key: cord-0964938-6rqqq0pl authors: Fontán-Vela, Mario; Gullón, Pedro; Padilla-Bernáldez, Javier title: Selective perimeter lockdowns in Madrid: a way to bend the COVID-19 curve? date: 2021-04-08 journal: Eur J Public Health DOI: 10.1093/eurpub/ckab061 sha: 2cb855f4bdb37890c66ecd8485bd9843dd422f0a doc_id: 964938 cord_uid: 6rqqq0pl Lockdowns have been widespread used to limit social interaction and bend the epidemic curve. However, their intensity and geographical delimitation have been variable across different countries. Madrid (Spain) implemented perimeter lockdowns in September with the purpose of bending the COVID-19 curve. In this paper we compared, using join point regressions, the evolution of COVID-19 cases in those areas where this intervention was implemented and those where it was not. According to our analysis, the decrease in the epidemic curve started before the impact of the perimeter lockdown could be reflected. The use of interventions aiming to limit mobility and economic activity has been a widespread non-pharmaceutical intervention (NPI) to control the transmission of SARS-CoV-2. 1 Lockdown interventions have been seen as one of the most effective ways to bend the epidemic curve. 2 However, their intensity and geographical delimitation have been variable. Thus, an individual evaluation of each strategy is required to determine their effectiveness in order to design future NPI to mitigate the effect of COVID-19 transmission. 3 Madrid is one of the most affected european regions throughout the pandemic, and the spanish region with the highest number of cases and deaths: until December 16th, 369,522 COVID-19 cases and 11,619 deaths have been declared, meaning an estimated seroprevalence of 18.6% since the beginning of the pandemic, as shown in a recent report 4 . One study showed that Madrid was the spanish region with the largest drop in weekly life expectancy during the first wave of the pandemic 4 . In September 21st, in the middle of the second wave of the pandemic, Madrid regional government implemented the so-called "perimeter lockdowns" in the Basic Health Zones (BHZ) 6 (the smallest health territorialization units, composed of a mean population of 22,290 inhabitants) where community transmission was reaching critical levels (14-day COVID-19 cases rate per 100,000 > 1,000 on September 15th). This perimeter lockdowns restricted mobility in-and-out of the BHZ except for essential purposes (work or academic activities, caring for dependent people or carrying out administrative procedures) as well as limited capacity of business but without closing restaurants or restricting other types of economic activity. This type of lockdown was only comparable to the "dynamic confinements" in Chile, whose effectiveness has been questioned 7 . Thus, we want to assess whether the trend of 14-day COVID-19 rate changed in the BHZ affected by perimetral lockdown with respect to other BHZ in the city of Madrid that were not affected by the perimetral lockdown. 3 We collected publicly available weekly data on the 14-day COVID-19 rate per 100,000 population between July 7th and 1st December from the open COVID-19 open portal dataset of the Health Department of the regional government. First, we selected all BHZ within the city of Madrid, excluding the rest of the region to avoid the potential bias of the rural areas. From this, we selected the 26 BHZ that were affected by the perimeter lockdowns and a control group of 29 BHZ that had a 14-day We conducted a descriptive analysis of the trends of 14-day COVID-19 rate of the city of In the city of Madrid, the 14-day COVID-19 rate increased in the second wave until a maximum peak of 770.5 cases per 100,000 inhabitants on September 22nd (Figure 1 ). In this analysis, we identified trend changes in the 14-day COVID-19 rate in Madrid using a joinpoint analysis in order to identify if these changes are associated with the selective perimeter lockdowns implemented on September 22nd. The change in the COVID-19 rate trend occurred around the 15th for the entire city of Madrid and the BHZ not affected by the perimeter lockdown, and around the 22th for those BHZ with perimeter lockdown implemented. Considering the incubation period and notification delay, it is required a minimum of 6-9 day lag period since the implementation of a NPI to detect any potential effect 8 . Our results showed that the decrease in the epidemic curve, both in the entire city and the BHZ of the analysis, started before the impact of the perimeter lockdown could be reflected. Also, our results showed that the perimeter lockdowns did not increase the velocity at which cases were decreasing. Thus, the observed decrease might be associated with other NPI implemented in previous weeks, such as social meetings limitation, night clubs' closure, or restaurant capacity limitation. There are many potential reasons why the implementation of perimeter lockdowns were not associated with a significant change in the epidemic curve. First, the curve started to http://mc.manuscriptcentral.com/ejph decrease before measures were taken. Second, perimeter lockdowns were focused on mobility rather than preventing high-risk situations (e.g. indoor activities 9 ). Third, mobility was allowed for essential activities such as working, which represent most of the mobility of residents living in the affected areas 10 . Fourth, boundaries of BHZ were unknown to many residents, as they are used just for healthcare territorialization. As it has been discussed regarding the Chilean experience, the "spillover" effect of neighboring areas that are not under lockdown limits its potential effectiveness in high-interconnected cities 7 . Other possibility is that people changed their behaviour as they anticipated the implementation of more restricting measures. This analysis has some limitations. First, our area-level analysis is sensible to ecological fallacy. Second, there are potential unmeasured confounders that were not under control (e.g. changes in personal awareness or personal attitudes). Third, there is a potential bias due to notification delay, which is more likely to happen when COVID-19 cases are higher. Fourth, our analysis is focused on the city of Madrid, so our conclusions could not be extended to other rural areas of the region. The perimeter lockdown implemented during the peak of the second wave had no effect in the COVID-19 rate trend. Evaluation of NPI is key to elucidate its effectiveness to control the SARS-CoV-2 transmission, and to gather relevant information to design policies during a public health emergency. http://mc.manuscriptcentral.com/ejph Effectiveness of isolation, testing, contact tracing, and physical distancing on reducing transmission of SARS-CoV-2 in different settings: a mathematical modelling study. The Lancet Infectious Diseases Estimating the effects of non-pharmaceutical interventions on COVID-19 in Europe Local lockdowns outperform global lockdown on the far side of the COVID-19 epidemic curve Estudio ENE-COVID: Cuarta ronda. Estudio de sero-epidemiología de la infección por SARS-CoV-2 en España Monitoring life expectancy levels during the COVID-19 pandemic: Example of the unequal impact of the first wave on Spanish regions de la Consejería de Sanidad, por la que se adoptan medidas específicas temporales y excepcionales por razón de salud pública para la contención del COVID-19 en núcleos de población correspondientes a determinadas zonas básicas de salud, como consecuencia de la evolución epidemiológica Effectiveness of Localized Lockdowns in the SARS-CoV-2 Pandemic The temporal association of introducing and lifting non-pharmaceutical interventions with the time-varying reproduction number (R) of SARS-CoV-2: a modelling study across 131 countries. The Lancet Infectious Diseases Mobility network models of COVID-19 explain inequities and inform reopening Movilidad y segunda ola en las áreas urbanas. ¿Dónde trabaja la población de las 37 zonas "confinadas" de Madrid?