Skip to content
chaleur

Settled science, open question.

That extreme heat, bushfire smoke and thunderstorm asthma raise illness and death is established science. In extreme heat, the evidence goes a step further: studies of who dies have repeatedly found that isolation is among the strongest risk factors, and that social contact during dangerous conditions is one of the few protective factors a community can influence. Whether the same social pattern decides who is harmed by smoke, or by the other hazards chaleur watches, has not been directly studied; extending it beyond heat is chaleur's design reasoning rather than a settled finding, and the events that shaped that reasoning are set out below. The open question, and the one a research-active clinical partner is best placed to answer, is whether prompting the person who already cares for someone to act at the moment conditions turn dangerous changes what happens next. That question is only partly answered by the existing literature, and the gaps are precisely where a pilot built on chaleur could contribute new evidence.

Two events

Paris, August 2003.

The heatwave that settled over France killed roughly 15,000 people in a matter of weeks. Richard Keller's study of the Paris deaths (Keller, 2015), building on Eric Klinenberg's social autopsy of the 1995 Chicago heatwave (Klinenberg, 2002), is precise about mechanism: the elderly died in such numbers not simply because their bodies were vulnerable to heat, but because their social networks had frayed, leaving them isolated and unable to reach the help they needed, often in small top-floor apartments whose zinc roofs turned the rooms into ovens. Keller's warning is the lesson most relevant to anyone designing a response: risk models built around the elderly are blunt instruments that obscure the role of poverty and isolation and overlook the thousands of other victims an age-based profile never captures.

Eastern Australia, 2019–20.

Through the Black Summer, bushfire smoke blanketed Sydney and much of eastern Australia for weeks, with population-weighted PM2.5 peaking at 98.5 µg/m³ on 14 January 2020, four times the national 24-hour standard. A preliminary evaluation in the Medical Journal of Australia attributed to that smoke, across the four eastern states alone, 417 excess deaths, 1,124 cardiovascular hospitalisations, 2,027 respiratory hospitalisations and 1,305 emergency presentations for asthma (Borchers Arriagada et al., 2020). The harm fell on people with cardiac and respiratory conditions.

The two events share no single climate driver. For extreme heat, the evidence that isolation decided who died is direct, and it is set out in the next section. For bushfire smoke the health burden is documented but the social mechanism has not been studied: whether isolation shapes who is harmed, as it demonstrably does in heat, is an open question, and one a pilot built on chaleur would be well placed to help answer.

The register, and its two failures

France's institutional answer to 2003, the plan canicule and its municipal register of vulnerable residents (established by the law of 30 June 2004 and the decree of 1 September 2004), was a serious and humane intervention, and it carries two limitations chaleur is designed around. The first is that it depends on a person being recognised as vulnerable and formally enrolled, whether by themselves or by a third party, which systematically misses the large group of people who manage independently, do not see themselves as at risk, and would never place their own name on a council list (the same population Abrahamson's interviews describe; see the third question below). The second is that it achieves coverage by assembling a centralised, name-identified register of health-vulnerable citizens, a privacy posture significant enough that France's own data-protection authority, the CNIL, had to insist these registers not become "population files". chaleur's design responds to both limits at once: the warning is addressed to the person who already worries, and the sensitive information stays on that carer's own device rather than in any central file. How that is built is set out on the Privacy page.

What is established

The risk chaleur is built around is well evidenced. Social isolation and living alone carry an increase in mortality risk comparable in magnitude to long-recognised factors such as smoking, physical inactivity and obesity (Holt-Lunstad et al., 2010; 2015), and that association holds after adjustment for prior health, which weakens the objection that it merely reflects sick people withdrawing from contact. In the specific setting of extreme heat, studies of who dies have repeatedly found that people who are isolated or rarely leave home are among those most at risk, and that social contact during dangerous conditions is one of the few protective factors a community can influence (Semenza et al., 1996; Bouchama et al., 2007). The premise that connection protects, and that isolation is a modifiable part of the risk, is therefore not in serious dispute.

What is not yet settled, and where a clinical partner could contribute

The evidence thins markedly once the question shifts from whether isolation is dangerous to whether acting on it in the way chaleur proposes actually helps. Three areas stand out, and each is a place where research conducted alongside a pilot would add something the field currently lacks.

Whether a prompted check-in changes outcomes.

The protective effect of social contact in heatwaves is drawn almost entirely from observational studies, which establish an association rather than test an intervention. Check-in and "buddy" schemes, in which neighbours or relatives are asked to look in on vulnerable people during dangerous conditions, are widely recommended and appear in many official heat plans, yet reviews of community heat interventions have noted that their effect on health outcomes has rarely been evaluated (Vu, Rutherford & Phung, 2019; Toloo et al., 2013). Whether a timely prompt to check in measurably reduces harm, and for whom, is close to an open question, and it is one a controlled pilot is well suited to address. Because the protective effect of contact rests on observational data, the natural next step is a controlled pilot with defined behavioural and check-in endpoints, which is the kind of study a clinical partner would help design.

Whether a personalised warning outperforms a general one.

Heat health warning systems have been associated with reduced deaths in several evaluations, but the same literature is consistent that the availability of a warning does not by itself change behaviour, and that whether a person acts depends heavily on whether they perceive the threat as applying to them or to someone they care about (Toloo et al., 2013). Warnings that carry a specific recommended action, and that connect the threat to the person, are associated with stronger risk perception and a greater intention to take protective steps than undifferentiated broadcast messaging (Heidenreich et al., 2024), though most of this evidence measures knowledge and intention rather than health outcomes. Personalisation is therefore promising rather than proven, and the size of any real-world benefit is unestablished.

Whether routing the warning through a carer reaches the people a general warning misses.

The recurring failure of public heat messaging is that the people most at risk frequently do not count themselves among them (Abrahamson et al., 2009, whose interviewees aged into their nineties largely did not regard themselves as old or at risk even when they carried qualifying chronic illness, while readily recognising that risk in others), so a warning addressed to the at-risk person reaches exactly the population least likely to act on it. chaleur's design responds by addressing the warning instead to the carer, who already regards the person as someone worth worrying about, and who therefore does not need to be persuaded of the risk before acting. A caution sits inside this same evidence: strong social networks can reinforce rather than puncture the belief that one will cope, so being connected does not by itself correct the misperception (Wolf et al., 2010). Whether routing the warning through a carer escapes the optimism the at-risk person holds about themselves, or merely inherits it, is itself an open and testable question, and one of the more interesting a pilot could answer. This is a coherent answer to a documented problem, but it is a design hypothesis rather than a finding, because no comparable tool has been trialled in this form. It is the part of chaleur most in need of, and most suited to, formal study.

A clinical partner would not be asked to endorse a finished product. They would be asked to help establish whether a promising idea works, to sign off the guidance the app relies on, and to shape the pilot that would test it.

References

  1. J. Holt-Lunstad, T. B. Smith, J. B. Layton, "Social Relationships and Mortality Risk: A Meta-analytic Review," PLoS Medicine 7(7):e1000316, 2010.
  2. J. Holt-Lunstad, T. B. Smith, M. Baker, T. Harris, D. Stephenson, "Loneliness and Social Isolation as Risk Factors for Mortality: A Meta-Analytic Review," Perspectives on Psychological Science 10(2):227–237, 2015.
  3. J. C. Semenza et al., "Heat-Related Deaths during the July 1995 Heat Wave in Chicago," New England Journal of Medicine 335(2):84–90, 1996.
  4. A. Bouchama et al., "Prognostic Factors in Heat Wave-Related Deaths: A Meta-analysis," Archives of Internal Medicine 167(20):2170–2176, 2007.
  5. M. S. Toloo et al., "Are heat warning systems effective?," Environmental Health 12:27, 2013.
  6. A. Heidenreich et al., "Individual heat adaptation: Analyzing risk communication, warnings, heat risk perception, and protective behavior in three German cities," Risk Analysis, 2024 (doi:10.1111/risa.14278).
  7. K. Abrahamson et al., "Perceptions of heatwave risks to health: interview-based study of older people in London and Norwich, UK," Journal of Public Health 31(1):119–126, 2009.
  8. J. Wolf, W. N. Adger, I. Lorenzoni et al., "Social capital, individual responses to heat waves and climate change adaptation: An empirical study of two UK cities," Global Environmental Change 20(1):44–52, 2010.
  9. A. Vu, S. Rutherford, D. Phung, "Heat Health Prevention Measures and Adaptation in Older Populations—A Systematic Review," International Journal of Environmental Research and Public Health 16(22):4370, 2019.
  10. Richard C. Keller, Fatal Isolation: The Devastating Paris Heat Wave of 2003 (University of Chicago Press, 2015).
  11. Eric Klinenberg, Heat Wave: A Social Autopsy of Disaster in Chicago (University of Chicago Press, 2002).
  12. N. Borchers Arriagada et al., "Unprecedented smoke-related health burden associated with the 2019–20 bushfires in eastern Australia," Medical Journal of Australia 213(6), 2020.
  13. Plan canicule and the registre nominatif des personnes vulnérables: Loi n° 2004-626 of 30 June 2004; Décret n° 2004-926 of 1 September 2004.
  14. Commission nationale de l'informatique et des libertés (CNIL), "Les registres communaux d'alerte et de protection des populations en cas de crise exceptionnelle," cnil.fr.