The short-term effects of air pollution on respiratory disease hospitalizations in 5 cities in Poland: comparison of time-series and case-crossover analyses

Alessandro Slama,Olga Turżańska-Wieczorek,Marcin Studnicki,Andrzej Śliwczyński,Bartłomiej Wiśnicki,Waldemar Wierzba,Jakub Kubajek,Jolanta Woźnica-Pyzikiewicz,Edward Franek,Maciej Zdrolik

doi:10.1007/s11356-020-08542-5

Abstract

Very few publications have compared different study designs investigating the short-term effects of air pollutants on healthcare visits and hospitalizations for respiratory tract diseases. This study describes, using two different study designs (a case-crossover design and a time-series analysis), the association of air pollutants and respiratory disease hospitalizations. The study has been conducted on 5 cities in Poland on a timeline of almost 4 years. DLNM and regression models were both used for the assessment of the short-term effects of air pollution peaks on respiratory hospitalizations. Both case-crossover and time-series studies equally revealed a positive association between air pollution peaks and hospitalization occurrences. Results were provided in the form of percentage increase of a respiratory visit/hospitalization, for each 10-μg/m3 increment in single pollutant level for both study designs. The most significant estimated % increases of hospitalizations linked to increase of 10 μg/m3 of pollutant have been recorded in general with particulate matter, with highest values for 24 h PM2.5 in Warsaw (6.4%, case-crossover; 4.5%, time series, respectively) and in Białystok (5.6%, case-crossover; 4.5%, time series, respectively). The case-crossover analysis results have shown a larger CI in comparison to the results of the time-series analysis, while the lag days were easier to identify with the case-crossover design. The trends and the overlap of the results occurring from both methods are good and show applicability of both study designs to air pollution effects on short-term hospitalizations.

Highlights

Responsible Editor: Philippe GarriguesAtmospheric pollution is representing globally the highest environmental risk for the human health
Very few studies have compared the output of the two methodologies (Lin et al 2002; Yuming et al 2010; Kayo et al 2009; Tong et al 2012; Meng et al 2017; Zheng et al 2015), in particular on the respiratory hospitalizations and healthcare visits associated with air pollution
This paper compares the already published results of the impact of air pollution on hospital admissions with a focus on respiratory diseases performed with a time-series multicity analysis (Slama et al 2019) with a case-crossover design performed on the same dataset

Summary

Introduction

Responsible Editor: Philippe GarriguesAtmospheric pollution is representing globally the highest environmental risk for the human health. “The time series, case-crossover, and panel studies are best suited for estimating the acute effects of air pollution” (Peng and Dominici 2008). When it comes to evaluating hospitalizations or healthcare visits or mortality as outcomes of a given study, the data structure and the type of effects that need to be studied typically lead to the model selection. While originally most of such studies (Fung et al 2003; Hajat 2003) used time-series analyses, progressively case-crossover studies have become more and more the alternative analytical approach (Duan et al 2016). This paper compares the already published results of the impact of air pollution on hospital admissions with a focus on respiratory diseases performed with a time-series multicity analysis (Slama et al 2019) with a case-crossover design performed on the same dataset

Methods

Results

Discussion

Conclusion