Allergic rhinitis is the most common IgE-mediated disease and is characterized by nasal symptoms such as nasal congestion, clear runny nose, itchy nose, and sneezing. Air pollutants are known to be an important factor affecting allergic diseases, and ...
Allergic rhinitis is the most common IgE-mediated disease and is characterized by nasal symptoms such as nasal congestion, clear runny nose, itchy nose, and sneezing. Air pollutants are known to be an important factor affecting allergic diseases, and there are several studies that allergic rhinitis is also related to the concentration of air pollutants. However, the association between allergic rhinitis and air pollutants cannot be accurately determined by simple regression analysis due to their monthly differences and seasonality. Therefore, this study aims to investigate the effect of air pollutant concentration on the medical cost of allergic rhinitis through time series analysis.
A time series analysis was conducted from January 2016 to December 2019 to analyze the effect of air pollutant concentration on the medical cost of allergic rhinitis using National Health Insurance data. Total medical costs were defined for medical management and outpatient prescriptions in 25 districts in Seoul. After calculating the monthly concentrations of particulate matter 10 (PM10), particulate matter 2.5 (PM2.5), ozone (O3), nitric dioxide (NO2), sulfur dioxide (SO2), and carbon monoxide (carbon oxide, CO), the effect of each air pollutant concentration on the medical cost of allergic rhinitis was evaluated through a time series analysis using seasonal autoregressive integrated moving average (SARIMA).
The annual number of patients with allergic rhinitis who visited clinics or hospitals in Seoul was 1.46, 1.47, 1.49, and
1.45 million for 2016, 2017, 2018, and 2019, respectively. When monthly differences and seasonality were adjusted through time series analysis, PM10, NO2, and CO concentrations were associated with increased medical costs for allergic rhinitis (p=0.048, p=0.001, and p=0.001, respectively). The increase in allergic rhinitis medical costs increased by 6.22% (95% confidence interval [CI]: 0.0–12.37%), 11.27% (95% CI: 6.03–16.50%) and 11.05% (95% CI: 7.09–15.01%) for one standard deviation increase in PM10, NO2, and CO concentrations.
From the above results, it was found that PM10, NO2, and CO were correlated with medical costs of allergic rhintiis when adjusting monthly differences and seasonality through time series analysis. This analysis revealed quantified economic losses related to the level of air pollutants, and these results may provide a better understanding of the health and economic effects of air pollutants on allergic rhinitis and useful insights for the determination of environmental policies.
In the above study, the association between allergic rhinitis and the air pollutants was evaluated after adjusting the seasonality and monthly differences. Although the seasonality and monthly difference that are important factors in both allergic rhinitis and air pollution were adjusted, the causal effect of air pollutants on allergic rhinitis was not evaluated. The association between air pollutants and chronic rhinitis outpatient visits has been widely assessed in previous studies, but causal inference has not been investigated. Therefore, I attempted to evaluate the causal effect of air pollutants on chronic rhinitis through instrumental variable analysis.
In the studies to evaluate associations, there are one or more confounding factors between the exposure (air pollution) and outcome (disease), which means that there are open causal paths between exposure and outcome. Therefore, in order to properly estimate the causal effect between exposure and outcome, potential confounding factors must be adjusted in the model or randomized through experiments. In studies that investigate environmental effects epidemiologically, it is virtually impossible to randomize, and instead, pseudo-randomization can be performed through statistical techniques such as instrumental variable analysis. Instrumental variable analysis is the evaluation of causal reasoning in outcomes using instrumental variables that can limit specific exposures. The important concept here is that the instrumental variable itself does not affect the outcome, but only through the exposure variable. TI affects chronic rhinitis only through air pollutants, and TI is used as an instrumental variable to analyze the causal relationship between air pollution and the number of outpatient visits for chronic rhinitis in this study.
This data was conducted based on environmental disease data provided by the National Health Insurance Service, and the subjects analyzed in the study were patients who visited hospitals in Seoul due to rhinitis from January 1, 2014 to December 31, 2017. The definition of rhinitis patients in the environmental cohort was based on the ICD-10 disease code. The air quality index (AQI) was calculated for the degree of air pollution using air pollutants PM10, PM2.5, O3, NO2, SO2, and CO, and the average AQI value was used as an exposure variable. The TI data measured at radiosonde in the Osan area were used as instrumental variables.
The number of outpatient visits due to chronic rhinitis was 81,210,447, and the analysis was performed based on this. The relative risk of one IQR increase in AQI was significant in lag 0,3,5,6 (p < 0.01) with the largest relative risk (RR) at lag 0, followed by lag 6 (lag 0: RR 1.078, 95% CI 1.045–1.113; lag 6: RR 1.043, 95% CI 1.037–1·087). In Subgroup analysis, the age group 10-19 was the most vulnerable group and had significant RR for lag 0-7 (RR 1.039-1.161). Negative control outcome and exposure analysis demonstrated instrument variable analysis provided a robust estimation of causal effect of AQI on the number of outpatient visits to chronic rhinitis.
This analysis using TI as an instrumental variable showed the causal relationship between the increase in air pollutant level and the number of chronic rhinitis outpatient visits. This analysis suggested a meaningful methodology as it can show the causal effect of air pollution, not merely the association between air pollutants and diseases that can be involved with many confounding variables.