Health & Place 67 (2021) 102294 Contents lists available at ScienceDirect Health and Place journal homepage: http://www.elsevier.com/locate/healthplace Energy poverty and health: Trends in the European Union before and during the economic crisis, 2007–2016 Laura Oliveras a, b, c, *, Andr�es Peralta a, d, e, Laia Pal�encia a, b, f, Merc�e Gotsens a, b, �pez a, b, c, f, Lucia Artazcoz a, b, c, f, Carme Borrell a, b, c, f, Marc Marí-Dell’Olmo a, b, f María Jos�e Lo a Ag�encia de Salut Pública de Barcelona, Lesseps 1, 08023, Barcelona, Catalonia, Spain b Institut d’Investigaci� o Biom�edica Sant Pau (IIB Sant Pau), Sant Quintí 77, 08041, Barcelona, Catalonia, Spain c Department of Experimental and Health Sciences, Universitat Pompeu Fabra, Doctor Aiguader 88, 08003, Barcelona, Catalonia, Spain d Health Inequalities Research Group, Employment Conditions Knowledge Network (GREDS-EMCONET). Department of Political and Social Sciences, Universitat Pompeu Fabra, Ramon Trias Fargas 25-27, 08005, Barcelona, Catalonia, Spain e Johns Hopkins University - Pompeu Fabra University Public Policy Centre, Ramon Trias Fargas 25-27, 08005, Barcelona, Catalonia, Spain f CIBER Epidemiología y Salud Pública (CIBERESP), Spain A B S T R A C T The aim of this study is to analyse the time trends in the European Union (EU) before and during the economic crisis in 1) the energy poverty (EP) prevalence; 2) the association between EP and health and 3) the impact of EP on health. We analyse trends among women and men in two EU macro regions, defined by a novel index of structural vulnerability to EP. The study shows how EP and its impact on health worsened during the economic crisis and identifies groups at higher risk such as women and people living in Mediterranean and Eastern European countries, which have been found to be countries with higher structural vulnerability to EP. 1. Introduction of the countries which determine the degree of protection that states provide to its population in the case of changes in the internal or external Energy for cooking, lighting, washing or achieving thermal comfort conditions that may drive households into EP. Structural vulnerability to at home is one of the essential elements of adequate housing, which EP is therefore closely associated with the percentage of people suffering is recognized as a human right by the United Nations (Office of the from EP in the countries (Recalde et al., 2019). United Nations, 2009). Energy poverty (EP) can be understood as the EP is a major social problem in Europe and is likely to further in­ inability of a household to secure a socially and materially required level crease in the coming years due to rising energy prices, continuing eco­ of energy services in the home (Bouzarovski, 2014). The main driver of nomic hardship, increasing inequalities, and greater climate variability EP is the interaction of high energy prices, energy-inefficient homes and due to climate change (Bouzarovski, 2014; Csiba et al., 2016; EPEE, appliances, low household incomes, and the specific energy needs of the 2009; Marmot Review Team, 2011). It is estimated that EP affects be­ household. These drivers are themselves embedded in more structural tween 50 and 125 million people in Europe (EPEE, 2009). Specifically, determinants such as energy, housing and labour policies and markets, in 2014, 10.2% of households in the European Union (EU) were unable economic policies, and welfare states (Bouzarovski, 2014; Csiba et al., to keep their home adequately warm during the cold months (Tirado 2016; Marí-Dell’Olmo et al., 2017). A broader and more dynamic con­ Herrero et al., 2016). However, as several studies have shown, EP is not ceptualisation of EP is energy vulnerability, which refers to the sus­ equally distributed socially and geographically. Socially, above-average ceptibility of a household to experience EP if there is a change in the rates of EP are experienced by the elderly, families with children internal conditions of the dwelling unit (such as a loss of employment) or (especially single parent families), households with people with in the external conditions (such as an economic crisis). In this case, EP disabilities or long-term illnesses, the unemployed, and people in can be considered as a temporary result of deprivation, meaning that poorly paid jobs (Bouzarovski, 2014; Csiba et al., 2016; EPEE, households can enter and exit the condition at specific moments (Tirado 2009; Herna �ndez, 2016; Gillard et al., 2017; Wright, 2004; Adam and Herrero et al., 2016). Another relevant concept is structural vulnera­ Monaghan, 2016; O’sullivan et al., 2016). Geographically, EP is far more bility to EP, which refers to the political and socio-economic conditions important in Mediterranean and Eastern EU Member States, than in the * Corresponding author. Institut d’Investigaci� o Biom� edica Sant Pau (IIB Sant Pau), Sant Quintí 77, 08041, Barcelona, Catalonia, Spain. E-mail address: loliveras@santpau.cat (L. Oliveras). https://doi.org/10.1016/j.healthplace.2020.102294 Received 22 March 2019; Received in revised form 24 December 2019; Accepted 31 January 2020 Available online 4 March 2020 1353-8292/© 2020 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). L. Oliveras et al. Health and Place 67 (2021) 102294 Central and Northern ones (Bouzarovski, 2014; Bouzarovski and time trend in the EU before and during the economic crisis in 1) the EP Tirado Herrero, 2017; Bouzarovski, 2018; Bollino and Botti, 2017). By prevalence; 2) the association between EP and health; and 3) the impact analysing spatial trends in EP at the national level, Bousarovski and of EP on health. We analysed trends among women and men in the two Tirado Herrero identified two differentiated macro regions in the EU, EU macro regions, defined by the degree of structural vulnerability to which they called the core and the periphery (Bouzarovski and Tirado EP. We calculated two different epidemiological measures, since asso­ Herrero, 2017). Going one step further, Recalde et al. studied the ciation measures allow us to know how EP and health are related and structural vulnerability of the EU-27 countries through the creation of impact measures tell us to what extent EP is an important public health an index that considers the structural determinants of EP, including problem (Gefeller, 1990). indicators such as the countries expenditure on social protection per inhabitant, the median income, the percentage of social rental stock or 2. Methods the annual electricity switching rate. Through a cluster analysis of the index, that study identified exactly the same two macro regions as 2.1. Design, study population, units of analysis, and information sources Bousarovski and Tirado Herrero, with the countries constituting the core being those with lower structural vulnerability to EP and those identi­ This individual-based trends study analysed three cross-sectional fied as belonging to the periphery being those with higher structural waves of the European Quality of Life Surveys, corresponding to the vulnerability to EP (Recalde et al., 2019). years 2007, 2012, and 2016 (European Foundation for the Improvement Several studies and reports have documented how EP negatively of Living and Working Conditions, 2018a). It covers 95 940 individuals affects people’s health and well-being (Csiba et al., 2016; (aged 18 years or older) from the 27 European countries constituting the Marí-Dell’Olmo et al., 2017; Marmot Review Team, 2011; Hern� andez, EU in the period 2007–2013 (EU27). The country samples are nationally 2016; Thomson et al., 2017a; Bosch et al., 2019; Healy, 2003; Liddell representative and range from 1000 to 3000 respondents per year, and Morris, 2010; Liddell and Guiney, 2015; Peralta et al., 2017; depending on country size (European Foundation for the Improvement Poortinga et al., 2017). So far, most studies have focused on the direct of Living and Working Conditions, 2018b). and indirect health effects of cold housing, with little evidence on other consequences of EP such as inadequate household temperature during 2.2. Study variables the hot months. The direct health effects of EP can be divided into increased morbidity rates and a higher risk of mortality. Cold temper­ We included three health outcomes as dependent variables. To assess atures are strongly related to cardiovascular and respiratory diseases, general health, we used self-reported health status measured by a single and double the risk of respiratory problems in children (Marmot Review item with five response categories that we dichotomised into two: good Team, 2011). Cold also supresses the immune system, increasing the risk self-reported health (when very good and good) and poor self-reported of infections and minor illnesses such as colds and flu (Liddell and health (when fair, bad, and very bad) (Quesnel-Vall�ee, 2007). To mea­ Morris, 2010). Living in EP also exacerbates existing conditions such as sure subjective psychological well-being, we used the 5-item World arthritis and rheumatism and negatively affects mental health by Health Organisation Well-being Index, which scores each item on a scale increasing the financial stress on households. The most extreme from 0 to 5 and the sum of the five-item score is multiplied by 4, to give a EP-related health outcome is mortality, which has been studied through final score ranging from 0 to 100. Higher scores indicate better comparison of excess winter mortality across European countries well-being. We created two dichotomous variables through the appli­ (Healy, 2003). Among indirect health impacts, the main findings are cation of two commonly used cut-off scores recommended to identify that EP impairs children’s educational attainment, emotional well-being people with reduced well-being (score<50) and clinical depression and resilience, reduces dietary opportunities and choices, and increases (score<28) (Topp et al., 2015). the risk of accidents and injuries at home (Marmot Review Team, 2011). The main independent variables were EP and time. The EP assess­ EP also hampers normal functioning in everyday areas such as work or ment was based on the consensual approach, consisting of subjective study and has social consequences such as stigmatisation and reduction self-reported indicators. To capture wider expression of EP, we included in social interaction. In addition, some groups are especially vulnerable two different indicators. The first indicator considers individuals to to the effects of EP, such as the elderly, children, and dependent or experience EP if they report that they cannot afford to keep their homes chronically ill people (Thomson et al., 2013). In such cases, individual adequately warm and the second one, if they report one or more arrears vulnerability is related to the larger number of hours spent at home. in utility bills such electricity, water, or gas during the past 12 months. Moreover, the impacts of EP can be aggravated by other situations of From here on we will designate the first indicator as EPT (where T stands vulnerability, such as a lack of social support (Marí-Dell’Olmo et al., for Temperature), and the second as EPA (where A stands for Arrears). 2017). Regarding time, we included the three last waves of the European Structural determinants and drivers of EP have been directly affected Quality of Life Surveys, which correspond to the years before (2007) and by the economic crisis that started in 2008 in Europe and the consequent during the crisis (2012 and 2016). implementation of austerity packages, which may have led energy The stratification variables were sex and country typologies. We used vulnerable households to go into EP. As some studies have shown, EP the two macro regions proposed by Recalde et al., which divided the 27 increased considerably during the economic crisis in countries such as EU member countries into two groups depending on their structural Greece (Dagoumas and Kitsios, 2014) and Spain (Tirado Herrero et al., vulnerability to EP: (1) countries with lower structural vulnerability 2016; Tirado Herrero and Jim�enez Meneses, 2016). Although recessions (LSVc) (Denmark, Netherlands, Sweden, Germany, Finland, Austria, pose risks to health, the interaction of fiscal austerity with economic Luxembourg, Belgium, France, United Kingdom, and Ireland) and (2) shocks and weak social protection is what ultimately seems to escalate countries with higher structural vulnerability (HSVc) (Czech Republic, health and social crises in Europe (Karanikolos et al., 2013). However, Malta, Slovenia, Portugal, Spain, Poland, Cyprus, Estonia, Hungary, despite the rising importance of EP in European policy and science Slovakia, Italy, Lithuania, Latvia, Greece, Bulgaria, and Romania). Age agendas, there is a lack of in-depth research on the trend in EP before was considered as an adjustment variable. and during the economic crisis at the European level, as well as on the possible aggravation of its health effects. Although some studies have 2.3. Data analysis shown that EP is associated with worse health outcomes, no studies have analysed the magnitude of this public health problem, that is, its impact To describe the study sample, we calculated the prevalence (per­ at the population level. centage) of EP and the prevalence of health outcomes among people The aim of this study was to fill this knowledge gap by analysing the living with and without EP and performed chi-square tests to assess 2 L. Oliveras et al. Health and Place 67 (2021) 102294 whether the prevalence changed over time. the supplementary information (S5 to S17). We analysed the association between EP and health through age- adjusted prevalence ratios (PR) and their 95% confidence intervals 3. Results (CI), obtained by fitting Poisson regression models with robust variance (Espelt et al., 2017; Barros and Hirakata, 2003). The dependent vari­ Of the 95 940 included individuals, 51.8% are women and 53.3% are ables were health outcomes, and the independent variables were EP, citizens of LSVc. Below, the results are presented separately for the two age, and year. To estimate changes in the association over time, we EP indicators, EPT and EPA, respectively. included the interaction between EP and survey year in the models and calculated PRs and CIs for each year. 3.1. Inability to keep home adequately warm (EPT) Finally, as a measure of impact, we calculated the population attributable risk percent (PAR%), that is, the proportion of negative 3.1.1. Prevalence of energy poverty health outcomes (e.g. poor self-reported health) in the population that is Descriptive statistics for EP and health outcomes stratified by sex, attributable to EP and that, theoretically, could be prevented by elimi­ region, and year are provided in Table 1. In the EU, 9.0% of the popu­ nating EP. Measures of impact are important, because they take into lation could not afford to keep their homes adequately warm in 2007. account the proportion of individuals exposed to the risk factor in the This percentage rose to 11.8% in 2012 and decreased to 9.0% in 2016. population, and not only the strength of the association between the risk This clear “peak” pattern (Ʌ) over time, in which EP increased from factor and the outcome. We obtained age-adjusted PAR% for each year 2007 to 2012 and decreased between 2012 and 2016 was observed in all directly from Poisson regression models by subtracting the expected groups, independently of sex and region (Fig. 1a). Throughout the study cases if EP were absent from the observed cases and dividing these by the period, EPT was noticeably higher in HSVc than in LSVc and women observed cases (Gefeller, 1990; Rückinger et al., 2009). To estimate experienced also higher EPT rates than men. The most affected popu­ changes over time, we calculated the PAR% differences between years. lation was therefore women living in HSVc, where EPT was 16.0%, For both PAR% and PAR% differences, the bootstrap resampling method 18.3%, and 14.8% in 2007, 2012, and 2016, respectively. was used to calculate their respective 95% CI. All analyses were performed for the three health outcomes separately 3.1.2. Association between energy poverty and health for both EP indicators, and were stratified by sex, country typologies, The association between health outcomes and EPT is shown in Fig. 2 and year. In all analyses we took into consideration the complex sample (subfigures a-c). Detailed data can be found in the supplementary in­ design by using sampling weights. We used the R software (version formation (S1). We found an overall statistically significant association 3.4.4, R Foundation for Statistical Computing). between the three health outcomes and EPT, but the strength of the association varied depending on the different health indicators and was 2.4. Supplementary analyses also affected by sex, region, and time. The health outcome most strongly associated with EPT was depres­ As a sensitivity analysis, we repeated all the analyses, selecting the sion. Although not statistically significant, PR values were generally individuals in each country that were in the lowest income quartile. higher for LSVc than for HSVc in all three indicators. For example, in Income quartiles were based on equivalised household income in pur­ LSVc, women with EPT had 3.24 (CI: 2.24–4.33) times more depression chasing power parities in euros, which make countries comparable. We than those without EPT in 2007 and in the case of men, in 2012 those also ran the analyses by country. Country-specific results can be found in with EPT had 4.16 (CI:3.14–5.50) times more depression than those Table 1 Weighted prevalence (%) of energy poverty and health outcomes by energy poverty over time stratified by sex and region. LSV: countries with lower structural vulnerability; HSV: countries with higher structural vulnerability. In parentheses, number of cases; in bold if chi-square test assessing change over time of prevalence is statistically significant (p-value < 0.05). 3 L. Oliveras et al. Health and Place 67 (2021) 102294 Fig. 1. Energy poverty prevalence over time strati­ fied by sex and region. EPT: Energy poverty indicator “Cannot afford to keep home adequately warm; EPA: Energy poverty indi­ cator “One or more arrears in utility bills during the past 12 months”; LSV: countries with lower struc­ tural vulnerability; HSV: countries with higher structural vulnerability. Chi-square tests assessing change over time of EP percentage is statistically significant (p-value<0.05) in all cases. without EPT (subfigure 2c). Regarding the time trend, from 2007 to We observed a clear tendency over time, as the PAR%s increased 2012, point estimates generally showed an increase in the strength of the between 2007 and 2012 and decreased between 2012 and 2016. This association between the selected health indicators and EPT, which was “peak” pattern became more evident in reduced well-being (subfigure more marked in reduced well-being and depression in men in HSVc, 2e) and even more so in depression (subfigure 2f), and particularly in where the increase was statistically significant (subfigures 2b-c, p-value HSVc, where differences between years were generally statistically sig­ of the year interaction were 0.011 and 0.02 respectively). Between 2012 nificant (S2). and 2016, PRs tended to decrease in men in LSVc, but tended to remain stable or increase in men in HSVc and women in both LSVc and HSVc. 3.2. Arrears in utility bills during the past 12 months (EPA) 3.1.3. Population impact of energy poverty on health 3.2.1. Prevalence of energy poverty The PAR% to EPT of the three health outcomes are also shown in The prevalence of EPA stratified by sex, region and year is shown in Fig. 2 (subfigures d-f). Detailed data can be found in the supplementary Fig. 1b. Overall, the proportion of people experiencing EP was higher data (S2). EPT significantly affected the population’s health and this when we used the EPA indicator. Again, HSVc and women were more impact was clearly stronger in HSVc than in LSVc. For example, in 2012, affected by EPA. Regarding time trends, although we still identified a the proportion of cases of depression due to EPT was 15.46% (CI: “peak” pattern (Ʌ) over time in all cases, the EPA increase between 2007 10.90%–20.35%) in men living in LSVc and 29.87% (CI: 23.84%– and 2012 in LSVc was less marked, but the decrease between 2012 and 36.30%) in those living in HSVc. In women, this proportion in the two 2016 was still pronounced. macro regions was 11.17% (CI: 7.94%–14.75%) and 21.28% (CI: 16.73%–236.37%), respectively (subfigure 2f). 4 L. Oliveras et al. Health and Place 67 (2021) 102294 Fig. 2. Age-adjusted prevalence ratio (PR) of poor self-reported health status, reduced well-being and likely depression in people with EPT versus people without EPT over time and respective age-adjusted population attributable risk percent (PAR%) of the three health outcomes. Results are stratified by sex and region. EPT: energy poverty indicator “Cannot afford to keep home adequately warm; PR: age-adjusted prevalence ratio; PAR%: age-adjusted population attributable risk percent; LSV: countries with lower structural vulnerability; HSV: countries with higher structural vulnerability. 3.2.2. Association between energy poverty and health throughout the study period. The association between health outcomes and EPA is shown in Fig. 3 (subfigures a-c). Detailed data can be found in the supplementary in­ 4. Discussion formation (S3). Once more, there was an overall statistically significant association between EPA and the selected health indicators. Depression 4.1. Principal findings was again the most closely associated health outcome with EPA and that whose PR values experienced the most oscillation over time. The results of this study show that EP clearly increased in the EU Between 2007 and 2012, time tendencies were less homogeneous during the economic crisis and decreased in more recent years. It also between the groups. The strength of the association between EPA and evidences regional and sex inequities in the distribution of EP across the health outcomes tended to decrease in men in HSVc and especially in EU, with higher EP prevalence in women and in HSVc. The main LSVc populations, where the decrease was statistically significant in contribution of this study is the evidence generated on the effects of EP poor self-reported health status and reduced well-being in women and in on health and well-being. We found a significant association between EP depression in men. In women in HSVc, point estimates generally showed and poor self-reported health status, reduced well-being, and depres­ a constant increase throughout the study period. The increase in the sion. Analysis of changes over time showed that the strength of the as­ strength of the association between 2012 and 2016 was observed in all sociation between EPT and health increased from 2007 to 2012. Between groups, but was particularly marked in women in both LSVc and HSVc 2012 and 2016, the association between EPT and health tended to and in men in HSVc (generally, p-values of the year interaction<0.05). decrease in men in LSVc, but not in women in LSVc or men and women in HSVc, where the association remained stable or became stronger. 3.2.3. Population impact of energy poverty on health Time tendencies of the association between EPA and health were less The PAR% to EPA of the three health outcomes is shown in Fig. 3 homogenous between 2007 and 2012. But, between 2012 and 2016 (subfigures d-f). Detailed data can be found in the supplementary data there was a general increase in the strength of the association, which was (S4). Again, EPA had a significant negative impact on all health in­ particularly significant in men in HSVc and in women in both LSVc and dicators, but this impact evolved differently over time according to HSVc. Although the strength of the association between EP and health macro region and sex. Point estimates suggested a general constant in­ tended to be greater in LSVc, the impact of EP on the three health out­ crease in the impact of EPA on health in HSVc. The trends in LSVc were comes was clearly greater in HSVc. The impact of EPT on health followed less evident. The impact of EPA on health in women in LSVc decreased a “peak” pattern over time, increasing between 2007 and 2012 and between 2007 and 2012 and tended to increase after 2012. In men, in decreasing between 2012 and 2016. Analysis of the impact of EPA on contrast, the impact of EPA on health remained stable or decreased health showed a tendency to constantly increase in women and men in 5 L. Oliveras et al. Health and Place 67 (2021) 102294 Fig. 3. Age-adjusted prevalence ratio (PR) of poor self-reported health status, reduced well-being and likely depression in people with EPA versus people without EPA over time and respective age-adjusted population attributable risk percent (PAR%) of the three health outcomes. Results are stratified by sex and region. EPA: energy poverty indicator “One or more arrears in utility bills during the past 12 months”; PR: age-adjusted prevalence ratio; PAR%: age-adjusted population attributable risk percent; LSV: countries with lower structural vulnerability; HSV: countries with higher structural vulnerability. HSVc. In LSVc, the tendencies were less evident; however, between 2012 less active energy markets, among other factors. This baseline or struc­ and 2016 the impact of EPA on health tended to decrease in men and to tural situation in HSVc could explain their higher percentage of EP as increase in women, particularly regarding depression. well as the greater impact of the consequences of economic crises, which together seem to have pushed a substantial number of energy-vulnerable households into EP. 4.2. Interpretation of results Consistent with other studies, we found a clear association between EP and negative health outcomes. This association was stronger and To our knowledge, this is the first study analysing the trend in EP at more susceptible to the effect of the economic crisis in the mental the European level for the period 2007–2016 and including structural health dimension, particularly in depression. This may be due to the vulnerability to EP. The general “peak” pattern observed is consistent multiple, and sometimes short-term, mechanisms and pathways with the hypothesis that the structural determinants and the main in which EP affects mental health, which could have an exponential drivers of EP were directly affected by the economic crisis and effect (Liddell and Guiney, 2015). Examples include chronic subsequent austerity-driven economic policies implemented in the EU, thermal discomfort, which has a direct impact on mental health increasing the number of households that were unable to meet their (Gilbertson et al., 2012), concern that cold is damaging physical health energy needs. These findings support those of single country studies in (Tod et al., 2012), worry about energy bills (Anderson et al., 2012), the Spain (Tirado Herrero et al., 2016; Tirado Herrero and Jim� enez experience of falling into debt (or fear of it) (Tod et al., 2012), stigma Meneses, 2016), Greece (Dagoumas and Kitsios, 2014), and England within the person’s community (Anderson et al., 2012), and the absence (Hills, 2012). On the one hand, the number of people at risk of poverty or of any solution or sense of control over the problem (Liddell and Guiney, social exclusion increased after the beginning of the crisis, reaching a 2015). EP often goes hand with hand with housing insecurity and peak in 2012, and decreased again slightly between 2012 and 2014. On household financial debt, which are also associated with worse mental the other hand, in the EU, average electricity prices increased by 29% health, as well as with worse physical health. Both situations increased between 2005 and 2011, a far larger increment than in other regions during the economic crisis across the EU, and particularly in some of the such as the United States of America, where average electricity prices HSVc, such as Greece, Portugal, Cyprus, and Spain (Va �squez-Vera et al., increased in the same period by 5% (Csiba et al., 2016). Moreover, as 2016; Sweet et al., 2013). The economic crisis also increased unem­ argued by Recalde et al., HSVc are characterised by lower expenditure ployment rates in many EU countries. Several studies have shown, how on social protection, higher long-term unemployment rates (and policies unemployment can have negative consequences on mental health, and to address them based on unemployment benefits rather than activation these can be harder in the case of long-term unemployment and in measures), increased in-work poverty, poor dwelling efficiency and lack countries with weak unemployment protection systems and without of adequate heating systems in the housing stock, as well as by having 6 L. Oliveras et al. Health and Place 67 (2021) 102294 active-labour-market programmes, as is the case in HSVc (Karanikolos as ours but just for 2012 concluded that EP produced the largest in­ et al., 2013; Uutela, 2010). Finally, the healthcare sector was also sub­ equalities in health and well-being in Sweden and Slovenia (Thomson ject to austerity measures which varied from freezes to severe cuts. Some et al., 2017a). Calculating PAR%, we observed that Sweden was one of of the HSVc like Greece, followed by Portugal and Italy and Spain to a the EU27 countries with a lesser impact of EP on health and that, among lesser extent, were the most affected countries (Serapioni, 2017; Simou HSVc, Slovenia also showed one of the less significant impacts. In and Koutsogeorgou, 2014). All this could contribute to the observed contrast, in 2012, the countries where EP had a greater impact on health association between EP and worse health outcomes and to its trend over were Cyprus and Greece (S14-16), which are both considered as HSVc. It time. should be noted that impact measures presume causality between risk Surprisingly, we found that generally the association between EP and factors and health outcomes. Cross-sectional studies, as is the case in the negative health outcomes may be stronger in LSVc, particularly when we present study, do not allow us to prove causal relationship. used the EPT indicator. These results are in line with the findings of The results of this study show how the impact of EPT on health fol­ previous research, reporting that Slovenia, Sweden, the Netherlands, lowed a clear “peak” pattern over time. This “peak” pattern was more Luxembourg, and Denmark experienced the greatest difference in health pronounced in HSVc, which again suggests an increase in health in­ outcomes among people with and without EP (Thomson et al., 2017a; equalities among regions during crises. These results are in line with the Healy, 2004). LSVc have higher levels of median income and income study by Toffolutti and Suhrcke, who argue that social protection ex­ equity (Eurostat. Quality of life, 2018) and therefore the absolute penditures, which are generally greater in LSVc, appear to help countries deprivation is lower. However, substantial research suggests that this to “smooth” the health response to recession (Toffolutti and Suhrcke, unexpected result could be partly explained by relative deprivation, that 2014). The impact of EPA on health varied differently over time, is, individual-level socioeconomic comparison (Adjaye-Gbewonyo and depending on region, but also on gender. We observed a constant in­ Kawachi, 2012). Furthermore, LSVc shape the centre and north of crease in women and men in HSVc, but different time patterns between Europe, where cold temperatures during winter are more extreme. This the sexes in LSVc, where after 2012 the health impact of EPA decreased fact could explain why, when people live in EP, the effects on health are in men and increased in women, as in women and men in HSVc. Inter­ likely to be greater and is consistent with our observation that the EPT sectionality may be a valuable analytical tool to understand the ways in indicator was more sensitive to this effect. This should alert us that the which both axes of inequality, territory and gender, intersect with each traditional focus on cold when studying EP and the lack of studies other, resulting in specific realities that may affect health in different analysing the effect of hot temperatures in people with EP, could un­ ways. Several reports suggest that women are more affected by EP than derestimate the already greater impact of EP in Mediterranean EU men due to economic, biological/physiological, and socio-cultural fac­ Member States, which are all included in HSVc. Including the effect of tors (Clancy et al., 2017). Women with low incomes are disproportion­ hot temperature when studing EP could increase even further the ately found as heads of households either as single parent families or due observed regional inequalities between LSVc and HSVc. to their greater longevity than men, living alone at pensionable age. In As expected, we observed a general increase in the strength of the addition, EP could have a greater impact on women’s health because association between EPT and health during the first years of the eco­ they are considered more sensitive to ambient temperature and because nomic crisis. However, between 2012 and 2016, the trend in the asso­ women still bear the burden of care and housework, tasks that are en­ ciation followed different patterns among groups. On the one hand, men ergy intensive and imply spending more time at home (Gonzalez Pijuan, in LSVc tended to recover after 2012, but not women in LSVc or men and 2018; Brodolini, 2011). Women are also responsible for energy man­ women in HSVc, where the association remained stable or became agement in the household. This puts them in conflict with other family stronger. These findings suggest an increase in health inequalities be­ members (a situation that is exacerbated in times of austerity) and tween regions and genders. Although counterintuitive, the observed makes them responsible for dealing with unpayable debts or requesting decrease in the strength of the association between EPA and health in subsidies. These situations, which are strongly related to stress, feelings women and men in LSVc and men in HSVc during the first years of the of helplessness, depression and stigma, have been reported in studies in crisis has been reported in other studies analysing health trends during both LSVc (Roehr, 2002; Carlsson-Kanyama and Lind�en, 2007) and the economic crisis (Bartoll et al., 2015; Malmusi et al., 2018). A HSVc (Gonzalez Pijuan, 2018; Petrova, 2017). possible explanation could be that, in times of recession, if more people experience difficulty in paying utility bills, the pool of people living in 4.3. Methodological considerations, strengths and limitations EPA tends to include a larger fraction of healthy people, leading to a decrease in aggregated levels of poor health outcomes rather than an EP is a complex phenomenon whose measurement provokes broad increase. This underlying idea is consistent with our finding that the debate. In this study we used the consensual approach and addressed the association between EPA and health became stronger after 2012, when affordability dimension of EP. The lack of information on other di­ the percentage of people living in EPA was significantly lower and mensions of EP such as access, flexibility, energy efficiency, needs, or therefore may represent a more vulnerabilised and affected population. practices, may hamper understanding of the various causes underlying This observed increase was more significant in women in LSVc and EP and its effects on health. Because the indicators used in this study are women and men in HSVc, supporting the hypotheses that health in­ dichotomous, we could not capture intermittent versus persistent ther­ equalities may have increased between regions and genders during the mal discomfort or arrears in utility bills. This approach does not allow economic crisis. measurement of the intensity of EP in households, which could confound Although the association between EP and negative health outcomes the association between EP and health. However, the use of subjective seems to be higher in LSVc, when we analysed the impact of EP, the indicators is believed to incorporate a more comprehensive perspective pattern was reversed, with HSVc clearly being the most affected coun­ of EP and to be less complex to collect, allowing their use at the Euro­ tries. This is because PAR% is dependent not only on the strength of the pean level (Thomson et al., 2017b). Moreover, we analysed the two association between EP and health, but also on the proportion of in­ indicators separately because they do not measure the same issue and, dividuals exposed to EP, and therefore provides a measure of the public therefore, the mechanisms and pathways mediating their association health impact of EP on the entire population. This is essential to assess with health may differ. Further studies should also overcome the the potential impact and the cost-effectiveness of preventive in­ dominant focus in EP research on heating and examine other energy terventions at the population level (Rückinger et al., 2009). PAR% are needs in greater depth such as cooling, lighting, and the use of modern infrequently reported, which may lead to misinterpretation of the real appliances such as computers or phones. magnitude of the problem and to overlook the existing inequities among EP is considered a distinct deprivation condition and not only a different groups. For example, another study analysing the same dataset symptom of a broader problem of poverty (EPEE, 2009; Bollino and 7 L. Oliveras et al. Health and Place 67 (2021) 102294 Botti, 2017; Hills, 2012). Although there is a clear overlap and linkages poor health, especially poor mental health, which, combined with a high between income poverty and EP, EP should be considered separately prevalence of EP, translates into a high impact at the population level. because 1) not everyone who is energy poor is income poor, 2) some of The results show that EP and its impact on health tend to worsen during the negative health outcomes of EP are more immediate than those the economic crisis and identifies vulnerable groups such as women and caused by income poverty and 3) EP is more amenable to change than people living in HSVc. Further research should investigate the increasing income poverty, since factors other than income can be tackled to reduce regional and gender inequities in the distribution of EP and its impact on EP (Marmot Review Team, 2011). One example is Barcelona’s Energy health. EP is a complex phenomenon and its solution requires tackling Advice Points (EAP), a service to guarantee the rights to energy and its structural determinants. It is essential to put people and their health basic supplies, as well as to improve the energy efficiency of homes in and well-being at the centre and rather than conceptualizing energy and the city of Barcelona. The EAPs offer personalized services such as housing, among others, as market goods, recognizing them instead as guidance on improving the management and energy efficiency of rights. Health must be mainstreamed into all policies at European, na­ households or the support and processing of grants, fines or procedures tional and local level. with companies to optimize bills (change of power, rate or company, elimination of unnecessary additional services, etc.) (Ecoserveis, 2011; Contributors Memoria d’actuacio �. Punts, 2018). Another example is the Warm Front scheme in England, a programme to improve energy efficiency of private LO and MM contributed in the conception and design of the study. households in or at risk of EP, which showed positive impacts on alle­ All the authors contributed to the acquisition and interpretation of data. viating EP and therefore improving mental health, alleviating respira­ LO and MM performed the statistical analyses. All the authors contrib­ tory problems in children and reducing deaths of older people uted in the interpretation and the discussion of the results. LO wrote the (Gilbertson and Green, 2008). However, EP disproportionately affects first draft of the paper. All the authors critically revised the manuscript low-income households. To ensure that low income is not the only EP and approved the final version. The corresponding author attests that all driver that explains the association found between EP and negative listed authors meet authorship criteria and that no others meeting the health outcomes, we conducted a sensitivity analysis and repeated all criteria have been omitted. analyses for the lowest income quartile. We decided to stratify rather than adjust the analyses by income because income is one of the un­ Funding derlying causes of EP and from a conceptual perspective, we do not consider it appropriate to soften the health effects of EP through income The study is part of the PENSA Project (“Pobreza energ�etica y salud adjustment. This extra analysis for the lowest income quartile confirmed desde multiples perspectivas: ana �lisis individual, geogra �fico y tempo­ that the low-income population experiences more EP than the general ral”), which received a research grant from the Carlos III Institute of population. Regarding time trends, EP followed the same “peak” pattern Health, Ministry of Economy and Competitiveness (Spain), awarded on in LSVc as in the general population, but this was not the case in HSVc, in the 2015 call under the Health Strategy Action 2013–2016, within the which EP affecting a high proportion of women tended to remain stable National Research Program oriented to Societal Challenges, within the and increased in men to women’s levels. Strikingly, the association be­ Technical, Scientific and Innovation Research National Plan 2013–2016, tween EP and the three health outcomes remained statistically signifi­ with reference PI15/02006, co-funded with European Union ERDF cant in this subpopulation. This extra analysis therefore reaffirms that funds (European Regional Development Fund)”. The funders were not removing the effect of low income does not remove the clear association involved in the research and preparation of the article, including study between EP and negative health outcomes. We would have liked to design; collection, analysis, and interpretation of data; writing of the conduct a more in-depth examination of how socioeconomic factors article; nor in the decision to submit it for publication. modulate the relationship between EP and health but were unable to do so because the income variable showed 25.2% missing values, which Ethics approval would have invalidated the results. For this reason, the results for the lowest income quartile should also be interpreted with caution and are This study has been approved by the Clinical Research Ethics Com­ therefore not published in this paper. mittee of Hospital del Mar (2015/6155/l). In an increasingly globalised and consumer-oriented world, health issues and their determinants increasingly require a transnational Transparency declaration perspective (Kickbusch, 2010; Kickbusch et al., 2016). Rethinking boundaries when studying public health issues helps to elucidate the The lead authors (LO and MM) affirm that the manuscript is an underlying global relations of social determinants of health. A strength honest, accurate, and transparent account of the study being reported; of this study is that we analysed the relationship between EP and health that no important aspects of the study have been omitted; and that any across two macro regions, which share key characteristics among the discrepancies from the study as planned (and, if relevant, registered) structural determinants of EP. This has enabled us to obtain a broader have been explained. perspective of the relationship between EP and health and of its sus­ ceptibility to external factors such as the economic crisis. At the same Data sharing time, the economic crisis did not affect all EU countries equally and we are aware that every country has its particularities, which could be Dataset is available free of charge to those who intend to use it for overlooked by aggregating data. Single-country data is also essential to non-commercial purposes at the UK Data Service. advocate actions at the country level. For all these reasons, the whole analyses by country can be found in the supplementary information (S5–S19). Due to the sample size, we could not stratify the analyses by Declaration of competing interest sex. Given the observed sex inequalities, in-depth research using the gender perspective is needed at the country level. None. 5. Conclusions and recommendations Appendix A. Supplementary data This study clearly recognises EP as an important public health Supplementary data to this article can be found online at https://doi. problem across the EU. We found a strong association between EP and org/10.1016/j.healthplace.2020.102294. 8 L. Oliveras et al. Health and Place 67 (2021) 102294 References Hern� andez, D., 2016. Understanding ‘energy insecurity’ and why it matters to health. Soc. Sci. Med. 167, 1–10. https://doi.org/10.1016/j.socscimed.2016.08.029. Hills, J., 2012. Getting the measure of fuel poverty - final report of the fuel poverty Adam, S., Monaghan, R., 2016. Fuel poverty: what it means for young parents and their review: summary and recommendations. ISSN 1465–3001. families. http://www.ncb.org.uk/media/1242633/ncb_fuel_poverty_final_for_websi Karanikolos, M., Mladovsky, P., Cylus, J., et al., 2013. Financial crisis, austerity, and te.pdf, 1-16. health in Europe. Lancet 381 (9874), 1323–1331. https://doi.org/10.1016/S0140- Adjaye-Gbewonyo, K., Kawachi, I., 2012. Use of the Yitzhaki Index as a test of relative 6736(13)60102-6. deprivation for health outcomes: a review of recent literature. Soc. Sci. Med. 75 (1), Kickbusch, I., 2010. In: How Global Is Health ? what Do We Mean by Global Health ;? 129–137. https://doi.org/10.1016/j.socscimed.2012.03.004. 28th Int Congr Med Women’s Int Assoc.. Anderson, W., White, V., Finney, A., 2012. Coping with low incomes and cold homes. Kickbusch, I., Allen, L., Franz, C., 2016. The commercial determinants of health. Lancet Energy Pol. 49, 40–52. https://doi.org/10.1016/j.enpol.2012.01.002. Glob Heal 4 (12), e895–e896. https://doi.org/10.1016/S2214-109X(16)30217-0. Barros, A.J., Hirakata, V.N., 2003. Alternatives for logistic regression in cross-sectional Liddell, C., Guiney, C., 2015. Living in a cold and damp home: frameworks for studies : an empirical comparison of models that directly estimate the prevalence understanding impacts on mental well-being. Publ. Health 129 (3), 191–199. ratio Alternatives for logistic regression in cross-sectional studies : an empirical https://doi.org/10.1016/j.puhe.2014.11.007. comparison of models that direct. BMC Med. Res. Methodol. 13, 1–13. Liddell, C., Morris, C., 2010. Fuel poverty and human health: a review of recent evidence. Bartoll, X., Toffolutti, V., Malmusi, D., Pal�encia, L., Borrell, C., Suhrcke, M., 2015. Health Energy Pol. 38 (6), 2987–2997. https://doi.org/10.1016/j.enpol.2010.01.037. and health behaviours before and during the Great Recession, overall and by Malmusi, D., Muntaner, C., Borrell, C., et al., 2018. Social and economic policies matter socioeconomic status, using data from four repeated cross-sectional health surveys in for health equity: conclusions of the SOPHIE Project. Int. J. Health Serv. 48 (3), Spain (2001-2012). BMC Publ. Health 15 (1), 1–12. https://doi.org/10.1186/ 417–434. https://doi.org/10.1177/0020731418779954. s12889-015-2204-5. Marí-Dell’Olmo, M., Novoa, A.M., Camprubí, L., et al., 2017. Housing policies and health Bollino, C.A., Botti, F., 2017. Energy poverty in Europe: a multidimensional approach. inequalities. Int. J. Health Serv. https://doi.org/10.1177/0020731416684292. PSL Q Rev. 70 (283), 473. http://roble.unizar.es:9090/login?url¼http://search.ebs Marmot Review Team, 2011. The Health Impacts of Cold Homes and Fuel Poverty, vol. cohost.com/login.aspx?direct¼true&db¼edsdoj&AN¼edsdoj.b72110b91dd42f38b 342. https://doi.org/10.1136/bmj.d2807. 2d4106af79060a&lang¼es&site¼eds-live. Memoria d’actuaci� o, 2018. Punts d’Assessorament Energ� etic de Barcelona. Servei de Bosch, J., Pal�encia, L., Malmusi, D., Marí-Dell’Olmo, M., Borrell, C., 2019. The impact of detecci�o i reducci�o de la pobresa ener�etica i millora de l’efic�encia de les llars de les fuel poverty upon self-reported health status among the low-income population in persones vulnerables. Europe. Hous. Stud. https://doi.org/10.1080/02673037.2019.1577954. Office of the United Nations High Commissioner for Human Rights, 2009. The Right to Bouzarovski, S., 2014. Energy poverty in the European union: landscapes of Adequate Housing. https://doi.org/10.1057/dev.2014.64. vulnerability. Wiley Interdiscip Rev Energy Environ 3 (3), 276–289. https://doi.org/ Peralta, A., Camprubí, L., Rodríguez-Sanz, M., Basaga~ na, X., Borrell, C., Marí- 10.1002/wene.89. Dell’Olmo, M., 2017. Impact of energy efficiency interventions in public housing Bouzarovski, S., 2018. Energy Poverty (Dis)Assembling Europe’s Infrastructural Divide. buildings on cold-related mortality: a case-crossover analysis. Int. J. Epidemiol. 46 Palgrave/Macmilan, Manchester, UK. (4), 1192–1201. https://doi.org/10.1093/ije/dyw335. Bouzarovski, S., Tirado Herrero, S., 2017. The energy divide: integrating energy Petrova, S., 2017. Illuminating austerity: lighting poverty as an agent and signifier of the transitions, regional inequalities and poverty trends in the European Union. Eur. Greek crisis. Eur. Urban Reg. Stud. https://doi.org/10.1177/0969776417720250, Urban Reg. Stud. 24 (1), 69–86. https://doi.org/10.1177/0969776415596449. 096977641772025. Brodolini, Fondazione, 2011. Gender Equality in Caring Responsibilities over the Poortinga, W., Jones, N., Lannon, S., Jenkins, H., 2017. Social and health outcomes Lifecycle, pp. 1–17. September. following upgrades to a national housing standard: a multilevel analysis of a five- Carlsson-Kanyama, A., Lind�en, A.L., 2007. Energy efficiency in residences-Challenges for wave repeated cross-sectional survey. BMC Publ. Health 17 (1), 927. https://doi. women and men in the North. Energy Pol. 35 (4), 2163–2172. https://doi.org/ org/10.1186/s12889-017-4928-x. 10.1016/j.enpol.2006.06.018. Quesnel-Vall� ee, A., 2007. Self-rated health: caught in the crossfire of the quest for “true” Clancy, J., Daskalova, V., Feenstra, M., Franceschelli, N., Sanz, M., 2017. Gender health? Int. J. Epidemiol. 36 (6), 1161–1164. https://doi.org/10.1093/ije/dym236. perspective on access to energy in the EU. http://www.europarl.europa. Recalde, M., Peralta, A., Oliveras, L., et al., 2019. Structural energy poverty vulnerability eu/RegData/etudes/STUD/2017/596816/IPOL_STU(2017)596816_EN.pdf. and excess winter mortality in the European union: exploring the association Csiba, K., Bajomi, A., Gosztonyi, A.,� 2016. Energy poverty handbook. In: Greens/EFA between structural determinants and health. Energy Pol. 133. group of the European Parliament (Ed.), Brussels, Belgium: Greens/EFA Group of the Roehr, U., 2002. Gender and energy in the north. Society 1–19. European Parliament. Rückinger, S., Von Kries, R., Toschke, A.M., 2009. An illustration of and programs Dagoumas A, Kitsios F. Assessing the impact of the economic crisis on energy poverty in estimating attributable fractions in large scale surveys considering multiple risk Greece. Sustain Cities Soc. 2014;13:267–278. doi:10.1016/J.SCS.2014.02.004. factors. BMC Med. Res. Methodol. 9 (1), 1–6. https://doi.org/10.1186/1471-2288-9- Ecoserveis. Atlas of energy poverty initiatives in Europe. State-by-State Review. htt 7. ps://www.ecoserveis.net/en/atlas-of-energy-poverty-initiatives-in-europe-state-by- Serapioni, M., 2017. Crise econ^ omica e desigualdades nos sistemas de saúde dos países state-review/. do Sul da Europa. Cad. Saúde Pública 33 (9). https://doi.org/10.1590/0102- EPEE, 2009. Tackling Fuel Poverty in Europe. Recommendations Guide for Policy 311X00170116. Makers. http://www.fuel-poverty.org/files/WP5_D15_EN.pdf. Simou, E., Koutsogeorgou, E., 2014. Effects of the economic crisis on health and Espelt, A., Marí-Dell’olmo, M., Penelo, E., Bosque-Prous, M., 2017. Estimaci� on de la healthcare in Greece in the literature from 2009 to 2013: a systematic review. Health Raz�on de Prevalencia con distintos modelos de Regresi� on: ejemplo de un estudio Pol. 115 (2–3), 111–119. https://doi.org/10.1016/j.healthpol.2014.02.002. internacional en investigaci� on de las adicciones. Adicciones 29 (2), 105–112. Sweet, E., Nandi, A., Adam, E.K., Mcdade, T.W., 2013. Social Science & Medicine the https://doi.org/10.20882/adicciones.823. high price of debt : household fi nancial debt and its impact on mental and physical European Foundation for the Improvement of Living and Working Conditions, 2018a. health. Soc. Sci. Med. 91, 94–100. https://doi.org/10.1016/j. European Quality of Life Survey Integrated Data File, 2003-2016. [data Collection], socscimed.2013.05.009. third ed. https://doi.org/10.5255/UKDA-SN-7348-3. UK Data Service. SN: Thomson, H., Thomas, S., Sellstrom, E., et al., 2013. Housing improvements for health European Foundation for the Improvement of Living and Working Conditions. European and associated socioeconomic outcomes (Review). Cochrane Database Syst. Rev. 3 quality of life surveys (EQLS). https://www.eurofound.europa.eu/surveys/europe https://doi.org/10.1002/14651858.CD008657.pub2.Copyright. an-quality-of-life-surveys. (Accessed 26 September 2018). Thomson, H., Snell, C., Bouzarovski, S., 2017a. Health, well-being and energy poverty in Eurostat. Quality of life - material living conditions. https://ec.europa. Europe: a comparative study of 32 European countries. Int. J. Environ. Res. Publ. eu/eurostat/statistics-explained/index.ph Health 14 (6). https://doi.org/10.3390/ijerph14060584. p/Quality_of_life_indicators_-_material_living_conditions#undefined. Published Thomson, H., Bouzarovski, S., Snell, C., 2017b. Rethinking the measurement of energy 2018. Accessed October 5, 2018. poverty in Europe: a critical analysis of indicators and data. Indoor Built Environ. 26 Gefeller, O., 1990. Theory and application of attributable risk estimation in cross- (7), 879–901. https://doi.org/10.1177/1420326X17699260. sectional studies. Stat. Appl. 2 (4), 323–331. Tirado Herrero, S., Jim� enez Meneses, L., 2016. Energy poverty, crisis and austerity in Gilbertson, J., Green, G., 2008. Warm Front - Better Health: Health Impact Evaluation of Spain. People Place and Policy Online 10 (1), 42–56. https://doi.org/10.3351/ the Warm Front Scheme. Sheffield, UK. http://www.apho.org.uk/resource/item.as ppp.0010.0001.0004. px?RID¼53281. Tirado Herrero, S., Jim� enez Meneses, L., L� opez Fern�andez, J.L., Perero Van Hove, E., Gilbertson, J., Grimsley, M., Green, G., Grp, W.F.S., 2012. Psychosocial routes from Irigoyen Hidalgo, V.M., Pobreza, Savary P., 2016. vulnerabilidad y desigualdad housing investment to health: evidence from England’s home energy efficiency energ� etica. Nuevos Enfoques de An� alisis. scheme. Energy Pol. 49, 122–133. https://doi.org/10.1016/j.enpol.2012.01.053. Tod, A.M., Lusambili, A., Homer, C., et al., 2012. Understanding factors influencing Gillard, R., Snell, C., Bevan, M., 2017. Advancing an energy justice perspective of fuel vulnerable older people keeping warm and well in winter: a qualitative study using poverty: household vulnerability and domestic retrofit policy in the United Kingdom. social marketing techniques. BMJ Open 2 (4), 1–12. https://doi.org/10.1136/ Energy Res Soc Sci. 29, 53–61. https://doi.org/10.1016/j.erss.2017.05.012. bmjopen-2012-000922. February. Toffolutti, V., Suhrcke, M., 2014. Assessing the short term health impact of the Great Gonzalez Pijuan, I., 2018. Gender Inequality and Energy Poverty. A Forgotten Risk Recession in the European Union: a cross-country panel analysis. Prev. Med. 64, Factor. https://esf-cat.org/desigualdad-genero-y-pobreza-energetica-presentem-le 54–62. https://doi.org/10.1016/j.ypmed.2014.03.028. studi-castella/. Topp, C.W., Østergaard, S.D., Søndergaard, S., Bech, P., 2015. The WHO-5 well-being Healy, J.D., 2003. Excess winter mortality in Europe: a cross country analysis identifying index: a systematic review of the literature. Psychother. Psychosom. 84 (3), key risk factors. J. Epidemiol. Community Health 57 (10), 784–789. https://doi.org/ 167–176. https://doi.org/10.1159/000376585. 10.1136/jech.57.10.784. Uutela, A., 2010. Economic crisis and mental health. Curr. Opin. Psychiatr. 23 (2), Healy, J.D., 2004. Housing, Fuel Poverty and Health: A Pan-European. Ashgate 127–130. https://doi.org/10.1097/YCO.0b013e328336657d. Publishing, Aldershot, UK. 9 L. Oliveras et al. Health and Place 67 (2021) 102294 V� asquez-Vera, H., Rodríguez-Sanz, M., Pal�encia, L., et al., 2016. Foreclosure and health O’sullivan, K., Barnard, L.T., Viggers, H., Howden-Chapman, P., 2016. Child and youth in southern Europe: results from the platform for people affected by mortgages. fuel poverty: assessing the known and unknown. People, Place Policy People. Place J. Urban Health 93 (2), 312–330. https://doi.org/10.1007/s11524-016-0030-4. Policy J Compil 10 (11), 77–87. https://doi.org/10.3351/ppp.0010.0001.0006. Wright, F., 2004. Old and cold: older people and policies failing to address fuel poverty. Soc. Pol. Adm. 38 (5), 488–503. https://doi.org/10.1111/j.1467-9515.2004.00403. x. 10