Exploring the health impacts of climate change: Challenges and considerations for health services research

The United Nations has declared a state of climate emergency regarding the unmitigated effects of global climate change, in large part due to its impacts on human safety and well-being.1 While climate scientists have worked to develop a deeper understanding of the influence of global warming on the environment, exploration of the human health dimensions are ongoing. One particular area of growing investigation is that of climate change's impacts on health services and outcomes. Indeed, the issue has been deemed critical enough such that the Agency for Healthcare Research and Quality (AHRQ) convened a roundtable discussion with the Department of Health and Human Services (HHS) in 2023 on the subject of measuring climate resiliency for healthcare delivery organizations.2 Underlying the relationship between environment and health are a series of methodological questions about how best to develop assessments of these relationships for exploring health services outcomes. The National Institutes of Health have also been working with stakeholders and the scientific community to create a strategic framework upon which to evaluate climate outcomes combined with an institute-wide initiative into this field. The goal of this endeavor appears to be reducing health threats from climate change by examining how climate drivers lead to populations being exposed to new environmental factors while identifying the health outcomes of these conditions, and also addressing the socioeconomic context within which these problems occur.3 Health services researchers are, and will continue to be, playing a major role in such efforts. In this commentary, we highlight three different manifestations of global warming: extreme precipitation events, wildfires, and heat waves, as illustrative examples of the challenges they can pose for health services researchers. For each of these phenomena, we briefly review what is known about their climate links and their relationship with human health, then provide some methodologic considerations in examining their effects on health systems and outcomes. While not a comprehensive review of the myriad ways in which our changing climate is impacting our well-being, we hope this discussion provides some points for contemplation among investigators and policymakers in this growing scientific field. Perhaps the most visible manifestation of climate change for the United States has been that of extreme precipitation events, chief among them worsened hurricane and storm seasons. Rising global temperatures are elevating surface ocean temperatures, resulting in more intense and extreme precipitation events: storms have presented with higher total rainfall, more powerful winds, and more severe storm surges.4 In addition to stronger sea-based storms, climate change is also affecting inland storms through fluctuant temperatures, leading to dangerous anomalous phenomena such as off-season tornadoes, polar vortexes, and precipitation-associated flash flooding.5 These natural disasters can have significant health impacts in a variety of ways. Traumatic injuries and deaths frequently follow their physical devastation, as has recently been tragically demonstrated by the death toll of Hurricane Helene in the eastern United States, with over 220 mortalities and rising at this time of this writing.6 Yet exposure to these weather events has also been found to be associated with lingering long-term health impacts at the community level.7 For example, studies conducted in New Orleans have shown that hospital admissions for heart attacks at the Tulane Medical Center increased threefold in the decade following Hurricane Katrina.8 These were found to be accompanied by a significant rise in key risk factors for cardiovascular disease such as high blood pressure, diabetes, elevated cholesterol levels, and tobacco smoking.8 In addition, mental health crises from sustained emotional trauma, stress, difficulty accessing resources, and a host of other healthcare delivery complications can create lasting medical problems during natural disasters.9 Regrettably, such weather catastrophes are not limited to the warm weather months. Winter blizzards and colder snaps have been shown to cause a rise in cardiovascular emergencies, with lower temperatures and large-scale temperature swings being correlated to increased heart attack and stroke rates.10-13 While anomalous weather events have the potential for massive impacts across large populations (there are currently more than 126 million people in North America living in regions at risk for a major hurricane), and discrete disasters are easy to define, measuring their resultant harms can be challenging.14 Such disasters can have destructive effects on healthcare infrastructure, forcing patients to divert to other health centers or to forgo scheduled appointments. Regional registries or hospital-based medical records may thus have interrupted or missing utilization data (particularly if power outages or evacuations limit primary data collection). Additionally, regional and national medication and medical supply shortages can have ripple effects, impacting the quality of care even at distant health centers. Finally, as noted previously, it is becoming increasingly obvious that the medical sequelae of major natural disasters can persist for years, and it is unknown if they may even compound over multiple severe storm seasons. Patients often forgo necessary healthcare during disasters, and attributing elevations in adverse outcomes to the weather event can be challenging, as it involves disentangling true exposure-related effects from delayed usual presentations, impacts of delaying needed care, and the so-called "harvesting effect." The last of these refers to a known epidemiologic phenomenon where anticipated deaths among chronically ill individuals are displaced forward in time by natural disasters, implying that some of the acute mortality burden following such events represents an acceleration of expected fatalities rather than those directly attributable to the disaster.15 Changes in maladaptive behavior in the wake of traumatic events can also be notoriously difficult to measure in vulnerable populations, particularly when there is subsequent migration away from the site of such injuries due to loss of housing or employment. The resultant compositional change could also introduce bias in community-level estimates of outcomes, necessitating careful risk adjustment. Lastly, we wish to emphasize that extreme precipitation events are not equivalent to the concept of climate change, as seasonal high-risk environmental threats such as hurricanes predate anthropogenic global warming. Rather, they are exacerbated by man-made greenhouse gas emissions, and as such, ascribing what fraction of the health impacts from a series of disasters can be directly attributed to the broader phenomenon of climate change is an ongoing field of debate and investigation.16 Regional wildland fires have become the face of climate change for much of Western North America, and their increase in frequency and severity have been tied to global warming and resultant changes in precipitation patterns.17 Recent US Environmental Protection Agency reports demonstrate that the annual extent of land burned by wildfires in the United States has increased since the 1980s, with the 10 most severe years (by total acreage burned) occurring since 2004.18 While wildfires can directly harm individuals in close proximity and destroy local infrastructure, the main threat they pose to human health is through resultant smoke plumes that contain an extraordinary collection of hazardous air pollutants. Such plumes may contain particulate matter, ground level ozone, carbon monoxide, nitrogen dioxide, and polycyclic aromatic hydrocarbons, among other dangerous airborne pollutants.19 Wildfire smoke and its component toxins can be carried for thousands of miles, impacting communities far from the site of the original fire.20 Exposure to these air pollutants has been linked to adverse health outcomes including short-term cardiovascular and respiratory events, chief among them pulmonary disease exacerbations such as asthma and reactive airway disease.21, 22 Short-term health outcomes have been measured from a variety of data sources including mortality registers, crisis line utilizations, hospital admission and billing records, emergency department visits, among other measures.23-25 Long-term impacts are harder to quantify, but associations between wildfires and poor birth outcomes and negative psychological impacts are likely to have far-reaching implications.23, 26-28 Despite this rapidly growing body of investigation, measurement of the health consequences of wildfires has been difficult, particularly in exposure definition. Health services researchers have defined wildfire exposure in a number of ways, including the National Oceanic and Atmospheric Administration's venerable hazard mapping system (consisting of satellite imaging of smoke plumes), ground-level regulatory air pollution monitors, nonregulatory investigative monitors, and air pollution models of varying complexity.29, 30 Fire smoke plumes can behave erratically and are highly dependent on local atmospheric conditions (e.g., temperature; humidity; and presence, speed, and direction of wind).31 Thus, even communities near major wildfires may not experience high air pollution days if the wind is blowing away from their location (conversely, those downwind may be heavily impacted even if further afield from the epicenter of the disaster). In fact, a recent study from our group identified that those living 150 km or further away from wildland fire centers may suffer greater mortality risks from long-term wildfire smoke exposure.32 Concurrently, rapidly changing wind patterns can quickly shift affected areas on daily and hourly bases, making individual exposures difficult to ascertain from remote monitoring techniques. Ground-level air monitors such as PurpleAir may be more helpful in estimating residential exposures, but cannot disambiguate wildfire from background air pollution. The importance of this factor becomes magnified because wildfire smoke often contains different components from that of usual residential/occupational air pollution, as the combustion of homes and industrial buildings can introduce other hazardous chemical compounds and heavy metals into plumes.33 Thus, analyses that model the exposure as individual pollutant fractions such as airborne particulate matter may underestimate the toxicity of the smoke. Such concerns have motivated higher dimensional machine-learning based models of wildfire air pollution exposure, which may hopefully expand to include multiple pollutants and paint a more complete picture of the hazard.34 Finally, a growing dimension of wildfire exposure research is that of the impacts of personal risk mitigation strategies. It is increasingly being recognized that people living in wildfire-prone regions are aware of the health risks posed by wildfire smoke, and may alter their behavior in response.35 Such actions may include beneficial activities such as avoidance of outdoor activities, installing HEPA filters, using N95 masks when spending time in smoky areas, and using air conditioning (or conversely risky behaviors due to misconceptions of benefits such as opening windows).36, 37 Thus, health services analyses of wildland fires based on remote sensing air pollution data may again incorrectly predict individual-level exposure levels, with the highest threat to health potentially posed by poor air quality days that are not perceived as such by affected populations. This phenomenon is further complicated by large-scale evacuations or local health center shutdowns, which may introduce additional measurement error into such investigations. Most topically, there has been a dramatic increase in global surface temperatures since the industrial revolution began burning fossil fuels, manifesting around the globe in extreme heat events. The previous decade constitutes the warmest years since temperature records were first collected in the late nineteenth century.38 Concerningly, heat waves have increased in both frequency and severity, with the resultant number of people impacted by extreme temperature events rising by 125 million between 2000 and 2016 alone.39 The impacts of high heat on the human body are well documented, with exposure to extreme temperatures engaging the body's thermoregulatory systems, the prolonged strain of which can lead to heat-related crises, such as heat exhaustion and heat stroke. More insidiously, high heat can stress the autoregulatory mechanisms of the circulatory system at large, inducing cardiovascular disease exacerbations and collapse.40, 41 While aggregate heat seasons can register sizable elevations in heat-related mortality, individual heat waves can have catastrophic effects as well. A severe heat event during the 2024 Hajj in Saudi Arabia alone is thought to have resulted in 1300 deaths.42 Elderly populations are at particularly higher risk of heat injury, due to blunted heat response mechanisms, higher rates of prevalent cardiovascular disease, and lack of risk mitigation resources.24, 43 Other studies have identified that vulnerable populations (children, elders, unhoused people, and persons of lower socioeconomic status) are at increased risk of deleterious health outcomes following heat waves, especially those living with preexisting chronic disease.39, 44 Challenges facing the assessment of the health effects of heat also include both exposure and outcome definition. Much like wildfire-associated air pollution, individual-level temperature exposure can vary by street and home, due to factors such as the amount of shade, type/quality of housing, and local greenspace availability. These factors also impact the duration and timing of high temperatures. For example, the Urban Heat Island phenomenon is observed where residences near more concrete and less greenery experience higher average temperatures and heat trapping at night.45 These factors can impact the exposure of individuals where lower evening temperatures can provide relief in some locations but not others. Along these lines, extreme heat is also associated with individual risk-mitigation behaviors such as the use of air conditioning or relocation to air conditioned spaces such as cooling facilities. These features are highly variable and depend on factors such as perceived personal risk, financial means, and cognitive/social access to transportation to access cooled spaces.46, 47 Lastly, there is strong evidence to suggest that the human body acclimates to levels of heat over time—as such, relative, rather than absolute, temperature elevations during heat waves may be more informative for understanding heat stress on affected populations.48 Regarding the outcomes of extreme temperatures, the aforementioned heat dissipation mechanisms of the body can tax the circulatory system. As such, simply quantifying diagnostic codes labeled as heat illnesses (e.g., heatstroke, heat exhaustion) may not capture the full extent of other heat-associated health impacts. Indeed, over half the deaths recorded during the 1995 and 1997 Milwaukee heat waves listed cardiovascular disease as the cause.49 Furthermore, the "harvesting effect" is most prominently encountered in extreme temperature events, necessitating researchers to maintain long enough outcome durations to capture the true magnitude of a heat wave's mortality impacts.50 Our understanding of the health impacts of climate change is growing, with health services research carrying promise for characterizing these effects due to the wide variety of available medical data sources and methodologies. Nevertheless, challenges and considerations remain in accomplishing this task. Care must be taken to advance our precision in measuring exposures at the individual and local levels, and to develop more comprehensive definitions of health impacts. Long-term outcomes are particularly difficult to ascertain, especially when evacuations, loss of service, and delayed onset diseases must be accounted for. While outside the scope of this discussion, health services researchers can also approach future analyses from the perspectives of health systems and workforces, which are integral in understanding the full breadth of the climate crisis.51 A warming climate will ensure that these extreme weather conditions will not only continue to occur but will most likely have progressively intense impacts in the future. As wildfires, hurricanes, and dramatic temperature events become more frequent and widespread, health services researchers will increasingly encounter such manifestations of global warming both as primary exposures in their work or complicating factors in it. Thus, investigators and policymakers in the field should become more familiar with their impacts, as those unaware of these environmental determinants of health will experience unexpected challenges attributable to such events in future research and strategy. None. The authors declare no conflicts of interest. Data sharing not applicable to this article as no datasets were generated or analysed during the current study.