Objectives The objective of this study was to use health administrative and environmental data to quantify the effects of ambient air pollution on health service use among those with chronic diseases. We hypothesised that health service use would be higher among those with more exposure to air pollution as measured by the Air Quality Health Index (AQHI).

Setting Health administrative data was used to quantify health service use at the primary (physician office visits) and secondary (emergency department visits, hospitalisations) level of care in Ontario, Canada.

Participants We included individuals who resided in Ontario, Canada, from 2003 to 2010, who were ever diagnosed with one of 11 major chronic diseases.

Outcome measures Rate ratios (RR) from Poisson regression models were used to estimate the short-term impact of incremental unit increases in AQHI, nitrogen dioxide (NO₂; 10 ppb), fine particulate matter (PM_2.5; 10 µg/m³) and ozone (O₃; 10 ppb) on health services use among individuals with each disease. We adjusted for age, sex, day of the week, temperature, season, year, socioeconomic status and region of residence.

Results Increases in outpatient visits ranged from 1% to 5% for every unit increase in the 10-point AQHI scale, corresponding to an increase of about 15 000 outpatient visits on a day with poor versus good air quality. The greatest increases in outpatient visits were for individuals with non-lung cancers (AQHI:RR=1.05; NO₂:RR=1.14; p<0.0001) and COPD (AQHI:RR=1.05; NO₂:RR=1.12; p<0.0001) and in hospitalisations, for individuals with diabetes (AQHI:RR=1.04; NO₂:RR=1.07; p<0.0001) and COPD (AQHI:RR=1.03; NO₂:RR=1.09; p<1.001). The impact remained 2 days after peak AQHI levels.

Conclusions Among individuals with chronic diseases, health service use increased with higher levels of exposure to air pollution, as measured by the AQHI. Future research would do well to measure the utility of targeted air quality advisories based on the AQHI to reduce associated health service use.