Midcentury

What does 2050 look like?

If today's temperature responses are applied to midcentury temperatures with nothing else changing (no adaptation, no new capacity), additional emergency department visits rise from roughly 198 thousand per decade in 2031–2040 to 317 thousand in 2051–2060, and the associated public cost increases from 58.4 to 93.8 million USD per decade.

These numbers are illustrative (ceteris-paribus)

The projections below apply today's estimated temperature responses to tomorrow's temperatures, holding behaviour, institutions, medical technology and infrastructure fixed. Nothing in them models air conditioning, changed hospital capacity, or public health response. They are an upper-bound illustration of what continued warming would do to a system that does not react, not a forecast of what will happen. Read them as a planning prompt, not a prediction.

Additional visits by decade
0100200Additional ED visits per decade (thousands)2031–2040198 decade total14552.42041–2050247 decade total18463.22051–2060317 decade total23878.9020406080Additional public cost per decade (million USD)2031–204058.4 decade total42.915.52041–205073.0 decade total54.318.72051–206093.8 decade total70.523.3
fewer cold daysmore hot daysCold days reduce visits, so as climate change makes them rarer, the low-visit days that cold used to cause disappear and visits rise. Heat adds visits directly, and climate change makes hot days more frequent, so visits from heat rise as well. Both components are therefore increases. Dots are the mean across five climate models; bars span the lowest and highest model.
Estimates behind this figure:m5_projections_range.csv
Full dataCode

Fewer cold days account for most of it, which is the counterintuitive part. Warming does not only add hot days, it removes cold ones, and since cold days suppress visits, removing them raises the total. In the earliest decade this component dominates the increase.

More hot days catch up but do not overtake. Their share grows across the decades while the fewer-cold-days component stays larger throughout, so planning that attends only to heatwaves would miss the bigger channel.

By age group
Decade:
0+20+40+60+80+100AgeAdditional ED visits per decade (thousands)<12+12.6−7.0+14.2−8.0+15.6−10.612–20+26.6+8.2+29.9+8.8+34.4+9.220–40+64.7+25.0+81.3+30.1+104.8+36.740–60+27.3+8.9+34.4+10.8+44.6+13.460–80+11.7+2.0+19.9+3.3+31.9+5.280+2.5+1.3+4.2+2.1+7.1+3.6
cold-drivenheat-driven
Bars run from zero. Heat-driven visits fall for children under 12, which the paper attributes to fewer cases of the common cold as temperatures rise.
Estimates behind these figures:m5_projections_by_age.csvm5_population.csv
Full dataCode
Estimated public cost by decade
0100200Additional ED visits per decade (thousands)2031–2040198 decade total14552.42041–2050247 decade total18463.22051–2060317 decade total23878.9020406080Additional public cost per decade (million USD)2031–204058.4 decade total42.915.52041–205073.0 decade total54.318.72051–206093.8 decade total70.523.3
fewer cold daysmore hot daysCold days reduce visits, so as climate change makes them rarer, the low-visit days that cold used to cause disappear and visits rise. Heat adds visits directly, and climate change makes hot days more frequent, so visits from heat rise as well. Both components are therefore increases. Dots are the mean across five climate models; bars span the lowest and highest model.

How the cost is estimated: each additional visit is valued at 207 USD, and the share of visits that end in hospitalisation, held at the historical rate, adds 691 USD each. Both unit costs are the Mexican Social Security Institute's official 2024 figures published in the Diario Oficial de la Federación (14 December 2023). Costs are summed over each decade and averaged across the five climate models. No adaptation, price change or capacity change is modelled.

Estimates behind this figure:m5_projections_range.csv
Full dataCode

Severity differs by channel

By 2051–2060, about 21 percent of the extra visits from more hot days are severe enough to require hospitalisation, against about 10 percent of those from fewer cold days. The historical rate across all visits is 13 percent. More hot days therefore produce a disproportionate share of severe cases, while the fewer-cold-days channel is concentrated in lower-severity visits, so the two channels are not interchangeable for capacity planning even where the visit counts look similar.

Estimates behind this figure:m5_projections_severity.csv
Full dataCode

Where the spread comes from

Temperature changes come from five CMIP6 global climate models, comparing a 1991–2010 baseline with future decades, aggregated to municipalities using population weights. The bars in the first chart span the lowest and highest model, so they describe model disagreement rather than statistical uncertainty.

Visit changes are converted to absolute numbers using SSP2 population projections by age group, adjusted for temperature-related mortality. Costs use official 2024 Mexican unit costs of 207 USD per visit and 691 USD per admission leading to hospitalisation.

Decade figures are totals for the whole decade. The paper's annual figure is the 2041–2060 decade average divided by ten.