Building a heat plan for 25% of South Africa's people, from 13 local studies

Most heat action plans are written by places that already have decades of local epidemiology telling them which temperatures kill whom. Gauteng — South Africa's most populated province, containing Johannesburg, Ekurhuleni and Pretoria and 25% of the national population — does not have that [s1].

A paper published in Annals of Global Health on 22 January describes what it looks like to build a plan anyway [s1]. The method is worth attention independently of the province it was built for, because most of the world's heat-exposed populations are in the same position.

The evidence that exists

The authors conducted a scoping review to establish baseline data on heat-related health impacts for Gauteng and for South Africa as a whole [s1].

Thirty-six studies met inclusion criteria. Thirteen of them included Gauteng data [s1]. All 36 showed impacts of heat on human health [s1].

Methodologically the literature was fairly uniform: most studies applied epidemiological time series linking meteorological exposure — temperature or heat indices — and, in some cases, air pollutants including PM2.5, PM10, NO₂ and O₃, with health outcomes [s1]. Exposure was assessed using remote sensing, reanalysis or station data, and analysed with regression or distributed lag models [s1].

Thirteen province-specific studies is a thin base on which to set activation thresholds for several million people. The honest position, which the paper takes, is that this is the evidence available and a plan has to be built from it rather than deferred until more arrives.

Borrowing structure from elsewhere

In place of local evidence the authors ran a benchmarking exercise, examining heat action plans from three jurisdictions: Maharashtra in India, Victoria in Australia, and Khyber Pakhtunkhwa in Pakistan [s1]. The choice is notable — two of the three are in low- and middle-income countries, which is not the usual reference set for adaptation planning.

Each contributed a different structural element [s1].

From Victoria, district-level thresholds. The authors' reasoning is that keeping activation simple and local suits Gauteng's heterogeneous microclimates across metros and townships [s1] — a single provincial threshold would either over-trigger in cooler areas or under-protect in hotter ones.

From Maharashtra, graded activation and clearly assigned departmental roles. The stated benefit is reducing ambiguity during multi-day heatwaves, which in Gauteng's case means aligning the Health, Infrastructure and Social Development departments [s1]. Multi-day events are precisely where unclear responsibility becomes visible.

From Khyber Pakhtunkhwa, the cooling-camp model — practical, low-cost interventions demonstrated in a low- and middle-income setting, which the authors suggest could be replicated at taxi ranks, clinics and malls during temperature peaks [s1]. Taxi ranks are a specific and sensible choice: they concentrate people who are outdoors, waiting, and often unable to leave.

The conclusion offered is that the literature and the international exemplars together provide an evidence base and adaptable models for a context-specific, multi-sectoral plan [s1].

What benchmarking cannot supply

Borrowing a structure is not the same as borrowing an effect size. None of the three exemplar plans is cited here with evidence that it reduced heat mortality in its own jurisdiction, and the paper does not claim otherwise — it identifies distinctive strengths of each design, which is a judgement about plan architecture rather than about outcomes [s1].

That gap is the subject of a PLOS Medicine piece published a week later, on 29 January [s2]. Its argument is short and direct: climate change is accelerating the frequency and severity of extreme weather events and increasingly threatening human health and life, particularly in low- and middle-income countries, and research on the effectiveness of climate adaptation interventions for human health — along with their desirability, implementation and financial viability — is urgently required [s2].

Four criteria, and effectiveness is only the first of them. A cooling camp that works but that no one uses, or that cannot be funded past its first season, has not solved anything. That framing sets a higher bar than the adaptation literature currently clears, and it applies directly to the Gauteng exercise: the plan being designed will be an intervention whose effectiveness is not yet established, in a province where the baseline epidemiology rests on 13 studies.

Why this is still the right way round

The alternative to building a plan from thin evidence is not building a better plan. It is having no plan while the evidence accumulates, during which the heat arrives regardless.

What the Gauteng paper does that is genuinely useful is make its own uncertainty legible. It states how many studies exist, how many are local, what methods they used, and exactly which elements of the plan are imported rather than derived [s1]. A plan documented that way can be evaluated later against what it borrowed and from where — which is what the PLOS Medicine argument asks for [s2].

Limits

The scoping review is a mapping exercise, not a meta-analysis; it does not pool effect estimates or grade study quality, and 36 studies showing heat impacts is a statement about the literature's direction rather than its magnitude [s1]. No heat action plan for Gauteng has been evaluated, because none has yet been implemented. The PLOS Medicine piece is an editorial statement of research priorities, not new evidence [s2].

What to watch

Whether the Gauteng plan is published with pre-specified evaluation criteria attached, and whether any of the three benchmark jurisdictions publishes outcome data that would let other provinces choose between their designs on evidence rather than on plausibility.

Sources

1. [s1] Temperature-Related Health Impacts: A Scoping Review and Benchmarking Exercise to Inform a Heat Action Plan. Annals of Global Health, 22 January 2026. https://doi.org/10.5334/aogh.5016
2. [s2] Intervention research to protect human health in the era of climate extremes. PLOS Medicine, 29 January 2026. https://doi.org/10.1371/journal.pmed.1004918