Is planned adaptation to heat reducing heat-related mortality and illness? A systematic review

Extreme heat is a public health risk [1‐3]. In 2013, 58.729 heat stroke diagnoses have been recorded for Japan [4], for example, and the United States Centers for Disease Control report an annual 659 cases (on average) of heat-related deaths between 1999 and 2009 [5]. These numbers are likely underestimated: as the physical effects of heat primarily exacerbate underlying conditions, diagnoses of death as heat-related are of varied quality [6]. Data availability on heat stroke incidence also depends on whether an emergency room or ambulance call occurs, as well as on active collection of such data. Heat increases the risk of dying of preexisting cardiovascular disease [6]; and heat stroke may lead to multiple organ failure [6‐8]. Heat-related morbidity and mortality are preventable. Older persons, people taking medications that impair thermoregulation [6], very young children, socially isolated elderly, and people physically active outdoors during very hot periods have been identified as particularly at risk [9‐12]. It has been argued that populations residing in urban centers are more vulnerable to heat events due to the urban heat island effect and higher population density [13‐17]. In recent years additional concerns have arisen about a contribution of global warming to an increased frequency of extreme temperature events [18, 19]. “Business as usual” climate change scenarios estimate that the incidence of heat events is likely to increase in the near future [20, 21]. As a result, it has been suggested that future health risks from heat might increase [22‐24]. In 2012, extreme temperature events classified as disasters by the WHO Collaborating Centre for Research on the Epidemiology of Disasters - CRED occurred 51 times worldwide, giving climatological disasters (temperature events, droughts and wildfires) an overall share of 23.8% of all 2012 disasters [25]. Recent severe heat waves occurred in Europe and Russia in 2003, 2006 and 2010, in the United States in 2012, in Australia in 2009 and 2013, and in Japan in 2010 and 2013, among others [26, 27]. Beyond these extreme cases, smaller scale heat waves occur frequently and pose risks to human health. Heat impacts on humans can be measured through thermal indices [28]. Various methods to calculate a heat index exist, and without adherence to a standard, comparability between measurements and studies is challenging [29, 30].

With a changing climate, populations of large cities in temperate regions, subtropical or tropical climates have all been characterized as vulnerable to heat [3, 31, 32]. Further measures may be needed to continually protect human health from adverse effects of heat on all continents. Adaptation to climate change has been defined as a “process of adjustment to actual or expected climate and its effects, in order to moderate harm or exploit beneficial opportunities” [33]. In this study, we are particularly interested in intentional, planned adaptation.

While we have conflicting information on risk perception of heat among populations [11, 34, 35], older persons have been characterized as especially susceptible to ill effects of heat [36‐38]. Heat warning systems have been introduced as a prevention measure [39‐41]. These usually combine information from weather stations based on a cutoff system with more or less targeted communication campaigns. Such heat warning systems can now be found across the planet, usually at city level [42].

Despite increased interest in climate change and its impacts, and a large number of heat prevention plans in place in higher-income countries to protect human health [39, 43], we have hardly any conclusive evidence on the effects of said adaptation measures [44]. Is climate change adaptation to heat reducing heat stroke incidence and heat-related mortality? This study uses a systematic review design in an attempt to answer this question.

We conducted a systematic literature review of peer-reviewed published literature. The PRISMA checklist, research protocol and the data extraction sheet can be found in the supplementary material (Additional files 1, 2 and 3). The scope of our review was as follows:

Because heat adaptation aims at preventing adverse health effects, we use the terms heat adaptation and heat prevention interchangeably in this review.

Context: International large urban centers^a.

The following outcomes were of interest:

The database search led to 5539 results, 2299 after removal of duplicates. After title and abstract screening 2252 articles were excluded because they did not concern human health or did not contain an evaluation. 47 articles were assessed as full texts. We excluded 29 articles after reading the full texts because no evaluation according to our criteria was described. Through additional sources such as reference lists we identified 12 studies. All in all, 30 articles were included in the review, as shown in the PRISMA flowchart (Figure 1).

The results of our review reveal difficulties in assessing adaptation effectiveness and are consistent with previous research. This suggests that issues of methodological rigor and what to measure when speaking about effectiveness of heat adaptation have not yet been resolved, despite increased interest in the matter.

Common themes in all studies were difficulties assessing adaptation effectiveness with standard epidemiological methods. This has been discussed particularly in the four systematic reviews. Specifically, the following issues in conducting rigorous studies to generate conclusive evidence of adaptation effects have been named:

Gupta et al. [75] call for experimental study designs to assess the effectiveness of using fans during a heat wave as they were unable to resolve conflicting information from observational studies in their Cochrane review. In our included studies a call for more rigorous methods was the standard solution to the above mentioned issues, without specific recommendations on how to achieve this. When trying to judge whether the information we gathered through the review is sufficient proof that heat adaptation reduces heat-related mortality and illness, we struggle with the following problems posed by the available studies:

Why is conducting experimental research of adaptation to heat so difficult? For one, defining the counterfactual, i.e. what would have happened in the absence of the adaptation measure, is problematic, because usually an entire city or even country is exposed to the adaptation measure. Choosing a different city as control would require careful matching. This is difficult for many reasons: for example, intercity microclimate variability could bias results, and to assess effects the control city would need to be exposed to a heat event of similar magnitude and length. Unlike other public health interventions, researchers and practitioners cannot limit exposure; they can only mediate it.

Second, heat prevention can occur at structural level, or at individual level through behavior change. Ethical concerns could be raised if structural prevention or a warning system were only available to an intervention group in one city. For instance, control populations could not be prevented from accessing public green spaces.

Third, heat by itself is not a new phenomenon. Much of the heat-related health advice provided by risk communication campaigns is common sense information: to stay hydrated, for example, or to seek shade and cool places [80]. Physical discomfort during heat makes it likely that people have followed such advice before official warnings were even issued. This might not only suggest absence of the classic control group for behavior, it is also more difficult to compare knowledge pre- and post-information campaigns. In light of future population aging, potential improvements to adaptation effects lie with targeting those elderly people who do not feel at risk through awareness raising interventions despite these difficulties. The use of innovative materials and social norms approaches could be evaluated.

While we argue that concrete evidence for the effectiveness of specific planned adaptation measures is lacking, our results show a mostly unanimous decline in sensitivity to heat over longer time periods. Alternative hypotheses for the causes of this decline should be investigated. Proposed alternatives have included biological adaptation [81], improvements to healthcare systems [82], technological advancements [83], adjustments to the urban built environment [84], and social progress [84]. The role each of the alternatives plays in declining heat sensitivity is debated [78].

Aware of these shortcomings, recent research projects into methods specifically for adaptation assessments have been designed [85], results are not yet available.

We were surprised to be unable to identify articles assessing infrastructural measures such as greening, or supply of air conditioning, although we had specifically intended to include these. Our focus on human health and our health–related search terms may have prevented us from finding articles on urban planning effects. Connecting specific urban planning to public health assessments might be a challenging but interesting future research topic.

Our results show that rigorous evaluation of adaptation is rare and difficult to conduct. The potential health effects of adaptation can currently not be measured conclusively. Up to now, we find limited intersectoral efforts between public health agencies and climate change adaptation policy. Such efforts might contribute to a reduction in adverse health effects of heat. In addition, involvement of the health sector in adaptation design, implementation and evaluation might increase chances of successful adaptation.

Current knowledge does not prove effectiveness of planned adaptation, yet a decline in sensitivity to heat hints at important developments. Recent articles published after the search period for this review observe a similar decline over long time periods [96‐98]. Whether biological adaptation, continuous improvements in healthcare, changes to the urban environment not declared “adaptation,” or a different unknown reason caused said decline is a matter of further interest. The seeming paradox between the observed decline in the examined studies and scholarly works referring to an expected increase in heat-related adverse health effects [99] needs to be assessed further as well. Low-regret adaptation options might be investigated while simultaneously increasing efforts to overcome methodological evaluation challenges with further research.

^aOriginally we had planned to include only cities with more than 500,000 inhabitants. Due to the limited study availability, however, we decided to broaden this criterion to cities of any size.

^b* = wildcard, all possible word endings included.