Study of the Effects of Air Pollutants on Human Health Based on Baidu Indices of Disease Symptoms and Air Quality Monitoring Data in Beijing, China

Shaobo Zhong, Zhichen Yu, Wei ZhuInternational Journal of Environmental Research and Public Health 2019 · DOI

TL;DR

This study leverages Baidu search trends as a proxy for public health concerns related to air pollution in Beijing. By combining this novel data source with traditional air quality and meteorological data, researchers developed a new health-focused air quality index (HAQI) that better reflects the health risks of air pollutants compared to the existing AQI.

What Problem Does This Paper Address?

Assessing the real-time health impacts of air pollution is crucial, but traditional methods relying on hospital admissions or mortality data face challenges. These include data acquisition difficulties, limited sampling, and potential delays in reporting. The authors aim to address this by using a more immediate indicator of public health concern: internet search behavior. Specifically, they want to quantify the short-term health effects of major air pollutants on respiratory and cardiovascular/cerebrovascular systems and propose a more health-sensitive air quality metric.

Methods

The researchers employed a Generalized Additive Model (GAM), a semiparametric approach well-suited for modeling complex, non-linear relationships. They used a GAM with a log link and Poisson error distribution, assuming that increased disease symptoms lead to increased internet searches. Key methodological aspects include:

  • Baidu Index as a Proxy: They utilized the Baidu Index, which reflects the popularity of search queries, as a surrogate for disease symptom reporting. They selected 25 keywords related to respiratory and cardio-cerebrovascular diseases.
  • Time Lag Analysis: Recognizing that health effects might not be immediate, they incorporated a three-day lag period for air pollutant exposure in their models.
  • Relative Risk (RR) Calculation: They calculated the relative risk of health outcomes associated with pollutant exposure, using the inter-quartile range (IQR) of pollutant concentrations as the unit of exposure.
  • Health Air Quality Index (HAQI) Development: Based on the calculated RRs, they proposed a new index, HAQI, designed to better represent the health burden of air pollution compared to the standard Air Quality Index (AQI).

Datasets

  • Air Quality Data: Daily concentrations of PM<sub>10</sub>, PM<sub>2.5</sub>, SO₂, O₃, NO₂, and CO, along with the daily Air Quality Index (AQI), were obtained from the Ministry of Environmental Protection (MEP) website for Beijing.
  • Meteorological Data: Hourly data for mean wind level, mean air temperature, and mean relative humidity were collected from meteorological stations in Beijing.
  • Baidu Indices: Daily search query frequencies for 25 keywords related to respiratory and cardio-cerebrovascular diseases were obtained from Baidu.com.

The study period covered March 1, 2015, to April 30, 2017.

Key Insights and Ideas

  • Internet Searches as a Health Indicator: The study demonstrates the potential of using aggregated internet search data (Baidu Index) as a timely and broad indicator of public health concerns related to air pollution, complementing traditional health data.
  • O₃ and NO₂ as Major Culprits: Ozone (O₃) and Nitrogen Dioxide (NO₂) were found to have the highest health impacts among the studied pollutants for both respiratory and cardio-cerebrovascular diseases.
  • Cardiovascular Vulnerability: Pollutants consistently showed a higher impact on cardiovascular diseases than on respiratory diseases.
  • HAQI's Superiority: The newly proposed Health Air Quality Index (HAQI), derived from relative risks, showed a stronger correlation with actual health risk (RR<sub>Total</sub>) than the standard AQI, suggesting it's a more accurate reflection of health impacts.
  • Meteorological Influence: Humidity, temperature, and wind speed all influence the relationship between air pollution and health outcomes, with humidity appearing to have the most significant impact on the respiratory index.

Key Results

  • Pollutant Impact Ranking: O₃ and NO₂ had the highest health impacts, followed by PM<sub>10</sub> and PM<sub>2.5</sub>.
  • RR Differences: The relative risks (RRs) for cardio- and cerebrovascular diseases were consistently higher than for respiratory diseases across all pollutants.
  • HAQI vs. AQI Correlation: The Pearson correlation coefficient between HAQI and RR<sub>Total</sub> was 0.95, significantly higher than that between AQI and RR<sub>Total</sub> (0.86), indicating HAQI's better performance in representing short-term air pollution risk.
  • Time Lag: A three-day lag period was found to be optimal for capturing the effects of air pollution on Baidu search indices.

Limitations and Future Work

  • Baidu Index Limitations: The authors acknowledge that the Baidu Index is a proxy and may have biases. It reflects general public awareness and search behavior, not necessarily diagnosed medical conditions. It also doesn't differentiate by age or sex.
  • Causality: The study establishes correlations but cannot definitively prove causality due to the data-driven nature of the analysis.
  • Keyword Selection: While efforts were made to select relevant keywords, the choice of search terms can influence results.
  • Future Work: The authors propose future research to obtain more comprehensive keyword data through network opinion analysis and big data mining, and to investigate search intentions more deeply. They also aim to conduct inter-validation studies comparing Baidu index data with contemporaneous outpatient data to assess the Baidu index's validity in predicting actual health outcomes.

Summary updated

Abstract

There is an increasing body of evidence showing the impact of air pollutants on human health such as on the respiratory, and cardio- and cerebrovascular systems. In China, as people begin to pay more attention to air quality, recent research focused on the quantitative assessment of the effects of air pollutants on human health. To assess the health effects of air pollutants and to construct an indicator placing emphasis on health impact, a generalized additive model was selected to assess the health burden caused by air pollution. We obtained Baidu indices (an evaluation indicator launched by Baidu Corporation to reflect the search popularity of keywords from its search engine) to assess daily query frequencies of 25 keywords considered associated with air pollution-related diseases. Moreover, we also calculated the daily concentrations of major air pollutants (including PM<sub>10</sub>, PM<sub>2.5</sub>, SO₂, O₃, NO₂, and CO) and the daily air quality index (AQI) values, and three meteorological factors: daily mean wind level, daily mean air temperature, and daily mean relative humidity. These data cover the area of Beijing from 1 March 2015 to 30 April 2017. Through the analysis, we produced the relative risks (RRs) of the six main air pollutants for respiratory, and cardio- and cerebrovascular diseases. The results showed that O₃ and NO₂ have the highest health impact, followed by PM<sub>10</sub> and PM<sub>2.5</sub>. The effects of any pollutant on cardiovascular diseases was consistently higher than on respiratory diseases. Furthermore, we evaluated the currently used AQI in China and proposed an RR-based index (health AQI, HAQI) that is intended for better indicating the effects of air pollutants on respiratory, and cardio- and cerebrovascular diseases than AQI. A higher Pearson correlation coefficient between HAQI and RR<sub>Total</sub> than that between AQI and RR<sub>Total</sub> endorsed our efforts.
BibTeX
@article{zhong2019study,
  author = {Shaobo Zhong and Zhichen Yu and Wei Zhu},
  title = {{Study of the Effects of Air Pollutants on Human Health Based on Baidu Indices of Disease Symptoms and Air Quality Monitoring Data in Beijing, China}},
  journal = {International Journal of Environmental Research and Public Health},
  year = {2019},
  doi = {10.3390/ijerph16061014},
}
Figure