Air Health Trend Indicator: How to characterize trend (linear vs. non-linear)?

The Air Health Indicator (AHTI) provides information on how the health risks from short-term exposures to air pollution change over time in relation to ground-level ozone (ozone) and fine particulate matter up to 2.5 micrometers in size (PM2.5). The AHTI measures day-to-day changes in non-accidental deaths and hospital admissions, for example, due to heart, circulatory and respiratory conditions a few days after exposure to smog (mostly ozone and PM2.5) during the warm season (from April to September), cold season (from October to March), and year-round (between January and December). Since 2011, the AHTI reports have been posted on the Environment and Climate Change Canada’s website and are accessible to all Canadians. This project will extend current AHTI in two areas: study areas (census divisions) and advanced model on trend detection.

There are three concerns with the current AHTI which we aim to address with this project: (1) representing urban cities only; (2) detecting linear trend (or monotonic change) only; and (3) modelling a single-pollutant. To address (1), this project will extend spatial coverage to include more urban cities and rural areas, each of which has a population of at least 40,000. This population size is necessary to have non-zero daily counts of mortality, which is required to have statistical power to estimate air pollution related health risk. To address (2), this project will apply a more advanced statistical model to detect non-linear trend. To address (3), this project will propose new models on multi-pollutant. Therefore, there are three objectives of this project:

1. To include rural areas in addition to urban centres,
2. To detect non-linear trend.
3. To consider multi-pollutants simultaneously in the trend.

Note: Urban centres usually have at least one NAPS monitoring station, whereas rural areas do not.