Vegni Ferdinando E, Castelli Beatrice, Auxilia Francesco, Wilkinson Paul
London School of Hygiene and Tropical Medicine, Public Health and Policy, Keppel st., London WC1E 7HT, UK.
Eur J Epidemiol. 2005;20(4):351-8. doi: 10.1007/s10654-005-0243-4.
STUDY OBJECTIVE AND SETTING: Particulate air pollution has a known negative effect on human respiratory health, often studied with hospital admissions, emergency room access, or mortality as health indicators. We evaluate respiratory drug dispensing data as health indicators for the effects of total suspended particles (TSP) air pollution in the city of Como (84,713 inhabitants).
Weekly count of individual patients with respiratory drug dispensed (Cases) and weekly dispensed daily defined doses (DDD) of drugs were modelled with weekly air mean concentrations of TSP using a Poisson regression model adjusted for long-term trends, seasonal variations, calendar variations due to holidays, and weather.
Relative risks (RR) were expressed for a variation from 10th to 90th percentile of TSP (29-92 microg/m3). Weekly aggregation was used in consideration of the complexity of drug dispensing data and potential biases of daily aggregation. For weekly mean concentrations of TSP, RR = 1.082 (95% Confidence level (CI) 1.002-1.169) for Cases and RR = 1.137 (95% CI 1.044-1.238) for DDD.
Our study concludes that both Cases and DDD of dispensed respiratory drugs could be useful for epidemiological surveillance of air pollutant health effects. Further investigation may routinely allow health and economics considerations, producing a stimulating new tool for health policy makers.
研究目的与背景：已知颗粒空气污染对人类呼吸健康有负面影响，通常以医院入院人数、急诊就诊情况或死亡率作为健康指标来进行研究。我们评估了呼吸药物配药数据，将其作为科莫市（84,713名居民）总悬浮颗粒物（TSP）空气污染影响的健康指标。
使用泊松回归模型，对每周呼吸药物配药的个体患者数量（病例数）和每周药物的每日限定剂量（DDD）配药量进行建模，该模型针对长期趋势、季节变化、节假日导致的日历变化以及天气进行了调整，模型的自变量为每周TSP的空气平均浓度。
TSP从第10百分位数到第90百分位数（29 - 92微克/立方米）变化时的相对风险（RR）。考虑到药物配药数据的复杂性和每日汇总可能存在的偏差，采用了每周汇总的方式。对于每周TSP平均浓度，病例数的RR = 1.082（95%置信区间（CI）1.002 - 1.169），DDD的RR = 1.137（95% CI 1.044 - 1.238）。
我们的研究得出结论，配给的呼吸药物的病例数和DDD都可用于空气污染对健康影响的流行病学监测。进一步的调查可能会经常纳入健康和经济方面的考虑因素，为卫生政策制定者提供一个令人振奋的新工具。
