La problématique de l’environnement-santé n’est pas nouvelle. De très nombreux auteurs (Beelen et al., 2014 ; DGS 2020 ; Zanobetti et al., 2008) se sont intéressés à l’effet – à plus ou moins long terme – sur la survenue de pathologies. Bien souvent, ces études travaillent à une échelle géographique nationale. Nous avons pris le parti de regarder les liens entre trois polluants atmosphériques à savoir le dioxyde d’azote, l’ozone et les particules (PM10) sur la survenue d’une exacerbation de bronchopneumopathies chroniques obstructives (BPCO) dans la ville de Nice, sur la période allant de janvier 2012 à décembre 2018, soulignant ainsi une accentuation des fragilités des humains évoluant sur ces territoires. Pour ce faire, nous nous sommes appuyés d’une part sur les données du PMSI et d’autre part sur des données recueillies par les différentes stations de surveillance de la qualité de l’air, provenant du réseau de surveillance de la qualité de l’air AtmoSud. Pour étudier ce lien statistique et ses conséquences, nous avons utilisé un modèle de régression de Poisson basé sur un modèle linéaire généralisé. Ce modèle a été essentiellement alimenté par deux indicateurs que nous avons construits. Le modèle de régression permet ainsi d’estimer le coefficient associé à notre variable explicative et de calculer un risque relatif. Ainsi, nous avons pu mettre en évidence un lien significatif entre deux de nos polluants et notre évènement sanitaire, à savoir la survenue d’une exacerbation de BPCO.
The environment-health issue is not new. Many authors (Beelen et al., 2014; DGS 2020; Zanobetti et al., 2008) have looked at the effect —in the long term— on the occurrence of pathologies. Very often, these studies work on a national geographical scale. We decided to look at the links between three atmospheric pollutants, namely nitrogen dioxide, ozone, and particulate matter (PM10), on the occurrence of an exacerbation of chronic obstructive pulmonary disease (COPD) in the city of Nice, over the period from January 2012 to December 2018, thus underlining an accentuation of the fragility of humans evolving in these territories. To do this, we relied on the one hand on PMSI data and on the other hand on data collected by the different air quality monitoring stations, from the AtmoSud air quality monitoring network. To study this statistical relationship and its consequences, we used a Poisson regression model based on a generalized linear model. This model was essentially fed by two indicators that we constructed. The regression model thus allows us to estimate the coefficient associated with our explanatory variable and to calculate a relative risk. Thus, we were able to show a significant link between two of our pollutants and our health event, namely the occurrence of a COPD exacerbation. From two databases, we built our statistical model to estimate a possible relationship between exposure to an air pollutant and a COPD exacerbation. The model chosen was a Generalized Linear Model (GLM) of the counting type (Poisson regression), taking into account overdispersion, i.e., when the variance of the observed data is greater than the mean of the variable.) Our unit of observation was the day. With this in mind, we constructed two indicators: - A health indicator: this is our variable to be explained. It corresponds to the number of daily admissions for a COPD exacerbation. The data for its construction were extracted from the PMSI files according to the main cause of hospitalization, recorded at hospital discharge. The model expresses the neperian logarithm of the expectation of our health indicator. - A pollution indicator: This is our explanatory variable. Initially, to model the population's exposure to ambient air pollution, our hypothesis was that the measurements recorded by AtmoSud are an unbiased estimate of the average of the individual values of the exposure of the individuals composing our study population. In practice, this assumption translates into the use of the arithmetic mean of the daily concentrations of the different pollutants recorded by the different measuring stations located in Nice (the daily values of a station being, themselves, the means of the 24 hourly measurements recorded by the station). This average constitutes the ambient exposure indicator which, each day, is attributed to the whole population. The indicator was constructed using data from daily measurements of pollutants from the various air quality monitoring stations in Nice. It is in micrograms per cubic meter (μg/m3). In a second step, this indicator underwent a transformation (of the binary type) to test the possible links with values taken as threshold (the 50 percentiles, the 75 percentile and the EU limit value for the protection of human life). The indicator was then set to “1” if it is above the threshold values and “0” if it is not. Our pollution indicator was tested on a set of lags (from 0 to 5 days, called Lag0 to Lag5) to assess possible statistical relationships on the same and previous days.Chronic obstructive pulmonary disease, a chronic inflammatory respiratory disease, is often not well known to the public. However, it leads to numerous hospitalizations and deaths each year due to its exacerbation. The main risk factor for this disease is smoking (active or passive), but other factors also increase the risk of developing this disease, notably indoor and outdoor air pollution. Our study highlighted this disease and established a link with certain air pollutants (the three main ones in the city). We found that increased concentrations of nitrogen dioxide could increase the number of admissions for COPD exacerbation. This link was established with concentration values well below the values defined by the EU for health protection. This could raise the question of re-evaluating these values or creating regionalized reference values. According to our results, ozone seems to have a protective effect on COPD exacerbations in our region. However, we could not establish a significant link between COPD exacerbation and particulate matter (PM10), although PM10 concentrations in Nice exceeded the European guideline values most of the time.
