Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

Climate Adaptation, Heat Risk, and Protective Behavior

Abstract. This fundamental research monograph examines climate adaptation as the transition from climatic exposure and hazard appraisal to feasible protective action, with extreme heat as the principal empirical focus. The analysis connects historical heat-health research with Protection Motivation Theory, the Protective Action Decision Model, private proactive adaptation, and models of efficacy and intention. It critically evaluates survey, longitudinal, occupational, epidemiological, quasi-experimental, and intervention evidence, including contradictory findings on warning effectiveness. The synthesis identifies the decisive roles of cooling access, housing, work organization, social support, communication, and inequity. Related flood, smoke, and compound-hazard decisions are examined to clarify limits of generalization. Scientific priorities include validated behavioral measures, causal intervention evaluation, multilevel exposure assessment, and equitable climate-health protection.

Keywords: climate adaptation; extreme heat; heat risk perception; protective behavior; heat-health warning systems; protection motivation; self-efficacy; occupational heat stress; heat action plans; cooling access; environmental psychology; climate health

Scientific Overview and Research Object

Climate adaptation is the process through which people, households, communities, and institutions adjust practices and environments in response to actual or anticipated climatic hazards. This research dossier examines the psychological organization of protective behavior under extreme heat, with connected analyses of flood, wildfire-smoke, and compound-hazard decisions. Its central object is the transition from exposure and warning to feasible action: noticing danger, interpreting its significance, selecting a protective response, coordinating with others, and sustaining that response under practical constraints. The analysis draws on environmental and health psychology, behavioral science, disaster studies, epidemiology, and occupational health. Contemporary temperature–mortality research demonstrates substantial health stakes, while explaining individual variation in response requires more specific behavioral evidence Gasparrini et al. (2015) Grothmann & Patt (2005).

A heat warning becomes protective only through an implementation pathway connecting credible information, perceived personal relevance, an actionable recommendation, material opportunity, and timely behavior. A person may understand a forecast yet remain in a hot dwelling because rent, employment, caregiving, mobility, or electricity costs restrict alternatives. Another may have cooling infrastructure but discount a hazard that has historically felt ordinary. These are distinct empirical mechanisms, and effective adaptation research must observe both. Behavioral protection therefore includes immediate actions such as finding cooling, modifying exertion, or checking on a neighbor, together with durable arrangements such as occupational scheduling, housing retrofits, care networks, and locally designed heat-health plans Lindell & Perry (2012) Johar et al. (2025).

This work uses a critical narrative synthesis of identified original studies, theoretical articles, systematic reviews, meta-analyses, and official public-health guidance. It is a literature-based research monograph, with no new participant recruitment, pooled effect estimate, systematic search protocol, or independent clinical review. The core questions are which psychological constructs predict action, which social resources make action possible, which interventions have demonstrated consequences for behavior or health, and how evidence should be judged when observational exposure, self-reported protective behavior, and mortality outcomes are conflated. Historical theoretical models provide explanations; contemporary intervention and quasi-experimental evaluations test parts of those explanations in specific settings Buller et al. (2026) Benmarhnia et al. (2016).

Definitions, Exposure, Vulnerability, and Adaptation

Extreme heat can be described through absolute temperature, apparent temperature, heat index, locally unusual conditions, duration, nighttime persistence, and epidemiologically derived thresholds. Heat stress describes the combination of environmental conditions, metabolic heat, clothing, and activity that challenges thermoregulation; heat strain describes the physiological burden that results. These variables are related but not interchangeable. A single outdoor temperature offers an incomplete estimate of personal risk because humidity, solar radiation, wind, exertion, acclimatization, building conditions, and access to cooling alter effective exposure. Occupational guidance explicitly treats workload and clothing as part of heat-risk assessment WHO & WMO (2025) Flouris et al. (2018).

Vulnerability is a relational property of a person in an environment. Physiological susceptibility, exposure duration, medical conditions, mobility, housing quality, social isolation, and institutional access jointly shape the risk of harm. Protective behavior is an observable or reportable action intended to lower exposure or consequences, such as moving into a cool space, rescheduling exertion, securing water, seeking assistance, or arranging a welfare check. Adaptation capacity includes both competence and the means to execute a response. A survey that asks whether a respondent knows what to do measures a different construct from one that verifies whether the respondent can change a work shift or reduce a measured indoor temperature CDC (2025b) Johar et al. (2025).

The same behavioral label can conceal different mechanisms. Staying home during a heat alert may be protective when the dwelling is cool and potentially hazardous when it retains heat. Using a fan may assist under some indoor conditions but has contextual thermal limits, so generic instructions require local qualification. Running air conditioning can lower exposure for a household while increasing cost, energy dependence, and sometimes vulnerability to outages. An adequate research definition links each act to the actual hazard pathway and resource conditions. Consequential outcomes include body temperature, symptoms, emergency attendance, mortality, and the distribution of benefits across populations; behavior frequency alone cannot establish the health effect of a measure CDC (2025b) Das et al. (2025).

Historical Formation of Heat-Health Research

Contemporary behavioral heat research developed at the intersection of disaster warning studies, risk perception, human thermophysiology, and city-level mortality surveillance. The catastrophic European summer of 2003 made the public-health significance of heat waves especially visible. Analysis of excess mortality in France documented the consequences of a prolonged event and established the urgency of preparedness, epidemiologic surveillance, and coordinated response. The 2003 event was a catalyst for heat-health plans, but it must be located within longer traditions of occupational hygiene, hazard warning, public-health prevention, and behavioral adaptation Fouillet et al. (2006) Lindell & Perry (2012).

Early evaluations raised the question of whether meteorological warning systems changed outcomes or merely predicted bad conditions. Ebi and colleagues evaluated estimated lives saved by Philadelphia's heat watch/warning system, while Kalkstein and Sheridan examined how Phoenix residents perceived heat and responded to alerts. These studies advanced two distinct levels of inquiry: population outcomes and reported behavioral reception. The distinction remains central because apparent knowledge of alerts cannot be used to infer preventable deaths, and aggregate mortality changes cannot identify which individual mechanism improved protection Ebi et al. (2004) Kalkstein & Sheridan (2007).

Theoretical integration gained clarity when Grothmann and Patt formulated a socio-cognitive model of private adaptation that combined risk appraisal with perceived ability to adapt. Later work incorporated protective action decision theory and self-efficacy into climate and heat research. The field has subsequently expanded from warning awareness to inequities in housing, work, urban infrastructure, and power. Recent systematic reviews evaluate not just information transfer but the quality of intervention evidence and measured behavior. This development makes the household a useful analytical unit while requiring institutional and community explanations alongside the individual Grothmann & Patt (2005) van Valkengoed et al. (2024) Buller et al. (2026).

Heat Exposure and Population Health Evidence

Temperature–mortality associations supply an essential background for behavioral intervention research. Gasparrini and colleagues analyzed daily mortality and temperature in 384 locations across thirteen countries, fitting local time-series models and pooling distributed-lag associations. This design demonstrates the importance of location-specific temperature-response functions and the temporal accumulation of exposure effects. It does not determine which particular residents failed to take a protective action. Epidemiological attributable fractions refer to specified populations, observation periods, temperature definitions, and models, so they cannot be treated as universal constants Gasparrini et al. (2015).

Health risks can arise at temperatures below a community's most spectacular extreme events, and heat effects can persist across consecutive days and warm nights. Regional acclimatization, mortality baselines, building design, and the temporal distribution of exposure influence comparisons between cities. Accordingly, a prevention program built only around a rare absolute temperature threshold may miss substantial burdens. HeatRisk and related health-based warning tools attempt to translate meteorological information into levels with different implications for susceptible people and the general public, yet a forecast remains a decision aid whose effect depends on uptake and available safeguards CDC (2025b) Gasparrini et al. (2015).

The methodological link from exposure science to psychological explanation requires measurement at appropriate scales. Outdoor monitoring stations may misrepresent a worker's radiant heat or an older adult's overnight bedroom exposure. Administrative records identify severe outcomes but contain limited information about perceived risk and behavior. Surveys provide beliefs and self-reported acts yet often lack physiological verification. A rigorous mixed design therefore links time-stamped warnings, measured indoor and outdoor conditions, individual capability, recorded actions, and health outcomes, with explicit attention to differential selection and attrition. This is a research-design implication drawn from existing evidence, rather than a claim that a single definitive study has already supplied every component Weinberger et al. (2018) Buller et al. (2026).

The Psychology of Risk Recognition

Perceived heat risk includes judgments of probability, severity, personal vulnerability, immediacy, and control. These judgments are shaped by lived experience, comparative optimism, familiarity, social cues, and the emotional salience of a hazard. Slovic's work established the importance of psychological dimensions in public risk judgments beyond formal estimates of danger. The inference for heat is that an externally measured risk ranking and a resident's everyday sense of threat can diverge systematically, particularly when dangerous conditions are normalized or the person compares them with prior summers rather than current health vulnerability Slovic (1987) Marlon et al. (2026).

Marlon and colleagues combined multiple years of United States survey data with the Centers for Disease Control and Prevention's Heat and Health Index to map differences between modeled public worry and assessed risk. Using data from 2018–2022, they identified substantial geographic misalignment, including regions where assessed vulnerability exceeded perceived risk. This is important evidence for the spatial distribution of the warning problem, but the study models risk perception at geographic scales. It does not demonstrate that a specific message, delivered to a specific person, caused or prevented a particular protective act Marlon et al. (2026).

A protective decision often begins with appraisal of the present situation rather than an abstract belief about climate change. People distinguish today's temperature, duration, humidity, indoor discomfort, family condition, and institutional alerts. In repeated hazards, they may rely on previous experience, leading to efficient routines or misleading familiarity. Survey measures should ask about the exact event, time window, setting, and threatened outcome, because a broad question about climate concern is an imperfect substitute for susceptibility to tonight's indoor heat or tomorrow's outdoor shift. Risk appraisal is an informative predictor, yet the strength of its association with action varies with response options and resources van Valkengoed et al. (2024) Kalkstein & Sheridan (2007).

Protection Motivation Theory and Self-Efficacy

Protection Motivation Theory, introduced by Rogers, describes the appraisal of threat and a proposed protective response. Its influential components include perceived severity and vulnerability, response efficacy, self-efficacy, and response costs. Maddux and Rogers sharpened the role of self-efficacy in protective motivation. Applied to extreme heat, the model asks whether a person regards the hazard as consequential, believes a cooling strategy can work, and believes that strategy is within personal reach. These appraisals can have different relationships with intentions and realized behavior, making them empirically distinguishable rather than interchangeable Rogers (1975) Maddux & Rogers (1983).

Bandura's conception of self-efficacy refers to beliefs about one's capacity to organize and execute action under given conditions. This concept matters acutely for heat adaptation because many recommended actions require access to a location, money, transportation, schedule flexibility, or another person's help. A resident who anticipates serious danger may still report low self-efficacy when the nearest cooling center is inaccessible. An intervention that improves knowledge but leaves transport unchanged may raise risk appraisal without increasing feasible protection. The theory thus directs attention toward both cognitive expectations and the opportunities through which beliefs can become action Bandura (1977) Grothmann & Patt (2005).

Longitudinal evidence provides a more demanding test of motivation than cross-sectional associations. Mehiriz and Cherqaoui followed older people and persons with chronic health conditions in Longueuil, Canada, using three surveys and individual- and time-fixed-effects analyses to study risk perception, response efficacy, and adaptation practices. The study directly addresses within-person change, but interpretation still depends on the timing and measurement of behavior and heat exposure. It illustrates the value of examining the credibility and feasibility of specific measures rather than assuming that raising concern alone will produce protection Mehiriz & Cherqaoui (2025).

Private Proactive Adaptation and Climate Cognition

Grothmann and Patt's Model of Private Proactive Adaptation to Climate Change foregrounds both perceived risk and the appraisal of adaptation options. It was developed to explain why resource availability alone could not fully account for adaptation differences. For household heat protection, the framework separates judgments about expected harm from judgments about the effectiveness and burden of potential responses. This distinction also helps analyze flood-proofing, evacuation preparation, or smoke filtration. The model is a theoretical organization of determinants, and its explanatory use calls for subsequent outcome-specific tests rather than automatic acceptance of any single directional pathway Grothmann & Patt (2005).

A longitudinal Dutch study by van Valkengoed, Perlaviciute, and Steg examined beliefs about climate change, risk perception, efficacy, and adaptation to flooding and heat stress. Their results distinguished intention from implemented behavior: perceptions of climate change and risk were associated with intentions, while efficacy beliefs related more directly to practical measures across the follow-up period. This finding complicates campaigns premised on the assumption that climate-change agreement is the decisive behavioral bottleneck. It suggests that the actionable unit for intervention is often a specific response, its anticipated outcome, and the capacity to perform it van Valkengoed et al. (2024).

Climate adaptation also involves competing goals. A household may have to balance thermal comfort against electricity bills, work obligations against rest breaks, or staying home against caregiving outside the home. Measuring response costs as a vague attitude can obscure materially different constraints. Costs can be financial, temporal, social, physical, and institutional; some are negotiable while others require policy change. A predictive model must therefore identify the level at which an intervention operates: individual learning, household coordination, organizational rules, or structural provision. Treating these levels as an integrated causal system provides a stronger empirical program than accumulating correlations between worry and generic willingness Grothmann & Patt (2005) Johar et al. (2025).

Protective Action Decision Model and Warning Processing

The Protective Action Decision Model developed by Lindell and Perry connects environmental cues, social signals, warning messages, attention, interpretation, perceived threat, possible actions, and action selection. Its value for heat is its insistence that communication takes place within an unfolding sequence. Recipients must encounter the message, believe the source, understand the relevance, identify a viable protective measure, and obtain resources before implementation. These steps can fail independently. A government alert can be technically accurate while practically ineffective when recipients do not know where to go or cannot interrupt paid work Lindell & Perry (2012) Weinberger et al. (2018).

A warning is more likely to enter the decision process when it is timely, comprehensible, location-specific, and associated with clear action options. However, empirical evidence on heat warnings is heterogeneous. Kalkstein and Sheridan's Phoenix survey connected higher risk perception with reported warning response among 201 participants; their design identified associations in a particular metropolitan context. The sample and cross-sectional measurement limit causal inference and generalization. Such studies are informative about reception, perceived danger, and possible communication gaps, while their mortality implications must be tested through independent outcome designs Kalkstein & Sheridan (2007) Lindell & Perry (2012).

Warnings for floods and smoke share elements with heat alerts but differ in response logistics. Flood evacuation may depend on route access, household members, pets, and rapidly changing water conditions. Smoke protection depends on air quality, building filtration, occupational exposure, and competing indoor heat. The common analytical process concerns the assessment of actionable threat and the organization of response; hazard-specific instructions should be based on specialized scientific and local emergency guidance. A single generic checklist cannot capture the tradeoffs produced by simultaneous heat, smoke, and power outages. The PADM provides a useful comparative structure without removing those distinct physical mechanisms Lindell & Perry (2012) CDC (2025b).

Attitudes, Norms, Intentions, and the Behavior Gap

The Theory of Planned Behavior relates behavioral intention to attitudes toward an act, perceived social expectations, and perceived behavioral control. Applied to heat protection, it encourages precise measurement of attitudes toward taking a cooling break, asking for assistance, or changing an outing rather than broad expressions of environmental concern. Perceived behavioral control may capture both confidence and expected constraint, making its operational definition important when financial or occupational limitations are substantial. Intentions are meaningful intermediate outcomes, yet they require separate validation against observed practice Ajzen (1991) van Valkengoed et al. (2024).

The intention–behavior gap is especially consequential for climate adaptation. People can accurately report what they plan to do and still remain exposed because of a sudden schedule change, a closed cooling center, inaccessible transportation, or a competing duty. The Dutch longitudinal findings underscore that beliefs and risk estimates associated with intentions need not predict completed measures with equivalent strength. Research should collect behaviors after an event, use time-stamped records where appropriate, and distinguish a newly adopted measure from a preexisting routine. Without these distinctions, an intervention can appear to succeed by increasing declared motivation while leaving the environmental dose unchanged van Valkengoed et al. (2024) Buller et al. (2026).

Social norms can support protection when supervisors encourage breaks, neighbors initiate check-ins, and community leaders publicly use cooling services. Norms can also suppress action if leaving a job site seems irresponsible or if requesting assistance is socially stigmatized. These mechanisms are hypotheses requiring context-specific observation. A behaviorally credible intervention considers the meaning of taking action within a person's social relationships and work role. The unit of analysis may be an individual intention, a mutually coordinated household plan, or an organizational practice; moving between units without measurement creates an ecological inference problem Ajzen (1991) WHO & WMO (2025).

Affect, Worry, Attention, and Normalization

Worry about heat can increase the salience of a hazard and motivate information seeking, but emotional concern does not uniquely determine protective action. Risk perceptions blend intuitive reactions with judgments about probability, severity, and controllability. Elevated distress may coincide with practical preparedness in some contexts and with helplessness in others. Research must specify whether an affective measure captures current fear, anticipated harm, trait anxiety, or concern about climate change as a broad social issue. The predictive role of affect should be evaluated alongside efficacy and resource conditions rather than as an isolated causal force Slovic (1987) Marlon et al. (2026).

Repeated experience with heat may generate learned competencies and habituated underestimation. Familiarity with high summer temperatures can make a warning less distinctive even when age, medication, humidity, or housing conditions increase personal vulnerability. Conversely, a previously severe episode may create lasting vigilance that facilitates rapid preparation. Cross-sectional survey patterns cannot determine which experiential process is dominant. Longitudinal within-person designs, comparable episodes, and records of protective actions can separate learning, seasonal acclimatization, changing exposure, and shifts in perceived danger Marlon et al. (2026) Mehiriz & Cherqaoui (2025).

Protective communication should therefore help people recognize specific thresholds, symptoms, and practical transitions in daily routines while maintaining a workable sense of agency. Exaggerating the emotional intensity of a message without improving response efficacy can leave the mechanism incomplete. Public-health agencies increasingly pair risk categories with suggested actions, including advance planning and identifying cool locations. Whether those suggestions translate to reduced morbidity should be evaluated through implementation and outcome studies. Behavioral science provides constructs for designing communication; the efficacy of any particular campaign is a separate empirical question CDC (2025b) Buller et al. (2026).

Household Decision-Making and Shared Protection

Households are decision systems in which age, care responsibilities, housing conditions, work shifts, language, mobility, and finances interact. A person may choose a safe place for an older relative while arranging child care or transportation. An early warning allows time to inventory needs, yet successful preparation depends on whether a household can coordinate multiple activities before the heat arrives. The household's risk should therefore be measured as the joint organization of protective opportunities, not merely the sum of individual attitudes. Studies that survey one respondent per dwelling may overlook negotiations and unequal distributions of labor Grothmann & Patt (2005) Johar et al. (2025).

Protective routines can be learned and rehearsed. Examples include knowing which rooms stay cooler, arranging check-ins with neighbors, documenting safe transport options, and discussing contingencies for electricity interruptions. These examples are mechanisms of planning, not a prescription that every household should use the same action. For some families, the primary need is access to a safe cooling environment or assistance with an energy bill; for others, the barrier is awareness of rising indoor temperatures. Effective program design begins with local conditions and uses intervention evaluation to determine whether the chosen support improves actual exposure and health Das et al. (2025) CDC (2025b).

A household approach creates an opportunity for equity-sensitive research. Adaptation costs may fall disproportionately on people who already provide unpaid care, have limited mobility, or cannot afford time away from work. Survey questions should identify whose action is reported, who bears a cost, and who receives the benefit. Community programs should assess spillovers: a neighbor's welfare check can protect one person without requiring the recipient to have personally sought a warning. This relational dimension explains why counting app downloads or household pamphlets can underrepresent the pathways that matter for vulnerable residents Lindell & Perry (2012) Johar et al. (2025).

Indoor Environments, Cooling, and Energy Security

Indoor heat is a crucial exposure pathway because many people spend long periods in buildings that accumulate heat and cool slowly overnight. Construction materials, insulation, shading, ventilation, occupant density, appliance use, and availability of air conditioning shape the indoor thermal environment. A forecast based on outdoor conditions is therefore an imperfect representation of individual dose. Community adaptation reviews examine interventions across built-environment changes, landscape strategies, and person-level practices; they show the breadth of possible intervention targets while emphasizing the importance of local design and evaluation Das et al. (2025).

Cooling access is simultaneously an engineering and behavioral question. A resident must know that safer conditions are needed, have a functioning source of cooling, be able to afford its use, and be willing and able to access it. In some settings a public cooling location is useful; in others transport barriers, opening hours, disability access, perceived safety, or caregiving duties reduce practical uptake. Effective evaluation therefore includes reach, utilization, indoor temperature, and health outcomes. A program should not infer success from the mere existence of cooling infrastructure or the number of people informed about its location Johar et al. (2025) CDC (2025b).

Electricity reliability adds a further layer. Heat waves can coincide with higher demand, while medically important equipment and temperature-sensitive medications may depend on power. Public-health guidance calls for anticipatory plans that account for these vulnerabilities, including discussion with appropriate clinicians when medical equipment or medicine is involved. Behavioral studies should record outages and alternative cooling options as time-varying exposures. Doing so prevents the analytically misleading assignment of low individual motivation to residents whose planned protection was interrupted by a system failure CDC (2025a) WHO & WMO (2025).

Older Adults, Chronic Illness, and Functional Capacity

Older adults and people with chronic medical conditions can face heightened harm from heat through physiological susceptibility, comorbidities, medication interactions, reduced mobility, and socially patterned living conditions. Risk is heterogeneous within age categories, making age an insufficient substitute for functional assessment. Household surveys should ask about living arrangements, access to cooling, physical limitations, chronic conditions when ethically and legally appropriate, and the capacity to obtain help. This emphasis on functional exposure aligns with research on protection motivation in populations whose adaptive options may be constrained Mehiriz & Cherqaoui (2025) CDC (2025a).

Protective behavior in this population may be jointly performed. A family member, caregiver, building manager, or community worker can interpret an alert, arrange transport, or check the indoor environment. That observation challenges research designs that attribute every action to a single respondent. Studies should distinguish independently performed behavior from assistance-dependent behavior and evaluate whether interventions reach persons who are socially isolated. Warning awareness is an important process measure but an insufficient substitute for behavior and preventable health outcomes, as intervention reviews repeatedly emphasize Johar et al. (2025) Buller et al. (2026).

Medication-related heat risk requires particularly careful scholarly treatment. The CDC identifies classes including some diuretics, anticholinergic drugs, and psychotropic medications as relevant to heat vulnerability, with mechanisms depending on the medicine and patient. Medication management is an individual clinical matter: changes require professional assessment, and abrupt stopping can itself cause harm. The psychology of protection involves recognizing the need for a clinician-supported hot-weather plan, preserving continuity of care, and reducing barriers to consultations. This monograph makes no medication dosing recommendations and does not equate population-level risk factors with personal clinical determinations CDC (2025a).

Occupational Heat, Control, and Worker Agency

Workplaces provide a decisive test of the difference between knowing how to prevent harm and possessing the power to act. Outdoor and indoor workers may face metabolic demands, protective clothing, direct radiation, limited rest opportunities, and productivity pressure. A systematic review and meta-analysis by Flouris and colleagues documented health and productivity consequences of occupational heat strain. Exposure depends on workload and working conditions, and a worker's protective behavior is constrained by rules, compensation, scheduling, and access to facilities. Individual education is a component of protection, but it cannot substitute for employer-controlled changes to exposure Flouris et al. (2018) WHO & WMO (2025).

Effective occupational strategies can include planned work–rest arrangements, cooled recovery environments, access to water, task rescheduling, heat acclimatization policies, and emergency recognition procedures, where locally appropriate and professionally established. Research should record whether these measures are available and whether workers can use them without penalty. Knowledge scores and self-reported hydration alone cannot demonstrate reduced heat illness. The WHO–WMO guidance emphasizes plans adapted to occupations and regions, stakeholder collaboration, and protection of workers with additional vulnerabilities, positioning prevention within institutional responsibility WHO & WMO (2025).

Buller and colleagues' 2026 systematic review of behavioral interventions for heat-illness prevention in community and occupational populations found a limited evidence base with substantial methodological constraints. The review indicates that information and behavioral programs may improve prevention knowledge and some practices but highlights the need for stronger comparison conditions and larger, more rigorous studies. The occupational lesson is not that behavior is irrelevant; it is that the observed behavior is partially produced by the organizational environment. Interventions should treat supervisor decisions, paid breaks, scheduling rights, and implementation fidelity as measurable causal components Buller et al. (2026).

Social Inequality, Housing, and Adaptation Constraints

Heat exposure is distributed across homes, neighborhoods, workplaces, and institutions that differ in physical quality and available resources. People with insecure housing, limited incomes, disabilities, inadequate transportation, or restricted access to medical care may have fewer ways to avoid prolonged exposure. These conditions enter the behavioral pathway as constraints on feasible choices. They also influence the meaning of survey responses: selecting 'I did not use cooling' does not establish lack of concern when the respondent had no affordable access to cooling. Research must identify capability before interpreting nonadoption as preference Marlon et al. (2026) Das et al. (2025).

Municipal heat action plans can have distributional consequences. Benmarhnia and colleagues evaluated Montreal's program using a difference-in-differences design applied to mortality records from summers before and after its 2004 implementation. Their findings were consistent with reduced mortality on hot days, with stronger estimated benefits in older populations and lower-education neighborhoods. As with any quasi-experiment, validity depends on comparison trends and possible concurrent changes. Nonetheless, the design supports a stronger causal argument than an unadjusted comparison of deaths before and after a policy Benmarhnia et al. (2016).

Equity-focused evaluation asks who receives an alert, who can execute its instructions, and who benefits in outcomes. It should assess reach separately for low-income tenants, informal workers, isolated residents, people with disabilities, and residents of heat-retaining neighborhoods. Administrative measures can obscure these differences if neighborhood averages are treated as individual conditions. Participatory design can uncover barriers that surveillance alone misses, while quantitative comparison can test whether resulting interventions improve exposure and health. These approaches are complementary parts of a rigorous evaluation program Johar et al. (2025) Benmarhnia et al. (2016).

Heat Warnings and the Evidence of Effectiveness

The causal chain for a heat-health warning spans meteorological prediction, agency issuance, message dissemination, comprehension, protective action, exposure reduction, and health outcomes. Evidence at one link cannot validate the entire chain. Ebi and colleagues estimated benefits associated with Philadelphia's heat watch/warning system, an important early attempt to connect public alerting with mortality consequences. Because such estimates depend on modeled counterfactual outcomes and policy context, their assumptions must be read alongside studies with other comparison strategies Ebi et al. (2004).

Weinberger and colleagues examined heat alerts in twenty United States cities during 2001–2006 using a case-crossover approach that compared days of similar observed heat index with and without alerts. Across the pooled cities, their estimate was a −0.5% change in mortality rate, with a 95% confidence interval from −2.8% to 1.9%; the association was not statistically distinguishable from zero. Philadelphia showed a beneficial association, while most individual cities did not. This result directly qualifies broad claims that issuing an alert necessarily prevents deaths. The analysis concerns its study period and alert systems rather than every subsequent program Weinberger et al. (2018).

The Montreal analysis, in contrast, evaluated a larger heat action plan incorporating interventions beyond a warning and detected mortality benefits under its quasi-experimental assumptions. Together, the studies suggest a scientifically consequential distinction between alert issuance and a comprehensive operational response. A warning may be effective when it activates outreach, cooling access, worker protection, surveillance, and medical preparedness. However, attributing the entire benefit to any one component requires a design that isolates that component or uses credible implementation variation Benmarhnia et al. (2016) Weinberger et al. (2018).

Community Interventions and Heat Literacy

Heat literacy encompasses recognition of heat as a health hazard, identification of susceptible circumstances, understanding of prevention options, and competence in recognizing when help is needed. Information programs often attempt to move these constructs together, but empirical outcomes should be measured separately. Johar and colleagues' 2025 systematic review selected ten community-based behavioral intervention studies from an initial retrieval of 1,266 articles; seven included studies came from high-income countries. Findings indicated improvements in knowledge and some symptoms, while behavioral and health effects were mixed and methods uneven Johar et al. (2025).

A major scientific contribution of the review was its critique of intervention design. It found that included studies had not systematically incorporated established heat-literacy frameworks, behavioral theory, or participatory community involvement across the research process. The resulting evidence cannot answer every question about implementation quality or generalizability. Research that uses a clearly stated theory should still assess whether its hypothesized mechanism changed: enhanced response efficacy, reduced barriers, stronger social coordination, or more accessible cooling. That distinction converts heat literacy from a broad educational slogan into testable constructs linked to outcomes Johar et al. (2025) Grothmann & Patt (2005).

Buller and colleagues examined a broader set of behavioral interventions in community and occupational contexts, highlighting the limited rigor of available studies. The synthesis suggests that education can influence proximal outcomes while leaving uncertainty about long-term health protection. Evaluation should separate immediate comprehension, retained knowledge, reported acts, verified environmental exposure, and clinical events. Programs adapted for language and local culture may improve acceptability, but their effectiveness must be demonstrated rather than inferred from the good intentions of communication designers Buller et al. (2026) Johar et al. (2025).

Landscape, Buildings, and Structural Adaptation

Climate adaptation includes physical changes to neighborhoods and buildings that alter the conditions in which individual decisions are made. Shade, vegetation, reflective or modified materials, ventilation, cooling infrastructure, and urban planning can affect surface or indoor temperature. Das and colleagues' 2025 systematic review evaluated a broad literature on community-based heat adaptation, covering landscape modifications, building interventions, and targeted individual-level measures. This breadth matters because changing exposure at its source can reduce the demands placed on personal vigilance during repeated events Das et al. (2025).

Built-environment measures require their own outcome hierarchy. A cooler pavement surface is not automatically evidence of lower indoor heat, fewer emergency visits, or improved mortality; these are connected yet different endpoints. Thermal performance depends on weather, spatial configuration, use patterns, maintenance, and the population actually exposed. Evaluations should identify possible rebound behavior, displacement of exposure, and unequal distribution of infrastructure. For residents with limited individual control, a reliable environmental change may be more consequential than an additional instruction to 'stay cool' Das et al. (2025) Gasparrini et al. (2015).

Adaptation policies can also create tradeoffs. Cooling equipment requires energy and maintenance; shaded public areas require safe access; changes to ventilation may interact with ambient air pollution or smoke. A scientifically responsible synthesis compares co-benefits and unintended consequences with explicit setting-specific data. Psychological relevance comes from how these interventions change perceived and actual choice sets. When the environment offers feasible safe alternatives, protective intentions are more likely to become observable action, a hypothesis consistent with efficacy-centered adaptation models that requires direct field testing Grothmann & Patt (2005) van Valkengoed et al. (2024).

Heat and Other Hazards as Compound Decisions

The broader scientific passport includes flooding, smoke, and chronic climate stress because households rarely face hazards in isolation. Protective actions can conflict: ventilating a building during a hot night may be incompatible with minimizing outdoor smoke exposure; sheltering in place during a flood may be unsafe when evacuation is needed; a power failure may remove both cooling and electronic warning access. Comparative disaster psychology can investigate how people select between incompatible recommendations, update decisions as conditions change, and coordinate multiple household members under uncertainty Lindell & Perry (2012) CDC (2025b).

Hazard-specific physical measurements remain indispensable. HeatRisk categories, air quality indicators, flood forecasts, and local emergency orders address different phenomena. An effective multihazard protocol specifies the exposure being reduced, the costs and contraindications of each action, and the authority of local warnings. The psychology of protective choice examines information conflict, attentional prioritization, perceived tradeoffs, and implementation barriers. It should not substitute a generalized climate-worry scale for the concrete risk management demanded by each hazard CDC (2025b) Slovic (1987).

Longitudinal climate adaptation studies can follow households across seasons and document changes in housing, work, weather, personal health, and social networks. Their advantage is the ability to distinguish temporary event response from durable preparation and to assess whether past hazard experience changes future capability. Such designs must also address attrition and time-varying confounding. A seemingly adaptive resident may have moved to a better-insulated dwelling, altered employment, or acquired family assistance. The resulting behavioral improvement should be interpreted in light of those changes rather than assigned wholly to a shift in perception van Valkengoed et al. (2024) Grothmann & Patt (2005).

Children, Caregiving, and Developmental Context

Children's exposure and protection are often mediated by adult routines, institutional schedules, and the design of schools and child-care environments. The relevant unit of analysis includes caregivers, educators, transportation providers, and the built setting. Infants and children differ in physiologic and behavioral capacities, but researchers should avoid treating all developmental stages as equivalent. Climate adaptation studies can examine whether adults recognize age-specific vulnerability, whether organizations provide cooling and schedule flexibility, and whether warning systems communicate clearly with the people who have practical responsibility for action WHO (2026a) WHO & WMO (2025).

Caregiving can generate difficult conflicts when a household has limited cooling, must travel to medical appointments, or contains several people with different needs. A parent who remains outdoors while arranging a safer environment for a child may appear behaviorally inconsistent if the survey records only individual exposure. Measuring who makes the decision, who executes it, and whose risk is reduced resolves some of this ambiguity. Protection may require both informed decisions and reliable institutional services, especially when transportation or school schedules cannot readily be changed Lindell & Perry (2012) Das et al. (2025).

Developmental research also offers opportunities to study learning. Repeated emergency drills, age-appropriate communication, and classroom instruction can establish knowledge and routines; their effectiveness for heat safety must be evaluated through actual preparedness and behavior rather than assumed from the presence of a curriculum. Studies should consider caregiver consent, privacy, accessibility, and protection from unsafe experimental heat exposure. The practical implication is to design studies in ordinary settings while preserving safeguards, rather than deliberately exposing vulnerable participants to hazardous temperatures Buller et al. (2026) Johar et al. (2025).

Mass Gatherings and Collective Risk Management

Heat at mass gatherings illustrates the interaction of personal knowledge with crowd-level constraints. Spectators and workers may stand for long periods, travel from unfamiliar climates, have limited access to water or shaded space, and face restrictions on movement. A person can recognize discomfort but remain unable to reach cooling or leave safely. Organizers control many relevant variables, including crowd flow, scheduling, access routes, sanitation, on-site care, and risk messaging. Consequently, intervention studies should measure venue design and operational decisions alongside individual self-protection WHO (2026a).

The World Health Organization's September 2026 evidence review examined cooling facilities, rest areas, breaks, hydration, integrated warnings, crowd management, sensing tools, and visitor knowledge through three systematic reviews. The companion implementation materials position heat safety as preparation, mitigation, and recovery across the event lifecycle. These are institutional syntheses and guidance materials, not evidence that every proposed intervention has already demonstrated the same magnitude of benefit. Their value for psychology lies in identifying conditions under which a person has actual access to a protective response WHO (2026a) WHO (2026b).

Mass events also expose the inadequacy of one-size-fits-all risk messages. Visitors differ in familiarity with local conditions, language, health status, economic resources, and the significance of activities they came to attend. Short, accessible, action-specific information can complement operational measures, but warnings cannot replace safe physical conditions. Rigorous evaluation should connect alert receipt with movement patterns, cooling utilization, physiological indicators when feasible, on-site health encounters, and any adverse effects of crowd-management choices WHO (2026a) Lindell & Perry (2012).

Artificial Systems, Forecasts, and Human Decision Support

Digital services increasingly mediate how people encounter heat forecasts, maps, notifications, and individual recommendations. The psychological research question concerns how such systems alter attention, comprehension, credibility, perceived control, and behavior. CDC HeatRisk is an example of a health-oriented forecast that translates anticipated conditions into a multilevel risk scale and corresponding protective suggestions. A technically sophisticated forecast can still fail when recipients cannot interpret the scale or act on its recommendations. Evaluation should test human–system comprehension and subsequent behavior, not just predictive accuracy CDC (2025b) Lindell & Perry (2012).

Artificial systems may support personalization by combining local environmental data with user-supplied constraints, provided privacy, accessibility, calibration, and appropriate medical oversight are protected. A decision-support interface could help a user locate available cooling, anticipate transport limitations, or coordinate welfare checks. These are prospective design possibilities, not established effects of artificial cognition on health. Research must compare algorithm-supported decisions with relevant nonalgorithmic alternatives and examine disparate error rates across neighborhoods, languages, and vulnerable groups. In this sense, the Artificial Era supplies a new context for the empirical study of distributed protective action Marlon et al. (2026) WHO (2026b).

A central conceptual distinction is between a system accurately classifying hazard and a household achieving safety. Forecast models operate on meteorological or epidemiological variables; people act within social arrangements, commitments, and environments. An interface can improve information while leaving the physical opportunity set unchanged. Artificial mediation should therefore be assessed through linked technical and behavioral outcomes: calibration, warning delivery, interpretation, feasible action, verified exposure reduction, and clinical benefit. Authorship's wider philosophical horizon does not replace these scientific tests or imply that conceptual claims about Artificial are independently established psychological evidence Grothmann & Patt (2005) Weinberger et al. (2018).

Experimental and Observational Research Designs

Research on climate protective behavior includes cross-sectional surveys, longitudinal panels, message experiments, natural experiments, quasi-experimental policy evaluations, exposure cohorts, time-series models, and mixed-methods studies. Each design answers different questions. A vignette can estimate the effect of message framing on stated intention but generally cannot establish how people will behave when the cost of action is real. A panel can describe within-person change yet remain vulnerable to time-varying confounding. A temperature–mortality study estimates population exposure associations but usually does not observe the individual protective decision van Valkengoed et al. (2024) Gasparrini et al. (2015).

Quasi-experimental evaluation is useful when warnings or heat action plans are implemented at identifiable times or thresholds. Benmarhnia and colleagues applied difference-in-differences to Montreal's heat plan and examined subgroup differences. Weinberger and colleagues exploited variation in alerts among days with similar observed heat index across twenty cities. Their different results reflect distinct policies, estimands, and assumptions, not necessarily a direct contradiction about an identical intervention. Researchers must explicitly define treatment, comparison, exposure window, confounders, and outcome and consider whether anticipatory behavior contaminates the nominal control condition Benmarhnia et al. (2016) Weinberger et al. (2018).

Clinical safety is essential for heat research. Controlled experiments should measure cognitive performance or perceptions only under ethically acceptable and medically supervised exposure conditions, with clear exclusion criteria and stopping rules. Field experiments can randomize communication or resource access without exposing participants to dangerous heat. To evaluate morbidity and mortality, researchers may use observational health databases under suitable ethics and data-governance controls. The relative evidentiary strength of a study emerges from the fit between its design and its claim, not from the mere presence of a statistically significant coefficient Buller et al. (2026) Johar et al. (2025).

Measurement, Psychometrics, and Behavioral Validity

A central measurement challenge is that protective behavior can refer to intention, self-report, observed implementation, or effective exposure reduction. These are sequential but nonidentical constructs. A respondent may report drinking more water while still experiencing substantial heat strain from occupational work; another may use a cooling center but only after prolonged exposure. Researchers should specify behavioral items, time windows, frequency, duration, and context. Psychometric tests should examine whether scales operate equivalently across cultural groups, ages, and disability-related response constraints before interpreting group differences as substantive Buller et al. (2026) Lindell & Perry (2012).

Risk perception itself has multiple dimensions: perceived likelihood, seriousness, worry, anticipated personal impact, perceived agency, response costs, and trust in messengers. Combining them into one global score can conceal distinct predictors and obscure intervention targets. The RAP framework developed by Marlon and colleagues is valuable for assessing a geographic alignment gap between public worry and assessed health risk, but its county-level outputs should not be interpreted as individual diagnostic scores. Cross-scale validity requires separate evaluation, especially when the assessed exposure and the survey response refer to different time horizons Marlon et al. (2026) Slovic (1987).

Objective measures improve inference but bring their own limitations. Indoor sensors can record temperature yet fail to capture an individual's movements; wearable devices can approximate thermal stress but pose acceptability and privacy challenges; mobility data can reveal access to cooling but cannot always establish actual exposure conditions. Administrative records may undercount less severe heat illness or reflect differential access to care. The most credible measurement strategy links sources through a prespecified causal model, reports missingness and calibration, and considers whether the act of monitoring changes behavior Gasparrini et al. (2015) WHO (2026a).

Review of Intervention Evidence and Its Limits

Recent evidence syntheses underscore that the volume of climate adaptation literature exceeds the number of robust causal evaluations of protective behavior. Johar and colleagues found ten eligible community intervention studies in their focused review, with uneven methodological rigor and mixed behavioral and health outcomes. Buller and colleagues similarly identified important design limitations in behavioral heat-illness prevention studies. These reviews support cautious optimism about education and outreach while resisting a generalized assertion that any single campaign reliably reduces deaths or illness Johar et al. (2025) Buller et al. (2026).

The broader review by Das and colleagues included 141 studies across landscape, building, and individual interventions. Its greater number does not mean it provides a direct meta-analytic estimate of a common psychological treatment effect. Cooling roofs, changing pavement, warning households, and modifying clothing target different physical pathways and have different outcome scales. Scientifically useful synthesis should distinguish interventions by mechanism, study design, baseline climate, participant characteristics, and outcome. A pooled effectiveness label can obscure crucial evidence when units and causal assumptions are incompatible Das et al. (2025).

A credible intervention synthesis therefore grades certainty at the claim level. Direct observations of mortality and heat-related hospitalization have high relevance but may suffer confounding. Randomized message studies can identify changes in comprehension but may provide limited evidence on health. Self-reported protective acts may be susceptible to recall and social desirability. Successful deployment of an alert platform proves technical functionality but says little about reach or impact. Program evaluation becomes more informative when it measures intermediate mechanisms and final outcomes within a common causal framework Weinberger et al. (2018) Benmarhnia et al. (2016).

Causal Identification and Interpretive Discipline

Researchers must distinguish association, prediction, and causation. A positive correlation between risk perception and protective behavior may arise because perceived danger motivates action, because an earlier adverse event caused both, or because people who act report heightened concern. Temporal ordering and control for plausible confounders improve interpretation, but survey adjustment cannot guarantee causal identification. Fixed-effects designs can remove stable unmeasured differences while leaving time-varying confounding. Therefore even longitudinal evidence should be discussed with its measured estimand and assumptions Mehiriz & Cherqaoui (2025) van Valkengoed et al. (2024).

At the policy level, difference-in-differences assumes an appropriate comparison trajectory in the absence of the intervention. Threshold-based alert studies depend on how alerts are issued, how forecast errors relate to actual risk, and whether other preventive responses differ across compared days. Measured reductions in mortality are an unusually consequential outcome, yet mechanisms may remain unidentified if several interventions begin together. The Montreal and twenty-city analyses should be read as complementary designs testing different policy packages rather than as simple votes for or against warning systems Benmarhnia et al. (2016) Weinberger et al. (2018).

Risk of bias also depends on who participates. Community surveys may underrepresent people without internet access, isolated older adults, mobile workers, or persons in informal housing. Intervention trials may recruit highly motivated volunteers and deliver programs more intensively than ordinary services can sustain. Meta-analyses can aggregate biased studies with great numerical precision. Evidence should therefore be evaluated for selection, loss to follow-up, missing outcomes, intervention fidelity, outcome measurement, and transportability. The leading reviews themselves acknowledge shortcomings in available heat behavior interventions Johar et al. (2025) Buller et al. (2026).

Geographic, Cultural, and Temporal Generalizability

Extreme heat is experienced against a local climatic baseline. A temperature considered unusual in one city may be routine in another, while indoor construction, humidity, power reliability, and cultural practices affect the practical meaning of that experience. Marlon and colleagues document pronounced geographic variation in alignment between assessed and perceived heat risk across the contiguous United States. Geographic patterns can inform local communication priorities but cannot automatically justify individual-level prescriptions or transfer the effectiveness of a strategy across climates without testing Marlon et al. (2026) Gasparrini et al. (2015).

The intervention literature also displays uneven geographic coverage. In Johar and colleagues' ten-study review, seven studies were conducted in high-income countries, limiting confidence about applicability where housing, public services, and health systems differ. A 2026 review of public heat perceptions further identified research gaps in geographic settings with substantial heat vulnerability. Such imbalances matter both scientifically and ethically because adaptation options are shaped by resources and institutions. Future studies should include diverse climatic zones, informal settlements, and populations exposed through work or caregiving, with local researchers involved in defining outcomes Johar et al. (2025) Marlon et al. (2026).

Time matters as much as place. Heat acclimatization, seasonal warning fatigue, technological adoption, household moves, aging, and public investment can change both vulnerability and protective capacity. A study undertaken during an exceptional summer may measure emergency response, while a study across ordinary years may capture gradual adaptation. Historical studies of 2001–2006 warning systems remain informative about their actual implementations, but should not be portrayed as direct assessments of present-day forecasting technologies. Longitudinal and repeated cross-city designs are needed to investigate changes without assuming stable causal relationships Weinberger et al. (2018) Mehiriz & Cherqaoui (2025).

Integrated Scientific Synthesis

The literature supports a multilevel interpretation of protection. Environmental risk creates potential harm through exposure, while perception influences the detection and interpretation of danger. Response appraisal and efficacy beliefs affect the selection of a protective measure. Implementation depends on material resources, institutional permissions, and interpersonal coordination. The chosen act must then produce sufficient exposure reduction to alter physiological and clinical outcomes. Each relation is theoretically plausible and supported by portions of the evidence, but complete pathway estimates remain scarce. A comprehensive account therefore combines behavioral theories with exposure measurement and outcome evaluation Grothmann & Patt (2005) Lindell & Perry (2012) Benmarhnia et al. (2016).

The strongest practical inference is that warning quality, individual capability, and structural protection should be designed together. An alert with no accessible cooling, transportation, worker protections, or medically appropriate contingencies provides an incomplete response system. A structural intervention without trusted public information may also fail to reach intended beneficiaries. Evidence from heat action plans and reviews of community interventions points toward bundled approaches, while uncertainty remains regarding which components cause the observed benefits in particular settings. Integration is a research hypothesis that must be tested through thoughtful comparative designs Das et al. (2025) Weinberger et al. (2018) WHO & WMO (2025).

This synthesis also clarifies why environmental psychology has an independent role within climate-health science. Epidemiology identifies when and where mortality risk increases; physiology explains heat strain; meteorology predicts hazardous conditions; and psychology investigates the organization of attention, appraisal, intention, coordination, and action. These components meet in the same real-world protective event. A research dossier focused on feasible action can connect them without reducing them to a single construct. Scientific progress lies in specifying the causal links and measuring them with the methods appropriate to each level Gasparrini et al. (2015) Buller et al. (2026).

Open Questions and Research Priorities

A major unresolved question concerns the conversion of warning receipt into verified protection. Research needs to identify what share of variation is explained by risk perception, response efficacy, housing and work constraints, and the timing of messages. Different populations may have different bottlenecks. Preregistered field experiments could compare action-oriented communication with communication plus transport support, workplace schedule changes, or guaranteed cooling access. Such studies should include prespecified health outcomes and distributional analyses where feasible, rather than treating increased information as the final endpoint Johar et al. (2025) Weinberger et al. (2018).

A second priority is better causal evidence on interventions and their sustainability. Short educational programs can shift immediate knowledge; long-term benefits may require continued resources and institutions. Researchers should measure exposure and health across multiple summers, investigate policy maintenance, and report financial and environmental costs. Natural experiments involving the expansion of cooling access or changes to labor rules may provide stronger causal leverage than another cross-sectional survey. Comparisons across cities should address baseline climate, urban form, demographic selection, and differing implementation fidelity Benmarhnia et al. (2016) Das et al. (2025).

A third priority addresses multihazard response and digitally mediated decisions. Heat, poor air quality, floods, and power failures can jointly constrain safe options. Artificial decision support may help organize information and local services, but must be evaluated for calibration, accessibility, privacy, equity, and downstream behavior. Particular attention should be given to disability, age, language, informal work, and locations poorly covered by published intervention studies. The central scientific objective is to determine when a technically available protective action becomes a socially and physically realizable one Marlon et al. (2026) WHO (2026a).

Conclusion

Climate adaptation under extreme heat is an organized behavioral process embedded in material and institutional environments. The evidence distinguishes assessed hazard, perceived risk, response efficacy, intention, feasible action, measured exposure, and health outcomes as separate components of a research program. Protection Motivation Theory, private proactive adaptation models, the Protective Action Decision Model, and intention-based accounts explain complementary parts of this sequence. Their usefulness depends on clear operational definitions and tests of specific protective actions under real-world constraints Rogers (1975) Grothmann & Patt (2005) Lindell & Perry (2012) Ajzen (1991).

Historical and contemporary studies establish both the seriousness of heat risk and the limits of generalized claims about warning effectiveness. Mortality studies document exposure consequences; population evaluations of alerts report mixed results; Montreal's heat action plan shows promising quasi-experimental evidence; and systematic reviews identify improvements in some educational outcomes amid persistent uncertainty about behavioral and clinical effects. The scientific synthesis therefore supports an integrated program of communication, accessible cooling, occupational protections, household preparedness, and outcome-oriented evaluation while preserving the distinction between tested effects and plausible mechanisms Gasparrini et al. (2015) Weinberger et al. (2018) Benmarhnia et al. (2016) Johar et al. (2025).

The durable research question is whether the social and physical conditions necessary for protection are present when a hazard occurs. A person acts within relations of care, work, housing, public infrastructure, and information systems. Future research can strengthen climate-health protection by measuring this complete pathway, reporting contradictory results, and evaluating interventions according to verified exposure reduction and equitable health benefit. The Artificial Era expands the means of forecasting and decision support, while the scientific standard remains the same: traceable claims, operationally defined mechanisms, appropriate comparison designs, and outcomes that matter to those exposed Buller et al. (2026) WHO (2026a).

References

  • Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179–211. https://doi.org/10.1016/0749-5978(91)90020-T
  • Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84(2), 191–215. https://doi.org/10.1037/0033-295X.84.2.191
  • Benmarhnia, T., Bailey, Z., Kaiser, D., Auger, N., King, N., & Kaufman, J. S. (2016). A difference-in-differences approach to assess the effect of a heat action plan on heat-related mortality, and differences in effectiveness according to sex, age, and socioeconomic status (Montreal, Quebec). Environmental Health Perspectives, 124(11), 1694–1699. https://doi.org/10.1289/EHP203
  • Buller, D. B., Adjei, I., Agarwal, R., Peterson, S. R., Escoffery, C., Buller, M. K., Walkosz, B. J., Anderson, K., & Morshed, A. B. (2026). Systematic review and narrative synthesis of behavioral interventions to improve heat-related illness prevention in community and occupational populations. BMC Public Health, 26, Article 367. https://doi.org/10.1186/s12889-025-25499-7
  • Centers for Disease Control and Prevention. (2025). Heat and medications: Guidance for clinicians. https://www.cdc.gov/heat-health/hcp/clinical-guidance/heat-and-medications-guidance-for-clinicians.html
  • Centers for Disease Control and Prevention. (2025). How to use the HeatRisk tool and Air Quality Index. https://www.cdc.gov/heat-health/hcp/clinical-guidance/how-to-use-the-heatrisk-tool-and-air-quality-index.html
  • Das, J. K., Kumari, B., Rahman, A. R., Azhar, M., Khan, S., Masood, U., Karan, D., Yasin, R., Burman, C., Augustin, J. L., Bonell, A., & Bhutta, Z. A. (2025). Evaluation of community-based heat adaptation interventions: A systematic review. BMJ Public Health, 3(2), e002332. https://doi.org/10.1136/bmjph-2024-002332
  • Ebi, K. L., Teisberg, T. J., Kalkstein, L. S., Robinson, L., & Weiher, R. F. (2004). Heat watch/warning systems save lives: Estimated costs and benefits for Philadelphia 1995–98. Bulletin of the American Meteorological Society, 85(8), 1067–1074. https://doi.org/10.1175/BAMS-85-8-1067
  • Flouris, A. D., Dinas, P. C., Ioannou, L. G., Nybo, L., Havenith, G., Kenny, G. P., & Kjellstrom, T. (2018). Workers' health and productivity under occupational heat strain: A systematic review and meta-analysis. The Lancet Planetary Health, 2(12), e521–e531. https://doi.org/10.1016/S2542-5196(18)30237-7
  • Fouillet, A., Rey, G., Laurent, F., Pavillon, G., Bellec, S., Guihenneuc-Jouyaux, C., Clavel, J., Jougla, E., & Hémon, D. (2006). Excess mortality related to the August 2003 heat wave in France. International Archives of Occupational and Environmental Health, 80(1), 16–24. https://doi.org/10.1007/s00420-006-0089-4
  • Gasparrini, A., Guo, Y., Hashizume, M., Lavigne, E., Zanobetti, A., Schwartz, J., Tobias, A., Tong, S., Rocklöv, J., Forsberg, B., Leone, M., De Sario, M., Bell, M. L., Guo, Y.-L. L., Wu, C.-f., Kan, H., Yi, S.-M., Coelho, M. S. Z. S., Saldiva, P. H. N., Honda, Y., Kim, H., & Armstrong, B. (2015). Mortality risk attributable to high and low ambient temperature: A multicountry observational study. The Lancet, 386(9991), 369–375. https://doi.org/10.1016/S0140-6736(14)62114-0
  • Grothmann, T., & Patt, A. (2005). Adaptive capacity and human cognition: The process of individual adaptation to climate change. Global Environmental Change, 15(3), 199–213. https://doi.org/10.1016/j.gloenvcha.2005.01.002
  • Johar, H., Abdulsalam, F. I., Guo, Y., Baernighausen, T., Jahan, N. K., Watterson, J., Leder, K., Gouwanda, D., Ramanathan, G. R. L., Lee, K. K. C., Mohamed, N., Zakaria, T. A., Barteit, S., & Su, T. T. (2025). Community-based heat adaptation interventions for improving heat literacy, behaviours, and health outcomes: A systematic review. The Lancet Planetary Health, 9(7), Article 101207. https://doi.org/10.1016/S2542-5196(25)00007-5
  • Kalkstein, A. J., & Sheridan, S. C. (2007). The social impacts of the heat-health watch/warning system in Phoenix, Arizona: Assessing the perceived risk and response of the public. International Journal of Biometeorology, 52(1), 43–55. https://doi.org/10.1007/s00484-006-0073-4
  • Lindell, M. K., & Perry, R. W. (2012). The protective action decision model: Theoretical modifications and additional evidence. Risk Analysis, 32(4), 616–632. https://doi.org/10.1111/j.1539-6924.2011.01647.x
  • Maddux, J. E., & Rogers, R. W. (1983). Protection motivation and self-efficacy: A revised theory of fear appeals and attitude change. Journal of Experimental Social Psychology, 19(5), 469–479. https://doi.org/10.1016/0022-1031(83)90023-9
  • Marlon, J. R., Begotka, N., Preston, A., Steifel, M., Keenan, J. M., Inman, C., Malmin, N. P., Carman, J., Goddard, E., Kotcher, J., Maibach, E., Rosenthal, S. A., & Leiserowitz, A. (2026). Experience-driven perceptions misalign with assessed heat risk in the United States. Nature Communications, 17, Article 909. https://doi.org/10.1038/s41467-025-67631-6
  • Mehiriz, K., & Cherqaoui, S. (2025). A longitudinal study of the effects of risk perception and response efficacy on adaptation of elderly individuals and individuals with chronic health conditions to heat waves. BMC Public Health, 25, Article 4295. https://doi.org/10.1186/s12889-025-25348-7
  • Rogers, R. W. (1975). A protection motivation theory of fear appeals and attitude change. The Journal of Psychology, 91(1), 93–114. https://doi.org/10.1080/00223980.1975.9915803
  • Slovic, P. (1987). Perception of risk. Science, 236(4799), 280–285. https://doi.org/10.1126/science.3563507
  • van Valkengoed, A. M., Perlaviciute, G., & Steg, L. (2024). From believing in climate change to adapting to climate change: The role of risk perception and efficacy beliefs. Risk Analysis, 44(3), 553–565. https://doi.org/10.1111/risa.14193
  • Weinberger, K. R., Zanobetti, A., Schwartz, J., & Wellenius, G. A. (2018). Effectiveness of National Weather Service heat alerts in preventing mortality in 20 US cities. Environment International, 116, 30–38. https://doi.org/10.1016/j.envint.2018.03.028
  • World Health Organization & World Meteorological Organization. (2025). Climate change and workplace heat stress: Technical report and guidance. World Health Organization. https://www.who.int/publications/i/9789240099814
  • World Health Organization. (2026). Prevention and management of heat-related illness at mass gatherings: Developing evidence-based communication messages. World Health Organization. https://www.who.int/publications/i/item/9789240124257
  • World Health Organization. (2026). Prevention and management of heat-related illness at mass gatherings: Systematic reviews. World Health Organization. https://www.who.int/publications/i/item/B09849