Future temperature-related mortality in the UK under climate change scenarios: Impact of population ageing and bias-corrected climate projections

Abstract

Background: Exposure to heat and cold poses a serious threat to human health. In the UK, hotter summers, milder winters and an ageing population will shift how populations experience temperature-related health burdens. Estimating future burdens can provide insights on the drivers of temperature-related health effects and removing biases in temperature projections is an essential step to generating these estimates, however, the impact of various methods of correction is not well examined.

Methods: We conducted a detailed health impact assessment by estimating mortality attributable to temperature at a baseline period (2007-2018) and in future decades (2030s, 2050s and 2070s). Epidemiological exposure-response relationships were derived for all England regions and UK countries, to quantify cold and heat risk, and temperature thresholds where mortality increases. UK climate projections 2018 (UKCP18)were bias-corrected using three techniques: correcting for mean bias (shift or SH), variability (bias-correction or BC) and extreme values (quantile mapping or QM). These were applied in the health impact assessment, alongside consideration of population ageing and growth to estimate future temperature-related mortality.

Findings: In the absence of adaptation and assuming a high-end emissions scenario (RCP8.5), annual UK temperature-related mortality is projected to increase, with substantial differences in raw vs. calibrated projections for heat-related mortality, but smaller differences for cold-related mortality. The BC approach gave an estimated 29 deaths per 100,000 in the 2070s, compared with 50 per 100,000 using uncorrected future temperatures. We also found population ageing may exert a bigger impact on future mortality totals than the impact from future increases in temperature alone. Estimating future health burdens associated with heat and cold is an important step towards equipping decision-makers to deliver suitable care to the changing population. Correcting inherent biases in temperature projections can improve the accuracy of projected health burdens to support health protection measures and long-term resilience planning.