Health impacts of climate change and mitigation policies in Ireland

Duffy, Katie et al. Research Report Health impacts of climate change and mitigation policies in Ireland Research Series, No. 188 Provided in Cooperation with: The Economic and Social Research Institute (ESRI), Dublin Suggested Citation: Duffy, Katie et al. (2024) : Health impacts of climate change and mitigation policies in Ireland, Research Series, No. 188, The Economic and Social Research Institute (ESRI), Dublin, https://doi.org/10.26504/rs188 This Version is available at: https://hdl.handle.net/10419/301930 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ HEALTH IMPACTS OF CLIMATE CHANGE AND MITIGATION POLICIES IN IRELAND KATIE DUFFY, KELLY DE BRUIN, LOÏC HENRY, CLEMENT KWEKU KYEI, ANNE NOLAN AND BRENDAN WALSH RESEARCH SERIES NUMBER 188 JULY 2024 E V I D E N C E F O R P O L I C Y HEALTH IMPACTS OF CLIMATE CHANGE AND MITIGATION POLICIES IN IRELAND Katie Duffy Kelly de Bruin Loïc Henry Clement Kweku Kyei Anne Nolan Brendan Walsh July 2024 RESEARCH SERIES NUMBER 188 Available to download from www.esri.ie © The Economic and Social Research Institute Whitaker Square, Sir John Rogerson’s Quay, Dublin 2 https://doi.org/10.26504/rs188 This Open Access work is licensed under a Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. ABOUT THE ESRI The Economic and Social Research Institute (ESRI) advances evidence-based policymaking that supports economic sustainability and social progress in Ireland. ESRI researchers apply the highest standards of academic excellence to challenges facing policymakers, focusing on ten areas of critical importance to 21st Century Ireland. The Institute was founded in 1960 by a group of senior civil servants led by Dr T.K. Whitaker, who identified the need for independent and in-depth research analysis. Since then, the Institute has remained committed to independent research and its work is free of any expressed ideology or political position. The Institute publishes all research reaching the appropriate academic standard, irrespective of its findings or who funds the research. The ESRI is a company limited by guarantee, answerable to its members and governed by a Council, comprising up to 14 representatives drawn from a crosssection of ESRI members from academia, civil services, state agencies, businesses and civil society. Funding for the ESRI comes from research programmes supported by government departments and agencies, public bodies, competitive research programmes, membership fees, and an annual grant-in-aid from the Department of Public Expenditure NDP Delivery and Reform. Further information is available at www.esri.ie. THE AUTHORS Kelly de Bruin and Brendan Walsh are Senior Research Officers at the Economic and Social Research Institute (ESRI) and Adjunct Associate Professors at Trinity College Dublin (TCD). Clement Kweku Kyei is a Postdoctoral Research Fellow at the ESRI. Loïc Henry is an Assistant Professor at LEDa, Université Paris-Dauphine, France. Anne Nolan is a Research Professor at the ESRI and Adjunct Professor at TCD. Katie Duffy was a Research Assistant at the ESRI when the research was conducted. ACKNOWLEDGEMENTS The research was funded by the Irish Heart Foundation and Irish Cancer Society on behalf of the Climate and Health Alliance under a programme of research on the ‘Health Effects of Climate Change and Mitigation Actions in Ireland’ carried out at the ESRI. This work was also supported by funding from the Environmental Protection Agency under the Climate Change Advisory Council Fellowship on Adaptation. The authors thank the members of the Steering Group (Emma Harte, Irish Cancer Society; Mark Murphy, Irish Heart Foundation; Tom McDermott, University of Galway; and Seamus McGuinness, ESRI) for helpful guidance and discussions throughout the project. We are very grateful to the Healthcare Pricing Office (HPO) and Met Éireann for providing the data used in the analyses in this report. This report has been accepted for publication by the Institute, which does not itself take institutional policy positions. All ESRI Research Series reports are peer reviewed prior to publication. The author(s) are solely responsible for the content and the views expressed. Table of contents | iii TABLE OF CONTENTS EXECUTIVE SUMMARY ........................................................................................................................ VIII 0.1 Introduction ....................................................................................................................viii 0.2 Methods ........................................................................................................................... ix 0.2.1 Climate change modelling ix 0.2.2 Health impacts ix 0.2.3 Health benefits and co-benefits x 0.3 Key findings ...................................................................................................................... xi 0.3.1 Climate change modelling xi 0.3.2 Health effects of temperature change xi 0.3.3 Health benefits and co-benefits of climate change mitigation measures xi 0.4 Discussion .........................................................................................................................xii CHAPTER 1: BACKGROUND .................................................................................................................... 1 1.1 Introduction ...................................................................................................................... 1 1.2 Conceptual framework (climate change and health) ....................................................... 2 1.3 Literature review (temperature change and health) ........................................................ 6 1.3.1 International literature on temperature change and health 7 1.3.2 Irish literature on temperature change and health 9 1.4 Conceptual framework (health benefits and co-benefits of climate change mitigation) ....................................................................................................................... 10 1.4.1 Air pollution 11 1.4.2 Sustainable transport and diet 12 1.5 Literature review (health benefits and co-benefits of climate change mitigation) ........ 13 1.5.1 Scenario-based analyses 13 1.5.2 Analyses of co-benefits of specific mitigation actions 15 1.6 Report structure .............................................................................................................. 17 CHAPTER 2: CLIMATE CHANGE MODELLING ....................................................................................... 18 2.1 Introduction .................................................................................................................... 18 2.2 Global and regional climate models ............................................................................... 18 2.3 Greenhouse gas emissions scenarios .............................................................................. 19 2.4 Future climate projections for Ireland ............................................................................ 21 2.5 Summary ......................................................................................................................... 25 CHAPTER 3: HEALTH EFFECTS OF TEMPERATURE CHANGE ................................................................ 26 3.1 Introduction .................................................................................................................... 26 Table of contents | iv 3.2 Data ................................................................................................................................. 27 3.2.1 Hospital In-Patient Enquiry (HIPE) 27 3.2.2 Population data 28 3.2.3 Meteorological data 28 3.2.4 Analytical sample 29 3.3 Methodology ................................................................................................................... 30 3.3.1 Outcome variables 30 3.3.2 Model specification 1: Temperature bins 30 3.3.3 Model specification 2: Mean deviation model 32 3.4 Results ............................................................................................................................. 34 3.4.1 Temperature bin analysis 35 3.4.2 Mean deviation model 41 3.5 Discussion ........................................................................................................................ 42 CHAPTER 4: HEALTH BENEFITS AND CO-BENEFITS OF CLIMATE CHANGE MITIGATION MEASURES ........................................................................................................................................... 44 4.1 Introduction .................................................................................................................... 44 4.2 Analyses of health benefits and co-benefits for Ireland ................................................. 44 4.3 Health effects of actions to mitigate temperature change in Ireland: Case study ......... 48 4.4 Summary ......................................................................................................................... 53 CHAPTER 5: SUMMARY, DISCUSSION AND POLICY IMPLICATIONS ..................................................... 55 5.1 Summary of main findings .............................................................................................. 55 5.2 Discussion and policy implications .................................................................................. 56 REFERENCES ......................................................................................................................................... 60 APPENDIX A – DATA, VARIABLES AND SUMMARY STATISTICS............................................................ 69 APPENDIX B – LAGGED EFFECTS OF TEMPERATURE BIN MODEL ........................................................ 72 APPENDIX C – MEAN DEVIATION MODEL: GROUP/DIAGNOSIS ANALYSIS AND FULL MODEL RESULTS ............................................................................................................................................... 73 APPENDIX D – PROJECTION ANALYSIS ................................................................................................. 76 Scenario RCP4.5 – most likely ................................................................................................. 76 List of tables | v LIST OF TABLES Table 2.1 The increase in global mean temperature compared to preindustrial level 21 Table 3.1 Temperature bins used in the analysis 31 Table 3.2 Illustration of mean deviation admissions variable 34 Table 3.3 Emergency in-patient hospital admissions (per 100,000 population) 2015–2019 35 Table 3.4 Temperature bin analysis 36 Table 3.5 Temperature bin analysis – age group analysis 38 Table 3.6 Temperature bin analysis – diagnostic group analysis 40 Table 3.7 Mean deviation model, quarters 2 and 3 41 Table 3.8 Numerical illustration 42 Table 4.1 Heat-attributable mortality in Ireland (additional deaths) 46 Table 4.2 Health and economic benefits of air pollution reductions (per annum) 48 Table 4.3 Illustration of temperature scenario data 49 Table 4.4 Number of days per quarter above the 90th threshold for the RCP4.5 (“most likely”) scenario 50 Table 4.5 Average quarterly threshold values for the 75th, 90th and 95th thresholds 50 Table 4.6 HIPE model results 51 Table 4.7 Projected increase in emergency hospitalisations by 20-year period (RCP4.5, 90th percentile) 52 Table 4.8 Lower bound estimates (RCP4.5, 95th percentile) 52 Table 4.9 Comparison of annual increases across RCP scenarios 53 Table A.1 Summary of variables used from HIPE 69 Table A.2 ICD-10-AM codes for diagnosis variables 70 Table B.1 Coefficient estimates for temperature bin analysis 72 Table C.1 Mean deviation model by age group 73 Table C.2 DIfferences model by diagnosis group 74 Table C.3 Coefficient estimates from the mean deviation model 75 Table D.1 Number of days per quarter above the 75th threshold (RCP4.5) 76 Table D.2 Projected increase in emergency hospitalisations (75th threshold and RCP4.5) 76 Table D.3 Number of days per quarter above the 95th threshold (RCP4.5) 76 List of figures | vi LIST OF FIGURES Figure 0.1 Infographic of effects of climate change on health x Figure 1.1 Infographic of effects of climate change on health 4 Figure 1.2 Infographic of effects of increasing temperatures on health 5 Figure 1.3 Infographic of health benefits and co-benefits of mitigation actions 11 Figure 2.1 Dublin County historical and future annual maximum temperature 22 Figure 2.2 Projections of temperature change for RCP4.5 (A) and RCP8.5 (B) scenarios 23 Figure 2.3 Mid-century seasonal projections of temperature change for RCP4.5 (a) and RCP8.5 (b) scenarios 24 Figure 3.1 Average daily maximum temperature (°C) and rainfall (mm) by county in 2019 29 Figure 3.2 Coefficient plot of temperature bins 37 Figure 4.2 Trend in heat-attributable deaths 46 Figure A.1 Emergency in-patient hospital admissions rate by county (2019) 71 Executive summary | xiii United States’ Gulf Coast. Finally, the report underlines the importance of appropriate data collection and availability in climate change research. Global Burden of Disease studies are only now starting to include temperature change as a risk factor for global disease and mortality, and are underpinned by numerous assumptions, many of which will not be transferable to an Irish (or moderate climate) context. The research in this report adopted a more direct approach, quantifying temperature change impacts on health (i.e., morbidity) via hospital admissions from the HIPE dataset. This approach allowed for the control of additional factors influencing hospital admissions and emphasises the value of making administrative health data like HIPE accessible to researchers and policymakers, and the linking of such data with other datasets (e.g., weather data), to fully harness its richness for evidence-based policy formulation. Background | 1 CHAPTER 1 Background 1.1 INTRODUCTION Recognition of the need to limit climate change has driven global negotiations concerning combined efforts to decrease greenhouse gas (GHG) emissions over the past decades within the United Nations Framework Convention on Climate Change (UNFCCC). In 2015, the Paris Agreement was adopted and to date has been ratified by 197 states and the European Union (EU); the global commitment to this agreement was reinforced in the Conference of Parties (COP) held in November 2021 and the resulting Glasgow Climate Pact. Under the Paris Agreement, the EU has submitted its EU-wide emissions targets (through Nationally Determined Contributions (NDCs)), which commit to a GHG emissions reduction goal of at least 55 per cent compared to 1990 levels by 2030 and net-zero emissions by 2050. These targets have been legislated through the EU Climate Law, making them legally binding. Ireland has shown its commitment to reducing emissions, where the Programme for Government 2020 included an annual emissions reduction target of 7 per cent, resulting in a 51 per cent reduction of emissions by 2030 (Government of Ireland, 2020). This target was made legally binding by the Climate Action and Low Carbon Development (Amendment) Act of 2021, and further commits to net-zero emissions by 2050. The government has also made significant strides over the past years to introduce and formulate policies needed to ensure these targets are met. The Climate Action Plans 2019 and 2021 provide a detailed plan of measures needed for this transition, with the third national update, the Climate Action Plan 2023, setting out the additional measures required to align with economy-wide carbon budgets and sectoral emission ceilings (Government of Ireland, 2021, 2022). However, the Environmental Protection Agency (EPA) has noted that Ireland is only on target to achieve a reduction of 29 per cent in GHG emissions by 2030 compared to a target of 51 per cent (EPA, 2023). When analysing the effects of climate change, the focus is often on the economic costs to society. For example, the European Environment Agency (EEA) estimated that the economic losses from weather- and climate-related extremes in Europe reached approximately half a trillion euro over the last 40 years (European Environment Agency, 2021), and €48bn in 2021 alone (van Daalen et al., 2022). For Germany, Karlsson and Ziebarth (2018) estimate that one additional hot day with temperatures above 30°C would create monetised health losses of between €750,000 to €5 million per 10 million population. In France, Adélaïde et al. (2022) estimate that the economic impact of health effects from heatwaves over the period 2015–2019 amounts to approximately €25.5 billion including mortality, minor restricted activity days and morbidity. Hensher (2023) highlights the fact Background | 2 that climate change can also affect the sustainability of the healthcare sector. He notes that it will be necessary to prepare the healthcare system for emerging health needs as well as the physical, economic and social impacts of continued climate change. Health impacts of climate change are important and remain relatively unexplored for Ireland. As climate change is predicted to worsen over time, climate changerelated health impacts are likely to become more pronounced even in temperate climates (Gibney et al., 2022). In addition, the health risks associated with climate change are more pronounced for vulnerable population groups, such as the older population and children, and those with pre-existing chronic diseases (Crimmins et al., 2016; EASAC, 2019; Flood et al., 2020; Romanello et al., 2022). More socioeconomically disadvantaged populations are also more vulnerable to the effects of climate change (Kaźmierczak et al., 2022). There is also little discussion of the avoided climate impacts in Ireland as a result of climate change mitigation measures. Mitigation measures limit the extent of climate change and hence limit the climate-related health impacts in Ireland. Certain mitigation measures also have concomitant health co-benefits, e.g., the shift to more bicycle-based commuting through increased cycle lanes or the switch to lower meat consumption can help reduce emissions and improve health outcomes. The research detailed in this report, carried out as part of a research programme funded by the Irish Heart Foundation and Irish Cancer Society on behalf of the Climate and Health Alliance, aims to help fill this gap. It focuses on the health impacts of climate change (and specifically temperature change) and how global and national commitments to limiting temperature change may reduce these impacts in Ireland. Therefore, it also aims to better understand some of the health benefits and co-benefits of climate change mitigation efforts. The following section (Section 1.2) sets out a conceptual framework that describes the various ways in which climate change may affect health. It provides a framework that underpins the analysis of the health effects of climate change in the Irish context (carried out in Chapter 3). Focusing on the health impacts of temperature change, Section 1.3 then discusses the national and international literature on the health effects of temperature change. Section 1.4 introduces the conceptual framework that underpins the analysis of health benefits and co-benefits of climate change mitigation actions (carried out in Chapter 4), while Section 1.5 discusses the national and international literature on the health benefits and co-benefits of climate mitigation actions. Section 1.6 provides a brief overview of the structure of the remainder of the report. 1.2 CONCEPTUAL FRAMEWORK (CLIMATE CHANGE AND HEALTH) When examining the effects of climate change on health, it is helpful to provide a framework in which to conceptualise the ways in which climate change affects health. International studies have shown that the pathways through which climate Background | 3 change affects health are complex (Crimmins et al., 2016; Romanello et al., 2022). There are three reasons underpinning this complexity: • the various events that could occur (e.g., increasing temperatures, heatwaves, flooding); • whether pathways are direct or indirect (i.e., mediated via other factors); • the various health outcomes that can be measured (e.g., specific diseases, mortality). The latest Assessment Report of the Intergovernmental Panel on Climate Change (IPCC) sets out the most up-to-date scientific information in relation to global climate change. Physical changes attributed to climate change include increases in hot extreme temperatures, upper ocean acidification, global sea level rise, glacier retreat, increase in heavy precipitation, increase in flooding, increase in fire weather and increase in agricultural and ecological drought (Cisse and McLeman, 2023). For Europe, extreme weather events (e.g., heatwaves), northward movement of diseases (e.g., malaria), drought, forest fires and soil moisture deficits have been highlighted as particular concerns (European Environment Agency, 2021). In terms of pathways, the literature that aids in the conceptualisation of climate change effects on health illustrates the effect pathways in the following way (Smith, 1999; Department of Health, 2019; EASAC, 2019; Romanello et al., 2022). The pathways are separated into two distinct channels: direct and indirect. Indirect pathways are further separated into effects mediated by (1) ecosystem/environment, and (2) institutions/infrastructure. Direct effects are not mediated by other factors and capture, for example, the effect that increasing temperatures or extreme weather events have on health outcomes. This could include flooding directly causing injury, but also heat exacerbating cardiovascular and respiratory diseases, especially among frailer individuals. Indirect pathways, by contrast, are mediated by other factors. Hotter temperatures can give rise to bacterial conditions in water that can lead to water-borne disease outbreaks, or wildfires that can be linked with respiratory and circulatory disease complications (Navarro et al. 2019). These pathways are summarised in the following infographic devised by the authors of this report. Background | 4 FIGURE 1.1 INFOGRAPHIC OF EFFECTS OF CLIMATE CHANGE ON HEALTH Source: Infographic devised by authors based on literature. Figure 1.1 shows that there are multiple health outcomes from even just one climate change event. For example, increases in temperature can cause cardiovascular and respiratory distress, while also contributing to longer warm seasons, aiding the transmission of insect-borne and water-borne diseases. According to research carried out as part of the 2019 Global Burden of Disease (GBD) study, the high temperature-related disability-adjusted life year (DALY) and death rates were the highest for lower respiratory infections, followed by stroke and diabetes mellitus (Song et al., 2021). The direct and indirect effects of climate change occur within a social and economic context. This context is important to consider because certain vulnerable groups in the population are more at risk. Literature has shown that these groups include, but are not limited to, those who are aged over 65 or under 5, those with chronic illnesses or disabilities and those who are socio-economically disadvantaged (Crimmins et al., 2016; EASAC, 2019; Romanello et al., 2022). For example, those in more disadvantaged social positions may be more likely to live in areas that are exposed to climate change (European Environment Agency, 2018). Vulnerable population groups may also be more vulnerable to the health-damaging effects of climate change such as air pollution, due to other characteristics such as poor housing conditions, chronic disease, etc. Certain occupational groups, such as outdoor workers, paramedics, firefighters and transport workers, as well as workers in hot indoor work environments, will be especially vulnerable to extreme heat (Flood et al., 2020). These complex interactions are illustrated for increasing temperatures in the following infographic devised by the authors based on reports from the IPCC (Smith et al., 2014; Cisse and McLeman, 2023), Department of Health (2019) and EASAC (2019). It would be possible to conceptualise other climate change events (e.g., flooding) in a similar framework. Background | 5 FIGURE 1.2 INFOGRAPHIC OF EFFECTS OF INCREASING TEMPERATURES ON HEALTH Source: Infographic devised by authors based on literature. This framework underpins the approach taken to structure the analysis in the first part of the report. The analysis will primarily examine the effects of increasing temperatures on health, which is identified by the EEA to be one of the key climatehealth threats facing Europe (Kaźmierczak et al., 2022). The EEA also notes that despite high average living standards, Europe’s ageing society and prevalence of chronic diseases make its population particularly vulnerable to heat (Kaźmierczak et al., 2022). While climate-health threats are likely to differ between northern and southern Europe, the Department of Health (2019) and Desmond et al. (2017) have identified increasing temperatures as one of the principal health threats facing Ireland with regard to climate change. In Ireland, mean air temperatures have increased by 0.8°C in the 1900–2011 period, with projections out to the midcentury estimating an increase of 1.1–1.6°C (Desmond et al., 2017). Desmond et al. (2017) also highlight the risk from increased precipitation or flooding which can increase the risk of water-borne diseases such as campylobacteriosis and cryptosporidiosis. Future research (discussed in Chapter 5) could examine the health impacts of these and other likely climate change events in Ireland, as well as the impacts on other outcomes such as worker productivity. In the following section, we provide a more detailed overview of the national and international literature that has examined the effect of temperature change on health, before moving on to discuss the conceptual framework and literature on the health benefits and co-benefits of climate change mitigation. Background | 6 1.3 LITERATURE REVIEW (TEMPERATURE CHANGE AND HEALTH) The available literature on the effects of temperature change on health covers a wide range of countries, time periods and modelling approaches. The climate in which temperature-health studies are conducted is important as countries or cities that are more acclimatised to heat may have already developed adaption measures to deal with hotter temperatures. Acclimatisation can occur through physical adaptation, housing characteristics, or behavioural patterns (e.g., staying indoors, changing work patterns) (Anderson and Bell, 2009). 1 Indeed Barreca et al. (2016) found that the diffusion of air conditioning explained nearly all of the decline in heat-related mortality observed in the US since 1960. This means that any identified effects of temperature on health from these studies may be muted or only exist for extreme temperatures, and may not therefore be generalisable to more temperate climates. Baccini et al. (2008) deal with this in their study of multiple cities by allowing the threshold temperature value to vary across different cities: they show that the threshold temperature for London is 23.9°C compared to Rome, which had a threshold value of 30.3°C. How temperature is characterised, and the resulting modelling approach, can also differ considerably across studies. For example, Karlsson and Ziebarth (2018) contrast the differing approaches implemented across epidemiological and economic studies. In some cases, temperature is characterised as a continuous variable, in others temperature is categorised into ‘temperature bins’ to allow for more flexible functional forms (Deschênes and Greenstone, 2011; White, 2017; Gibney et al., 2022; Liao et al., 2023), while in others threshold or percentile values are constructed (Hajat et al., 2006; Baccini et al., 2008; Breitner et al., 2014). Many studies also take account of the potential for lagged effects in the response of health outcomes (e.g., mortality) to changes in temperature (Goodman et al., 2004; Baccini et al., 2008; Anderson and Bell, 2009; Zeka et al., 2014; Liao et al., 2023). Furthermore, previous research uses a variety of health outcome measures to assess the health impacts of temperature change. Indeed, a recent overview of systematic reviews of the literature on the effects of climate change on health categorised health impacts into ten broad groups (Rocque et al., 2021). These groups covered outcomes such as hospitalisations, mortality, infectious diseases and respiratory, cardiovascular and neurological disease, and mental health and wellbeing (Rocque et al., 2021). Mortality is generally measured by using the number of deaths (all-cause and in some cases, cause-specific) or the agestandardised mortality rate. Mortality is the most common outcome examined in existing studies of the effects of temperature on health (Hajat et al., 2006; Baccini et al., 2008; Deschênes and Greenstone, 2011; Breitner et al., 2014; Gasparrini et al., 2022; Liao et al., 2023). However, while mortality measures deaths in a population, morbidity is also a key outcome. Morbidity focuses on the prevalence 1 It has been suggested the same is true for cold weather. Deschênes and Greenstone (2011), however, find very weak evidence in support of this hypothesis for cold weather. Background | 7 and impact of diseases and health conditions, including non-fatal conditions. Morbidity allows for the broader impact of disease, or the causes of disease (e.g., temperature increases), on quality of life and healthcare systems to be examined. Morbidity is commonly proxied by calculating healthcare utilisation, such as in-patient hospital admission rates or the rate of attendances to emergency departments (EDs). In Section 3.2 in Chapter 3, we describe the data and methods we use in this analysis to model the impact of temperature change on health in Ireland. In the following sections (Sections 1.3.1 and 1.3.2), we survey the relevant national and international literature on temperature change and health. 1.3.1 International literature on temperature change and health The 2019 GBD study provides a comprehensive picture of mortality and disability across countries, time, age, and sex. It quantifies health loss from hundreds of diseases, injuries and risk factors. As a result of multiple requests to begin capturing important dimensions of climate change into the GBD study, the direct relationship between high and low non-optimal temperatures on all GBD disease and injury outcomes was modelled in the 2019 study (Murray et al., 2020). 2 Globally, low temperature or cold was the 12th-leading level 3 3 risk factor for deaths in 2019, contributing to 2.9 per cent of all deaths, or 1.65 million deaths. Globally, 308,000 deaths in 2019 were attributable to exposure to high temperatures, concentrated mainly in south Asia, north Africa, the Middle East and Sub-Saharan Africa. A related study for the 2019 GBD focused specifically on high temperature, and evaluated the disease burden attributable to high temperature. The results show that in 2019, 589 deaths (or a rate of 0.06 per 100,000 population) in Western Europe were attributed to high ambient temperature. The analysis also showed that the disease burden attributable to high temperature varied spatially, with the heaviest burden in regions with low socio-demographic index (SDI) and the lightest burden in regions with high SDI (Song et al., 2021). There are two key pan-European studies that examined heat effects on mortality in cities. While Baccini et al. (2008) examined the effects of temperature, Hajat et al. (2006) analysed whether there is an added heatwave effect on mortality. Overall, positive associations between temperature and mortality were identified. Baccini et al. (2008) estimated the observed mortality effect past a city-specific threshold at about 3.1 per cent for Mediterranean cities and 1.8 per cent for northern Europe for each 1°C increase above the threshold. Hajat et al. (2006) found that for London, mortality increased by 5.1 per cent for every 1°C increase above the identified threshold. The studies found that overall summertime mortality burden is more important to examine than the acute effects of 2 However, other climate-related relationships, such as between precipitation or humidity and health outcomes, have not yet been evaluated. 3 The GBD methodology has a risk factor hierarchy. Level 1 risk factors are behavioural, environmental and occupational, and metabolic; Level 2 risk factors include 20 risks or clusters of risks (e.g., non-optimal temperature); Level 3 includes 52 risk factors or clusters of risks (e.g., high temperature) (Murray et al., 2020). Background | 8 heatwaves. Their findings also showed greater associations for respiratory deaths, even when air quality was controlled for. Other studies have examined the temperature-heat relationship within one country. Breitner et al. (2014) conducted a study on three German cities, using time-series analysis to investigate the association between daily air temperature and cause-specific mortality. They found that an increase from the 90th to the 99th percentile of 2-day mean temperature led to an increase in non-accidental mortality by 11.4 per cent. Alternatively, a decrease from the 10th to the 1st percentile in a 15-day mean temperature led to an increase in mortality by 6.2 per cent. They found that the population aged over 85 were the most susceptible to excess heat. Adélaïde et al. (2022) analysed the health effect of heatwaves in France using ED visits and out-patient clinic visits. They found a significant effect of temperature on ED visits for heat-related symptoms. Liao et al. (2023) examined the relationship between extreme heat and mortality using county-level data for China over the period 2000–2015. They found that an additional day with a maximum temperature of 38°C or above was associated with a 1.7 per cent net increase in the monthly mortality rate (relative to if that day’s maximum had been in the 16–21°C range). The use of ED data to estimate temperature-related health effects is also applied in a study for the UK (Gibney et al., 2022). This study draws on a similar methodology as White (2017), who investigated the relationship between temperature and ED attendances in California. There are similar findings from the two studies, but the analysis in the UK study is more relevant to the analysis in this report. Gibney et al. (2022) found an immediate temperature effect on heatrelated morbidity, as measured by ED visits. In contrast, cold-related morbidity has a lagged response of up to three weeks after the temperature shock, with a greater cumulative effect. Other health outcomes have also been examined; for example, Graff Zivin et al. (2018) exploited variation in survey interview dates with young people from the US National Longitudinal Survey of Youth to examine the effect of temperature on cognitive performance. They find that maths performance declines linearly above 21°C, with the effect statistically significant beyond 26°C (the effects of temperature fluctuations on assessments of reading recognition and comprehension were non-significant). Mullins and White (2019) analysed the effect of temperature fluctuations on a variety of mental health outcomes (including ED visits for mental health conditions, and suicides) in the US, and found that cold temperatures reduce mental health symptoms while hot temperatures increase them. A recent systematic review found evidence of associations between heat and preterm birth (Bekkar et al., 2020). Yu et al. (2023) note that climate change will widen inequities in cancer incidence and treatment through its complex connections with modifiable risk factors, such as ambient and household air pollution. While the evidence base for associations between high temperatures and cancer is still developing, increasing ultraviolet radiation (UV) exposure is associated with increased risks of melanoma and other skin cancers (e.g., Background | 9 squamous cell skin cancer). Recent research from the US has also found that firefighters who deal with wildfires have increased risks of lung cancer and cardiovascular disease mortality (Navarro et al. 2019). A few studies examining temperature effects on health look at distributional impacts and cause-specific impacts. Rizmie et al. (2022) used a similar methodology to both White (2017) and Gibney et al. (2022) for UK ED attendance data, but subsequently stratified the analysis by age and socio-economic deprivation. They also stratified by type of diagnosis. They found that older and deprived populations were the most at risk of adverse health effects from temperature-related illness, and in particular for admissions due to metabolic disease and injuries. Gasparrini et al. (2022) also examined the distribution of vulnerability to health risks from temperature for England using small-area data. While cold-related excess mortality over the period 2000–2019 was substantially higher than heat-related excess mortality, they found that there was increased risk of heat-related mortality in more socio-economically deprived and urban areas. They found an increased risk of cold-related mortality in northern regions of England. Using data from New York City, Lin et al. (2009) found that extreme high temperatures increase hospital admissions for cardiovascular and respiratory disorders, with older and Hispanic residents particularly vulnerable to the temperature effects on respiratory illnesses. 1.3.2 Irish literature on temperature change and health Irish studies examining the link between temperature change and health 4 are similar in methodology to the European studies mentioned in the previous section, although all of the Irish studies summarised in this section use mortality as the outcome of interest. The Irish literature shows that a broader analysis of both hot and cold weather is necessary, as both are shown to have significant associations with mortality. There are two pan-European studies in which data from Ireland is included. Healy, 2003 used a multi-country analysis using 14 European countries to examine the effect of cold weather on mortality. Baccini et al. (2008) used a similar approach but with multiple European cities and examined the effect of heat on mortality. Healy (2003) found that Ireland had the third highest rates of excess winter mortality after Portugal and Spain for the period 1988 to 1997. Baccini et al. (2008) found that there was no significant association between heat and mortality for Dublin between 1990 and 2000. Another study, Pascal et al. (2013), examined the mortality effects of five heatwaves in Ireland from 1981 to 2006. Overall, they found that 294 excess deaths were attributed across all of the heatwaves. The authors highlight the urban-rural divide showing that during the summer months, 4 While not examined in this report, an emerging literature is examining the link between other aspects of climate change and health in Ireland (see for example, Musacchio et al. (2021) who quantify the capacity of private well users in Ireland to cope with flood-triggered contamination risks). Background | 16 have major health benefits. Compared with the reference scenario, they project that adoption of global dietary guidelines would result in 5.1 million avoided deaths per year and 79 million years of life saved. The equivalent figures for the vegetarian diet are 7.3 million avoided deaths and 114 million life years saved, and for the vegan diet 8.1 million avoided deaths and 129 million life years saved. Milner et al. (2015) estimate that if the average UK dietary intake were optimised to comply with the WHO recommendations 9 , there could be an incidental reduction of 17 per cent in GHG emissions. Adherence to such a diet could save almost 7 million years of life lost prematurely in the UK over a 30-year period and increase average life expectancy by over 8 months. Focusing on carbon taxation, Vandenberghe and Albrecht (2018) simulate three carbon tax scenarios in the energy and food sector in Belgium and assess the resulting health-related co-benefits (i.e., reduction in PM air pollution, and consumption of animal products). They find that the carbon tax could prevent 42,300–78,800 DALYs in Belgium, or save 0.6–1.1 per cent of total healthcare expenditure and an additional 0.06–0.12 per cent of Belgian GDP. Many of the above studies rely on data and estimates from the epidemiological literature that examine the impact of behaviours that are consistent with climate change mitigation actions, such as increased active travel and more plant-based diets, on health outcomes (for summaries, see Saunders et al., 2013; Godfray et al., 2018). These estimates are then used to simulate the impact of specific climate change mitigation actions on health (for example, see Shaw et al., (2011) for active travel and Farchi et al. (2017) for diet). However, establishing causality in the underlying relationships is difficult (e.g., does active travel lead to better health outcomes, or are those in better health more likely to use active travel?) (Kroesen and De Vos, 2020). These estimates tend not to take into account the environmental and health consequences (both positive and negative) of broader behavioural changes (e.g., a switch from meat- to plant-based diets is likely to lead to increases in the consumption of other food groups such as nuts and seeds) (Springmann et al., 2016). In addition, the timing of health co-benefits is likely to differ, making quantification of co-benefits into the future difficult. For example, benefits from climate change mitigation actions include likely immediate reductions in acute respiratory infections in children from decreases in air pollution (particularly in low-income countries), short-term and medium-term reductions in cardiovascular disease incidence and mortality that might occur over a period of years, and reductions in cancer incidence and mortality related to obesity that might take place over decades (Haines et al., 2009). 9 In order to conform to the WHO nutritional recommendations, the UK diet would need to contain less red meat, dairy products, eggs and sweet and savoury snacks, but more cereals, fruit and vegetables. Background | 17 1.6 REPORT STRUCTURE The remainder of the report is structured as follows: in Chapter 2, we provide an overview of climate modelling and the climate change projection process and scenarios used throughout this report; in Chapter 3 we outline the results of our analysis of the effects of temperature changes over the period 2015–2019 on emergency in-patient hospital admissions; in Chapter 4 we examine the health benefits and co-benefits of climate change mitigation measures on mortality, morbidity and selected economic outcomes. We also illustrate the potential effects on morbidity using a simulation that predicts the impacts on emergency in-patient hospital admissions of different temperature paths for Ireland out to 2100. Chapter 5 summarises and discusses the results. Climate change modelling | 18 CHAPTER 2 Climate change modelling 2.1 INTRODUCTION The Earth’s climate system is highly complex and involves a multitude of interaction mechanisms. To understand the Earth’s climate system, climate models have been developed and improved significantly over the past decades. Climate models also give insights into the evolution of the climate and enable the prediction of future changes to the climate. The Earth’s climate system consists of the atmosphere (i.e., the layers of gases that envelop the Earth), the hydrosphere (i.e., water), the cryosphere (i.e., ice and snow), the land surface (i.e., soil and rocks), and the biosphere (i.e., animals and plants), all influenced by various external forcing mechanisms such as solar and orbital variations (IPCC, 2013). Climate change modelling is an essential tool for understanding the Earth’s climate system and hence how it will impact society. This chapter provides an overview of climate change modelling, including global and regional climate models, greenhouse gas emissions (GHG) scenarios, and Irish climate projections based on simulations performed by the Irish Centre for High-End Computing (ICHEC). 2.2 GLOBAL AND REGIONAL CLIMATE MODELS Global Climate Models (GCMs) are used to understand past climate variations, reproduce historical climate patterns, and predict the characteristics of the future climate. They are complex mathematical representations of the components of the Earth’s climate system (atmosphere, cryosphere, hydrosphere, land surface, and biosphere) and their interactions. By representing the Earth’s climate system as a set of mathematical equations that are numerically solved using established mathematical techniques, GCMs capture the interactions between the various components and their behaviour over time, enabling climate scientists to study the long-term impacts on the climate (IPCC, 2013; Auffhammer, 2018). GCMs require a large amount of computational power to solve the complex equations that control the climate system. Using supercomputers, they divide the Earth into a three-dimensional grid and calculate interactions between grid cells at different latitudes and locations. These calculations are performed over short periods of time, or “time steps”, allowing the model to mimic weather phenomena or climatic events on timescales ranging from hours to hundreds of years (IPCC, 2013). While GCMs have greatly improved our understanding of the Earth’s climate system and its response to external influences such as solar variations and GHG emissions, they are not without flaws. Model parameters are prone to uncertainty, small-scale processes cannot be fully resolved, and it is challenging to precisely represent feedback mechanisms, cloud formation, and regional climate dynamics. Climate change modelling | 19 Climate scientists continue to refine and validate these models using observational data. Also, multiple models or ensemble simulations are used to account for uncertainties and present a range of potential future climate scenarios. The Coupled Model Intercomparison Project (CMIP) is paramount in this regard, as it brings together climate modelling groups from around the world to compare and analyse climate model simulation results. Like GCMs, regional climate models (RCMs) are used to simulate and project future climate conditions. However, RCMs provide more detailed information on the climate in a particular region, such as a country, a town, or a river basin. They can capture regional features, topography, and local climate drivers that may not be adequately resolved in GCMs. In other words, the primary distinction between RCMs and GCMs lies in their spatial resolution and the scope of the areas they cover. RCMs often use the outputs from larger-scale GCMs as initial and boundary conditions. That is, they take the coarse-scale information provided by GCMs and refine it to provide more localised climate projections. This downscaling makes it possible to provide climate data at scales that are important for regional planning, impact assessments, and decision-making. It is important to note that RCMs have the same limitations and uncertainties as GCMs, including parametrisation and model biases. Moreover, the quality of RCM projections is influenced by how accurately GCMs supply the boundary conditions (IPCC, 2013). 2.3 GREENHOUSE GAS EMISSIONS SCENARIOS Human behaviour continues to increase the level of GHG emissions. GHG accumulates in the atmosphere, reflecting the outgoing radiation from the sun back to the Earth, essentially trapping the sun’s heat. As GHG emissions are the main driver of observed changes in our climate, understanding the future path of GHG emissions is essential to estimate future climate change. Given the uncertainty of what future GHG emissions will be, different scenarios have been developed by the international community of climate researchers. These scenarios of GHG emissions are an integral part of climate change modelling and are useful for several purposes, including understanding and predicting future climate change. They help in establishing a connection between atmospheric GHG concentrations and changes in global temperature and other climate variables. By simulating various emissions scenarios in climate models, climate scientists can assess the climate system’s sensitivity to various amounts of greenhouse gases. To ensure consistency across research applying future climate change scenarios, the IPCC developed a Special Report on Emissions (SRE) with concomitant scenarios (SRES) in 2000. These scenarios were replaced by the Representative Concentration Pathway (RCP) scenarios for the IPCC fifth Assessment Report (AR5) in 2014. RCPs represent the different future trajectories of GHG concentrations in the atmosphere based on a wide range of assumptions regarding population growth, economic development, technological innovation and attitudes to social and environmental sustainability (IPCC, 2014). For instance, all RCPs include the Climate change modelling | 20 assumption that air pollution control becomes more stringent over time as a result of rising income levels (van Vuuren et al., 2011). There are four main RCPs with numerical values 2.6, 4.5, 6.0, and 8.5. These numbers represent the radiative forcing (i.e., the difference between the incoming and outgoing energy from the sun) values in the year 2100. The four RCPs comprise a mitigation scenario (RCP2.6) that results in a very low forcing level, two stabilisation scenarios (RCP4.5 and RCP6.0), and a scenario (RCP8.5) that has extremely high GHG emissions. In other words, RCP2.6 represents a pathway where GHG emissions are significantly reduced, leading to an estimated 1.6°C increase in global average temperature by 2100 relative to the pre-industrial period (1850–1900). In what follows, we will refer to this as the “Paris Agreement” Scenario. This is an ambitious interpretation of the Paris Agreement, reflecting the goals of the agreement, and does not refer to the current pledges under the Paris Agreement which would result in significantly higher concentrations. RCP8.5 is a pathway where GHG emissions continue to grow unmitigated, resulting in a best estimate global average temperature rise of 4.3°C by 2100. We will refer to this pathway as the “no mitigation” pathway, acknowledging that this is an extreme interpretation of no climate action and refers to a worst case scenario. RCP4.5 and RCP6.0 are two medium stabilisation pathways, with varying levels of mitigation (Met Office, 2018). RCP4.5 is referred to as the “most likely” scenario in the context of this report. The increase in global mean temperature predicted by the RCP pathways for the late 21st century is shown in Table 2.1. RCPs and Shared Socio-economic Pathways (SSPs) are conceptually related, which is crucial to highlight. SSPs are a set of scenarios aimed at providing a consistent framework for examining the interaction between socio-economic development and climate change. There are five SSPs, and they represent the range of possible futures based on different assumptions of future societal trends. See below for a summary of SSP narratives (Riahi et al., 2017). SSP1 (Sustainability – taking the green road) This represents a future characterised by gradual but pervasive shifts towards a more sustainable path, emphasising more inclusive development that respects perceived environmental boundaries. This implies a world with a balanced use of resources and a transition to renewable energy sources. SSP2 (Middle of the road) This represents a future that does not differ significantly from historical trends in the areas of social, economic, and technological development. This means a world where societal and environmental policies evolve slowly, with some improvements but also persistent challenges. SSP3 (Regional rivalry – a rocky road) This represents a future marked by resurgent nationalism, concerns about competitiveness and security, and regional conflicts that push countries to increasingly focus on domestic or, at most, regional issues. This means a world with growing inequalities, regional conflicts, and uneven access to resources. Climate change modelling | 21 SSP4 (Inequality – a divided road) This represents a future characterised by highly unequal investments in human capital, combined with increasing disparities in economic opportunity and power with the resultant consequences being increasing inequalities and stratification both across and within countries. This means a world with significant social disparities, resource depletion, and environmental degradation. SSP5 (Fossil-fuelled development – taking the highway) This represents a future where the push for economic and social development is coupled with the exploitation of abundant fossil fuel resources and the adoption of resource- and energy-intensive lifestyles. This means a world heavily reliant on fossil fuels, with high GHG emissions and limited climate change mitigation efforts. The previous IPCC SRES scenarios were developed in a sequential fashion where the socio-economic assumptions were translated into resulting radiative forcing and temperature change. In the new process, the RCPs and SSPs were developed separately and later linked. This has led to a lack of a clear connection between the SSPs and RCPs, were an SSP could be linked to different RCPs. Table 2.1 displays the SSPs linked to the four main RCP scenarios and the resulting estimated temperature change. TABLE 2.1 THE INCREASE IN GLOBAL MEAN TEMPERATURE COMPARED TO PREINDUSTRIAL LEVEL SSP Scenario RCP Scenario Change in temperature (oC) by 2081–2100 SSP 1 RCP2.6 “Paris Agreement” 1.6 (0.9 to 2.3) SSP 2 RCP4.5 “most likely” 2.4 (1.7 to 3.2) RCP6.0 2.8 (2.0 to 3.7) SSP 5 RCP8.5 “no mitigation” 4.3 (3.2 to 5.4) Source: Met Office (2018), based on Table 12.3 of IPCC AR5 Working Group One. Notes: Numbers in parentheses indicate the likely range. 2.4 FUTURE CLIMATE PROJECTIONS FOR IRELAND The approach of regional climate modelling was employed by climate researchers at the ICHEC to accurately model the Irish climate, capture its distinctive features, and provide high spatial resolution (with grids of 4km2) climate information for the evaluation of the local effects of climate change. Two RCMs, the Consortium for Small-scale Modelling-Climate Limited-area Modelling (COSMO-CLM) and the Weather Research and Forecasting (WRF), were used to downscale five CMIP– Project 5 (CMIP5) GCM datasets: CNRM-CM5, EC-EARTH (four ensemble members), HadGEM2-ES, MIROC5, and MPI-ESM-LR. The simulations were run for the base period 1981–2000 and the future period 2041–2060. The difference between the two periods provides a measure of climate change. To account for the uncertainty in future GHG emissions, the future climate was simulated under both Climate change modelling | 22 the RCP4.5 (“most likely”) and RCP8.5 (“no mitigation”) scenarios (Nolan and Flanagan, 2020). According to their simulation analysis, Ireland will be exposed to higher mean temperatures in the future. The mean annual temperature is projected to increase by 1–1.6°C by the middle of the century (i.e., 2041–2060) compared to the reference period 1981–2000, under the RCP4.5 (“most likely”) climate scenario. With this warming will come hotter days and nights. In comparison to the baseline period, the warmest 5 per cent of daily maximum temperatures are projected to increase by 1–2.2°C while the coldest 5 per cent of daily minimum temperatures are projected to rise by 1–2.4°C. However, these projections show variations in temperature across the country that are expected to be marked by increased temperatures in eastern regions. For instance, Figure 1 demonstrates that the annual maximum temperature in Dublin County has been rising between 1961 and 2021. The ICHEC simulations show that such an upward trend is more likely to persist, especially in the high-risk scenario implied by the RCP8.5 (“no mitigation”) emission trajectory. Note that the yellow line has a considerable degree of variability because it represents the yearly maximum temperature that was observed in Dublin County. FIGURE 2.1 DUBLIN COUNTY HISTORICAL AND FUTURE ANNUAL MAXIMUM TEMPERATURE Source: Met Éireann gridded weather dataset for historical observations (1961–2021), ICHEC simulations for future projections (Nolan and Flanagan, 2020). Authors’ calculation for aggregation at the Dublin County level. Notes: Yellow line represents the observed annual maximum temperature in Dublin County. The orange and red lines give the projections for the annual maximum temperature in Dublin County for RCP4.5 (orange) and RCP8.5 (red), averaged over a 10-year rolling window among the five GCMs considered in ICHEC simulations. Climate change modelling | 23 The projected annual and seasonal temperatures for Ireland are shown in Figures 2.2 and 2.3, respectively. It should be noted that in each figure and scenario, the future period, 2041–2060, is compared with the 1981–2000 period. Also, the numbers included in each plot are the minimum and maximum projected changes, displayed at their locations. FIGURE 2.2 PROJECTIONS OF TEMPERATURE CHANGE FOR RCP4.5 “MOST LIKELY” (A) AND RCP8.5 “NO MITIGATION” (B) SCENARIOS Source: Nolan and Flanagan (2020) Climate change modelling | 24 FIGURE 2.3 MID-CENTURY SEASONAL PROJECTIONS OF TEMPERATURE CHANGE FOR RCP4.5 “MOST LIKELY” (A) AND RCP8.5 “NO MITIGATION” (B) SCENARIOS Source: Nolan and Flanagan (2020) Heatwaves are also predicted to become more frequent by the middle of the century, with the southeast experiencing the biggest increases. Table 2.2 presents the projected increases over the 20-year period 2041–2060 for the “mostly likely” and “no mitigation” scenarios. TABLE 2.2 PROJECTED INCREASE IN THE NUMBER OF HEATWAVE EVENTS OVER THE PERIOD 2041–2060 Scenario Projected increase RCP4.5 (“most likely”) 1 to 8 RCP8.5 (“no mitigation”) 3 to 15 Source: Nolan and Flanagan (2020) In the RCP4.5 (“most likely”) and RCP8.5 (“no mitigation”) scenarios, the number of “frost days”, or days with a minimum temperature below 0oC, is expected to decline by 45 per cent and 58 per cent, respectively. Additionally, under the RCP4.5 (“most likely”) and RCP8.5 (“no mitigation”) scenarios, respectively, it is predicted that the proportion of ice days (days with a maximum temperature colder than 0oC) will decline by 68 per cent and 78 per cent. For precipitation, a substantial decrease is projected for the summer months, although non-summer months are projected to record marginal changes. Overall, it is projected that precipitation will Climate change modelling | 25 display more variability by the middle of this century as a result of an increasing frequency of droughts and heavy rainfall events. More so, the projected increase in heatwaves will directly impact public health and mortality, but this may be offset by the projected decrease in frost and ice days (Nolan and Flanagan, 2020). 2.5 SUMMARY This chapter gave an overview of the climate modelling behind the climate projections used in this report. Projecting future climate change and social and economic developments remains a complex task with high levels of uncertainty. Applying an Irish-specific climate model that regionalises global projections is the most robust way to approach projections of the future climate for Ireland. In the following analysis, we focus on three RCP scenarios, namely the RCP8.5 (“no mitigation”), RCP4.5 (“most likely”), and the RCP2.6 (“Paris Agreement”). Health effects of temperature change | 32 The dependent variable, 𝑌𝑐,𝑤,𝑦, represents the emergency in-patient admission rate per 100,000 population in a particular county, week and year. The temperature variable, 𝑡𝑒𝑚𝑝, is characterised in a number of different ways in order to test various specifications for the relationship between temperature and emergency hospital admissions (described in greater detail below). Lagged temperature variable is also included in the specification because there can be residual effects on hospital admissions from temperature in previous weeks. 13 Specifically, 𝑡𝑒𝑚𝑝𝑐,𝑤,𝑦 represents the indicator variable for temperatures in bin j for the current week. Similarly, 𝑡𝑒𝑚𝑝𝑐,𝑤−𝑘,𝑦 represents the indicator variable for temperatures in bin j for the k-th lagged week (or the previous week), and 𝑡𝑒𝑚𝑝𝑐,𝑤−𝑙,𝑦 represents the indicator variable for temperatures in bin j for the l-th lagged week (or the two previous weeks). The coefficients associated with these variables are represented by 𝛽𝑗, 𝜙𝑗,𝑘, and 𝛿𝑗,𝑙 respectively. In addition, the diagnosis categories are represented by 𝜂𝑐,𝑤,𝑦, year trend by 𝜏, county fixed effect by 𝜑, and month fixed effect by 𝑚. The year trend and month fixed effects are included to control for annual and seasonal factors respectively. Rainfall is used to account for the effects of humidity along with its lagged values, as is done in previous studies (White, 2017). However, for brevity, it has been excluded from equation (4.1) but was controlled for during the estimation process. We include a number of control variables to account for the casemix of patients, socio-economic status, and case severity which may differ across hospital, years, and weeks within a given year. The vector, 𝑿𝑐,𝑤,𝑦, denotes the socio-demographic variables that we control for in our regression analysis. These controls variables capture the mean composition of admitted patients by age and sex for each county, week and year, as well as the mean marital status composition, mean proportion of patients that are admitted as private patients, and the mean proportion of patients with a medical card. Additionally, the variable, 𝐶𝐼𝑐,𝑤,𝑦, is the mean Charlson co-morbidity index taken for each county, week and year. The Charlson co-morbidity index is a score based on all of a patient’s diagnoses (using ICD codes) and serves as a predictor of mortality risk within a year following hospitalisation (Charlson et al., 1987). 3.3.3 Model specification 2: Mean deviation model The previous temperature bins specification examines all months and explores the impact of both colder and hotter weather on emergency in-patient hospitalisations for our patient groups. However, a key question in this study is the implications of increases in hotter days on hospitalisations, as hotter days are likely to become more prevalent due to climate change in the future. Therefore, in this analysis, we focus only on quarter 2 (April, May, June) and quarter 3 (July, August, September), the periods where hot weather in Ireland occurs. 13 This is common practice in similar literature and the standard number of lags is 30 days or three weeks (Hajat et al., 2006; Baccini et al., 2008; Breitner et al., 2014; White, 2017; Gibney et al., 2022). Health effects of temperature change | 33 In this specification, we construct the dependent and temperature variable of interest using a methodology that estimates deviations from a multi-year (2015– 2019) mean. For each county, week, and year (denoted as 𝑐,𝑤,𝑦), the number of admissions (𝑎𝑑𝑚𝑐,𝑤,𝑦) and the temperature (𝑡𝑒𝑚𝑝𝑐,𝑤,𝑦) are compared against the respective five-year averages (𝑎𝑑𝑚       𝑐,𝑤 for admissions and 𝑡𝑒𝑚𝑝        𝑐,𝑤 for temperature): 𝑑𝑖𝑓𝑓(𝑎𝑑𝑚)𝑐,𝑤,𝑦 =𝑎𝑑𝑚𝑐,𝑤,𝑦−𝑎𝑑𝑚        𝑐,𝑤 𝑎𝑑𝑚        𝑐,𝑤 , where 𝑎𝑑𝑚𝑐,𝑤,𝑦 is the number of emergency in-patient hospital admissions by county, week and year. 𝑎𝑑𝑚       𝑐,𝑤,𝑦 represents the average number of admissions per county and week over the five-year span from 2015 to 2019. This approach allows us to measure the admissions for a specific county, week, and year as a proportion of its five-year average. A similar approach is applied to the temperature variable of interest: 𝑑𝑖𝑓𝑓(𝑡𝑒𝑚𝑝)𝑐,𝑤,𝑦 =𝑡𝑒𝑚𝑝𝑐,𝑤,𝑦−𝑡𝑒𝑚𝑝         𝑐,𝑤 𝑡𝑒𝑚𝑝         𝑐,𝑤 , where 𝑡𝑒𝑚𝑝𝑐,𝑤,𝑦 represents the average temperature for a county, week, year (in °C) and 𝑡𝑒𝑚𝑝        𝑐,𝑤 represents the average temperature per county and week over the five-year span from 2015 to 2019. Table 3.2 provides a numerical illustration of this process, using week 27 in North Dublin as an example. A mean level of admissions is calculated for each week across the five years (column 3), indicated above by the 𝑎𝑑𝑚       𝑐,𝑤 variable. A difference is taken between the actual level of admissions (𝑎𝑑𝑚𝑐,𝑤,𝑦) for that week and the five-year mean (column 4). This difference is then calculated as a percentage (column 5). This is done for each county, week, and year observation for both hospital admissions and maximum temperature. Health effects of temperature change | 34 TABLE 3.2 ILLUSTRATION OF MEAN DEVIATION ADMISSIONS VARIABLE Year Total Admissions Mean Admissions, 2015–2019 Difference % Difference 2015 252 274.46 -22.46 -8.18 2016 265 274.46 -9.46 -3.45 2017 258 274.46 -16.46 -6.00 2018 283 274.46 8.54 3.11 2019 307 274.46 32.54 11.86 Source: Authors’ analysis. The specification is as follows: 𝑑𝑖𝑓𝑓(𝑎𝑑𝑚)𝑐,𝑤,𝑦=𝛼+𝛽𝑗𝑑𝑖𝑓𝑓(𝑡𝑒𝑚𝑝)𝑐,𝑤,𝑦+𝜙𝑗𝑑𝑖𝑓𝑓(𝑡𝑒𝑚𝑝)2𝑐,𝑤,𝑦+ 𝑿𝑐,𝑤,𝑦 +𝐶𝐼𝑐,𝑤,𝑦+𝜂𝑐,𝑤,𝑦+𝜏+𝜑+𝑚+𝜀𝑐,𝑤,𝑦 (3.2) where c represents the county of residence, w represents the week, and y the year. The dependent variable, 𝑎𝑑𝑚𝑐,𝑤,𝑦, represents the number of admissions compared against the respective five-year average (𝑎𝑑𝑚       𝑐,𝑤). 𝑡𝑒𝑚𝑝𝑐,𝑤,𝑦 represents the number of admissions compared against the respective five-year average (𝑡𝑒𝑚𝑝        𝑐,𝑤). Once more, the vector, 𝑿𝑐,𝑤,𝑦 denotes the socio-demographic and casemix variables that we control for in our regression analysis, and 𝐶𝐼𝑐,𝑤,𝑦 is the mean Charlson co-morbidity index taken for each county, week and year. Diagnosis categories are represented by 𝜂𝑐,𝑤,𝑦, year trend by 𝜏, county fixed effect by 𝜑, and month fixed effect by 𝑚. The year trend and month fixed effects are included to control for annual and seasonal factors respectively. Rainfall is used to account for the effects of humidity along with its lagged values. 3.4 RESULTS Before presenting the results for the main model specifications (temperature bins and mean deviation), in Table 3.3 we show how emergency in-patient hospital admissions (per 100,000 population) varied over the period 2015–2019. This table shows the average weekly emergency hospital admissions rate across counties for each year-quarter. Health effects of temperature change | 35 TABLE 3.3 EMERGENCY IN-PATIENT HOSPITAL ADMISSIONS (PER 100,000 POPULATION) 2015–2019 2015 2016 2017 2018 2019 Quarter 1 57.8 58.6 58.5 64.9 64.9 Quarter 2 53.3 56.3 58.3 59.2 59.8 Quarter 3 48.3 52.3 53.6 55.0 55.4 Quarter 4 53.3 61.0 61.4 62.1 59.9 Source: Authors’ analysis. Figure A.1 in Appendix A shows how emergency hospital in-patient admissions (per 100,000 population) vary by county of residence (using 2019 as an example). The data show that admissions are highest in the west and midlands. This could be due to an ageing population in those areas. 3.4.1 Temperature bin analysis Table 3.4 shows the results from the temperature bin analysis. The reference category used is the temperature bin 10–13°C. This temperature bin has the greatest proportion of week-county observations. All other coefficients are interpreted with respect to this temperature bin. These coefficients are also plotted in Figure 3.2. The results show that for temperatures greater than 16°C, there are greater rates of emergency hospital admissions. There is a statistically significant relationship between temperatures of 16–25°C and emergency hospital admissions. The coefficient plot illustrates how the coefficient gets larger at higher temperatures. The two highest temperature bins show no statistically significant relationship, which is likely due to two reasons: there are much fewer observations in these temperature bins (see also Table 3.1), and at higher temperatures, individuals are more likely to take precautionary measures that protect them from high temperatures (e.g., staying indoors during the hottest period of the day). The coefficients in Table 3.4 can be interpreted as changes per 100,000 admissions. For example, at temperatures between 22°C and 25°C, there was an increase in emergency hospital admissions of 4.71 per 100,000 population compared to when temperatures are in the reference category (10–13°C). These coefficients can also be interpreted as percentage changes if they are interpreted with respect to the mean admissions rate (55.61 per 100,000 population). In this case, there is an increase in emergency admissions of 8.5 per cent 14 when the temperature is between 22°C and 25°C compared to 10–13°C. Full model results can be found in Appendix B. 14 Calculated as a proportion of the weekly mean admissions rate: ((4.71/55.61)*100). Health effects of temperature change | 36 TABLE 3.4 TEMPERATURE BIN ANALYSIS Admissions Rate Temperature Bin: 1–4°C -5.17** (2.28) Temperature Bin: 4–7°C 1.73** (0.83) Temperature Bin: 7–10°C -1.93*** (0.47) Temperature Bin: 10–13°C Ref. - Temperature Bin: 13–16°C 0.91 (0.58) Temperature Bin: 16–19°C 1.97** (0.78) Temperature Bin: 19–22°C 3.29*** (0.91) Temperature Bin: 22–25°C 4.71*** (1.30) Temperature Bin: 25–28°C 3.94 (2.75) Temperature Bin: 28°C+ 2.78 (3.76) County Fixed Effects Yes Year Trend Yes Month Fixed Effects Yes Lagged Temperature Bins Yes Socio-Demographic Controls Yes Mean dependent variable 55.61 N 6,88915 R2 0.56 Source: Authors’ analysis. Notes: Standard errors in parentheses. * p<0.05, ** p<0.01, *** p<0.001 15 In order to calculate an admissions rate, North Dublin and South Dublin were aggregated – therefore, the final sample size is 6,889 (5 years x 53 weeks x 26 counties – 1 week Carlow 2017). Health effects of temperature change | 37 FIGURE 3.2 COEFFICIENT PLOT OF TEMPERATURE BINS Source: Authors’ analysis. The analysis was also carried out separately for each age group; the results are presented in Table 3.5. Results show that the effect of temperature has a larger impact on hospitalisation for children (0–14) compared to other age groups. A statistically significant effect of temperature emerges above 16°C, similar to the main analysis. At temperatures 22–25oC there was an increase in emergency hospital admissions of 7.02 per 100,000 population compared to when temperatures are in the reference category. This equates to an increase in emergency admissions of 12.2 16 per cent among children. For the working age group (15–64), the effects are smaller in magnitude, and similarly significant only between 16 and 25 degrees. At temperatures 22–25oC there was a 7.8 17 per cent increase in emergency hospital admissions compared to when temperatures are in the reference category (10–13°C). Hospital admissions for the older (65+) age group exhibit a similar pattern to the results in Table 3.4 although none of the effects (with the exception of temperatures between 4–7oC) are statistically significant. While the results may indicate no effect of temperature change on emergency in-patient hospitalisations for the older population over the period, these results could also indicate adaptive behaviour on the part of older 16 (7.02/57.64)*100 17 (2.14/24.45)*100 Health effects of temperature change | 38 people to higher temperatures in Ireland, an effect that may kick in at lower temperatures than for other population groups. TABLE 3.5 TEMPERATURE BIN ANALYSIS – AGE GROUP ANALYSIS Children (0–14) Working Age Group (15–64) Older Adults (65+) Temperature Bin: 1–4°C -6.79 (4.27) -1.80 (1.63)) -13.20 (9.32) Temperature Bin: 4–7°C 1.03 (1.55) -0.05 (0.59) 17.58*** (3.27) Temperature Bin: 7–10°C -2.15** (0.89) -1.16*** (0.34) -0.26 (1.82) Temperature Bin: 10–13°C Ref. - Ref. - Ref. - Temperature Bin: 13–16°C 1.75 (1.10) 0.24 (0.42) 1.73 (2.32) Temperature Bin: 16–19°C 3.41** (1.46) 1.15** (0.56) -3.19 (3.03) Temperature Bin: 19–22°C 6.26*** (1.72) 1.52** (0.66) -1.93 (3.52) Temperature Bin: 22–25°C 7.02*** (2.43) 2.14** (0.93) 3.25 (4.88) Temperature Bin: 25–28°C 0.26 (5.25) 0.60 (2.01) 11.47 (10.57)7) Temperature Bin: 28°C+ -5.29 (7.26) 2.71 (2.78) -1.52 (15.37) County Fixed Effects Yes Yes Yes Year Trend Yes Yes Yes Month Fixed Effects Yes Yes Yes Lagged Temperature Bins Yes Yes Yes Socio- Demographic Controls Yes Yes Yes Mean Dependent Variable 57.64 27.45 179.45 N1 6,853 6,886 6,884 R2 0.41 0.34 0.51 Source: Authors’ analysis. Notes: Standard errors in parentheses. 1 The sample sizes differ across specifications because some county-week-year units will have no HIPE observations for that age group. * p<0.05, ** p<0.01, *** p<0.001 Health effects of temperature change | 39 This analysis is also carried out for each of the five diagnosis groups and Table 3.6 shows these results. The results for the diagnosis analysis show very similar patterns for the whole temperature bin analysis, except that there are also statistically significant results for the coefficients at lower temperatures (apart from metabolic diseases, which are significant only at lower temperatures). At temperatures between 1–4°C, it can be seen that there is a decrease in the emergency hospital admissions rate across all diagnostic groups. As pointed out by previous literature, this is likely a behavioural effect where the very cold weather prevents people from seeking medical care (Gibney et al., 2022). For all of the chapters on diseases (except metabolic diseases), we can see that there are statistically significant effects of both cold and warm temperatures on emergency hospital admissions. At lower temperatures, there is a decrease in emergency hospital admissions. This is consistent with the literature in this field which posits that at cold temperatures, there is an avoidant health-seeking behaviour as people do not want to venture out in unsafe conditions (Gibney et al., 2022). At warmer temperatures, certain health conditions tend to be exacerbated such as circulatory, respiratory and infectious diseases as well as injuries. Our analysis shows that emergency admissions due to metabolic diseases are not statistically significantly responsive to cold temperatures. However, they are statistically significantly responsive to warmer temperatures. Some of the international literature also posits that the effect of rising temperatures on infectious diseases (including E.coli VTEC) may be due to contaminated water sources, where the warm weather allows bacteria to live longer and enter drinking water streams (Romanello et al., 2022). Health effects of temperature change | 40 TABLE 3.6 TEMPERATURE BIN ANALYSIS – DIAGNOSTIC GROUP ANALYSIS (1) Circulatory Disease (2) Respiratory Disease (3) Metabolic Diseases (4) Infectious Diseases (5) Injuries Temperature Bin: 1–4°C -5.34** (2.25) -5.22** (2.29) -4.15* (2.21) -5.38** (2.23) -5.07** (2.25) Temperature Bin: 4–7°C 1.91** (0.82) 1.76** (0.83) 1.03 (0.83) 2.01** (0.80) 1.46* (0.82) Temperature Bin: 7–10°C -1.65*** (0.46) -1.87*** (0.47) -0.91** (0.45) -1.28*** (0.45) -1.65*** (0.46) Temperature Bin: 10–13°C Ref. - Ref. - Ref. - Ref. - Ref. - Temperature Bin: 13–16°C 0.87 (0.58) 0.79 (0.59) 0.89 (0.56) 0.76 (0.57) 0.82 (0.58) Temperature Bin: 16–19°C 1.81** (0.77) 1.67** (0.78) 1.86** (0.76) 1.74** (0.76) 1.79** (0.77) Temperature Bin: 19–22°C 3.05*** (0.90) 2.86*** (0.92) 2.72*** (0.88) 2.96*** (0.89) 3.23*** (0.91) Temperature Bin: 22–25°C 4.44*** (1.27) 4.36*** (1.30) 4.42*** (1.25) 4.42*** (1.25) 4.64*** (1.27) Temperature Bin: 25–28°C 3.89 (2.71) 3.89 (2.76) 6.22** (2.73) 4.71* (2.71) 3.72 (2.72) Temperature Bin: 28°C+ 1.69 (3.71) 1.88 (3.76) 2.44 (3.61) 1.66 (3.61) 1.69 (3.71) County Fixed Effects Yes Yes Yes Yes Yes Year Trend Yes Yes Yes Yes Yes Month Fixed Effects Yes Yes Yes Yes Yes Lagged Temperature Bins Yes Yes Yes Yes Yes Socio-demographic controls Yes Yes Yes Yes Yes Mean dependent variable 56.30 56.22 57.63 56.54 56.28 N1 6,869 6,887 5,969 6,743 6,874 R2 0.57 0.56 0.61 0.58 0.57 Source: Authors’ analysis. Notes: Standard errors in parentheses. 1 The sample size differs by diagnosis group because for some county-week-year units there were no HIPE observations for certain diagnosis groups. * p<0.05, ** p<0.01, *** p<0.001 Health effects of temperature change | 41 3.4.2 Mean deviation model Table 3.7 presents the results from the mean deviation model, which is estimated for the months April–September (quarters 2 and 3) only. The effect size of the differences model is 0.18, which means that a 1 percentage point deviation in maximum weekly temperature from the mean induces a 0.18 percentage point deviation in weekly hospital admissions from the mean. TABLE 3.7 MEAN DEVIATION MODEL, QUARTERS 2 AND 3 %Δ In-Patient Hospital Admissions % Difference in Maximum Temperature 0.18*** (0.03) % Difference in Maximum Temperature2 -0.64*** (0.18) Temperature Lag 1 -0.04 (0.03) Temperature Lag 2 -0.01 (0.03) County Fixed Effects Yes Year Trend Yes Month Fixed Effects Yes Diagnosis Fixed Effects Yes Socio-demographic Controls Yes N 3,518 R2 0.11 Source: Authors’ analysis. Notes: Standard errors in parentheses. * p<0.05, ** p<0.01, *** p<0.001 Table 3.7 clearly shows that temperature increases are related to higher rates of hospitalisation. To provide a more intuitive illustration of findings to highlight the size of this effect, Table 3.8 applies the results from Table 3.7 to the first week of July in North Dublin as a linear predictor. In this county-week observation, the mean temperature across the five-year period is 19.3°C and the mean number of weekly admissions is 278. We estimate the predicted change in emergency in-patient hospitalisations for 1°C, 5°C, and 10°C increase in temperature. It is clear that for very large changes, a number of additional admissions (23, or 8.3% per admissions) results. Health benefits and co-benefits of climate change mitigation measures | 48 TABLE 4.2 HEALTH AND ECONOMIC BENEFITS OF AIR POLLUTION REDUCTIONS (PER ANNUM) Central Estimate Lower Bound Upper Bound Death averted from all (natural) causes in adults (≥ 30 years) 66 50 73 Cardiovascular hospital admissions (all ages) 12 2 21 Respiratory hospital admissions (all ages) 26 0 54 Restricted activity days (all ages) 77,883 72,754 83,012 Lost workdays in the employed population (18–65 years) 17,470 14,863 20,059 Life years gained 659 498 739 Total economic benefit ($)1 214,173,681 98,834,955 336,129,634 Source: CLIMAQ-H (WHO, 2023) Note: 1 Assesses the total economic benefit in $US 2020 prices assuming a 5% discount rate. 4.3 HEALTH EFFECTS OF ACTIONS TO MITIGATE TEMPERATURE CHANGE IN IRELAND: CASE STUDY Chapter 3 highlighted a relationship between temperature fluctuations and emergency in-patient hospital utilisation in Ireland, which can be interpreted as a proxy for morbidity. This raises concerns about how continued climate change will affect population health, the Irish healthcare system and infrastructure in the future, and how different temperature paths (and associated mitigation effort) may lead to greater or lesser impacts on emergency hospitalisation utilisation. In this section, we conduct an illustrative projection analysis to estimate the effect that different scenarios for increased temperatures would have on emergency in-patient hospitalisation in the future. We follow the broad approach outlined above by using various scenarios of future temperature change to simulate the projected impact on emergency in-patient hospitalisations, using parameter estimates from models estimated on the HIPE data used in Chapter 3. Data The data applied in this analysis concerns average daily maximum temperatures, temperature thresholds and projected number of days above these thresholds over the period 2020–2100 for three RCP scenarios. The RCP scenarios have been described in detail in Chapter 2. We apply the RCP4.5 (“most likely”) scenario in this analysis. 19 19 RCP4.5 – intermediate emissions. CO2 emissions increase only slightly before decline commences around 2040 (consistent with full implementation of all current policies). Health benefits and co-benefits of climate change mitigation measures | 49 We apply two key variables within this projection analysis: 1. The average daily maximum temperature per quarter per 20-year projection period up to 2100; 2. The proportion of days per quarter per 20-year period above the 75th, 90th and 95th temperature percentiles. Each of these variables has values for each RCP, each 20-year period, each quarter and each county. This is illustrated in Table 4.3 for the RCP4.5 (“most likely”) scenario. TABLE 4.3 ILLUSTRATION OF TEMPERATURE SCENARIO DATA RCP Scenario Baseline: 1980–2000 Q1 – 4 Counties + Dublin PCs1 RCP4.5 Baseline: 1980–2000 Q1 – 4 Counties + Dublin PCs 2021–2040 Q1 – 4 Counties + Dublin PCs 2041–2060 Q1 – 4 Counties + Dublin PCs 2061–2080 Q1 – 4 Counties + Dublin PCs 2081–2100 Q1 – 4 Counties + Dublin PCs Note: 1 PC is an acronym for post code. It is difficult to explicitly project extreme weather events, and for example the number of very hot days in a given period in the future. In the projection analyses, we include three threshold values that are equivalent to the 75th percentile temperature, the 90th percentile temperature and the 95th percentile temperature. These threshold values are based on the baseline data for the period 1980–2000. These thresholds are used to proxy for very high temperature days in which the temperature exceeded each different threshold in the baseline periods. The data from the RCP scenarios also contain projections of the number of days per quarter per 20-year period above each of these thresholds. Given the proportion of days projected to be above a certain threshold in a quarter, we multiplied this proportion by the number of days in that quarter. The projections of the number of days per quarter per 20-year period above each of these thresholds differ across RCP scenarios. Table 4.4 presents the number of days above the 90th percentile threshold for each quarter and 20-year projection period for the RCP4.5 (“most likely”) scenario. This scenario projects that in quarter 3 in the 2020–2040 and 2040–2060 periods, there will be approximately 22–25 days in which the temperature exceeds the 90th percentile temperature from the 1980–2000 baseline period. This decreases in subsequent periods as climate change mitigation measures are adopted. Health benefits and co-benefits of climate change mitigation measures | 50 TABLE 4.4 NUMBER OF DAYS PER QUARTER ABOVE THE 90TH THRESHOLD FOR THE RCP4.5 (“MOST LIKELY”) SCENARIO 2020–2040 2040–2060 2060–2080 2080–2100 Quarter 1 11.3 10.8 24.3 19.6 Quarter 2 11.9 19.4 18.7 17.1 Quarter 3 22.2 24.9 15.3 21.6 Quarter 4 25.9 16.1 22.9 32.8 Methodology In order to align the climate projections data with the HIPE data used in Chapter 3, the HIPE model was estimated at the county, quarter and year level. Using the temperature data for 2015–2019, it was possible to calculate whether a day was above the 75th, 90th or 95th temperature thresholds. The number of days were aggregated up to the quarter level for each county and year. Table 4.5 provides an illustration of the average temperature threshold for each quarter based on the historical 1980–2000 data. TABLE 4.5 AVERAGE QUARTERLY THRESHOLD VALUES FOR THE 75TH, 90TH AND 95TH THRESHOLDS Mean temperature thresholds by quarter for Ireland 75th Threshold 90th Threshold 95th Threshold Quarter 1 10.80 11.97 12.64 Quarter 2 16.48 18.53 19.94 Quarter 3 19.95 21.85 23.03 Quarter 4 12.56 13.88 14.57 This data was then used to estimate a model (one for each temperature threshold) that regressed the number of days per county per quarter per year above the respective threshold on the rate of emergency in-patient hospital admissions per county per quarter per year. As before, year and county fixed effects were included. The coefficient estimates from this model indicate the effect that one extra day above the threshold in a quarter has on emergency hospital admissions. The model is estimated only for quarter 2 and quarter 3 and is applied only to quarter 2 and quarter 3 projections in this analysis. We do not apply the model estimates to all quarters, as higher than average temperatures in winter months (quarter 4 and quarter 1) would possibly result in fewer emergency hospitalisations. 20 Therefore, our number of observations in the estimated model is 260 (26 counties, 2 quarters and 5 years of data). 20 When the model was estimated for quarter 1 and quarter 4 only, the coefficient (for the number of days per quarter above the relevant threshold) was negative (and not statistically significant). When the model was estimated for quarter 2 and quarter 3 only, the coefficient was positive and statistically significant. Health benefits and co-benefits of climate change mitigation measures | 51 Next, we apply these coefficient estimates to the projected quarterly estimates of the number of days above the respective thresholds (see Table 4.4 for these data for each 20-year projection period for the RCP4.5 (“most likely”) scenario for the 90th percentile). For each temperature threshold, RCP scenario and 20-year projection window, this gives us a projection of the average number of extra hospital admissions in a quarter that we would expect. Finally, we aggregate the quarterly projections to derive an annual estimate of emergency in-patient hospitalisations. This procedure results in 36 possible projections (3 RCP scenarios x 3 temperature thresholds x 4 20-year projection windows). In the discussion that follows, we mainly focus on the results using the 90th percentile threshold, but full results for the RCP4.5 (“most likely”) scenario are available in Appendix E. Results Table 4.6 presents the coefficient estimate results for the analysis; each column is for a different temperature threshold. The dependent variable is the emergency hospital admissions rate (per 100,000 population). The coefficient estimates the effect that one extra day in a quarter above the threshold has on the emergency hospital admissions rate. The model is estimated for quarter 2 and quarter 3 only. TABLE 4.6 EMERGENCY IN-PATIENT MODEL RESULTS, HIPE 2015–2019 75th Percentile Threshold 90th Percentile Threshold 95th Percentile Threshold Coefficient estimate 1.79** (0.52) 2.85*** (0.71) 3.22*** (0.90) Observations 260 260 260 R2 0.91 0.92 0.91 Mean dependent variablea 1,036 1,036 1,036 Note: a relates to emergency in-patient hospitalisations per 100,000 population. Controls include socio-demographic characteristics, medical casemix, county fixed effects, year fixed effects. *p<0.05, **p<0.01, ***p<0.001 Our results above suggest that an additional day in quarter 2 or quarter 3 above the 90th temperature percentile increases the rate of emergency hospital admissions by 2.85 per 100,000 population. We apply the estimate above for the 90th threshold to the number of projected days above the 90th percentile in both quarter 2 and quarter 3. These estimates are also totalled to give an annual estimate. Table 4.7 shows the estimated number of additional emergency hospitalisations per quarter, averaged over the 20-year period. These results give us an indication of what the average quarterly increase in hospital admissions would be under the RCP4.5 (“most likely”) scenario for each projected 20-year period. Our results suggest that there would be an increase of 97.2 per 100,000 each year in the period Health benefits and co-benefits of climate change mitigation measures | 52 2021–2040. This figure increases to 126 per 100,000 in the period 2041–2060. As a proportion of the mean level of admissions, this is indicative of a 9.4 per cent 21 increase in emergency hospital admissions annually in the period 2021–2040. This increases to 12.2 per cent annually for the period 2041–2060. TABLE 4.7 PROJECTED INCREASE IN EMERGENCY IN-PATIENT HOSPITALISATIONS BY 20-YEAR PERIOD (RCP4.5, 90TH PERCENTILE) 2021–2040 2041–2060 2061–2080 2081–2100 Quarter 2 34.0 55.2 53.4 48.8 Quarter 3 63.2 71.1 43.5 61.6 Total 97.2 126.3 97.0 110.4 Climate change adaption scenario It is unlikely that there will be no adaption to a warmer climate undertaken by both individuals and the healthcare system. As such, we use the 95th percentile threshold as an estimate of the lower bound in this projection analysis. It will provide a preliminary indication of what the burden on the healthcare system might be if the population and healthcare system adapt to higher temperatures. If this occurs, then a higher temperature threshold would be more indicative of the effect of climate change on emergency hospitalisations in the future. Using the coefficient estimate in the final column of Table 4.6, we obtain the following results that we can use as our lower bound. TABLE 4.8 LOWER BOUND ESTIMATES (RCP4.5, 95TH PERCENTILE) 2021–2040 2041–2060 2061–2080 2081–2100 Quarter 2 22.0 39.3 37.3 34.5 Quarter 3 53.4 57.0 26.9 46.0 Total 75.5 96.3 64.2 80.5 In this case, the annual increase in emergency hospital admissions for the 2021–2040 period is 75.5 per 100,000 population. This is equivalent to a 7.3 per cent annual increase. For the period 2041–2060, this annual increase becomes 9.3 per cent. Although these are lower bound estimates, they still indicate a relatively large increase in annual emergency hospital admissions. This is particularly true for a healthcare system that is currently functioning at full capacity. These estimates suggest the need for climate change adaption in the healthcare system to prevent future capacity problems from arising. Table 4.9 compares the projection analysis and lower bounds under the RCP2.6, RCP4.5 and RCP8.5 scenarios. 21 We take the annual estimate (97.2) as a proportion of the mean dependent variable (1,036). Health benefits and co-benefits of climate change mitigation measures | 53 TABLE 4.9 COMPARISON OF ANNUAL INCREASES ACROSS RCP SCENARIOS RCP2.6 – Paris Agreement RCP4.5 – most likely RCP8.5 – no mitigation 90th percentile 95th percentile 90th percentile 95th percentile 90th percentile 95th percentile 2021–2040 10.7% 9.1% 9.4% 7.3% 10.2% 7.9% 2041–2060 14.1% 10.4% 12.2% 9.3% 13.0% 10.0% 2061–2080 7.3% 5.0% 9.4% 6.2% 13.5% 9.5% 2081–2100 9.1% 7.2% 10.7% 7.8% 18.3% 14.1% These findings predict a significant increase in the number of high temperature days in Ireland under all RCP scenarios. Even under the most optimistic RCP scenario, the projected rise in emergency in-patient hospitalisations during the 2021–2040 and 2041–2060 periods is stark. However, the more optimistic RCP scenarios do project a lesser effect on hospital demand in later periods. When considered alongside international evidence, these results emphasise the potential need for policymakers to implement adaptive measures and increase capacity to accommodate the higher hospital demand from higher temperatures. The outcomes in this section should be interpreted as average annual increases within each projected period. However, the surges in hospital demand due to high temperatures are likely to be concentrated around specific years, dates, and locations. The increased demand for hospital care due to higher temperatures is expected to be most acute in the summer months, with quarter 3 recording the highest number of days exceeding each threshold. Furthermore, geographical variations in high-temperature days may also influence the hospitals that might experience the highest demand for care. Therefore, the health system’s adaptability to handle higher temperatures at particular times and in specific regions will be crucial in adjusting to the elevated temperatures in the coming years. 4.4 SUMMARY In this chapter, we first provided an overview of the health benefits and co-benefits of climate change mitigation in Ireland, using approaches that simulate the effects of various temperature path scenarios on health, thereby deriving an assessment of the potential health benefits and co-benefits of different scenarios. The COACCH project results show that mitigation in line with the Paris Agreement (consistent with RCP2.6) would still result in approximately 200 heat-attributed deaths in the last decade of the 21st century. A future consistent with the RCP8.5 (“no mitigation”) scenario would result in more than 1,000 additional deaths in the same decade compared to RCP2.6 (“Paris Agreement”). Focusing specifically on air pollution mitigation measures, the results of the WHO CLIMAQ-H model for Health benefits and co-benefits of climate change mitigation measures | 54 Ireland suggest that 66 deaths per annum could be averted, with considerable additional social and economic benefits (e.g., in terms of worker productivity). In the second part of the chapter, we applied climate projections to the HIPE data used in Chapter 3. This analysis shows that, under the most benign climate change scenario (RCP2.6, Paris Agreement), emergency in-patient admissions in quarter 2 and quarter 3 could increase by an average of 10 per cent per annum over the period 2020–2040. If climate change mitigation actions were implemented fully in accordance with the Paris Agreement (RCP2.6), the impact on emergency in-patient hospitalisations could be lower in the latter half of this century. However, if no policy changes are implemented to reduce emissions (i.e., a future consistent with the RCP8.5 scenario), there could be a steep increase in annual emergency in-patient admissions across quarter 2 and quarter 3, of up to 18.3 per cent in the period 2080–2100 (depending on the temperature threshold used). Simulation exercises such as these are subject to numerous assumptions that are necessary in order to simplify the analysis but which may induce error in projections. In particular, in this application, the model estimated for HIPE data over the period 2015–2019 assumes that the relationship between the average number of days per quarter above the various temperature thresholds and emergency in-patient hospitalisations holds over the entire projection period, i.e., up to 2100. As discussed in Chapter 1, previous literature has suggested that individuals may adopt adaptive strategies to cope with hotter temperatures; therefore, the behaviours we are assuming for 2015–2019 may represent an upper bound on the simulated effects over time. Summary, discussion and policy implications | 55 CHAPTER 5 Summary, discussion and policy implications 5.1 SUMMARY OF MAIN FINDINGS Recognition of the need to limit climate change has driven global negotiations concerning combined efforts to decrease greenhouse gas (GHG) emissions. In Ireland, the Climate Action and Low Carbon Development (Amendment) Act of 2021 commits to the achievement of an annual emissions reduction target of 7 per cent, resulting in a 51 per cent reduction of emissions by 2030, and further commits to net-zero emissions by 2050. The Climate Action Plans provide a detailed plan of measures needed for this transition. When analysing the effects of climate change, the focus is often on the economic costs to society. Population health impacts of climate change are also important and remain relatively unexplored for Ireland. In addition, there is also little discussion of the potential benefits and co-benefits of emission reduction policies for population health. The research detailed in this report, carried out as part of a research programme funded by the Irish Heart Foundation and Irish Cancer Society on behalf of the Climate and Health Alliance, aimed to help fill this gap. It focuses on the health impacts of climate change (and specifically temperature change) and how global and national commitments to limiting temperature change may reduce these impacts in Ireland. Therefore, it also aims to better understand the health benefits and co-benefits of climate change mitigation efforts. According to the climate simulation analysis, Ireland will be exposed to higher mean temperatures in the future. The mean annual temperature is projected to increase by 1–1.6°C by the middle of the century (i.e., 2041–2060) compared to the reference period 1981–2000, under the RCP4.5 climate scenario (the most likely scenario). With this warming will come hotter days and nights. The implications of higher temperatures on morbidity were then assessed by combining detailed meteorological data with data on acute public hospital emergency in-patient hospitalisations from the Hospital In-Patient Enquiry (HIPE) system over the period 2015–2019. The results showed that temperatures between 22°C and 25°C are associated with an 8.5 per cent increase in emergency in-patient hospitalisations, with children particularly affected. In terms of health benefits and co-benefits, the research highlighted findings from the COACCH project which estimate that annual mortality under the most pessimistic scenario (RCP8.5) could be around 1,400 additional deaths by the end of the 21st century, in contrast to 216 under the most optimistic scenario (RCP2.6). Focusing specifically on air pollution mitigation measures, the results of the WHO CLIMAQ-H model for Ireland suggest that 66 deaths per annum could be averted, with considerable additional social and economic benefits (e.g., in terms of worker productivity). In terms of morbidity, an analysis of the HIPE data showed that an Summary, discussion and policy implications | 56 additional day in summer months above the 90th temperature percentile in the period 2081–2100 could be associated with a 10.7 percentage increase in annual emergency in-patient hospitalisations under the RCP4.5 scenario (“most likely”), and as high as 18.3 percentage increase under the most pessimistic (“no mitigation”) RCP8.5 scenario. 5.2 DISCUSSION AND POLICY IMPLICATIONS The analysis in this report has highlighted a number of key findings for Ireland in relation to climate change in health, in terms of future temperature paths, impacts on morbidity and projections of future health impacts of selected temperature paths. The analysis is necessarily limited in scope, and does not consider a number of other important issues that would need to be considered to assess a) the full health effects of climate change in Ireland and b) the full health benefits and co-benefits of climate change mitigation actions. In particular, as illustrated in Figure 1.1, the impact of climate change on health is complex, covering multiple pathways that link a variety of climate change features (e.g., rising temperatures, rising sea levels, more extreme events, etc.) with a myriad of health outcomes (e.g., mortality, chronic disease incidence, mental health, etc.). In this report, we focused on the impact of the feature of climate change (i.e., increasing temperature) that is considered one of the most likely and harmful climate change features for Ireland (Desmond et al., 2017). Future work in Ireland could consider the impact of other climate change events on health in Ireland, a broader set of health outcomes, and the potential for health benefits and co-benefits on other dimensions of health in addition to mortality and emergency hospital admissions. In addition, while the analysis in this report has focused on the potential health co-benefits of mitigation action in terms of various projected future temperature paths (consistent with different mitigation scenarios), future work could assess the health benefits and co-benefits of selected mitigation actions (e.g., more sustainable transport). 22 A number of implications for policy arise from the analysis in this report. Policy responses to climate change fall broadly into two categories: mitigation (preventing or reducing the scale of future harm), and adaptation, implementing a wide range of adaptation solutions, including effective heat health action plans, urban greening, appropriate building design and construction, and adjusting working times (Hensher, 2023; Kaźmierczak et al., 2022). The analysis in this report has focused primarily on mitigation, and the potential health benefits and co-benefits of limiting future temperature rises and impacts. An important conclusion for policy is that there are considerable health benefits and co-benefits from mitigation that should be considered in policymaking. The broader literature 22 This type of analysis was outside of the scope of the current report due to the data and methodological requirements. For example, in assessing the health benefits/co-benefits of increased cycling, an assessment is needed of a) how much extra cycling, b) how cycling is linked to health and c) what aspect of health is affected, and how is that quantified? Summary, discussion and policy implications | 57 highlights the importance of careful consideration of mitigation and adaptation measures, including an assessment of the potential for unintended consequences. For example, past decades of efforts to reduce CO2 emissions in Europe without adequate consideration of health include promoting diesel over gasoline-powered vehicles, and the promotion of biomass for residential heating, both of which resulted in considerable emissions of health-damaging air pollutants (van Daalen et al., 2022). Similar concerns can be raised over adaptation technologies; for example, Deschenes (2022) notes that more attention needs to be devoted to increasing opportunities and finding solutions to protect human health from extreme heat while at the same time minimising the damages from the local and global externalities caused by the electricity generation necessary for meeting the increased cooling demand that climate change will bring. Nonetheless, the findings highlight the continued importance of policy measures to achieve the targets set out in the Climate Action Plans. In terms of air quality, for example, the recent Clean Air Strategy commits to achieving the final WHO ACQ values by 2040 (Government of Ireland, 2023), and at EU level, the proposed revision to the Ambient Air Quality Directive will set interim 2030 EU air quality standards, aligned more closely with WHO guidelines, and set Europe on a trajectory to achieve zero pollution for air by 2050. Policy measures to mitigate the impacts of climate change, such as decarbonising home heating, promoting active travel and transitioning to electric vehicles, will be an important component of the policy response, and will also have concomitant benefits for population health (van Daalen et al., 2022). The findings in relation to temperature effects on emergency hospital admissions highlight the importance of considering the broader impacts of climate change on the health sector. The EEA note that improving the resilience of healthcare facilities across Europe is necessary not only due to the pressure on their capacity to deliver patient care during heatwaves or diagnostics during outbreaks of climate-sensitive infectious diseases, but also due to the fact that they tend to be located in urban areas that are more prone to the ‘urban heat island’ effect (Kaźmierczak et al., 2022). Many healthcare systems have also introduced plans to adapt to climate change. In October 2020, the English National Health Service (NHS) became the world’s first national health system to commit to becoming ‘net zero’, pledging to reduce its carbon emissions to net zero by 2040, including in its facilities and buildings. 23 The HSE has now followed suit. In June 2023, the HSE launched its Climate Action Strategy 2023–2050. 24 This is a health service-wide strategy that aims to reduce the impacts of climate change on the health service and deliver healthcare in a more environmental and socially sustainable manner. A key goal is to achieve net-zero emissions for the HSE by 2050. 23 NHS Net Zero Building Standard www.england.nhs.uk/estates/nhs-net-zero-building- standard/#:~:text=The%20NHS%20Net%20Zero%20Building,now%20and%20in%20the%20future. 24 www.hse.ie/eng/about/who/healthbusinessservices/national-health-sustainability-office/climate-change-and- health/hse-climate-action-strategy-2023-50.pdf. References | 64 Hao, G. et al. (2022). ‘Associations of PM2.5 and road traffic noise with mental health: Evidence from UK Biobank’, Environmental Research, Vol. 207, p.112221, https://doi.org/10.1016/j.envres.2021.112221. Healy, J.D. (2003). ‘Excess winter mortality in Europe: A cross country analysis identifying key risk factors’, Journal of Epidemiology & Community Health, Vol. 57, No. 10, pp.784–789, https://doi.org/10.1136/jech.57.10.784. Hensher, M. (2023). ‘Climate change, health and sustainable healthcare: The role of health economics’, Health Economics [Preprint], https://doi.org/10.1002/hec.4656. IPCC (2013). Summary for Policymakers. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, edited by T.F. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S.K, Cambridge University Press. IPCC (2014). Climate Change 2014: Synthesis Report, Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team, R.K. Pachauri, and L.A. Meyer (eds.)], IPCC, Geneva, Switzerland, p.151. Karlsson, M. and Ziebarth, N.R. (2018). ‘Population health effects and health-related costs of extreme temperatures: Comprehensive evidence from Germany’, Journal of Environmental Economics and Management, Vol. 91, pp.93–117, https://doi.org/10.1016/j.jeem.2018.06.004. Kaźmierczak, A. et al. (2022). Climate change as a threat to health and well-being in Europe: Focus on heat and infectious diseases. Luxembourg: Publications Office of the European Union (EEA Report, 7), www.eea.europa.eu/publications/climatechange-impacts-on-health. Keegan, C. et al. (2020). Projections of expenditure for public hospitals in Ireland, 2018–2035, based on the Hippocrates Model, ESRI Research Series 117, Dublin: ESRI, https://doi.org/10.26504/rs117. Kroesen, M. and De Vos, J. (2020). ‘Does active travel make people healthier, or are healthy people more inclined to travel actively?’, Journal of Transport & Health, Vol. 16, p.100844, https://doi.org/10.1016/j.jth.2020.100844. Liao, H. et al. (2023). ‘Extreme temperatures, mortality, and adaptation: Evidence from the county level in China’, Health Economics [Preprint], https://doi.org/10.1002/hec.4649. Lin, S. et al. (2009). ‘Extreme high temperatures and hospital admissions for respiratory and cardiovascular diseases’, Epidemiology (Cambridge, Mass.), Vol. 20, No. 5, pp.738–746, https://doi.org/10.1097/EDE.0b013e3181ad5522. Lyons, S. et al. (2023). ‘The Effect of “Smoky” Coal Bans on Chronic Lung Disease Among Older People in Ireland’, Economics & Human Biology, Vol. 50, p.101275, https://doi.org/10.1016/j.ehb.2023.101275. Markandya, A., Sampedro, J., Smith, S.J., Van Dingenen, R., Pizarro-Irizar, C., Arto, I., and González-Eguino, M. (2018). ‘Health co-benefits from air pollution and mitigation costs of the Paris Agreement: A modelling study’, The Lancet Planetary Health, Vol. 2, No. 3, pp.e126-e133, https://doi.org/10.1016/S2542-5196(18)30029-9. References | 65 Met Office (2018). UKCP18 Guideline: Representative Concentration Pathways, www.metoffice.gov.uk/binaries/content/assets/metofficegovuk/pdf/research/ukc p/ukcp18-guidance---representative-concentration-pathways.pdf. Milner, J. et al. (2015). ‘Health effects of adopting low greenhouse gas emission diets in the UK’, BMJ Open, Vol. 5, No. 4, p.e007364, https://doi.org/10.1136/bmjopen-2014- 007364. Mullins, J. and White, C. (2019). ‘Temperature and mental health: Evidence from the spectrum of mental health outcomes’, Journal of Health Economics, Vol. 68, p.102240, https://doi.org/10.1016/j.jhealeco.2019.102240. Murray, C.J.L. et al. (2020). ‘Global burden of 87 risk factors in 204 countries and territories, 1990–2019: A systematic analysis for the Global Burden of Disease Study 2019’, The Lancet, Vol. 396, No. 10258, pp.1223–1249, https://doi.org/10.1016/S0140- 6736(20)30752-2. Musacchio, A., Andrade, L., O’Neill, E., Re, V., O’Dwyer, J., and Hynds, P.D. (2021). ‘Planning for the health impacts of climate change: Flooding, private groundwater contamination and waterborne infection – A cross-sectional study of risk perception, experience and behaviours in the Republic of Ireland’, Environmental Research, Vol. 194, https://doi.org/10.1016/j.envres.2021.110707. Navarro, K.M., Kleinman, M.T., Mackay, C.E., Reinhardt, T.E., Balmes, J.R., Broyles, G.A., Ottmar, R.D., Naher, L.P. and Domitrovich, J.W. (2019). ‘Wildland firefighter smoke exposure and risk of lung cancer and cardiovascular disease mortality’, Environmental Research, Vol. 173, pp.462–468, https://doi.org/10.1016/j.envres.2019.03.060. Nolan, P. and Flanagan, J. (2020). ‘High-resolution climate projections for Ireland – A multi-model ensemble approach’, EPA Report Research 339, https://www.epa.ie/publications/research/climatechange/Research_Report_339_Part1.pdf. OECD (2016). The Economic Consequences of Outdoor Air Pollution. Paris: OECD, https://doi.org/10.1787/9789264257474-en. OECD (2020). Health at a Glance: Europe 2020. Paris: OECD, www.oecd-ilibrary.org/socialissues-migration-health/health-at-a-glance-europe-2020_82129230-en. Orru, K. et al. (2016). ‘Well-being and environmental quality: Does pollution affect life satisfaction?’, Quality of Life Research, Vol. 25, No. 3, pp.699–705, https://doi.org/10.1007/s11136-015-1104-6. Pascal, M. et al. (2013). ‘Heatwaves and mortality in Ireland, planning for the future’, Irish Geography, Vol. 46, No. 3, pp.203–211, https://doi.org/10.1080/00750778.2014.898125. Peters, R. et al. (2019). ‘Air pollution and dementia: A systematic review’, Journal of Alzheimer’s Disease, Vol. 70, No. s1, pp.S145–S163, https://doi.org/10.3233/JAD-180631. Pimpin, L. et al. (2018). ‘Estimating the costs of air pollution to the National Health Service and social care: An assessment and forecast up to 2035’, PLOS Medicine, Vol. 15, No. 7, p.e1002602, https://doi.org/10.1371/journal.pmed.1002602. References | 66 Power, M.C. et al. (2015). ‘The relation between past exposure to fine particulate air pollution and prevalent anxiety: Observational cohort study’, BMJ: British Medical Journal, Vol. 350, p.h1111, https://doi.org/10.1136/bmj.h1111. Riahi, K. et al. (2017). ‘The shared socioeconomic pathways and their energy, land use, and greenhouse gas emissions implications: An overview’, Global Environmental Change, Vol. 42, pp.153–168, https://doi.org/10.1016/j.gloenvcha.2016.05.009. Rizmie, D. et al. (2022). ‘Impact of extreme temperatures on emergency hospital admissions by age and socio-economic deprivation in England’, Social Science & Medicine, Vol. 308, p.115193, https://doi.org/10.1016/j.socscimed.2022.115193. Rocque, R.J. et al. (2021). ‘Health effects of climate change: An overview of systematic reviews’, BMJ Open, Vol. 11, No. 6, p.e046333, https://doi.org/10.1136/bmjopen- 2020-046333. Romanello, M. et al. (2022). ‘The 2022 report of The Lancet Countdown on health and climate change: Health at the mercy of fossil fuels’, The Lancet, Vol. 400, No. 10363, pp.1619–1654, https://doi.org/10.1016/S0140-6736(22)01540-9. Sampedro, J. et al. (2020). ‘Health co-benefits and mitigation costs as per the Paris Agreement under different technological pathways for energy supply’, Environment International, Vol. 136, p.105513, https://doi.org/10.1016/j.envint.2020.105513. Saunders, L.E. et al. (2013). ‘What are the health benefits of active travel? A systematic review of trials and cohort studies’, PLOS One, Vol. 8, No. 8, p.e69912, https://doi.org/10.1371/journal.pone.0069912. Ščasný, M. et al. (2019). ‘Non-market impacts–health: Report D2. 6 for the project CO-designing the Assessment of Climate CHange costs (COACCH)’, Vrije University Amsterdam, https://research.vu.nl/en/publications/non-market-impacts-health- report-d26-for-the-project-codesigning-. Sera, F. et al. (2019). ‘How urban characteristics affect vulnerability to heat and cold: A multi-country analysis’, International Journal of Epidemiology, Vol. 48, No. 4, pp.1101–1112, https://doi.org/10.1093/ije/dyz008. Shaw, C., Blakely, T. and Tobias, M. (2011). ‘Mortality among the working age population receiving incapacity benefits in New Zealand, 1981–2004’, Social Science & Medicine, Vol. 73, No. 4, pp.568–575, https://doi.org/10.1016/j.socscimed.2011.06.003. Smith, J. (1999). ‘Healthy bodies and thick wallets: The dual relation between health and economic status’, Journal of Economic Perspectives, Vol. 13, No. 2, pp.145–166, https://doi.org/10.1257/jep.13.2.145. Smith, K. et al. (2014). ‘Human health: Impacts, adaptation, and co-benefits’, Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects, Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge, UK: Cambridge University Press, pp.709–754, www.ipcc.ch/site/assets/uploads/2018/02/WGIIAR5-Chap11_FINAL.pdf. Song, Jian et al. (2021). ‘Ambient high temperature exposure and global disease burden during 1990–2019: An analysis of the Global Burden of Disease Study 2019’, The Science of the Total Environment, Vol. 787, p.147540, https://doi.org/10.1016/j.scitotenv.2021.147540. References | 67 Springmann, M. et al. (2016). ‘Analysis and valuation of the health and climate change cobenefits of dietary change’, Proceedings of the National Academy of Sciences, Vol. 113, No. 15, pp.4146–4151, https://doi.org/10.1073/pnas.1523119113. Tarabochia-Gast, A.T., Michanowicz, D.R. and Bernstein, A.S. (2022). ‘Flood risk to hospitals on the United States Atlantic and Gulf coasts from hurricanes and sea level rise’, GeoHealth, Vol. 6, No. 10, p.e2022GH000651, https://doi.org/10.1029/2022GH000651. van Daalen, K.R. et al. (2022). ‘The 2022 Europe report of The Lancet Countdown on health and climate change: Towards a climate resilient future’, The Lancet Public Health, Vol. 7, No. 11, pp.e942–e965, https://doi.org/10.1016/S2468-2667(22)00197-9. van Vuuren, D.P., Edmonds, J.A., Kainuma, M., Riahi, K., and Weyant, J. (2011). ‘A special issue on the RCPs’, Climatic Change, Vol. 109, 1-4, https://doi.org/10.1007/s10584- 011-0157-y. Vandenberghe, D. and Albrecht, J. (2018). ‘Tackling the chronic disease burden: Are there co-benefits from climate policy measures?’, The European Journal of Health Economics: HEPAC: Health Economics in Prevention and Care, Vol. 19, No. 9, pp.1259–1283, https://doi.org/10.1007/s10198-018-0972-4. Vos, T. et al. (2020). ‘Global burden of 369 diseases and injuries in 204 countries and territories, 1990–2019: A systematic analysis for the Global Burden of Disease Study 2019’, The Lancet, Vol. 396, No. 10258, pp.1204–1222, https://doi.org/10.1016/S0140-6736(20)30925-9. West, J.J. et al. (2013). ‘Co-benefits of global greenhouse gas mitigation for future air quality and human health’, Nature Climate Change, Vol. 3, No. 10, pp.885–889, https://doi.org/10.1038/NCLIMATE2009. Weuve, J. et al. (2012). ‘Exposure to particulate air pollution and cognitive decline in older women’, Archives of Internal Medicine, Vol. 172, No. 3, pp.219–227, https://doi.org/10.1001/archinternmed.2011.683. White, C. (2017). ‘The dynamic relationship between temperature and morbidity’, Journal of the Association of Environmental and Resource Economists, Vol. 4, No. 4, pp.1155–1198, https://doi.org/10.1086/692098. WHO (2021). WHO global air quality guidelines: particulate matter (PM2.5 and PM10), ozone, nitrogen dioxide, sulphur dioxide and carbon monoxide, Geneva: World Health Organization. WHO (2023). Achieving health benefits from carbon reductions. Manual for use of the climate change mitigation, air quality and health tool, Copenhagen: WHO Regional Office for Europe. Williams, M.L. et al. (2018). ‘The Lancet Countdown on health benefits from the UK Climate Change Act: A modelling study for Great Britain’, The Lancet: Planetary Health, Vol. 2, No. 5, pp.e202–e213, https://doi.org/10.1016/S2542-5196(18)30067-6. Wood, D. et al. (2022). ‘Exposure to ambient air pollution and the incidence of dementia in the elderly of England: The ELSA cohort’, International Journal of Environmental Research and Public Health, Vol. 19, No. 23, https://doi.org/10.3390/ijerph192315889. References | 68 Yu, P. et al. (2023). ‘Cancer and ongoing climate change: Who are the most affected?’, ACS Environmental Au, Vol. 3, No. 1, pp.5–11, https://doi.org/10.1021/acsenvironau.2c00012. Zeka, A. et al. (2014). ‘The association of cold weather and all-cause and cause-specific mortality in the island of Ireland between 1984 and 2007’, Environmental Health, Vol. 13, No. 1, p.104, https://doi.org/10.1186/1476-069X-13-104. Appendix | 69 APPENDIX A – DATA, VARIABLES AND SUMMARY STATISTICS Table A.1 provides a summary of the variables used in our analysis from the HIPE dataset. TABLE A.1 SUMMARY OF VARIABLES USED FROM HIPE Variable Notes County All counties in the Republic of Ireland; incl.: North Dublin; South Dublin Tipperary North; Tipperary South Waterford County; Waterford City Cork County; Cork City Limerick County; Limerick City Galway County; Galway City Not Ireland No fixed abode Age (5yr Brackets) From 0–4 to 100+ Sex Male/Female Public or Private Status Public/Private Medical Card Status No Medical Card Medical Card Unknown Marital Status Single Married Widowed Other (incl. separated) Unknown Divorced Civil Partner Former Civil Partner Surviving Civil Partner Length of Stay Measured in days, minimum value is 0.5 Diagnosis Codes From primary diagnosis to 30th diagnosis code Uses ICD-10-AM Mode of Emergency Admission Emergency Department AMAU-In-patient Other Unknown AMAU Only Local Injury Unit ASAU-In-patient ASAU Only Appendix | 70 Elective/Emergency/Maternity Indicates whether an admission is elective, emergency or maternity Admission Month January–December Admission Week Weeks 0–52 Weeks start on the first Sunday of each year. 2017 has weeks numbered 1–53 because it starts on a Sunday TABLE A.2 ICD-10-AM CODES FOR DIAGNOSIS VARIABLES Diagnosis ICD-10-AM Codes Circulatory Diseases I00 – I99 Respiratory Diseases J00 – J99 Metabolic Diseases E00 – E99 Infectious Diseases A00 – A99 & B00 – B99 Injuries T00.1 – T00.9; T01.0 – T01.4; Appendix | 71 Figure A.1 shows the emergency in-patient hospital admissions rate (per 100,000 population) by county for 2019. The admissions rate is calculated for every week of the year and these weekly admission rates are averaged across the whole year (for 2019). The figure below shows the variation in emergency hospital admissions rates across county of residence. There are higher rates of emergency hospital admissions among people who live in the west of Ireland and in the midlands. FIGURE A.1 EMERGENCY IN-PATIENT HOSPITAL ADMISSIONS RATE BY COUNTY (2019) Appendix | 72 APPENDIX B – LAGGED EFFECTS OF TEMPERATURE BIN MODEL Table B.1 shows the coefficient estimates for the control variables of the temperature bin analysis. TABLE B.1 COEFFICIENT ESTIMATES FOR TEMPERATURE BIN ANALYSIS Hospital Admissions Rate Mean of Charlson Co-Morbidity Index 0.59 (0.71) Mean Number of Medical Card 7.02*** (1.90) Mean Public/Private Status 6.12*** (2.39) Mean Married or Cohabiting -1.52 (1.06) Male – Age 0–9 -5.91 (7.88) Male – Age 10–19 -16.88* (8.74) Male – Age 20–29 -12.74** (9.27) Male – Age 30–49 -24.01*** (8.19) Male – Age 50–69 -13.85* (7.75) Male – Age 70–79 -17.51** (7.82) Male – Age 80–89 -11.10 (8.01) Male – Age 90+ Ref - Female – Age 0–9 -18.42** (8.01) Female – Age 10–19 -4.69 (9.29) Female – Age 20–29 -24.11** (10.60) Female – Age 30–49 -18.72** (8.55) Female – Age 50–69 -12.55 (7.90) Female – Age 70–79 -13.47* (7.96) Female – Age 80–89 -8.78 (8.02) Female – Age 90+ -14.19 (9.12) Appendix | 73 APPENDIX C – MEAN DEVIATION MODEL: GROUP/DIAGNOSIS ANALYSIS AND FULL MODEL RESULTS Table C.1 shows the results of the mean deviation model when it estimated by age group. The age groups are those that are used in the temperature bin analysis and are grouped accordingly: (1) children (ages 0–14); (2) working (ages 15–64); and (3) older (ages 65+). The analysis shows that there are significant effects of maximum temperature on emergency in-patient hospitalisations for the children’s age group and the working age group for the differences model. This is likely to be a reflection of how the model is constructed mathematically using a deviation from the mean level of hospital admissions by age group. The older age group would have a relatively high average number of hospital admissions, so any deviations from this mean would be quite small in proportion. TABLE C.1 MEAN DEVIATION MODEL BY AGE GROUP (1) Child (2) Working (3) Retired % Difference in Maximum Temperature 0.30*** (0.07) 0.14*** (0.05) 0.07 (0.05) % Difference in Maximum Temperature2 -0.92** (0.39) -0.58* (0.30) -0.38 (0.29) Temperature Lags Yes Yes Yes County Fixed Effects Yes Yes Yes Year Trend Yes Yes Yes Month Fixed Effects Yes Yes Yes Socio-demographic Controls Yes Yes Yes N 3,521 3,532 3,522 R2 0.05 0.04 0.19 Note: Analysis includes precipitation controls, including lagged precipitation for three weeks. Standard errors in parentheses. * p<0.05, ** p<0.01, *** p<0.001 Table C.2 illustrates the results of the model when it is estimated for each diagnosis group. The analysis shows that for injuries, the relationship between temperature and admissions is statistically significant. Previous literature supports the fact that risky and aggressive behaviour becomes more common during hot spells of weather – this could also be due to the fact that people are more likely to consume alcohol during hotter weather, increasing the likelihood of accidents (Hagström,