Electoral processes and COVID-19 infections in Japan

Election campaigns and polling stations might be areas of potential infections; however, it is unknown whether elections can impact the spread of coronavirus disease (COVID-19). This study aimed to evaluate the association between election process and COVID-19 infection in Japan. Quasi-experimental design using time-series COVID-19 cases data with an election intervention. We chose the cities where elections for major and/or city/ward assembly members were held in 2021 January in Japan. The study period for each city spans one weeks before the start date of the election campaign and three weeks after the voting day (36 days). The daily number of patients testing positive for COVID-19. We used Poisson regression analysis with sandwich estimator to evaluate the association between election and the spread of COVID-19 infections. For Miyakojima City, we assessed the relationship between election and COVID-19 infections by using the Box-Jenkins autoregressive integrated moving average (ARIMA) model. We also estimated the instantaneous reproduction number (Rt) in the city. There were 17 cities that met the inclusion criteria. In all models with three types of different lag effects (i.e., 4, 9, and 14 days), election was not a significant predictor of COVID-19 infections in the 17 cities. For Miyakojima City, the autoregressive integrated moving average (0, 0, 0) with a lag of 14 (=14) was the best model. The partial coefficient ({omega}) of the election was 11.91 (95% confidence interval: 3.10-20.72, P<0.001), indicating that the election was associated with an increase in the number of COVID-19 cases 14 days after the election campaign (mean: 12 cases/day). Rt hovered far above 1 during and after the election campaign in Miyakojima. Although elections were associated with an increased number of COVID-19 cases in Miyakojima, this association was not verified in the analysis including all 17 cities. Therefore, if preventive measures prescribed by election guidelines are followed, elections do not necessarily relate to a spread of COVID-19 as long as the election process does not involve activities necessitating close contact.