Climate change and global health outcome indicators: A scoping review

Abstract

Background

The impact of climate change on human health is not evenly distributed and is affected by regional geography and vulnerability of the local population. Official statistics that report these uneven impacts are needed to facilitate strategic planning and resource allocation.

Purpose

Identify globally defined indicators of the impacts of climate change on human health to inform the design of official statistics.

Methods

We followed recognized methods guidance for scoping reviews.

Results

Reviewing 4415 unique records, we extracted 73 unique and 33 repeated indicators from 20 sources. Temperature-related indicators were the most common (27%, 29/106), but many were repeated. Injury or illness indicators were more frequent than mortality indicators, with 59% (43/73) and 37% (27/73) respectively. Following breakdown of the categories into smaller, more specific outcomes, mortality from extreme weather events (n = 10) and illness due to zoonoses/vector-borne diseases (n = 9) were the most prevalent indicators. There was an absence/gap of indicators for five secondary categories.

Conclusion

Synthesis of climate-sensitive health indicators is crucial for establishing a cohesive official statistics framework to monitor the health impacts of climate change-related events. The abundance (and gaps) of indicators across categories of health effects aids in prioritization of developing new indicators and improving data availability.

1. Introduction

Official statistics have a vital role in supporting risk assessment and decision-making by government, business, civil society organizations, researchers and members of the public. ¹ One area of urgent and increasing concern for the development and application of official statistics and indicators is climate change. ² Strong climate change action requires support from sophisticated statistical frameworks with comprehensive indicators. Indicators, with their potential to reveal relative positions or show positive or negative change, are crucial for setting targets and monitoring achievement.^1,2

Given the large magnitude of human health impacts of climate change, ³ corresponding official statistics and related data infrastructure to monitor prevalence and temporal changes are required. Data must be robust enough to meet the following decision-making challenges: the need to consider decisions and their impacts over multiple overlapping time-scales; the inherent uncertainty in the rate, magnitude and pattern of climate change for any one location; and the fact that relevant data may be unavailable, or imprecise, especially at the local scale at which many health-system decisions are made.^4,5

Health-system decision makers are progressively becoming aware that climate change poses increasing threats to human health. Climate change leads to increasing levels of carbon dioxide and short-lived climate pollutants; rising temperatures; rising sea levels; and increases in the intensity, frequency, duration and seasonality of extreme weather events. ⁶ Related health impacts can be direct; for example, extreme heat disproportionately affects certain population groups, such as young children, the elderly, and those living in substandard housing or who are unhoused.^7,8 The impacts may also be indirect, with delivery of quality care and health facility infrastructure being impaired by a changing climate. For example, extreme weather events have particular impacts on people whose health depends on regularly scheduled and uninterrupted access to health services and those with technology dependencies. ⁹ Monitoring human health impact and risk from climate change requires reliable, long-term data sets. ¹⁰

Recognizing the need to support decision-making in this complex and important area, inter-governmental organizations such as the World Health Organization (WHO) and research teams, notably the Lancet Countdown collaboration, have developed indicators for climate-health outcomes. This review supports the Official Statistics on Climate-Health Interactions (SOSCHI) project, led by the UK Office for National Statistics and funded by Wellcome. ¹¹ The goal of SOSCHI is to develop a transparent, globally generalizable, and spatially scalable (from national to regional) framework and technical platform for official statistics on climate change, the environment, and health. ¹¹ The partners on the SOSCHI project include the African Institute for Mathematical Sciences (AIMS) Rwanda, the Regional Institute for Population Studies (RIPS) at the University of Ghana, the UK Health Security Agency and the Cochrane Planetary Health Thematic Group. ¹¹ This project will develop methods and tools that can be applied by analysts across a broad range of contexts, to produce regular indicators that can inform policy and practice. The outputs of the project will help stakeholders produce high-quality data and statistics and communicate with a range of audiences. ¹¹

To support the development of a unified indicator set, the scoping review reported in this article was undertaken by staff from the Cochrane Planetary Health Thematic Group, working closely with the ONS, to collate published indicators and statistical approaches that measure the impact of climate change on human health. The primary research question this scoping review aims to address is, what indicator sets or statistical approaches have already been published for climate-health related data?

2. Methods

Our preliminary scan of relevant literature revealed that mapping the available indicator sets would require collating information from a wide range of evidence types. Therefore, we used a scoping review methodology. This approach is suitable for examining the extent, range, and nature of research activity in a particular area, summarizing, and disseminating research findings, and identifying gaps in the available literature. Similar to a systematic review, a scoping review uses systematic and rigorous methods to collate and present the research evidence relevant to a particular question. In contrast to a systematic review, however, a scoping review addresses a broad topic for which many study designs may be applicable and does not include quality assessment of included studies. A scoping review approach is highly suitable for a broad area that has not been previously mapped, as is the case with the current state of global climate-health indicators. This scoping review was guided by the methodology framework outlined by Arksey and O’Malley and the more recent PRISMA extension for scoping reviews.^12,13 The protocol was registered in Open Science Framework (https://osf.io/swx9h/).

2.1. Inclusion criteria

Documents included in this scoping review discuss or report the use of globally defined indicators for health outcomes in the context of climate change. For this project, globally defined indicators were defined as having the following characteristics:

1. A quantitative, comparable measure of a health outcome or health status (i.e., mortality, illness, hospitalizations, emergency department presentations). Indicators reporting vulnerability or exposure were excluded.

2. Directly linked, or having some level of attribution, to climate change or climate change-related events as reported by the authors.

3. Contained in a document for which at least one author reports an affiliation with, or the indicator is used by centers recognized as, global health authorities.

The definition of global authority was developed with input from multiple stakeholders. A global authority needed to have a worldwide scope or represent a consortium of multiple countries such as the European Union. National organizations from the United States (US) such as the US Centers for Disease Control (CDC) and the US Environmental Protection Agency (EPA) were included due to the significant impact and uptake of their work globally. Evidence syntheses (systematic, scoping and narrative reviews) were included as well as white papers, reports, position statements/recommendations, policies, working group reports, and frameworks from global health authorities. Primary research articles (observational studies, RCTs, case reports, case series) that described an association between climate change and a specific health outcome, but not how to measure that association, were excluded. Also excluded were letters, opinion pieces, abstracts and clinical guidelines.

2.2. Search strategy

The comprehensive database search strategy was developed by an information specialist in consultation with project members from the UK Office for National Statistics (ONS) and the Cochrane Planetary Health Thematic Group (CPHTG). The strategy was based on the concepts of climate change and health indicators using key words such as climate change, global warming, environmental health, health, health status, indicator, measure, metric, and statistic. The search was executed in the following biomedical databases from inception to May 2023: EBSCO, Embase, Epistemonikos, MEDLINE, ProQuest, Google Scholar, Elsevier Scopus, Clarivate Web of Science. The full search strategy is available in Appendix A.

An additional search of grey literature was also undertaken through systematic screening of relevant websites of accepted global authorities on climate change and human health. The websites were identified by the ONS and CPHTG members and included the National Institutes of Health (NIH) Climate Change and Human Health Literature Portal, the WHO, the CDC, the United Nations (UN) and the European Environment Agency (EEA Climate-ADAPT) (see Appendix B for the complete list). Each website was reviewed independently by two review authors for relevant page titles and those with built-in search engines were searched using synonyms and different configurations of the text words “indicator,” “health” and “climate change.”

2.3. Evidence selection

Records identified by the database search were de-duplicated and uploaded into the COVIDENCE tool (Veritas Health Innovation). Two review authors screened at the title and abstract level following a small (n = 20) pilot screening to ensure consistency. Following the first level screening, potentially relevant publications were subject to a further round of independent screening at the full text level by the same two review authors. Any disagreements were resolved by consensus or consultation with a third author.

Records identified by the website search were recorded in Excel (Microsoft). Documents available in PDF were added to the COVIDENCE tool at the full text level. Information available only on a website underwent additional full screening using Excel to track decisions. Any disagreements were resolved by consensus or consultation with a third author.

2.4. Data extraction

Data were extracted independently by two review authors using a pilot-tested Excel spreadsheet. All data variables were pre-defined in an extraction code book to ensure consistency. In addition to basic descriptive data, key information extracted included the name of the indicator, the verbatim definition of the indicator, spatial scale (global, national, regional), data source, and the authors’ determination of the indicator's proximal or distal linkage to climate change. Any disagreements were resolved by consensus or consultation with a third author.

2.5. Analysis of evidence

Categorization

Measures of climate-sensitive health outcomes identified were sorted into eight primary categories based on climate change impacts on health illustrated by the WHO (Table 1).^6,14 Categorization was undertaken independently by two review authors. Any indicators that fit in more than one category were classified under the most relevant outcome as determined by the reviewers. For example, the indicator ‘number of cases of Giardiasis’: Giardiasis is mainly a water-borne disease but can also be food-borne; based on the above sorting methods this indicator was classified as water-borne.

Table 1.

Primary categories of climate change related health outcomes.

• Extreme weather event-related injury and mortality	• Zoonoses & vector-borne diseases
• Temperature-related illness and mortality	• Malnutrition and food-borne diseases
• Air quality-related illness and mortality	• Mental and psychosocial health
• Water-borne diseases and other water-related health impacts	• Other (e.g., non-communicable diseases)

Secondary categorization of the indicators involved breaking down each primary category into smaller, more specific outcomes based on the measure (e.g., mortality vs illness) or other identifiers. Examples are shown in Table 2; the full list can be found in Figure 4.

Table 2.

Examples of secondary category development

Primary category	Secondary categories
Extreme weather event-related injury and mortality	Injury due to extreme weather events	Mortality due to extreme weather events
Temperature-related illness and mortality	Heat-related illness	Heat-related mortality
	Cold-related illness	Cold-related mortality

Figure 4.

Frequency of unique indicators in each secondary category.

Unique indicators

To determine if an indicator was unique, the review authors examined both the indicator name and reported definition. Indicators with similar names and definitions were grouped together under one unique indicator (e.g., heat-related mortality and deaths from extreme heat would fall under one unique indicator). When this occurred the most straightforward indicator name would be used as the unique indicator title while the remaining similar indicators were referred to as repeated indicators. All the distinct definitions for the indicators were retained for the description. Indicators were not separated into unique categories based on location (e.g., the indicators ‘flood-related deaths in Europe’ and ‘flood-related deaths in the US’ would both be listed under the unique indicator ‘flood-related deaths’), or on units of measurement (e.g., the indicators ‘total number of deaths from heat’ and ‘death rate per 100,000 due to heat’) would both be listed under the indicator ‘heat-related mortality’.

3. Results

From both the database and website searches 4415 unique records were screened at the title and abstract level; 154 underwent further full text screening and 73 unique indicators and 33 repeated indicators were extracted from 23 publications (20 papers and 3 websites totalling 13 pages). Figure 1 illustrates the flow of sources through the inclusion process. Three of the papers were part of the Lancet Countdown's ongoing publications of climate change indicators and reported duplicate or out-of-date indicators. These three papers were excluded from the analysis, leaving a total of 17 papers. Table 3 summarizes the characteristics of included studies. Sources that were excluded based on our definition of global authority are available in Appendix C.

Figure 1.

PRISMA diagramn.

Table 3.

Characteristics of included studies.

First			Global
Author,	Data	Climate change health	organization or	How data were
Year	origin	indicator categories addressed	author affiliation	identified/synthesized	Data sources
Berko 2014 15	USA	Extreme weather-related injury and mortality, Temperature-related illness and mortality	CDC	Vital statistics were used to measure weather-related health events	National Center for Health Statistics
Bernard 2001 16	USA	Air quality-related illness and mortality	CDC, US EPA	Health effects for indicators obtained from an American Thoracic Society review based on extensive evidence drawn from animal and in vitro toxicology, human clinical exposure studies, field exposure studies, and epidemiologic studies.	US EPA
CDC's Climate and Health Program (2023 Website) 17	USA	Temperature-related illness and mortality	CDC's Climate and Health Program	Indicators were developed in collaboration with national, state, and local environmental health partners. ‡	National Syndromic Surveillance Program, Government bureau department data
CSTE 2018 18	USA	Extreme weather-related injury and mortality, Temperature-related illness and mortality, Air quality-related illness and mortality, Zoonoses & vector-borne diseases	CDC	Development of the indicators included a comprehensive literature review, identification of longitudinal data sets that were applicable at the state and community level, and pilot studies to improve and finalize implementation instructions. ‡	Statewide hospital discharge records, SHELDUS, CDC Lyme Stats tables and Statewide Data
Dora 2015 19	90 countries	Air quality-related illness and mortality, Malnutrition and food-borne diseases	WHO	Indicators were identified through a process of scoping review and expert and stakeholder consultation	WHO Global Household Air Pollution Measurement database and WHO indoor air quality guidelines
English 2009 20	USA	Extreme weather-related injury and mortality, Temperature-related illness and mortality, Air quality-related illness and mortality, Zoonoses & vector-borne diseases, Malnutrition and food-borne diseases	CDC, US EPA	Indicators selected following review of governmental and nongovernmental documents and published literature ‡	National Center for Health Statistics, EM-DAT, ArboNET surveillance system, CDC Division of Vector-borne Infectious Diseases
European Climate and Health Observatory – Climate ADAPT (2022 Website) 21	Europe	Extreme weather-related injury and mortality, Temperature-related illness and mortality	European Climate and Health Observatory, Climate ADAPT	Indicators drawn from Lancet Countdown and Copernicus Climate Change Service ‡	ERA5-Land, EURO-CORDEX, Copernicus Climate Change Service, Copernicus Atmosphere Monitoring Service, Lancet Countdown, CATDAT
European Environment Agency (2016 Website) 22	Europe	Extreme weather-related injury and mortality	EEA	NR ‡	WHO, Eurostat, EM-DAT
European Environment Agency 2022 23	Europe	Temperature-related illness and mortality, Water-borne diseases and other water-related health impacts, Zoonoses & vector-borne diseases, Malnutrition and food-borne diseases	EEA	Indicators drawn from Lancet Countdown in Europe and Copernicus Climate Change Service, as well as case studies throughout Europe and national policy analyses	European Centre for Disease Control and Prevention database
Jung 2018 24	Western Pacific	Extreme weather-related injury and mortality, Zoonoses & vector-borne diseases, Malnutrition and food-borne diseases	WHO (Western Pacific Region)	United Nations Sustainable Development Goals used as framework	World Bank Data Bank
Lay 2018 25	USA	Temperature-related illness and mortality	CDC, US EPA	Used ICD codes to measure temperature-related health events ‡	Truven Health Research database, North American Land Data Assimilation System
Romanello 2022 26	Global	Extreme weather-related injury and mortality, Temperature-related illness and mortality, Air quality-related illness and mortality	Lancet Countdown, UN	Indicators selected following international collaboration and consensus ‡	European Centre for Medium-range Weather Forecasts, EM-DAT, Human Development Index at the UN Development Programme
Swedish International Development Cooperation Agency (SIDA) 2011 27	Global	Extreme weather-related injury and mortality, Zoonoses & vector-borne diseases, Malnutrition and food-borne diseases	SIDA	Indicators selected from those in use by partner countries and internationally accepted ones	Commission for Sustainable Development, Millennium Development Goals, Joint Monitoring Programme, WHO, EM-DAT
United Nations Economic and Social Commission for Western Asia 2017 28	Middle East	Extreme weather-related injury and mortality, Temperature-related illness and mortality, Water-borne diseases and other water-related health impacts, Zoonoses & vector-borne diseases	UN ESCWA	Proposed indicators were discussed at expert group meeting, from which final indicators were selected	UNISDR, Global Health Data Exchange
United States Environmental Protection Agency 2016 29	USA	Temperature-related illness and mortality, Zoonoses & vector-borne diseases	US EPA	Indicators screened and selected based on usefulness, data quality, and relevance from peer-reviewed, publicly available data ‡	CDC National Vital Statistics system and Environmental Public Health Tracking Program
Van Daalen 2022 30	Europe	Temperature-related illness and mortality	Lancet Countdown, UN	Indicators selected following international collaboration and consensus ‡	Eurostat, European Centre for Medium-range Weather Forecasts, Geostat
World Health Organization 2018 31	Global	Extreme weather-related injury and mortality	WHO	Indicator selection guided by priority global monitoring requirements related to human health, with organizations such as the UN and World Health Assembly ‡	Ministries of Health, National Disaster Loss Database
World Health Organization 2020 32	Global	Air quality-related illness and mortality, Water-borne diseases and other water-related health impacts, Zoonoses & vector-borne diseases	WHO	A draft global strategy, including climate change elements, was submitted to WHO's regional committees and Executive Board	NR
World Meteorology Organization 2021 33	Global	Extreme weather-related injury and mortality, Temperature-related illness and mortality	WMO	NR	EM-DAT
Zhang 2018 34	Western Pacific	Extreme weather-related injury and mortality, Temperature-related illness and mortality, Mental and psychosocial health	Lancet Countdown, UN	Lancet Countdown indicators were assessed in the Australian context ‡	EM-DAT

‡ Detailed statistical methods are reported.

Three of the analyzed sources were from the Lancet Countdown (one each from the Global Countdown, ²⁶ the European Countdown, ³⁰ and the Australian Countdown. ³⁴ The organizations producing or associated with the rest of the sources were: the US CDC (n = 6),^{15–18,20,25} the WHO (n = 4),^19,24,31,32 the US EPA (n = 4),^16,20,25,29 the United Nations (UN) (n = 3),^26,30,34 the European Environmental Agency (n = 2),^22,23 the European Climate and Health Observatory (n = 1), ²¹ the Swedish International Development Agency (SIDA) (n = 1),²⁷ the UN Economic and Social Commission for Western Asia (ESCWA) (n = 1), ²⁸ and the World Meteorological Organization (WMO) (n = 1). ³³ Some sources had more than one global affiliation. The papers were published between 2001 and 2023 with the majority (85%, n = 17) published after 2014.

3.1. Primary categorization

From the 20 analyzed sources, 106 indicators were extracted and sorted into primary categories (Table 1), allocating one category per indicator. Nine indicators initially fit into more than one primary category. For the categories listed in Table 1, indicators for temperature-related illness and mortality were most frequent (27%, n = 29/106), followed by indicators for extreme weather event-related injury and mortality (22%, n = 23/106) and air quality-related illness and mortality (20%, n = 21/106) (Figure 2). No indicators were extracted for the category of ‘Other outcomes (e.g., non-communicable diseases)’.

Figure 2.

Frequency of indicators in each primary category.

Dark boxes represent the number of unique indicators for that primary health outcome category (n = 73). Light boxes represent repeated indicators (n = 33).

3.2. Unique indicators

Seventy-three indicators were identified as unique and are listed along with all their reported definitions in Table 4, organized by primary category. The largest number of unique indicators was found for the health outcomes of air quality-related illness and mortality (28%, n = 29/73), followed by indicators for extreme weather event-related injury and mortality (19%, n = 14/73) and temperature-related illness and mortality (16%, n = 12/73) (Figure 2). Four of the primary categories contained only unique indicators (mental and psychosocial health, water-borne diseases and other water-related health impacts, malnutrition and food-borne diseases, and air quality-related illness and mortality), while the remaining three had a range of 7 to 17 repeated indicators (zoonoses & vector-borne diseases (n = 7), extreme weather event-related injury and mortality (n = 9) and temperature-related illness and mortality (n = 17) (Figure 2). Regarding health outcome severity, for all categories unique indicators measuring injury or illness were most frequent (59%, n = 43/73) compared to indicators measuring mortality (37%, n = 27/73) or measuring both mortality and morbidity (4%, n = 3/73) (Figure 3).

Table 4.

List of unique indicators (n = 73) and related indicators (n = 33) for climate change impacts on human health, grouped by primary category.

	Related Indicators		Global organization
Unique Indicator	(if applicable)	Definition	or author affiliation	Data origin
Extreme weather event-related injury and mortality
All weather-related deaths	Not applicable	Deaths due to weather (extreme natural heat, extreme natural cold, floods, storms, and lightning) may be identified by using both the underlying and contributing causes of death.	CDC	USA
Flood-storm-lightning-related deaths	Not applicable	Flood-storm ICD–10 codes X38 (victim of flood), X37 (victim of cataclysmic storm), or X33 (victim of lightning) on death certificate.	CDC	USA
Total number of victims (killed or affected) of natural disasters (women/men)	Not applicable	Measures total numbers of victims killed or affected by natural disaster and is thus an indicator for vulnerability to natural disasters, including geophysical, hydrological meteorological, climatological, and biological disasters. In case of limited disaster preparedness and weak resilience more people will be affected.	SIDA	Global
	Number of deaths, missing persons, and directly affected persons attributed to disasters per 100,000 population	Number of deaths, missing persons, and directly affected persons attributed to disasters per 100,000 population.	WHO	Western Pacific
Deaths due to drought	Not applicable	Numbers of reported deaths due to drought from 1970–2019.	WMO	Global
Deaths due to storms	Not applicable	Numbers of reported deaths due to storms from 1970–2019.	WMO	Global
Deaths due to flood	Not applicable	Numbers of reported deaths due to flood from 1970–2019.	WMO	Global
	Fatalities associated with floods	Number of fatalities associated with floods.	European Climate and Health Observatory Climate ADAPT	Europe
	Deaths related to flooding in Europe	Deaths per million inhabitants related to flooding in Europe.	EEA	Europe
Number of reported deaths due to landslide	Not applicable	Number of reported deaths due to landslide from 1970–2019.	WMO	Global
Number of reported deaths due to wildfire	Not applicable	Number of reported deaths due to wildfire from 1970–2019.	WMO	Global
	Fatalities associated with wildfires	Number of fatalities associated with wildfires.	European Climate and Health Observatory Climate ADAPT	Europe
Deaths due to extreme weather events	Not applicable	Deaths (per 100,000 population) from wildfires, flooding, and storms.	CDC	USA
	Lethality of extreme weather events	Deaths, as proxy of the lethality of weather-related disasters, are defined as the number of people who lost their life because the disaster happened.	Lancet Countdown, UN	Global
	Number of mortalities from extreme weather events	Number of mortalities from extreme weather events (flooding and storms).	CDC, US EPA	USA
	Number of deaths attributed to disasters	Number of people who died during the disaster, or directly after, as a direct result of the hazardous event.	WHO	Global
Number of deaths attributed to hydrometeorological disasters	Not applicable	This indicator measures the number of people who died during the disaster, or directly after, as a direct result of the hazardous event (per 100,00 population).	UN ESCWA	Middle East
Number of missing persons attributed to hydrometeorological disasters	Not applicable	Missing: The number of people whose whereabouts is unknown since the hazardous event. It includes people who are presumed dead although there is no physical evidence (per 100,000 population).	UN ESCWA	Middle East
	Number of missing persons by disaster per 100,000 people	Missing: number of people whose whereabouts is unknown since the hazardous event. It includes people who are presumed dead although there is no physical evidence. The data on number of deaths and number of missing are mutually exclusive.	WMO	Global
Injuries due to extreme weather events	Not applicable	Injuries (per 100,000 population) from wildfires, flooding, and storms.	CDC	USA
	Number of injuries from extreme weather events	Number of injuries from extreme weather events.	CDC, US EPA	USA
People affected by weather-related disasters	Not applicable	People affected are defined as those requiring immediate assistance during a period of emergency; hence requiring basic survival needs such as food, water, shelter, sanitation and immediate medical assistance.	Lancet Countdown, UN	Global
Lethality of weather-related disasters	Not applicable	Died from drought, floods, storms, bushfires and extreme temperature.	Lancet, UN	Western Pacific
Temperature-related illness and mortality
Heat-related deaths	Not applicable	ICD–10 codes X30 (exposure to excessive natural heat), T67 (heatstroke or sunstroke), or both on death certificate.	CDC	USA
	Heat related deaths	Annual rate for deaths classified by medical professionals as “heat-related” each year in the United States.	US EPA	USA
	Deaths due to extreme temperature	Numbers of reported deaths due to extreme temperature from 1970–2019.	WMO	Global
	Number of heat-related deaths	This indicator measures the annual number of deaths attributable to high temperatures.	UN ESCWA	Middle East
	Heat-related mortality	Annual deaths per million inhabitants.	EEA	Europe
	Heat-related mortality	Heat-related mortality	Lancet	Europe
	Health effects of temperature change	Association between summer daily mean temperature and non-accidental mortality.	Lancet	Western Pacific
	Summer heat-related mortality	Heat death: ICD-10 codes X30 or T67.0 – T67.9, EXCLUDING cases having ICD-10 code W92 (man-made source of heat) in any cause of death from May through September.	CDC	USA
	Heat-related mortality	Heat-related mortality incidence (annual deaths per million inhabitants) in Europe.	European Climate and Health Observatory Climate ADAPT	Europe
	Excess mortality due to extreme heat	Death as heat-related when heat is the underlying or contributing cause of death. Excess mortality can be calculated by comparing the number of deaths during an extreme temperature event with those during a reference period that has been matched by day of the week and other potentially confounding factors, or by using a time-series approach.	CDC, US EPA	USA
Deaths due to heat and cardiovascular disease	Not applicable	heat-related cardiovascular disease deaths, which include deaths due to heart attacks, strokes, and other diseases related to the circulatory system.	US EPA	USA
	Heat-related deaths	Deaths result from heat stroke and related conditions, cardiovascular disease, respiratory disease, and cerebrovascular disease.	CDC's Climate and Health Program	USA
Heat-related mortality (over 65)	Not applicable	Global heat-related mortality for populations older than 65 years.	European Climate and Health Observatory Climate ADAPT	Europe
	Heat-related mortality	Tracks the global total number and spatial pattern of heat-related mortality from 2000 to 2019 of the 65-and-older population.	Lancet Countdown, UN	Global
Health effects of temperature change	Not applicable	Association between summer daily mean temperature and non-accidental mortality	Lancet	Western Pacific
Cold-related deaths	Not applicable	ICD–10 codes X31 (exposure to excessive natural cold), T68 (hypothermia), or both on death certificate.	CDC	USA
Hyperthermia emergency department (ED) visits	Not applicable	(ICD9-CM codes 992.0–992.9).	CDC, US EPA	USA
Emergency department (ED) visits associated with heat-related illness (HRI)	Not applicable	Emergency department (ED) visits associated with heat-related illness (HRI) per 100,000 ED visits.	CDC's Climate and Health Program	USA
	Summer heat-related morbidity -ED visits	ED visits due to heat. From May 1, 2000 to September 30, 2015 use ICD-9-CM codes (992.0 to 992.9, E900.0, E900.9). From May 1, 2016 to present, use ICD-10-CM codes (X30, T67.0-T67.9).	CDC	USA
Cardiovascular emergency department (ED) visits	Not applicable	ICD9-CM codes 460–519.	CDC, US EPA	USA
Myocardial infarction (MI) emergency department (ED) visits	Not applicable	ICD9-CM code 410.	CDC, US EPA	USA
Heat-related hospitalizations	Not applicable	Rates for hospital admissions due to “heat-related” illnesses such as heat exhaustion, heat cramps, mild heat edema (swelling in the legs and hands), heat syncope (fainting), and heat stroke. It is based on hospital discharge records, which include a diagnosis determined by a physician or other medical professional (per 100,000 people).	US EPA	USA
	Summer heat-related morbidity- hospitalizations	Hospitalizations due to heat. From May 1, 2000 to September 30, 2015 use ICD-9-CM codes (992.0 to 992.9, E900.0, E900.9). From May 1, 2016 to present, use ICD-10-CM codes (X30, T67.0-T67.9).	CDC	USA
Heat-related illness	Not applicable	Includes cardiovascular and respiratory complications, renal failure, electrolyte imbalance, kidney stones, negative impacts on fetal health, and preterm birth.	CDC's Climate and Health Program	USA
	Excess morbidity due to extreme heat	Excess morbidity due to extreme heat.	CDC, US EPA	USA
Annual worker heat-related illness cases with days away from work	Not applicable	Annual worker heat-related illness cases with days away from work.	CDC's Climate and Health Program	USA
	Annual rate of heat-related illness with days away from work per 10,000 full-time workers	Annual rate of heat-related illness with days away from work per 10,000 full-time workers.	CDC's Climate and Health Program	USA
Annual worker days lost due to heat-related illness	Not applicable	Annual worker days lost due to heat-related illness.	CDC's Climate and Health Program	USA
	Annual median days of work lost due to heat-related illness	Annual median days of work lost due to heat-related illness.	CDC's Climate and Health Program	USA
Air quality-related illness and mortality
Premature cardiorespiratory mortality	Not applicable	Deaths from heart or lung disease in excess of number expected.	CDC, US EPA	USA
Increased health care use	Not applicable	Increased hospitalizations, physician visits, emergency department visits.	CDC, US EPA	USA
Asthma exacerbations	Not applicable	Increased physician and emergency department visits, medication use, symptom reporting. Decreased peak flow measurements.	CDC, US EPA	USA
Increased respiratory illness	Not applicable	Increased respiratory infections, physician visits, episodic symptoms.	CDC, US EPA	USA
Mortality rate attributed to household and ambient air pollution	Not applicable	Mortality rate attributed to household and ambient air pollution.	WHO	Global
Estimated burden of disease from urban ambient air pollution	Not applicable	Estimates of burden of disease from ambient air pollution in cities are generated by WHO air quality data (from monitoring of stations, satellite remote sensing, and air pollution chemical transport models) and scientific evidence from population exposure–response relationships.	WHO	Global
Mortality attributed to household air pollution	Not applicable	Modelled estimates of personal household air pollution exposure from the WHO Global Household Air Pollution Measurement database and WHO indoor air quality guidelines for household fuel combustion, which summarises evidence from exposure-response studies.	WHO	Global
Morbidity attributed to household air pollution	Not applicable	Modelled estimates of personal household air pollution exposure from the WHO Global Household Air Pollution Measurement database and WHO indoor air quality guidelines for household fuel combustion, which summarises evidence from exposure-response studies.	WHO	Global
Global mortality attributed to outdoorambient air pollution at levels above WHO guidelines (annualmean PM 2·5 or 10 μg/m³).	Not applicable	Population exposure to ambient air pollution concentrations exceeding WHO Air Quality Guidelines based on ground level urban air pollution monitoring of data collected by WHO and complemented by other methods such as remote sensing using satellite data and air pollution (chemical) transport models, to cover areas with little data. Mortality estimates based on population exposure to air pollution from monitoring of data in association with evidence on exposure-response, as summarised in WHO Air Quality Guidelines.	WHO	Global
Global morbidity attributed to outdoorambient air pollution at levels above WHO guidelines (annualmean PM 2·5 or 10 μg/m³).	Not applicable	Population exposure to ambient air pollution concentrations exceeding WHO Air Quality Guidelines based on ground level urban air pollution monitoring of data collected by WHO and complemented by other methods such as remote sensing using satellite data and air pollution (chemical) transport models, to cover areas with little data. Morbidity estimates based on population exposure to air pollution from monitoring of data in association with evidence on exposure-response, as summarised in WHO Air Quality Guidelines.	WHO	Global
Mortality from ambient air pollution by sector	Not applicable	This indicator estimates the mortality attributable to ambient PM2·5, combining atmospheric modelling with information about activity in emitting sectors.	Lancet Countdown, UN	Global
Premature deaths from ambient PM 2.5 exposure due to electricity use for air conditioning	Not applicable	To estimate country/region-specific premature deaths from ambient PM 2.5 exposure due to electricity use for air conditioning the proportion of total electricity final consumption used for air conditioning was multiplied by the estimated country/region-specific premature deaths due to PM 2.5 emissions from electric power plants.	Lancet Countdown, UN	Global
Asthma related hospital admissions	Not applicable	The number of hospitalization for the following disease groups by stratifying data: a. Asthma (493 or J45). Number and rate per 100,000 population	CDC	USA
Asthma related ED visits	Not applicable	Principal (primary) discharge diagnosis:a. Asthma (493 or J45). Number and rate per 100,000 population	CDC	USA
Allergic disease related hospital admissions	Not applicable	The number of hospitalization for the following disease groups by stratifying data:a. Allergic rhinitis (hay fever) (477 or J30.1)b. Atopic dermatitis (eczema) (691.8 or L20)c. Allergic urticaria (hives) (708.0 or L50.0)d. Allergic conjunctivitis (372.14 or H10.1). Number and rate per 100,000 population	CDC	USA
Allergic disease related ED visits	Not applicable	Principal (primary) discharge diagnosis:a. Allergic rhinitis (hay fever) (477 or J30.1)b. Atopic dermatitis (eczema) (691.8 or L20)c. Allergic urticaria (hives) (708.0 or L50.0)d. Allergic conjunctivitis (372.14 or H10.1). Number and rate per 100,000 population	CDC	USA
Human cases of coccidioidomycosis	Not applicable	Human cases of coccidioidomycosis.	CDC, US EPA	USA
Respiratory/allergic disease related to increased air pollution	Not applicable	Respiratory/allergic disease related to increased air pollution (O3 and other pollutants).	CDC, US EPA	USA
Respiratory/allergic mortality related to increased air pollution	Not applicable	Respiratory/allergic disease related to increased air pollution (O3 and other pollutants).	CDC, US EPA	USA
Respiratory/allergic disease related to increased pollens	Not applicable	Respiratory/allergic disease related to increased pollens.	CDC, US EPA	USA
Respiratory/allergic mortality related to increased pollens	Not applicable	Respiratory/allergic mortality related to increased pollens.	CDC, US EPA	USA
Water-borne diseases and other water-related health impacts
[Cases of] Leptospirosis	Not applicable	Confirmed cases per 100,000 population.	EEA	Europe
[Cases of] Shigellosis	Not applicable	Confirmed cases per 100,000 population.	EEA	Europe
[Cases of] Cryptosporidiosis	Not applicable	Confirmed cases per 100,000 population.	EEA	Europe
[Cases of] Giardiasis	Not applicable	Confirmed cases per 100,000 population.	EEA	Europe
Mortality rate attributed to unsafe water, unsafe sanitation and lack of hygiene (exposure to unsafe Water, Sanitation and Hygiene for All (WASH) services	Not applicable	Mortality rate attributed to unsafe water, unsafe sanitation and lack of hygiene (exposure to unsafe Water, Sanitation and Hygiene for All (WASH) services.	WHO	Global
Incidence of waterborne diseases	Not applicable	Incidence of waterborne diseases.	UN ESCWA	Middle East
[Cases of] Legionnaires’ disease	Not applicable	Confirmed cases per 100,000 population.	EEA	Europe
Zoonoses & vector-borne diseases
Mortality from climate-sensitive diseases	Not applicable	Mortality from climate-sensitive diseases.	WHO	Global
Incidence of vector-borne and waterborne diseases	Not applicable	This indicator tracks the annual incidence of vector- borne diseases (number of cases).	UN ESCWA	Middle East
Deaths among children under five years of age due to malaria (%) (girls/boys)	Not applicable	Measures the share of children under five years of age that die, where malaria is the underlying cause of death. Underlying cause-of-death information is needed.	SIDA	Global
Dengue infections	Not applicable	Confirmed cases per 100,000 population.	EEA	Europe
	Human cases of dengue fever	Human cases of dengue fever.	CDC, US EPA	USA
Malaria incidence	Not applicable	Malaria incidence per 1000 population.	WHO	Global
	[Cases of] Malaria	Confirmed cases per 100,000 population.	EEA	Europe
[Cases of] Chikungunya	Not applicable	Confirmed cases per 100,000 population.	EEA	Europe
[Cases of] Tick-borne encephalitis	Not applicable	Confirmed cases per 100,000 population.	EEA	Europe
Human cases of Lyme disease	Not applicable	Human cases of Lyme disease per 100,000 population.	CDC	USA
	[Cases of] Lyme disease	Confirmed cases per 100,000 population.	EEA	Europe
	Human cases of Lyme disease	Human cases of Lyme disease.	CDC, US EPA	USA
	[Cases of] Lyme disease	Rate of confirmed Lyme cases nationwide. Diagnosed by health care providers.	US EPA	USA
Human cases of West Nile Virus Disease	Not applicable	Human cases of West Nile Virus Disease, per 100,000 population.	CDC	USA
	Human cases of West Nile virus	Human cases of West Nile virus.	CDC, US EPA	USA
	West Nile Virus Infection	Confirmed cases per 100,000 population.	EEA	Europe
Cases of neuroinvasive West Nile virus disease	Not applicable	Rate of reported neuroinvasive West Nile virus disease cases across the United States. Neuroinvasive cases, which occur for less than 1 percent of people infected with West Nile virus, are those that affect the brain or cause neurologic dysfunction.	US EPA	USA
Human cases of hantavirus cardiopulmonary syndrome (HCPS)	Not applicable	Human cases of hantavirus cardiopulmonary syndrome (HCPS).	CDC, US EPA	USA
Malnutrition and food-borne diseases
[Cases of] Campylobacteriosis	Not applicable	Confirmed cases per 100,000 population.	EEA	Europe
Vibriosis infection	Not applicable	Confirmed cases per 100,000 population.	EEA	Europe
Percentage of populationundernourished (urban/rural;women/men)	Not applicable	A measure of food security. The percentageof population undernourishedprovides information on thenumber of people within a population whose dietary energy intake lies below their minimum requirements.	SIDA	Global
Prevalence of stunting in children younger than 5 years	Not applicable	Prevalence of stunting in children younger than 5 years (i.e., percentage of children who fell below -2 SD from the length-for-age or height-for-age WHO Child Growth Standards median), data for whom are most readily available, or younger than 2 years, which is the group most sensitive to the interventions.	WHO	Global
[Cases of] Salmonellosis	Not applicable	Confirmed cases per 100,000 population.	EEA	Europe
[Cases of] STEC/VTEC (toxin-producing Escherichia coli)	Not applicable	Confirmed cases per 100,000 population.	EEA	Europe
Human shellfish poisonings	Not applicable	Human shellfish poisonings due to harmful algae blooms.	CDC, US EPA	USA
Mental and psychosocial health
Mental health	Not applicable	Association between mean annual maximum temperatures and suicide rates.	Lancet, UN	Western Pacific

Figure 3.

Frequency of severity measures.

Dark boxes represent indicators that measure mortality; light boxes represent indicators that measure injury or illness; medium boxes at the end represent indicators that measure both mortality and morbidity.

3.3. Secondary categorization

From the seven primary categories with indicators reported above, 23 secondary categories were identified (Figure 4). The 73 unique indicators were each classified into the single most relevant secondary category and mapped to illustrate the variation in use of indicators between the different areas of climate change. The most prevalent secondary category indicators were mortality from extreme weather events and illness from zoonoses/vector-borne diseases with 10 and 9 unique indicators respectively. No indicators were identified for the following five secondary categories: illness from mental and psychological health, cold-related illness, mortality from food-borne disease, mortality from malnutrition and mortality from water-borne diseases.

4. Discussion

The main purpose of this scoping review was to map the currently available indicators relevant to human health outcomes that are directly influenced by climate change, to support the development of a transparent, globally generalizable framework and technical platform for official statistics on climate change and health. The scoping review was rigorously carried out and designed in close collaboration with end users; it was conducted according to a registered protocol and with the support of an experienced information specialist.

This review included 20 diverse sources that were, in most cases, syntheses of previous indicator sets and/or incorporated expert opinion. These data were sourced from seven North American, six global (including low- and middle- income countries), four European, two Western Pacific, and one Middle Eastern datasets. Therefore, the overview presented here is comprehensive. We have accounted for duplication and heterogeneity in the included indicators by identifying unique indicators and grouping them with repeated ones, along with multiple definitions (see Table 4). This review makes an important contribution to the work of establishing standardized indicators across countries, which will enable vital work such as analyzing patterns of change over time and supporting the application of methods for the detection and attribution of climate change impacts. ³⁵

The main categories of indicators retrieved were: temperature-related illness and mortality; extreme weather event-related injury and mortality; and air quality-related illness and mortality. There is increasing strength of scientific attribution of the increased intensity, frequency and duration of heatwaves and other extreme weather events, and changes in air quality, to climate change.^36,37 Therefore, as climate change intensifies, we can expect more health impacts in these three categories, and consequently the range of available indicators will be useful for tracking the changing burden of disease and associated policy needs.

Five secondary categories did not have climate-sensitive health indicators and understanding this gap is helpful to inform respective official statistics. These indicator gaps were for cold-related illness, mortality from food-borne diseases, mortality from malnutrition, mortality from water-borne disease and illness from mental and psychosocial health. It is interesting to speculate why these secondary categories might have indicator gaps: i) there is limited evidence of attribution from climate change to human health effects, ii) operationalising the concept is challenging and iii) data availability barriers. For instance, focusing on psychosocial health it is surprising there is an absence of indicator as climate change is projected to be the greatest mental health threat of the twenty-first century. ³⁸ Yet, mental health impacts of climate change are under-researched compared with other health areas. ³⁸ Further, while diagnostic codes (e.g., ICD10 codes) capture some aspects of mental health, there is limited globally standardised data on sub-clinical mental wellbeing as it is less readily observable.

We encountered the following challenges with this review:

• The review was designed to include indicators related to the direct impacts of climate change on human health. In practice, there were sometimes challenges determining whether an indicator measured direct vs indirect impacts of climate change; for example, the indicator of “pollen days” versus “increase in allergy symptoms” track the same process by which climate change increases the pollen generated by trees, thereby increasing symptoms for those allergic to pollen, but measure different points in the process.

• We found that the authors of included sources differed in their decisions about attributing specific indicators to climate change, or classifying impacts of certain climate change-related events, such as wildfires, as being health-related.

4.1. Limitations

Limitations of this scoping review include the following;

• The decision to restrict inclusion to global authorities was made on grounds of feasibility. Due to the fact that the included sources were largely syntheses and/or compilations of expert opinion, the potential limitation of this decision was offset by the comprehensiveness of the included sources.

• The majority of the included sources came from high-income settings (see Table 3. Characteristics of included studies). This reflects what is available, but highlights the fundamental need for more health research focused on low- and middle-income countries. ³⁹

• The determination as to whether an indicator was unique was limited by reported details and entailed subjective assessments by the review team.

4.2. Future work

This scoping review synthesized one category of indicators: direct climate change impacts on population health. It would be valuable to amplify this work with future syntheses of indicators of health-system impacts; indirect human health impacts; vulnerability, exposure and risk; adaptation processes; health-system resilience; institutional engagement; and learning and knowledge management.^4,39 Analyzing indirect and/or qualitative indicators could be particularly useful for mental health topics. It would also be valuable to include indicators from outside the health system, relevant to the provision of core health. ⁴ As well, the indicators we retrieved were designed mainly with data at the national and large regional level; further assessment would be useful to establish the extent to which these indicators are relevant at local or smaller regional levels at various reporting cadences. ¹⁰ The additional burden of national reporting of official statistics data to global authorities such as the UN and WHO is minimal in most cases, given that the great majority of the data required is already collected by national governments for their own use while governments which are unable to collect such data for their own statistics are unlikely to do so for international purposes. Future work could delve more deeply into reporting burden, such as consideration of the cadence of reporting, data collection costs and analytical complexity. Lastly, further work to investigate the degree to which the indicators collated here are being used in practice by national statistics organizations would be valuable.

The purpose of this paper is to provide a comprehensive and reproducible shortlist of candidate indicators. The larger project of which this scoping review forms a part involves a wide international collaboration to develop these indicators, including partnering policy makers and statisticians from the UK, Rwanda and Ghana to discuss feasibility. Our approach is to identify the currently most valuable measurement approaches and make these available in a form that can be used by official statistics organisations, not inventing new indicators unless absolutely necessary to meet identified needs. The problems of statistical capacity and limited data availability in LMICs are central considerations, and future work would be beneficial to dive into on-the-ground feasibility across different contexts.

For the task of creating official statistics frameworks, the findings of this scoping review, and future syntheses addressing the gaps listed above, should be amplified with an expert consensus process such as the Delphi method in order to draw on the nuances and networks that subject-matter experts can provide.^40,43

5. Conclusion

The goal of policymakers, as climate change intensifies, must be to protect human health and health systems from a range of impacts. To this end, relevant and efficient indicators can track progress and highlight where and when corrections are needed for future planning and to increase the effectiveness of adaptation-related policy-making. ⁴ Indicators are only as useful as the data available to support them; in countries for which extensive data is available to support analyses by age, gender, socio-economic status and other demographic criteria, decision-making can be more nuanced, which has equity implications given that less-affluent countries often have less-detailed data available.

Our findings make an important contribution to supporting global climate-health policymaking by presenting a comprehensive and up-to-date overview of indicators relevant to the health impacts of climate change. The collection and synthesis of measures of climate-sensitive health outcomes undertaken in this review will aid in provision of a framework and platform for future monitoring of, and adaptation to, the health impacts of climate change-related events. Official statistics can contribute to the measurement and monitoring of various aspects of climate change and support bringing together current and future statistical work into a cohesive framework. Indicators, with their potential to reveal relative positions or show positive or negative change, are crucial for setting targets and monitoring achievement. The list of indicators presented here will have value for official statistics and long-term policy and planning purposes.

Official statistics are produced by governmental organisations and (in most countries) are independent of immediate political control, giving them authority and credibility. Presently there is a lack of harmonized approaches across countries’ official statistics in regard to climate change and health. Addressing this gap, and supporting the required actions with the needed funding, will help to support national monitoring and evidence-based policy-making, always recognizing that policymaking processes are influenced by many factors beyond data availability. ⁴¹ The frontiers of climate change science are continually moving. Over time, to help governments meet the urgent imperatives of effective climate action, official statistics offices may come to incorporate new data sources such as satellite observation or processed data published by academic institutions or non-governmental organizations. The incorporation of ‘non-official’ data sources presents a dilemma which many governmental organisations are already grappling with, and it requires finding a principled balance between data quality and known provenance on the one hand, and increasing pressures for timeliness and granularity on the other. The ability of official statistics offices, especially in low- and middle-income countries, to monitor the effects of climate change will be increased by the provision of practical, coherent standards and open-source tools. Both producers and consumers of climate change research will benefit from faster study development based on shared approaches and a more consistent “language” for communication of findings. Ultimately these benefits will translate into a better-informed international community and more effective, evidence-based climate change interventions. ⁴²