• This review of 53 studies explores how Air Quality Health Index (AQHI) methodologies have evolved over time. • Most models use mortality data and single-pollutant approaches for risk estimation. • Future AQHI should include multi-pollutant models and real-time health indicators. • A structured framework is proposed to support more effective public health alerts. Environmental pollution presents an escalating threat to public health, highlighting the urgent need for accessible tools to assess air pollution-related risks—particularly among vulnerable populations. The Air Quality Health Index (AQHI), which links ambient air pollutant levels to health outcomes, offers a scalable strategy for real-time risk communication and mitigation of health impacts. To rationally optimize AQHI construction methodologies, we conducted this systematic review. This systematic review examined AQHI construction methodologies across 53 articles retrieved from eight databases, including PubMed, Embase, and CNKI (China National Knowledge Infrastructure), covering publications from March 1, 2008, to March 17, 2025. The majority of articles (66.0%) were published after 2021, with 88.7% conducted in China. Mortality (56.6%), outpatient visits (22.6%), and hospitalizations (13.2%)—primarily due to respiratory and cardiovascular conditions—were the most frequently assessed health outcomes. Six major pollutants (PM₂.₅, PM₁₀, SO₂, NO₂, CO, and O₃) were commonly used as independent variables. Methodologically, 64.2% of articles employed single-pollutant models, while 39.6% used multi-pollutant approaches. Validation primarily involved (73.6%) comparing excess risk estimates between AQHI and the conventional Air Quality Index (AQI), and 39.6% of articles conducted stratified analyses by demographic factors such as gender and age. The mean QATSDD score was 26, indicating moderate methodological quality. These findings provide practical insights for AQHI development and support its integration into air quality warning systems to better protect at-risk populations.
Liao et al. (Fri,) studied this question.