Methods for Quantifying, Projecting, and Managing the Health Risks of Climate Change

A rapidly increasing literature base is quantifying associations between weather and climate patterns and climate-sensitive health outcomes. In North America, between January 1, 2013 and February 14, 2020, the most studied health outcomes were heat-related morbidity and mortality (35.9%), respiratory health outcomes (25.1%), and general causes of morbidity and mortality (24.2%).⁸ Temperature was the most common weather variable investigated (76.2%), followed by precipitation (30.2%). Heat events were the most common climate hazard (32.8%), followed by changes in air quality (30.0%). Other climate-sensitive health outcomes were much less studied.

Identifying weather variables of interest should be based on understanding of disease etiology (particularly for infectious diseases). Epidemiology studies show that exposure to high ambient temperatures increases morbidity and mortality, reduces worker productivity, increases adverse pregnancy outcomes, and negatively affects mental health.^9,10 For example, self-reported mental health in the United States is improved on cooler days, whereas higher temperatures increase the probability of reporting poor mental health.¹¹ Daily or weekly associations between temperature or precipitation and health outcomes are increasingly described using distributed lag nonlinear models to estimate trends over time, accounting for potential lags between exposure and health responses.¹² For example, the role of temperature in seasonal variations in mortality varies across cities and countries in tropical, dry, temperate, and continental climate zones, with the highest impacts in temperate and continental climate zones.¹³ Lags vary by health outcome, with a lag of about 24 hours between the start of a heat wave and increased mortality¹⁴; the lag between heavy precipitation events and increased diarrheal disease is several weeks.¹⁵

Research and practice have identified populations and regions particularly vulnerable to changing weather patterns.² A helpful tool for decision-makers is mapping hotspots of vulnerability and projecting how these could change over time as climate-sensitive health outcomes shift their geographic range, to help communicate where impacts are most likely to occur and to prioritize actions.¹⁶ Mapping vulnerabilities should be done at the finest geographic scale possible, to facilitate community understanding of the current challenges associated with climate change and how those could alter over time. However, these analyses are often hindered by the difficulty in obtaining fine-scale health outcome data due to laws intended to ensure confidentiality of these data.

Statistical methods developed by climate scientists, termed detection and attribution, are quantifying the extent to which a single event or a longer-term trend could be attributed to climate change.¹⁷ Models or observations are used to determine whether a system has changed, such as the number or intensity of heat waves, or the number of hospitalizations and deaths in a heat wave.¹⁸ When a change is detected, the extent to which climate change could have caused that change is quantified. Limited detection and attribution studies for health impacts have mostly focused on heat wave–related mortality. An example is the heat dome that settled over the U.S. Pacific Northwest and western Canada at the end of June 2021.¹⁹ A detection and attribution study determined that the heat wave intensity was virtually impossible in the absence of anthropogenic climate change. The heat wave caused more than 1000 excess human deaths and a 69-fold increase in heat-related emergency department visits compared with the same period in 2019; other critical infrastructure was affected. In some coastal areas, high intertidal temperatures combined with low tides resulted in the deaths of over a million shellfish, affecting livelihoods, food security, and recreation. A modeling study estimated that 37% of heat-related deaths in 43 countries over recent decades were caused by climate change.²⁰

The health risks of climate change result from the intersection of the hazards created by changing weather patterns, the extent to which populations are exposed to those hazards, the vulnerability of those so exposed, and the capacity of individuals and health systems to prepare for and manage changing health risks.²¹ Quantitative methods can be used to estimate relationships and extrapolate future burdens and risks. Expert judgment and development of qualitative scenarios also can be used by health professionals to estimate future impacts.⁷ Qualitative projections consider recent trends in weather variables, extreme weather events, and/or sea level rise; trends in the burden of climate-sensitive health outcomes (e.g., in cases of dengue fever); and planned modifications to policies and programs (e.g., increasing resources for the dengue control program). Expert judgment is used to combine these sources of information into statements about the extent to which climate change could alter disease burdens over coming decades.²²

Quantitative projections of health risks use exposure-response relationships, generally based on short-term associations between weather variables and health outcomes, and projected changes in weather and climate trends (or extremes) to estimate changes over time in the burden of disease of interest. Two basic types of models are used: empirical (statistical) and process-based (biological or mechanistic), often reaching different conclusions. Empirical (statistical) models use observations to establish relationships between a health outcome and one or more weather and non-meteorological (e.g., socioeconomic or demographic) variables.²³ Projected changes in weather variables are used to estimate the future distribution or burden of a health outcome, generally assuming linear relationships. Challenges include data availability, accuracy, and length of time series. For vector-borne disease models, observations of current disease patterns reflect the effectiveness of control programs and other variables. Because validated absence data are often not available, these models may not accurately reflect the environmental suitability of a location for transmission.

Process-based models are built from equations describing the dynamics of disease etiology; they generally simulate the impact of weather variables on the health outcomes of interest using equations describing disease burdens at daily or weekly time steps.²⁴ Projected changes in weather variables under different climate change scenarios are used to estimate the future distribution and burden of the health outcome. Challenges to process-based modeling include poorly constrained values for parameters in the equations, the assumption of a causal relationship, and the possibility of unknown processes not described in the model.

Because the current magnitude and pattern of health burdens is at least partially the result of past investments in health systems and development choices such as urbanization plans, the degree to which current exposure-response relationships are predictive of future patterns of climate-sensitive ill health is unclear.²⁵ Projecting risks requires using scenarios that describe not just temperature change over this century, but also the evolution of the social and environmental determinants of health, specific adaptation policies, and other factors that can influence the incidence and geographical range of climate-sensitive health outcomes.²⁴

Climate-sensitive health outcomes are among the leading causes globally of current morbidity and mortality, particularly for children.² Health ministries and departments and other health organizations have a wide range of options to manage these outcomes, although few policies and programs explicitly incorporate climate change. Therefore, health systems may not be adequately prepared for climate-related shocks and stresses. Proactive assessment of the effectiveness, strengths, and weaknesses of these programs can provide critical insights for identifying possible modifications to increase the capacity to manage the additional health burden due to climate change.⁷

Assessments of the capacity of health systems to deal with climate-related shocks and stresses focus on the required capacities for each building block of climate-resilient and environmentally sustainable health systems, as outlined in the World Health Organization operational framework (Fig. 2).²⁶ The capacity assessment should include specific climate change effects (e.g., floods), the specific health outcomes of interest (e.g., diarrheal diseases), and broader impacts (e.g., disruption of water and energy access in health care facilities). For example, reducing the burden of temperature-sensitive waterborne diseases requires improved access to safe water and improved sanitation, in addition to explicitly including climate change into health system policies and programs.

Within a health system, a range of organizations have individual or joint responsibility for managing climate-sensitive health outcomes, including the ministry or department of health, civil organizations, private-sector actors, and others.⁷ For example, health departments typically have responsibility for vector-borne disease surveillance and control programs, while disaster risk reduction activities may be joint across health, emergency management, and other departments and may include the Red Cross or other organizations. A capacity assessment should consult with all relevant organizations and institutions to determine what is working well, what could be improved, and the capacity of the programs to address possible increases in the incidence or changes in seasonality or geographical range of the health outcomes of concern.

This information can then be used to identify opportunities for increasing resilience by, for example, improving monitoring and surveillance, developing and deploying early warning and response systems, augmenting education and training for health care providers and the public, modifying regulations for water- and foodborne diseases to account for projected changes in access to clean water, and improving emergency response plans.⁷ These changes should be structured such that their effectiveness can be evaluated in a continuous cycle of learning.

Comprehensive estimates of the costs and benefits of actions to reduce the health risks of climate change include the following: costs for health systems to respond to a climate-related event; the costs of injuries, illnesses, and deaths and of lost work time or reduced productivity; the costs of increasing the resilience of health systems to future climate change; the costs associated with failures to adapt to or mitigate the risks of climate change; and the costs of managing residual risks, acknowledging that health systems are unlikely to be fully resilient to increasing impacts of climate change. These estimates should include resource costs (e.g., medical care costs), opportunity costs (e.g., lost productivity), and welfare costs (e.g., pain and suffering).^23,27

For example, the health-related costs from 10 climate-sensitive events, including wildfires, high ozone, infectious disease outbreaks, pollen, harmful algal blooms, and extreme weather, in 11 U.S. states in 2012 were estimated at $10.0 billion from 917 deaths, 20,568 hospitalizations, and 17,857 emergency department visits.²⁸ Between 2015 and 2019, the economic impacts of the health effects from heat waves in France were €25.5 billion, from mortality (€23.2 billion), minor restricted activity days (€2.3 billion), and morbidity (€0.3 billion).²⁹

Future health care costs from exposure to extreme heat in 2041 to 2070 in Michigan were projected to be $280 million for nonaccidental mortality and $14 million for heat-associated emergency department visits.³⁰ Better quantification of these significant health costs can justify investments in adaptation and mitigation.

International agreements and national and state mitigation actions are aiming to rapidly decrease GHG emissions to reduce the magnitude and destructiveness of future climate variability and change. Economic valuations of the associated avoided hospitalizations and premature deaths indicate that they generally offset a significant proportion, if not all, of the costs of implementing mitigation policies, providing further incentives for aggressive policy action. That is, mitigation policies should be implemented because of the significant improvements to population health in addition to the benefits for achieving climate policy. Further estimates would help policy makers prioritize investments based on mitigation potential and expected health benefits.

Mitigation policies benefit health through the following: reducing exposure to air pollutants from coal-fired power plants and other stationary sources, increasing physical activity from walking and biking, and encouraging dietary changes that increase consumption of whole grains, vegetables, and fruit over red meat.³¹ Benefits from these policies accrue quickly and locally. Examples include $350 million in benefits to child health over the period from 2009 to 2014 from the U.S. Regional Greenhouse Gas Initiative.³² Another example is the projected health benefits of mitigation policies in energy, food and agriculture, and transport sectors across nine countries in 2040.³³ Benefits were attributable to the mitigation of direct GHG emissions and the commensurate actions that reduce exposure to harmful pollutants as well as to improved diets and safe physical activity. Compared with the current pathway, a sustainable pathway consistent with the goal of the Paris Agreement and the Sustainable Development Goals projected an annual reduction of 1.18 million air pollution–related deaths, 5.86 million diet-related deaths, and 1.15 million deaths due to physical inactivity. Adopting a more ambitious health-in-all climate policies scenario could result in a further reduction of 462,000 annual deaths attributable to air pollution, 572,000 annual deaths attributable to diet, and 943,000 annual deaths attributable to physical inactivity.

To strengthen the policy relevance of health cobenefits research, an expert panel of researchers developed recommendations for conducting and reporting estimates, including the following: through harmonization of approaches for stakeholder engagement, modeling inputs and assumptions, parameterization and reporting, approaches to uncertainty and sensitivity analysis, accounting for policy uptake, and discounting.³¹

About 8.3% of the U.S. GHG emissions are from health care, indicating significant scope for greening the health sector through increasing energy efficiency of health care facilities, using renewable energy, and reducing emissions throughout the supply chain, particularly when GHG emissions are not correlated with health systems quality.³⁴ The estimated disease burden in 2018 from these emissions was about 388,000 disability-adjusted life years; this is similar to the years of lives lost from preventable medical errors.