. We aimed to examine the differences in the effect of localised restrictions on COVID-19 cases and how these differences varied by deprivation. Methods We examined the transmission impact of tier 3 restrictions using data on weekly reported COVID-19 cases, adjusted for case-detection rates for 7201 neighbourhoods in England. We identified areas that entered tier 3 restrictions in October and December, 2020, constructed a synthetic control group of places under tier 2 restrictions, and compared changes in weekly infections over a 4-week period. Sufficiently granular data on deaths were not available to investigate excess mortality. We analysed whether this effect varied by level of deprivation and the prevalence of a new variant (B.1.1.7), by stratifying the synthetic control weighting by subgroups and then including an interaction term between subgroup and intervention in the regression model. We used the English Indices of Multiple Deprivation and its tertiles in the stratification to measure deprivation. We tested the spatial spillover effects excluding tier 2 areas within 20 km of tier 3 areas. Ethics approval was not needed. Findings The introduction of tier 3 restrictions was associated with a reduction in infections of 14% (95% CI 10–19) in October and of 20% (13–29) in December, or with a reduction in absolute number of total infections of 3536 (95% CI 2880–4192) in October and of 92 732 (49 776–135 688) in December, compared with what would have been expected under tier 2 restrictions. The effects were similar across levels of deprivation and by the prevalence of the new variant. We found smaller effects with high non-significant p values when excluding boundary areas. Interpretation Compared to tier 2 restrictions, restrictions on hospitality and meeting outdoors in tier 3 areas had a moderate effect on transmission and these restrictions did not appear to increase inequalities in COVID-19 cases. Limitations include a lack of specificity as to which of the main restrictions contributed to this effect, potentially biases from the crude measure of case-detection rates applied, and the lack of controls for individual or household characteristics in this ecological analysis. Funding National Institute for Health Research (NIHR).

Evaluating the impacts of tiered restrictions introduced in England, during October and December 2020 on COVID-19 cases: a synthetic control study*

BackgroundIn 2020, a second wave of COVID-19 cases unevenly affected places in England leading to the introduction of a tiered system of controls with different geographical areas subject to different levels of restrictions. Whilst previous research has examined the impact of national lockdowns on transmission, there has been limited research examining the marginal effect of differences in localised restrictions or how these effects vary between socioeconomic contexts. We therefore examined how Tier 3 restrictions in England implemented between October-December 2020, which included additional restrictions on the hospitality sector and people meeting outdoors affected COVID-19 case rates, compared to Tier 2 restrictions, and how these effects varied by level of deprivation.MethodsWe used data on weekly reported COVID-19 cases for 7201 neighbourhoods in England and adjusted these for changing case-detection rates to provide an estimate of weekly SARS-CoV-2 infections in each neighbourhood. We identified those areas that entered Tier 3 restrictions at two time points in October and December, and constructed a synthetic control group of similar places that had entered Tier 2 restrictions, using calibration weights to match them on a wide range of covariates that may influence transmission. We then compared the change in weekly infections between those entering Tier 3 to the synthetic control group to estimate the proportional reduction of cases resulting from Tier 3 restrictions compared to Tier 2 restrictions, over a 4-week period. We further used interaction analysis to estimate whether this effect differed based on the level of socioeconomic deprivation in each neighbourhood and whether effects were modified by the prevalence of a new more infectious variant of SARS-CoV-2 (B.1.1.7) in each area.ResultsThe introduction of Tier 3 restrictions in October and December was associated with a 14% (95% CI 10% to 19%) and 20% (95% CI 13% to 29%) reduction in infections respectively, compared to the rates expected if only Tier 2 restrictions had been in place in those areas. We found that effects were similar across levels of deprivation and limited evidence that Tier 3 restrictions had a greater effect in areas where the new more infectious variant was more prevalent.DiscussionCompared to Tier 2 restrictions, additional restrictions on hospitality and meeting outdoors introduced in Tier 3 areas in England had a moderate effect on transmission and these restrictions did not appear to increase inequalities.