![Sample our Bioscience journals, sign in here to start your access, Latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5925/TOC-Subject-Sample-2021-Bioscience.jpg"})
Abstract
Two statistical regression models-the autoregressive integrated moving average (ARIMA) and the generalized additive model (GAM)-were compared in terms of their ability to assess the association between ambient particulate air pollution concentrations (PM10) and daily hospital admissions for chronic obstructive pulmonary disease (COPD) in Reno-Sparks, Nevada, during the period 1990 to 1994. The study involved 3115 admissions for COPD. The daily average concentration of ambient PM10 was 36.55 μg/m3. After being adjusted for the effects of weather, day of the week, season, and time trend, both the ARIMA and the GAM methods consistently found that PM10 is a statistically significant predictor of daily hospital admissions for COPD. The percentage increase in hospital admissions for COPD for an interquartile increase (26.6 μg/m3) of the 24-h average of PM10 on the 14 prior days is 4.29% (95% CI 1.22∼7.36%) with ARIMA analysis and 5.62% (95% CI 2.16∼9.08%) with GAM analysis. The percentage increase in hospital admissions for COPD for an increase of 26.6 μg/m3 of the 24-h average of PM10 on the concurrent day was 4.73% (95% CI 0.88∼8.58%) with GAM analysis. Comparisons of the two methods showed that GAM analysis was more sensitive than ARIMA analysis in predicting correlations between PM10 levels and hospital admissions for COPD. However, the ARIMA method is a more powerful technique for dealing with the problems of high-order autocorrelation errors.