Download:
-
Air pollution was among the leading ten health risk factors in 1990 and remained an important factor as of 2016 (1). Since air pollution control is a long-term challenge, China is likely to encounter more air pollution related health issues going forward. In environmental health risk communication aimed at protecting public health, an air quality index will play a significant role.
The United States Environmental Protection Agency (EPA) started using the Pollutant Standards Index (PSI) to report the daily air quality index in 1976. Since 1999, the EPA replaced the PSI with the air quality index (AQI) (2). This standard is now commonly adopted worldwide. In communicating air pollution health risks, the AQI avoids confusion caused by listing various pollutant concentrations. However, AQI is not intended to indicate health risks of air quality, so the usefulness of the AQI is inherently limited. Briefly, the AQI calculates air quality based on limits for different pollutants. Air quality standards are based on a variety of factors, including technical and economic accessibility as air quality management objectives. Nevertheless, AQI could not reflect the non-threshold concentration-response relationship between air pollutant(s) and health consequence(s). Moreover, AQI does not reflect the combined health effects of many air pollutants.
To better communicate health risks of air pollution to the general public, Canada considered some health risk parameters and introduced a new index system called the air quality health index (AQHI) (3). Based on the Canadian approach, Hong Kong, China developed an AQHI reporting system in 2013 (4). More recently, some studies in China discussed the construction of an AQHI from scientific perspectives for a single city, multiple cities, and nationwide. These studies suggested that the AQHI could better reflect the health risks of air pollution in China. Nevertheless, these studies covered few areas or were insufficiently validated because of data scarcity and most have not released to the public.
In order to develop a Chinese AQHI reporting system, the National Health Commission (NHC) and China CDC commissioned the National Institute of Environmental Health to study the development of a Chinese AQHI. This report introduced the construction method of Chinese AQHI to guide unified application of AQHI throughout China.
Chinese AQHI was constructed by referencing and improving on the approach of Stieb et al. (3). During the construction of AQHI, three key points were the included pollutants, preferred health outcomes, and chosen scaling parameter.
The air pollutants of AQHI in this report include PM2.5, O3, NO2, and SO2 (the rationale is discussed in the Supplementary Material). The descriptive analysis of national county-level pollutants was shown in Table 1.
Variable Effective days Mean ± SD Min P25 Median P75 Max Pollution (μg/m3) PM2.5 1,332,272 51.1 ± 43.3 3.0 23.9 39.0 63.7 806.6 NO2 1,333,267 34.3 ± 20.7 1.0 19.1 29.9 44.9 471.1 SO2 1,335,313 22.9 ± 26.7 2.9 8.7 14.7 26.4 715.7 O3 1,326,722 59.6 ± 31.8 2.1 35.7 55.4 79.0 625.3 AQHI 1,307,118 4.0 ± 1.7 1 3 4 5 10+ Abbreviation: SD, Standard deviation; Min, minimum during the study period; P25, first quartile; P75, third quartile; Max, maximum during the study period. Table 1. Descriptive analysis of county-level pollutants and air quality health index (AQHI) of China, 2013–2018.
Mortality was chosen as the health outcome due to its severity and because mortality data are more objective and have better quality in China. The burden of disease attributable to air pollution in the Global Burden of Disease study and the WHO air quality guidelines for particulate matter were also calculated based on mortality data (5). The chosen outcome was non-accidental mortality to reflect the overall health impact.
The AQHIs were scaled by dividing by the 99th percentile for the time period, instead of dividing by the maximum value like other studies. Pollution levels vary widely from region to region in China, and some regions had incidents of heavy pollution events. If divided by the maximum value, the indices in many areas would present low risk, which would mislead the users of this index.
Based on the above consideration, the Chinese AQHI was constructed. Hourly concentrations of PM2.5, O3, SO2, and NO2 were collected from 2013 to 2018 in 769 counties of China (Figure 1). The regression coefficient β was derived from the regression models relating air pollutants (PM2.5, O3, SO2, and NO2) (6-9) to mortality in China (Table 2).
Figure 1.Distribution of 769 counties of the national air quality health index study in China from 2013 to 2018.
Pollutants β value PM2.5 0.00022 O3 0.00024 NO2 0.00090 SO2 0.00059 Table 2. Exposure-response relationships of air pollution and non-accidental mortality.
Daily excess percentage risks of mortality relative to zero concentration, for non-threshold exposure-response relationship between air pollutant(s) and health consequence(s), of exposure to each pollutant in each day from 2013 to 2018 were calculated (Formula 1). Then the sum of daily excess percentage risks of mortality of the 4 pollutants at t day was calculated, then the 99th percentile at each study site was calculated (Formula 2). AQHI was separated into 11 levels of 1–10 and 10+ at each county at any day t in the past and the future (Formula 3).
$$ {ER}_{ijt}=100\times \left[exp\left({\beta }_{i}\times {x}_{ijt}\right)-1\right] $$ (1) where
$ {ER}_{ijt} $ represents the daily excess percentage risk of mortality associated with the pollutant i in j county at t day,$ {\beta }_{i} $ was the regression coefficient of pollutant i from previous studies,$ {x}_{ijt} $ was the mean concentration of pollutant i in j county at t day.$$ E{R}_{P_{99}}=P_{99}{\begin{array}{c}j=1 \cdots n\\ t=1\cdots m\end{array}}\left[{\Sigma }_{i=1\cdots q}\right(E{R}_{ijt}\left)\right] $$ (2) where
${ER}_{P_{99}}$ represents the 99th percentile of daily sum of excess percentage risk of mortality associated with PM2.5, O3, NO2, and SO2 in each county of all 769 counties at each day from 2013 to 2018. The result of${ER}_{P_{99}}$ was 20.04.$$ {AQHI}_{jt}=(E{R}_{jt}\times 10)/E{R}_{P_{99}} $$ (3) where
$ {ER}_{jt} $ represents the daily sum of excess percentage risks of mortality associated with the PM2.5, O3, NO2, and SO2 in j county at t day.Table 1 summarized the descriptive analysis of county-level pollutants and AQHI during 2013−2018 in China. The averaged levels of AQHI were 4. The frequency distribution of AQHI in 36 cities (all municipalities directly under the central government, cities specifically designated in the state plan, and provincial-level capitals of China) from 2013–2018 was shown in Supplementary Table S1.
For the convenience of public communication, AQHI values (1–10+) were grouped into 4 levels: low health risk (1–3), moderate health risk (4–6), high health risk (7–10), and very high health risk (10+). According to different health risk levels (including low, moderate, high, and very high health risk levels), health protective messages will be provided to the general population, patients with cardiopulmonary disease, and sensitive populations (including the elderly, pregnant woman, and children), respectively.
HTML
-
This report introduced the method of calculating the Chinese AQHI. This AQHI was established based on daily PM2.5, O3, NO2, and SO2 concentration of 769 counties and the exposure-response relationship between air pollution and mortality. The AQHI should be an effective tool to reflect and communicate the health risk of air pollution with the public in China.
Compared with research from Du et al. (10), this report had several differences: different research regions, different included pollutants, and different index of standardization. This report included almost all counties with national air pollution monitoring stations in China, which provided higher regional resolution and better data. Considering the concentration of SO2 is still relatively high in some regions of China, our report also included this pollutant to be a better indicator of health risks.
While several factors were considered based on the current data, this method has several limitations. First, a uniform set of parameters and formulas across the nation could cause bias in some regions. Using regional exposure-response relationships and combining certain pollutants could provide a more accurate indication of the health risks of short-term exposure to air pollution. Nonetheless, in public communication, getting the index understood and avoiding conflicting results in different regions is more important than finding the best index. Hence, an AQHI was constructed with the idea of simplifying the information. Second, the β value used in this study was extracted from ecological studies which might contain inherent biases. In future studies, the index should be updated with better exposure-response relationships.
Additional validation of the AQHI is required. More relevant health data will be collected to further verify, evaluate, and optimize the index. Furthermore, more attention should be focused on the evaluation of the effectiveness of the AQHI application. To evaluate the effectiveness of the AQHI implementation, relevant data of differences-in-differences estimates before and after the AQHI application in county with and without published AQHI data will be collected in the next few years. The index will be adjusted and optimized based on the validation results
Fundings: National Key Research and Development Program of China (2016YFC0206500); National High-level Talents Special Support Plan of China for Young Talents; and Beijing Natural Science Foundation (7172145).
| Citation: |
|