Heat sends Australians to emergency. Bushfire smoke sends them to hospital

Heat and bushfire smoke are usually discussed together in Australia, because they arrive together. A study published in the Medical Journal of Australia on 6 August pulls them apart, and finds they push people into different parts of the health system [s1].

The analysis was built to support Australia's first National Climate Risk Assessment [s1]. It is a cross-sectional population study covering people accessing public health services between 2014 and 2020 for cardiovascular, respiratory and mental health conditions — the condition groups carrying the highest burden of disease [s1]. Rather than looking at deaths, it looks at four utilisation measures at once: emergency department presentations, inpatient admissions, Medicare Benefits Schedule claims in primary care, and Pharmaceutical Benefits Scheme prescriptions dispensed [s1].

That four-channel design is what makes the result readable. Most heat-and-health studies pick one endpoint. Following the same exposure through four different doors shows where in the system the load actually lands.

Heat shows up at the front door

Heat was most strongly associated with emergency department presentations [s1]. Per degree of temperature change, the relative risk of an ED presentation was 1.033 (95% CI 1.023–1.043; p < 0.001) for respiratory conditions and 1.023 (95% CI 1.016–1.031; p < 0.001) for mental health conditions [s1].

Those are small per-unit numbers, and it is worth being precise about what they mean. A relative risk of 1.033 per degree is a roughly 3% increase in respiratory ED presentations for each degree of temperature change — not a 3% increase from a heatwave. Across the several degrees a heatwave adds above baseline, and across a national population, the multiplication is what makes it a system problem rather than a rounding error.

Smoke shows up one day later, upstairs

Fine particulate matter behaved differently. PM2.5 at a one-day lag had the highest association with hospital admissions, with a relative risk of 1.014 per 1 μg/m³ change in particulate density for both cardiovascular (p = 0.001) and respiratory (p = 0.002) conditions [s1].

Two things in that sentence matter. The first is the lag: the effect is strongest a day after the exposure, not during it, which is consistent with smoke exposure producing a deterioration that takes time to become bad enough to require admission. The second is the destination. Heat's strongest signal was in the emergency department; smoke's strongest signal was in inpatient admission. An ED presentation and an admission are different resources, different costs, and different planning problems.

Mental health was sensitive to both, everywhere

The finding the authors single out is that mental health outcomes showed vulnerability across all domains [s1]. Primary care visits for mental health conditions were strongly associated with both exposures: a relative risk of 1.049 (95% CI 1.040–1.058; p < 0.001) per unit of PM2.5, and 1.027 (95% CI 1.014–1.039; p < 0.001) per degree of temperature [s1].

That PM2.5 figure for mental health primary care is, notably, larger than the PM2.5 figures reported for cardiovascular or respiratory admissions. Smoke is not usually framed as a mental health exposure. Here it produced one of the study's stronger associations, in the part of the system — general practice — that is least visible in most climate-and-health reporting.

Medication use rose across respiratory, cardiovascular and mental health categories in response to both exposures, with particulate matter showing the stronger effects for all three prescription categories [s1].

What this design can and cannot establish

This is a cross-sectional population study of associations, not a causal experiment, and the paper is framed as characterising associations rather than proving mechanism [s1]. Short-term lags between exposure and utilisation were modelled [s1]; longer-horizon consequences are outside what this design addresses.

The 2014–2020 window is also worth noting. It includes the 2019–2020 Black Summer fire season, which is both the strength and the limitation: it means the smoke exposures analysed include genuinely extreme national events rather than only routine ones, but it also means a single exceptional season carries weight in the estimates. The study covers people accessing public health services [s1], so care delivered entirely privately sits outside the frame.

Nothing here tells an individual reader what their own risk is on a smoky day. The unit of analysis is a health system, and the finding is about where load appears in it.

What it is actually for

The paper's own conclusion is a planning argument: because heat and particulate matter load different parts of the system, and because mental health services are sensitive to both, air quality management and heat-health policies need to be integrated rather than run as separate programmes [s1].

That is a more specific claim than the usual call for climate preparedness. It says that a hospital network preparing only for heat-driven ED surges will be caught out by a smoke-driven admissions surge a day later, and that both will show up in general practice as mental health presentations that neither plan accounted for. Whether Australia's National Climate Risk Assessment translates that into differentiated service planning is the thing to watch next.

Sources

• [s1] The Impact of Heat and Bushfire Smoke on Health System Utilisation in Australia, Medical Journal of Australia, published online 6 August 2026. https://doi.org/10.5694/mja2.70261