TY - JOUR
T1 - Impact of non-pharmaceutical interventions on COVID-19 incidence and deaths
T2 - cross-national natural experiment in 32 European countries
AU - Costa, Diogo
AU - Rohleder, Sven
AU - Bozorgmehr, Kayvan
N1 - Publisher Copyright:
© The Author(s) 2024.
PY - 2024/8/28
Y1 - 2024/8/28
N2 - Purpose Non-pharmaceutical interventions (NPIs) have been the cornerstone of COVID-19 pandemic control, but evidence on their effectiveness varies according to the methods and approaches taken to empirical analysis. We analysed the impact of NPIs on incident SARS-CoV-2 across 32 European countries (March-December 2020) using two NPI trackers: the Corona Virus Pandemic Policy Monitor – COV-PPM, and the Oxford Covid-19 Government Response Tracker – OxCGRT. Methods NPIs were summarized through principal component analysis into three sets, stratified by two waves (C1-C3, weeks 5–25, and C4-C6, weeks 35–52). Longitudinal, multi-level mixed-effects negative binomial regression models were fitted to estimate incidence rate ratios for cases and deaths considering different time-lags and reverse causation (i.e. changing incidence causing NPIs), stratified by waves and geographical regions (Western, Eastern, Northern, Southern, Others). Results During the first wave, restrictions on movement/mobility, public transport, public events, and public spaces (C1) and healthcare system improvements, border closures and restrictions to public institutions (C2) were associated with a reduction in SARS-CoV-2 incidence after 28 and 35-days. Mask policies (C3) were associated with a reduction in SARS-CoV-2 incidence (except after 35-days). During wave 1, C1 and C2 were associated with a decrease in deaths after 49-days and C3 after 21, 28 and 35-days. During wave 2, restrictions on movement/mobility, public transport and healthcare system improvements (C5) were also associated with a decrease in SARS-CoV-2 cases and deaths across all countries. Conclusion In the absence of pre-existing immunity, vaccines or treatment options, our results suggest that the observed implementation of different categories of NPIs, showed varied associations with SARS-CoV-2 incidence and deaths across regions, and varied associations across waves. These relationships were consistent across components of NPIs derived from two policy trackers (CoV-PPM and OxCGRT).
AB - Purpose Non-pharmaceutical interventions (NPIs) have been the cornerstone of COVID-19 pandemic control, but evidence on their effectiveness varies according to the methods and approaches taken to empirical analysis. We analysed the impact of NPIs on incident SARS-CoV-2 across 32 European countries (March-December 2020) using two NPI trackers: the Corona Virus Pandemic Policy Monitor – COV-PPM, and the Oxford Covid-19 Government Response Tracker – OxCGRT. Methods NPIs were summarized through principal component analysis into three sets, stratified by two waves (C1-C3, weeks 5–25, and C4-C6, weeks 35–52). Longitudinal, multi-level mixed-effects negative binomial regression models were fitted to estimate incidence rate ratios for cases and deaths considering different time-lags and reverse causation (i.e. changing incidence causing NPIs), stratified by waves and geographical regions (Western, Eastern, Northern, Southern, Others). Results During the first wave, restrictions on movement/mobility, public transport, public events, and public spaces (C1) and healthcare system improvements, border closures and restrictions to public institutions (C2) were associated with a reduction in SARS-CoV-2 incidence after 28 and 35-days. Mask policies (C3) were associated with a reduction in SARS-CoV-2 incidence (except after 35-days). During wave 1, C1 and C2 were associated with a decrease in deaths after 49-days and C3 after 21, 28 and 35-days. During wave 2, restrictions on movement/mobility, public transport and healthcare system improvements (C5) were also associated with a decrease in SARS-CoV-2 cases and deaths across all countries. Conclusion In the absence of pre-existing immunity, vaccines or treatment options, our results suggest that the observed implementation of different categories of NPIs, showed varied associations with SARS-CoV-2 incidence and deaths across regions, and varied associations across waves. These relationships were consistent across components of NPIs derived from two policy trackers (CoV-PPM and OxCGRT).
KW - Non-pharmaceutical interventions
KW - Covid-19
KW - Europe
KW - Natural experiment
KW - Multi-level modelling
KW - Infectious diseases
UR - http://www.scopus.com/inward/record.url?scp=85202620448&partnerID=8YFLogxK
U2 - 10.1186/s12889-024-19799-7
DO - 10.1186/s12889-024-19799-7
M3 - Article
SN - 1471-2458
VL - 24
JO - BMC Public Health
JF - BMC Public Health
IS - 1
M1 - 2341
ER -