key: cord-0711988-lixejzu2 authors: Gayat, Etienne; Raux, Mathieu title: From descriptive epidemiology to interventional epidemiology: The central role of epidemiologists in COVID-19 crisis management date: 2022-05-03 journal: Anaesth Crit Care Pain Med DOI: 10.1016/j.accpm.2022.101056 sha: 08a71f6030684b0c227a56d64b95dca77751e6b2 doc_id: 711988 cord_uid: lixejzu2 nan From descriptive epidemiology to interventional epidemiology: The central role of epidemiologists in COVID-19 crisis management The topic of epidemiological modelling in the context of the COVID-19 pandemic has not only been central to many scientific developments but also, and perhaps most importantly, has had very strong direct implications for crisis management. The search in the PubMed database with the keyword ''epidemiological model COVID'' retrieved 1669 scientific publications published between September 2020 and February 2022! In the April issue of Anaesthesia Critical Care & Pain Medicine is published a series of updates describing the extent to which clinical epidemiology has been an indispensable tool in the management of the health crisis in France, in Europe and more generally, around the world. The COVID-19 pandemic has certainly been exceptional, but it is not the first global health crisis in our modern history. Indeed, the H1N1 pandemic in 2009 [1] and the Ebola epidemic in West Africa in 2013-2016 [2] were recent opportunities to increase the interest in epidemiological modelling, as mentioned by Cré pey et al. in their contribution to this issue [3] . The compelling need to develop epidemiological models was first identified in a report by scientists at Imperial College at the end of February 2020 [4] , as discussed by Sofonea et al. in this issue [5] . During the COVID-19 pandemic, epidemiological models have become the cornerstone of crisis management, as they made it possible to better understand the spread of SARS-CoV-2, to predict future risk and to assess control strategies. These models have allowed us to stay one step ahead and thus to better adapt our capabilities for the management of COVID-19 patients. From then on, epidemiological tools have been used not only to describe the pandemic but also to support crisis management decisions. Table 1 summarises the main fields of interventions for epidemiological modelling in the context of the management of the COVID-19 pandemic. Numerous epidemiological studies have focused on determining the exposures associated with COVID-19 infections [6] , with the idea of obtaining decision support tools for enhanced and targeted control of these exposures to limit the spread of the virus. As pointed out by Colosi et al., school closures have been a widely used measure in the first year of the pandemic for this purpose [7] . However, the impact of these school closures has not been negligible for children's well-being and mental health, and modelling work specifically studying school closures has helped determine a balance between controlling the spread of the virus and maintaining a satisfactory educational offer. COVID-19 pandemic's impact on hospitals was threefold: an influx of patients with COVID-19, a significant risk of nosocomial infections, and an impact on health care professionals, causing absenteeism. Particular efforts regarding modelling nosocomial transmission have been made, allowing for a better understanding as discussed by Opatowski et al. in their contribution to this issue [8] . Other epidemiological studies have rapidly alerted researchers to the risk of nosocomial infections, particularly in patients admitted for surgery, where the acquisition of a COVID-19 infection in the hospital is associated with an increase in morbidity and mortality [9] . These results led to recommendations for a preoperative screening strategy [10] . Finally, mathematical modelling was also used to unravel some key aspects of host-pathogen interactions [11] . Within-host models of SARS-CoV-2 have been used to model viral kinetics and its impact on virus spread and the theoretical effects of new treatments. However, it should be noted that epidemiological models have limitations. The main limitation is that they are all based on assumptions. Even if these assumptions are based on previous knowledge, data and observations, they may be partially incorrect, requiring an update of the model. Thus, the main assumptions considered were the contagiousness and severity of the virus strain, the effects of mitigation measures, the effect of natural immunity, the effect of immunity induced by vaccination and the effectiveness of medical management. It has been clearly demonstrated that all these variables have evolved over time, requiring constant adaptations of the models. 2. Open access to (good) data: the key to success! Building an epidemiological model requires access to robust epidemiological data. To be useful, the data that is made available must be of good quality, i.e., unbiased, understandable and frequently updated. Very quickly, many countries gave access to the data produced by their information systems. France has been particularly active in this area, making hospitalisation data (SIVIC database), test data (SI-DEP database) and vaccination data (COVID-19 vaccine database) available via the https://www. data.gouv.fr website. Thus, many teams in France and teams around the world have contributed to the epidemiological modelling of COVID-19. Far from being an unproductive competition, a worldwide collaboration has been created, even including the creation of a space for sharing different modelling approaches (for example, the European COVID-19 Forecast Hub, https://covid19forecasthub.eu). Beyond epidemiological modelling, open science has also led to the sharing of technological solutions facilitating crisis management, particularly through decision-making tools for health care structures, as discussed by Garaix et al. [12] . After the initial shock of the first waves successively affecting different continents, epidemiological modelling was used to model the maintenance of the non-COVID-19 health care offerings. Indeed, the management of the first waves of COVID-19 patients was largely to the detriment of the management of non-COVID-19related pathologies. The impact on the prognosis of non-COVID-19 patients has been modelled; the example of the delay in diagnosis of breast cancer is particularly noteworthy. Using the theoretical frameworks of three established cancer intervention and surveillance modelling network breast cancer models, Alagoz et al. predicted that the initial pandemic-related disruptions in breast cancer care will have a small but significant long-term cumulative impact on breast cancer mortality [13] . Here, again, epidemiological modelling has contributed to the reflection on the adaptation of care offerings with the goal of maintaining non-COVID-19 care offerings in parallel with the adaptation of the care system for the management of patients infected by SARS-CoV-2. Epidemiological modelling provides reliable and objective information for decision-makers to consider the different options. However, as pointed out by Crepey et al. [3] , ''making the choice remains a political decision''. In fact, the decision-making process not only considers the need to face the pandemic and to take the best possible care of our patients but must also consider other realities: social and psychological consequences, consequences on the management of non-COVID-19 diseases, etc. To put it another way, epidemiological modelling was one of the tools that allowed decision-makers to manage the constant uncertainty during this crisis. More than two years after the beginning of the pandemic, our health system response has improved thanks to progress in treatments, testing strategies, prevention (through social distancing and especially through vaccination) and the improvement of our capacity to model the pandemic. It is in this respect that epidemiologists, such as physicians, nurses and health system managers, have made major contributions to the fight against COVID-19. The authors declare that they have no competing interest. Definition of a threshold for triggering the plan Definition of the number of people to be evacuated Decision support for COVID-19 vaccination policies Determining priority audiences Development of a booster dose administration strategy To predict the effects of pharmacological interventions using within-host models of SARS-CoV-2 Modeling the epidemic waves of AH1N1/09 influenza around the world Ebola virus disease in West Africa -the first 9 months of the epidemic and forward projections Challenges for mathematical epidemiological modelling Report 9: impact of non-pharmaceutical interventions (NPIs) to reduce COVID19 mortality and healthcare demand Epidemic models: why and how to use them Exposures associated with SARS-CoV-2 infection in France: a nationwide online casecontrol study Modeling COVID-19 in school settings to evaluate prevention and control protocols Contributions of modeling for the control of COVID-19 nosocomial transmission Mortality and pulmonary complications in patients undergoing surgery with perioperative SARS-CoV-2 infection: an international cohort study Guidelines: anaesthesia in the context of COVID-19 pandemic Within-host models of SARS-CoV-2: what can it teach us on the biological factors driving virus pathogenesis and transmission? Vé ber A. Decision-making tools for healthcare structures in times of pandemic Impact of the COVID-19 pandemic on breast cancer mortality in the US: estimates from collaborative simulation modeling b Department of Anesthesia and Critical Care Medicine, AP-HP, Hôpital Lariboisie`re None.