SSM - Population Health 7 (2019) 100367 Contents lists available at ScienceDirect SSM - Population Health journal homepage: www.elsevier.com/locate/ssmph Article Inequalities in health and health behaviours between couple and lone T mothers before and during the financial crisis in Spain (2003–2012) Sara Trujillo-Alemána,b,1, Glòria Péreza,b,c, , Vanessa Puig-Barrachinaa, Mercè Gotsensa, ⁎ Jillian Reynoldsd, Silvia Ruedae, Carme Borrella,b,c a Agència de Salut Pública de Barcelona, Pl. Lesseps, 1, Barcelona 08023, Spain b Universitat Pompeu Fabra, C/ Dr. Aiguader, 88, Barcelona 08003, Spain c CIBER Epidemiología y Salud Pública (CIBERESP), C/ Monforte de Lemos, 3-5, Madrid 28029, Spain d Agency for Health and Quality Assessment of Catalonia (AQuAS), C/ Roc Boronat, 81-95, Barcelona 08005, Spain e DEP Institut, C/Aragó, 631-633, Barcelona 08026, Spain ARTICLE INFO ABSTRACT Keywords: Lone mothers report worse health and adopt more risky health behaviours than couple mothers, as largely Spain documented in several European countries, but not deeply in Spain. The primary aim of this study was to identify Lone mothers the possible existence of inequalities in health and health behaviours between couple and lone mothers in Spain Health inequalities by occupational social class and employment status. A second aim was to explore whether any inequalities were Health behaviours influenced by the economic crisis beginning in 2008, analysing changes in inequalities between 2003–2004 and Crisis 2011–2012. Social class Two waves of the cross-sectional Spanish National Health Survey data were used. Analyses were restricted to mothers aged 16–64 years, with at least one child aged 18 years or younger. The sample consisted of 2982 mothers in 2003–2004 and 3070 in 2011–2012, representing more than 80% of couple mothers. Two health outcomes and two health behaviour measurements were used. Robust Poisson regression was run to estimate inequalities between couple and lone mothers, calculating prevalence ratios adjusted by age and stratified by social class and employment status. We found inequalities in health and health behaviours between couple and lone mothers in Spain amongst the manual social class, with lone mothers reporting a more than 30% higher prevalence of poor self-perceived health and being smoker in both time points of study compared with couple mothers. Furthermore, lone mothers were at 50% higher risk of having at least one selected chronic condition and 86% higher probability of sleeping less than 6 hours/day in 2011–2012 This study could not confirm that inequalities between couple and lone mothers changed in Spain during the study period, although some patterns were noticeable. Inequalities pointed towards an increase amongst mothers in the manual social class with paid employment, while inequalities amongst unemployed mothers (both manual and non-manual social class) pointed towards a decrease. 1. Background et al., 2012; Rousou, Kouta, Middleton, & Karanikola, 2013; Whitehead et al., 2000). Changes in health inequalities between couple and lone Lone mothers are one of the most disadvantaged groups from both mothers have also been studied across Europe (Burström, Diderichsen, the social and the health perspective (Whitehead, Burström, & Shouls, & Whitehead, 1999; Fritzell & Burström, 2006; Fritzell, Diderichsen, 2000). They are at higher risk of poverty, unemployment Ringbäck Weitoft, Fritzell, & Burström, 2007). However, no scientific and social exclusion (Di Nella, Almeda Samaranch, & Ortiz Monera, studies have focused on Spanish lone mothers or analysed changes in 2014; Fritzell & Burström, 2006; Whitehead et al., 2000), and they inequalities during a period of financial crisis. show worse health outcomes than couple mothers, as largely docu- mented in several European countries (Burström et al., 2010; Fritzell Corresponding author at: Agència de Salut Pública de Barcelona, Pl. Lesseps, 1, 08023 Barcelona, Spain. ⁎ E-mail addresses: struale@gobiernodecanarias.org (S. Trujillo-Alemán), gperez@aspb.cat (G. Pérez), vpuig@aspb.cat (V. Puig-Barrachina), mgotsens@aspb.cat (M. Gotsens), jmreynolds@gencat.cat (J. Reynolds), srueda@dep.net (S. Rueda), cborrell@aspb.cat (C. Borrell). 1 Present address: Dirección General de Programas Asistenciales, Plaza Dr. Juan Bosch Millares, 1, Las Palmas de Gran Canaria, 35004, Spain. https://doi.org/10.1016/j.ssmph.2019.100367 Received 19 November 2018; Received in revised form 23 January 2019; Accepted 27 January 2019 2352-8273/ © 2019 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/). S. Trujillo-Alemán et al. SSM - Population Health 7 (2019) 100367 1.1. Spanish socioeconomic and political context 26.7% vs 25.6%) (European Commission, 2018). Initially, the con- struction sector was more affected by the economic crisis, what gen- Although the one-parent family model is still less widespread in erated an increase in the unemployment rate amongst men, but then, Spain than in most other European countries, the number of households the increase in the unemployment rate within the service sector affected constituted by this type of family model has increased by roughly three- women to a greater extent (Instituto Nacional de Estadística, 2018). In fold in the last 20 years (European Commission, 2018). This phenom- Spain, part-time jobs are usually precarious and mainly related to the enon could be related to the lower marriage rate, the higher divorce service sector. Most of jobs in Spain are full-time jobs (70.5%), with rate and the increasing number of births outside marriage. Another differences between men (79.6%) and women (59.3%) (Servicio factor is the progressive incorporation of women into the labour market Público de Empleo Estatal, 2011). However, part-time jobs show the in the last few decades. Paid work provides economic empowerment, opposite situation, while amongst men the proportion of part-time jobs thus improving the balance of power between men and women and is of 19.7%, for women is of 39.2% (Servicio Público de Empleo Estatal, providing women with better chances to end a relationship through 2011). In addition, the social protection system in Spain only provides separation or divorce. In addition, employment enables economically coverage to the 77.4% of people who are unemployed (Servicio Público independent women to be lone mothers without having a stable partner de Empleo Estatal, 2011). Furthermore, in 2013, 46.4% of Spanish (Artazcoz et al., 2014). These are lone mothers by choice, who are workers earned less than €1000 per month, being younger workers the probably in an advantaged socioeconomic position and have the fi- most affected by these low salaries, illustrating the impact that the nancial resources to decide to have a child without a partner. However, economic crisis has had on working conditions in Spain (Eurofound, employment may also increase the difficulty of balancing work and 2015). family life, especially if family support policies are scarce, as in Spain, Much research has focused on describing and analysing inequalities where state welfare policies are less developed than in other European in health by social class in Spain. People belonging to disadvantaged countries, and social protection is largely based on family support. In social classes have shown worse living and working conditions, which addition, when there is limited co-responsibility from the welfare state, translates into poor health (Comisión para Reducir las Desigualdades the ability to maintain a family, especially for lone mothers, will de- Sociales en Salud en España, 2012). The negative impact of un- pend on their socioeconomic position or social class and their own re- employment on overall health and mental health has also been studied sources. in the Spanish working-age population, this impact being particularly According to Korpi, Spain is considered to have a ‘traditional family high for the long-term unemployed (Urbanos-Garrido & Lopez- policy model’ characterised by supporting the father as the main Valcarcel, 2015). In addition, the psychological effects of unemploy- breadwinner and encouraging the mother to stay at home to care for ment have been aggravated by the economic crisis (Urbanos-Garrido & children (Korpi, 2000). It consists of a solidarity model, with high levels Lopez-Valcarcel, 2015). When economic conditions worsen, health also of family support and low levels of social protection for children and deteriorates. In times of economic crisis, a deterioration in health and families. The welfare state only intervenes to tackle highly critical si- an increase in health inequalities is expected, especially amongst more tuations and the nuclear and extended family is expected to provide disadvantaged groups (Bacigalupe & Escolar-Pujolar, 2014). The rise in support in childcare (Korpi, 2000). Korpi’s description of Spain as a social inequalities in health during crises could be explained by in- ‘traditional family policy model’ is exemplified by expenditure on social creasing social inequality and poverty, greater job insecurity and higher protection. In Spain in 2014, the total expenditure on social protection unemployment rates. per inhabitant in terms of purchasing power standard (PPS)2 per in- habitant was almost 2000 PPS lower than the European Union average 1.2. Lone mothers: a disadvantaged group (6177.0 versus 7912.8), and was much lower than in countries with more developed welfare states such as Sweden (9856.1) or Norway Lone mothers, considered as those mothers not living together with (12123.0). Considering only the expenditure on social benefits for fa- a partner, are one of the most disadvantaged social groups in many milies and children, Spain spent 318.0 PPS per capita, while the Eur- countries (Rousou et al., 2013). They are at higher risk of poverty and opean average was more than double this figure (652.2) (European are more frequently unemployed than couple mothers (Fritzell & Commission, 2018). Although Korpi’s model was described many years Burström, 2006; Whitehead et al., 2000). From a health perspective, ago (Korpi, 2000), the welfare state has not expanded in Spain in recent lone mothers show worse health outcomes than those mothers who are years. This can be partly attributed to the financial crisis in the country either married or living with a partner (couple mothers). They report since the second quarter of 2008 and the consequent cuts in social worse self-perceived health (Burström et al., 2010; Fritzell et al., 2012; services (Bacigalupe, Shahidi, Muntaner, Martín, & Borrell, 2016). Lahelma, Arber, Kivelä, & Roos, 2002) and experience more limiting Socioeconomic and political context influence the pathways for the long-standing illnesses (Fritzell et al., 2007). Lone mothers also have generation of health inequalities amongst different social groups more mental health problems (Bull & Mittelmark, 2009; Dziak, Janzen, (Diderichsen, Evans, & Whitehead, 2001; Whitehead et al., 2000). The & Muhajarine, 2010) and often adopt risky health behaviours. Poverty unemployment rate for both men and women has increased during the and unemployment influence mothers’ living conditions and what Spanish economic crisis. While in 2007, before the economic crisis, the choices they make in life, both in relation to standard of living and total unemployment rate in Spain was 8.2%; in 2009, after the starting health behaviours (Fritzell et al., 2007). Lone mothers are more fre- of the economic crisis, this figure reached 17.9%, experiencing the quently smokers, and more often indicate drug or alcohol abuse in higher rate in 2013 (26.1%). In addition, unemployment rate has been comparison to couple mothers (Fritzell et al., 2012, 2007; Sperlich, persistently higher amongst Spanish women in comparison to men, Maina, & Noeres, 2013). either before (2007: 8.6% vs 6.4%) or during the economic crisis (2013: These inequalities could be partially explained by the socio- economic disadvantage, financial hardship and lack of social support possibly experienced by lone mothers (Bull & Mittelmark, 2009; Rousou 2 et al., 2013). Separated or divorced lone mothers could receive child The purchasing power standard (PPS) is an artificial currency unit. Prices support from ex-partners, helping them to face lone motherhood. are different across countries, so different amounts of national currency units are needed for the same goods and services according to the country. However, However, many lone mothers are the main providers and experience an one PPS allows to buy the same amount of goods and services in each country increased risk of poverty since they may rely on a single income making it possible to compare the total expenditure on social protection or the (Fritzell & Burström, 2006; Fritzell et al., 2007). Lone mothers are also expenditure on social benefits for families and children, across different the main carers, since there may not be co-responsibility from a partner European countries. in the domestic sphere. Therefore, as the sole providers and carers in 2 S. Trujillo-Alemán et al. SSM - Population Health 7 (2019) 100367 the household, lone mothers may depend on support systems from the years, with at least one child aged 18 years or younger cohabiting in the welfare state to a greater extent than couple mothers (Fritzell & same household. The overall study population included 2951 mothers Burström, 2006; Fritzell et al., 2007). Public policies in Spain are still in 2003–2004 and 2698 in 2011–2012 (weighted subsample: 2982 and too focused on heterosexual couples, with only one breadwinner, and 3070, respectively). Lone mothers who were not household heads were based on the sexual division of work. This reinforces the difficulties of then excluded, because of their different sociodemographic character- balancing personal time, paid and unpaid work time and family time istics and the very small sample size. Analyses stratified by social class (Almeda & Di Nella, 2011). were restricted to those mothers without missing values in this variable, Several studies on health inequalities between couple and lone while analyses stratified by social class and employment status were mothers have been published in different European countries con- also restricted to those who were employed, unemployed or house- cerning different types of welfare regimes (Berkman et al., 2015; wives. Burström et al., 2010; Roos, Burström, Saastamoinen, & Lahelma, 2005; Van de Velde, Bambra, Van der Bracht, Eikemo, & Bracke, 2014). Some 2.2. Health outcomes and health behaviours of them have also considered changes in inequalities in times of eco- nomic downturn (Burström et al., 1999; Fritzell & Burström, 2006; Four outcome measurements were used: two health outcomes and Fritzell et al., 2007). However, no published scientific studies have two health behaviours. focused on mothers living in Spain, except for a report by Reynolds et al. in the context of the EUROTHINE Project (Reynolds, Rueda, & Lahoz, 2007). They observed small differences in health outcomes be- 2.2.1. Self-perceived health tween lone mothers-household heads and couple mothers, probably This variable has been demonstrated to be a reliable measure of influenced by an overrepresentation of lone mothers in the higher so- objective health, as well as a good predictor of mortality (Idler & cioeconomic positions. Large differences between couple and lone Benyamini, 1997). The respondents were asked whether they assess mothers of lower socioeconomic positions were found in contrast to their health as ‘very good, ‘good’, ‘fair’, ‘poor’ or ‘very poor’. The three small or inexistent inequalities between lone and couple mothers with latter categories were combined to yield a measure of ‘poor self-per- more socioeconomic resources (Reynolds et al., 2007). In addition, in- ceived health’. equalities between lone mothers-household heads and lone mothers- non-household heads were found in part due to differences in the age 2.2.2. Selected chronic conditions distribution, the level of responsibility of maintaining the household The questionnaire listed a wide range of illnesses and participants and the degree in which they receive support from other member in the were asked whether they have suffered any in the last 12 months, ac- household (Reynolds et al., 2007). Considering the emergence of the cording to a medical diagnosis. This variable was constructed by se- economic crisis in Spain in 2008 and the consequent increase in un- lecting diabetes, peptic ulcer, heart diseases and bronchitis, taking into employment rate, it would be interesting to explore whether these in- account the analyses conducted by Rius et al. which confirmed the equalities have changed taking into account not only the occupational association between comorbidity and the risk of death based on a co- social class, but also the employment status. Therefore, the primary aim morbidity index (Rius et al., 2004). Illnesses were selected according to of the present study was to determine the possible existence of in- their inclusion in the work of Rius et al. and considering if these ill- equalities in health and health behaviours between couple and lone nesses were listed in both waves of the Spanish National Health Survey mothers in Spain, by occupational social class and employment status. analysed in this study (2003–2004 and 2011–2012) (Instituto Nacional A second objective was to explore whether any inequalities were in- de Estadística, 2017). If the answer was affirmative to any of these fluenced by the economic crisis beginning in 2008, analysing changes in questions, the respondent was considered to have ‘at least one selected inequalities between 2003–2004 and 2011–2012. chronic condition’. 2. Methods 2.2.3. Smoking behaviour 2.1. Study design and sample Participants were asked whether they currently smoke, with the category of ‘smokers’ both daily and non-daily smokers. Although self- A comparative study was undertaken analysing data from two reported smoking tends to underestimate smoking status, both have waves of the Spanish National Health Survey, taking place respectively been shown to have a satisfactory correlation (Gorber, Schofield- in 2003–2004 (before the financial crisis) and 2011–2012 (during the Hurwitz, Hardt, Levasseur, & Tremblay, 2009). crisis). This survey is a cross-sectional nationwide survey that examines a representative sample of the Spanish non-institutionalised population. 2.2.4. Sleeping pattern The survey provides self-reported information on health status, mor- This variable was constructed by asking the respondents how many bidity, health behaviours and use of healthcare services, including so- hours they usually sleep per day. No clear cut-off point was found for ciodemographic data. The sampling method consisted of a stratified categorisation in the existing literature, and so participants were di- sampling in three stages. First stage units were the census sections. vided into those sleeping ‘6 or more hours/day’ and those sleeping ‘less Second stage units were the main family homes. Third stage units were than 6 hours/day’ (Alvarez & Ayas, 2004). selected considering the list of people in the household, since in each family home one adult was selected and one child in case there were any. It achieved a sample size of 21650 adults in 2003–2004 and 26502 2.3. Type of motherhood: main explanatory variable adults in 2011–2012. The global response rate was 67% in 2003–2004 and 71% in 2011–2012, although those households with no response A ‘couple mother’ was defined as a mother who was either married were replaced. Therefore, the total effective sample was 98% and 89% or cohabiting with a partner. A ‘lone mother-household head’ did not of the theoretical sample in 2003–2004 and 2011–2012, respectively. cohabit with a partner and was the reference person, meaning that she More details of the methodology of the survey are described elsewhere was the household member who most regularly (not occasionally) (Instituto Nacional de Estadística, 2017). No consent statement from contributed to the household budget to cover expenses. A ‘lone mother- participants was necessary since all microdata were anonymised and non-household head’ did not cohabit with a partner and was not the openly available online. head of the household, since the main breadwinner in the household For this study, a subsample was selected of all mothers aged 16–64 could be other relatives such as her parents. 3 S. Trujillo-Alemán et al. SSM - Population Health 7 (2019) 100367 2.4. Sociodemographic variables and those from the manual social class was observed amongst lone mothers-non-household heads. The proportion of mothers in the Age was considered a categorical variable in the descriptive analysis manual social class decreased for couple mothers (from 61.9% to (mothers aged 16–34 and 35–64 years) and as an independent con- 56.7%) between the two data points and increased for lone mothers- tinuous variable for robust Poisson regression analyses. household heads (from 46.6% to 58.4%) and lone mothers-non-house- Occupational social class, assigned according to current or most re- hold heads (from 72.2% to 84.7%). A high proportion of lone mothers- cent occupation (Fritzell & Burström, 2006), was collapsed into ‘non- household heads were employed in 2003–2004 (81.1%) compared with manual social class’ and ‘manual social class’ with a widely used couple mothers (47.2%). Unemployment rates increased in 2011–2012 Spanish adaptation of the British classification system (Domingo- for the three groups of mothers, and to a greater extent for lone mo- Salvany et al., 2013; Domingo-Salvany & Marcos Alonso, 1989). thers, both household heads (from 8.6% to 24.0%) and non-household Employment status was based on a person’s current main activity. heads (from 15.7% to 43.6%). ‘Employed’, ‘unemployed’, ‘housewives’ and ‘other situations’ were Some differences were noted amongst lone mothers. While ap- distinguished. ‘Employed’ included women in paid employment, while proximately 20% of lone mothers-household heads were young in both ‘unemployed’ referred to those women who were not in paid employ- 2003–2004 and 2011–2012, this figure reached 60–70% for lone mo- ment but in active employment search. The category of ‘housewives’ thers-non-household heads. Most lone mothers-non-household heads represented all those women whose main activity was housework and belonged to the manual social class both in 2003–2004 (72.2%) and ‘other situations’ included those who had retired or had taken early 2011–2012 (84.7%), in a proportion around 25% points higher than retirement, were studying and unable to work or who could not be lone mothers-household heads (46.6% and 58.4%, respectively in classified into the other categories. This category included women who, 2003–2004 and 2011–2012). Finally, lone mothers-non-household without exercising any paid activity and who were not self-employed, heads were also more frequently unemployed and were housewives, received income from renting a property and/or from other invest- showing large differences in employment rate, especially in 2011–2012, ments; women temporarily deprived of their freedom, and those who, when almost 70% of lone mothers-household heads were employed without exercising economic activity, received public or private assis- compared with approximately 30% of lone mothers-non-household tance. heads. Age was the adjustment variable, while occupational social class and employment status were stratification variables. 3.2. Health and health behaviours 2.5. Statistical analyses The prevalence of poor health outcomes and health behaviours is shown in Table 2. For both time periods, the prevalence was mostly First, a bivariate analysis was carried out to test for differences higher amongst lone mothers-household heads, except for the pre- amongst the three types of motherhood in sociodemographic and out- valence of smoking, which was highest amongst lone mothers-non- come variables in both 2003–2004 and 2011–2012. Chi-square tests household heads. Compared with the first study period, the prevalence with corresponding p-values were used. To estimate inequalities in of poor self-perceived health was lower for couple mothers and lone health and health behaviours between lone and couple mothers, age- mothers-household heads in 2011–2012, but not for lone mothers-non- adjusted prevalence ratios (aPR) with 95% confident intervals (95% CI) household heads. The proportion of mothers having at least one se- were calculated using robust Poisson regression, stratified first by oc- lected chronic condition and sleeping less than 6 hours/day was higher cupational social class, and then by both occupational social class and in 2011–2012 than in 2003–2004 for all three groups of mothers. In employment status. Poisson regression models with robust variance are addition, for the latter period, the prevalence of smokers was lower in based on the Huber sandwich estimator implemented in the statistical all groups of mothers, with lone mothers-non-household heads showing software Stata®. This variance estimator has shown to be correct and higher differences in prevalence between 2003–2004 and 2011–2012 robust in different experimental situations such as in the use of different (from 78.1% to 42.3%). However, it should be noticed that most of the values of prevalence (low, moderate or high prevalence) or estimating results are not statistically significant. several models (crude and adjusted) (Barros & Hirakata, 2003; Espelt, Marí-Dell’Olmo, Penelo, & Bosque-Prous, 2016). Interaction terms be- 3.3. Inequalities in health and health behaviours between couple and lone tween motherhood type and survey year were included to identify mothers: Interaction amongst social classes possible changes in inequalities in health and health behaviours be- tween couple and lone mothers between the two time periods. Sig- Different patterns of inequalities in health and health behaviours nificant interaction would indicate that the associations between type were found between couple and lone mothers-household heads in of motherhood and outcome measures changed during the study years. Spain, depending on occupational social class, since statistically sig- Sampling weights were applied in all the calculations to achieve po- nificant results were found amongst the manual social class, but not pulation representativeness. All analyses were performed using Stata® amongst the non-manual social class (Table 3). From now on, when v.13. describing results for lone mothers, we will be referring to lone mo- thers-household heads, unless otherwise specified. Analyses stratified 3. Results by occupational social class are presented below. Amongst mothers belonging to the manual social class, lone mothers 3.1. General description of the population showed statistically significant worse health outcomes and health be- haviours than couple mothers. In 2003–2004, lone mothers reported a The sociodemographic characteristics of the three types of mothers 30% higher prevalence of poor self-perceived health (aPR = 1.33; 95% are shown in Table 1. In both periods (2003–2004 and 2011–2012), the CI = 1.06–1.66) and being smokers (aPR = 1.30; 95% CI = 1.08–1.57) distribution showed the same pattern, with a higher proportion of compared with couple mothers. Comparing the two time points of couple mothers (around 84%) than lone mothers overall (around 16%). study, changes in inequalities between lone and couple mothers were Differences between the two groups of lone mothers were observed over not observed amongst the manual social class in 2011–2012 (aPR = time, with a decrease in the proportion of lone mothers-household 1.36; 95% CI = 1.05–1.77, for self-perceived health; and aPR = 1.34; heads (from 14.4% to 12.1%), and an increase in lone mothers-non- 95% CI = 1.08–1.66, for smoking behaviour). On the other hand, household heads (from 1.7% to 3.7%). having at least one selected chronic condition and sleeping less than The highest proportion of young participants, unemployed mothers 6 hours/day showed a higher increase in inequalities between lone and 4 S. Trujillo-Alemán et al. SSM - Population Health 7 (2019) 100367 Table 1 General description of the study population, by type of motherhood: Spanish Health Survey, 2003–2004 and 2011–2012. Survey year 2003–2004, % SURVEY YEAR 2011–2012, % Couple Lone mothers- Lone mothers-non- Couple Lone mothers- Lone mothers-non- mothers household heads household heads mothers household heads household heads (n=2502) (n=429) (n=51) p-value (n=2586) (n=370) (n=114) p-value Age groups (years) 16–34 28.1 19.5 60.1 < 0.001 24.7 23.0 68.4 < 0.001 35–64 71.9 80.5 39.9 75.3 77.0 31.6 Employment status Employed 47.2 81.1 68.3 < 0.001 57.7 67.5 33.8 < 0.001 Unemployed 11.3 8.6 15.7 18.9 24.0 43.6 Housewife 39.9 6.5 10.2 21.8 3.8 14.9 Other situationsa 1.6 3.8 5.8 1.6 4.7 7.7 Occupational social class Non-manual 37.7 50.8 27.8 < 0.001 42.2 35.0 13.4 < 0.001 Manual 61.9 46.6 72.2 56.7 58.4 84.7 Missing values 0.4 2.5 0.0 1.1 6.5 1.9 a This category includes mothers who were retired or had taken early retirement, were studying and unable to work or who could not be classified into the other categories, especially, women who, without exercising any salaried activity and who were not self-employed, received income from property rents and/or from other investments; women temporarily deprived of their freedom; and those who, without exercising economic activity, received public or private assistance. couple mothers amongst the manual social group. During the first time 3.4. Inequalities in health and health behaviours between couple and lone period, lone mothers had almost the same risk of having at least one mothers: Interactions amongst social classes and employment statuses selected condition (aPR = 1.04; 95% CI = 0.62–1.76) but were at 50% higher risk in 2011–2012 (aPR = 1.51; 95% CI = 1.04–2.20) compared The patterns of association between type of motherhood and health with couple mothers. Similarly, the higher probability amongst lone outcomes or health behaviours became more complex when, in addition mothers of sleeping less than 6 hours/day increased from 65% (aPR = to occupational social class, employment status was considered 1.65; 95% CI = 1.03–2.64) to 86% (aPR = 1.86; 95% CI = 1.12–3.08). (Table 4). However, changes in inequalities were not statistically significant (p- On the one hand, amongst mothers in the manual social class, lone value for interaction terms between type of motherhood and survey mothers from the three groups of employment status showed higher year > 0.05). risks of poor health outcomes and health behaviours than couple mo- Some non-statistically significant patterns were observed between thers. Statistically significant results were observed mainly amongst couple and lone mothers belonging to the non-manual social class. On mothers from the manual social class who were employed. the one hand, lone mothers showed a higher risk of poor health beha- On the other hand, amongst mothers from the non-manual social viours than couple mothers. They were approximately 25% more likely class, lone mothers also showed worse health behaviours in all three to be smokers both in 2003–2004 (aPR = 1.26; 95% CI = 1.00–1.59) groups of employment status. However, different patterns were found and in 2011–2012 (aPR = 1.23; 95% CI = 0.95–1.58). They were also for health outcomes. Lone mothers from the non-manual social class more likely to sleep less than 6 hours/day, showing different patterns with paid employment reported better self-perceived health and a lower between the two time periods. While lone mothers belonging to the risk of having at least one selected chronic condition during both the non-manual social class were 50% more likely to sleep less than time periods compared with couple mothers. 6 hours/day than couple mothers in 2003–2004 (aPR = 1.48; 95% CI Considering changes in inequalities between 2003–2004 and = 0.67–3.24), this risk was less than 10% higher in 2011–2012 (aPR = 2011–2012, some patterns were found in inequalities over time, al- 1.08; 95% CI = 0.54–2.17). However, lone mothers did not show a though there were non-statistically significant results for the interaction higher risk for adverse health outcomes both in 2003–2004 and between type of motherhood and survey year. First, inequalities be- 2011–2012. Compared with couple mothers, lone mothers reported tween lone and couple mothers showed a non-significant increase better self-perceived health in 2003–2004 (aPR = 0.79; 95% CI = amongst mothers from the manual social class in paid employment for 0.56–1.13; no differences in 2011–2012) and a lower risk of having at all outcomes. In contrast, no clear changes were discernible amongst least one selected chronic condition in 2011–2012 (aPR = 0.59; 95% CI mothers from the non-manual social class in paid employment as in- = 0.32–1.07; no differences in 2003–2004). equalities increased, decreased or remained stable, depending on the Table 2 Prevalence of poor health outcomes and health behaviours, by type of motherhood: Spanish Health Survey, 2003–2004 and 2011–2012. Survey year 2003–2004, % Survey year 2011–2012, % Couple Lone mothers- Lone mothers-non- Couple Lone mothers- Lone mothers-non- mothers household heads household heads mothers household heads household heads (n=2502) (n=429) (n=51) p-value (n=2586) (n=370) (n=114) p-value Poor self-perceived 26.4 28.8 14.8 0.215 20.3 27.5 19.1 0.055 health At least one selected 7.1 7.4 3.8 0.677 11.6 14.5 17.3 0.228 chronic condition Smoker 35.4 42.7 78.1 < 0.001 29.5 39.1 42.3 0.004 Sleeping less than 6.0 9.1 7.6 0.186 6.3 11.0 11.0 0.024 6 hours/day 5 S. Trujillo-Alemán et al. SSM - Population Health 7 (2019) 100367 Table 3 Associations between type of motherhood and poor health outcomes and health behaviours, by occupational social class: Spanish Health Survey, 2003–2004 and 2011–2012. Changes in inequalities between couple and lone mothers-household heads. Non-manual social class Manual social class 2003–2004 2011–2012 2003–2004 2011–2012 (n = 1162) (n = 1221) (n = 1748) n = (1683) % aPR (95% CI) % aPR (95% CI) p-value inter % aPR (95% CI) % aPR (95% CI) p-value inter Poor self-perceived health Couple mothers 20.9 1.00 14.9 1.00 29.5 1.00 24.6 1.00 Lone mothers-household heads 17.0 0.79 (0.56–1.13) 15.2 1.00 (0.66–1.52) 0.408 42.2 1.33 (1.06–1.66)* 34.0 1.36 (1.05–1.77)* 0.883 At least one selected chronic condition Couple mothers 5.5 1.00 10.4 1.00 7.8 1.00 12.7 1.00 Lone mothers-household heads 5.6 1.00 (0.52–1.91) 6.2 0.59 (0.32–1.07) 0.243 8.9 1.04 (0.62–1.76) 19.5 1.51 (1.04–2.20)* 0.253 Smoker Couple mothers 29.5 1.00 27.3 1.00 39.2 1.00 31.2 1.00 Lone mothers-household heads 37.3 1.26 (1.00–1.59) 33.5 1.23 (0.95–1.58) 0.879 49.4 1.30 (1.08–1.57)** 41.3 1.34 (1.08–1.66)** 0.862 Sleeping less than 6 hours/day Couple mothers 3.2 1.00 5.4 1.00 7.8 1.00 7.0 1.00 Lone mothers-household heads 4.8 1.48 (0.67–3.24) 5.9 1.08 (0.54–2.17) 0.559 14.0 1.65 (1.03–2.64)* 13.3 1.86 (1.12–3.08)* 0.735 %: prevalence (in percentage) of health and health behaviours for couple mothers and lone mothers-household heads. aPR (95% CI): age-adjusted prevalence ratio amongst lone mothers-household heads compared to couple mothers (95% confidence interval). p-value inter: p-value for the interaction between type of mo- therhood and survey year. * p-value < 0.05 ** p-value < 0.01 outcome analysed. On the other hand, inequalities between un- 2003–2004 compared with that amongst couple mothers, suggesting employed lone and couple mothers, from both non-manual and manual the importance of paid employment in providing lone mothers with the social classes, showed a non-significant decrease over time for all out- economic resources that allow them to become independent. This could comes. Amongst housewives, inequalities in health between lone and also explain the decrease in lone mothers-household heads in couple mothers increased mostly for health outcomes and decreased for 2011–2012 because, if these women are unemployed, they will not be health behaviours (except for sleeping less than 6 hours/day amongst able to afford lone motherhood on their own. Therefore, greater vul- mothers from the manual social class). Finally, inequalities in smoking nerability to unemployment could be experienced by lone mothers- behaviour between lone and couple mothers showed a non-significant household heads, who could be more severely affected by the economic decrease amongst unemployed mothers and housewives (both the non- crisis, since, as the main breadwinners in the household, they probably manual and manual social class), while a non-significant increase was need a job to a greater extent. Thus, they are more dependent on social observed amongst employed mothers from the manual social class. protection policies for families and children, which are scarce in Spain. The decision to divide lone mothers into household heads and non- 4. Discussion household heads was initially taken to reflect the characteristics of the Spanish ‘traditional family policy model’ and the peculiar dependence The main results of this study show that inequalities in health and of mothers in general and lone mothers in particular on their relatives health behaviours between couple and lone mothers could exist in to take care of their children (Korpi, 2000). However, due to the very Spain, especially for those belonging to manual social class. Changes in small sample size and differences in sociodemographic characteristics, inequalities between 2003–2004 and 2011–2012 could not be con- lone mothers-non-household heads were excluded from further ana- firmed, although some patterns were detected. lyses. On the other hand, some specifications on the definition of the oc- cupational social class and employment status should be mentioned. 4.1. Differences in sociodemographic characteristics Authors decided to differentiate between those women who are not in paid employment but are in active employment search (unemployed), The proportion of lone mothers-household heads decreased, while from those women who are not either in paid employment or in active that of lone mothers-non-household heads increased. An explanation employment search (housewives). Unemployed women are within the for this finding could be that, in economic downturns, a serious option labour market and they might receive some unemployment benefits for many lone mothers to avoid child poverty is to return to the parental from the welfare state if they have previously been in paid employment. home and receive financial support from their families to make ends While housewives are not within the labour market, are not entitled to meet. New lone mothers may also have more frequently felt unable to any unemployment benefits since they have not been in paid employ- afford a household in 2011–2012 than in 2003–2004 and so they may ment, and their main activity declared was housework. As different have remained in the parental home. Additionally, given the economic categories of employment status could overlap since both employed and difficulties during a crisis, fewer couple mothers may have been able to unemployed women usually are housewives too, it is worth-mentioning afford to separate from their partners, due to their economic depen- that this variable was based on the women’s current main activity and dence. always the categories ‘employed’ and ‘unemployed’ prevailed over The unemployment rate increased for all three groups of mothers ‘housewives’. The distinction between unemployed and housewives was but was especially high for lone mothers, both household heads and also made because an important proportion of housewives was found in non-household heads. This finding highlights the disadvantaged posi- our sample since they represent an important part of the women po- tion of lone mothers and could be explained by their greater vulner- pulation in Spain, a country characterised by a ‘traditional family policy ability and exposure to unemployment (Artazcoz et al., 2014). Of note, model’ where the incorporation of women into the labour force is still employment was high amongst lone mothers-household heads in 6 S. Trujillo-Alemán et al. SSM - Population Health 7 (2019) 100367 Table 4 Associations between type of motherhood and poor health outcomes and health behaviours, by occupational social class and employment status: Spanish Health Survey, 2003–2004 and 2011–2012. Changes in inequalities between couple and lone mothers-household heads. Non-manual social class Manual social class 2003–2004 2011–2012 2003–2004 2011–2012 (n = 1143) (n = 1207) (n = 1712) (n = 1655) % aPR (95% CI) % aPR (95% CI) p-value inter % aPR (95% CI) % aPR (95% CI) p-value inter Poor self-perceived health Employed Couple mothers 20.0 1.00 12.5 1.00 24.3 1.00 21.2 1.00 Lone mothers-household heads 17.3 0.82 (0.55–1.20) 12.8 0.96 (0.56–1.64) 0.625 36.8 1.43 (1.04–1.95)* 35.3 1.61 (1.13–2.31)** 0.613 Unemployed Couple mothers 14.7 1.00 20.4 1.00 26.8 1.00 26.3 1.00 Lone mothers-household heads 20.0 1.36 (0.32–5.69) 22.3 1.09 (0.50–2.37) 0.790 41.7 1.37 (0.75–2.50) 27.8 1.07 (0.68–1.68) 0.516 Housewife Couple mothers 22.6 1.00 20.7 1.00 32.9 1.00 25.7 1.00 Lone mothers-household heads 16.4 0.60 (0.08–4.48) 19.6 1.12 (0.19–6.61) 0.650 60.6 1.68 (0.92–3.04) 53.2 2.14 (1.05–4.37)* 0.608 At least one selected chronic condition Employed Couple mothers 5.0 1.00 9.5 1.00 6.9 1.00 9.9 1.00 Lone mothers-household heads 5.0 0.99 (0.46–2.13) 4.9 0.51 (0.24–1.10) 0.226 7.6 1.06 (0.51–2.18) 15.9 1.57 (0.89–2.76) 0.403 Unemployed Couple mothers 7.6 1.00 15.2 1.00 4.7 1.00 15.1 1.00 Lone mothers-household heads 11.0 1.51 (0.23–9.98) 10.9 0.78 (0.27–2.29) 0.553 15.3 2.91 (0.86–9.91) 28.3 1.89 (1.09–3.30)* 0.528 Housewife Couple mothers 4.9 1.00 11.4 1.00 8.9 1.00 13.7 1.00 Lone mothers-household heads …a 23.8 2.54 (0.41–15.9) 10.1 10.1 0.97 (0.23–3.42) 16.4 1.30 (0.38–4.41) 0.748 Smoker Employed Couple mothers 30.4 1.00 27.1 1.00 40.7 1.00 31.8 1.00 Lone mothers-household heads 35.4 1.16 (0.89–1.51) 32.1 1.18 (0.88–1.59) 0.927 44.9 1.12 (0.87–1.44) 41.9 1.33 (1.00–1.76)* 0.370 Unemployed Couple mothers 31.1 1.00 27.5 1.00 51.3 1.00 34.0 1.00 Lone mothers-household heads 78.0 2.51 (1.43–4.42) 41.3 1.52 (0.87–2.65) 0.218 71.6 1.44 (1.07–1.95)* 42.9 1.26 (0.88–1.81) 0.570 Housewife Couple mothers 27.8 1.00 28.5 1.00 34.7 1.00 28.6 1.00 Lone mothers-household heads 57.9 1.93 (0.73–5.13) 49.5 1.85 (0.68–5.03) 0.954 53.7 1.65 (0.94–2.90) 18.2 0.63 (0.17–2.38) 0.192 Sleeping less than 6 hours/day Employed Couple mothers 3.2 1.00 5.3 1.00 6.8 1.00 7.8 1.00 Lone mothers-household heads 3.7 1.11 (0.41–3.04) 5.9 1.08 (0.49–2.42) 0.966 12.0 1.63 (0.89–2.97) 15.8 1.95 (1.05–3.65)* 0.683 Unemployed Couple mothers 5.8 1.00 6.8 1.00 9.8 1.00 7.1 1.00 Lone mothers-household heads 13.9 2.40 (0.38–15.3) …a 12.8 1.22 (0.33–4.55) 8.8 1.24 (0.45–3.41) 0.989 Housewife Couple mothers 2.2 1.00 2.7 1.00 7.7 1.00 5.5 1.00 Lone mothers-household heads 54.7 18.3 (4.4–76.6)** 19.6 9.49 (1.21–74.3)* 0.622 7.6 0.82 (0.20–3.28) 10.0 2.02 (0.32–12.79) 0.442 %: prevalence (in percentage) of health and health behaviours for couple mothers and lone mothers-household heads. aPR (95% CI): age-adjusted prevalence ratio amongst lone mothers-household heads compared to couple mothers (95% confidence interval). p-value inter: p-value for the interaction between type of mo- therhood and survey year. *: p-value < 0.05; **: p-value < 0.01. a : insufficient number of respondents. very low in comparison with other European countries (Artazcoz et al., health improved, a phenomenon also found by other authors in the 2014; Korpi, 2000). In addition, it should be noticed that both un- same context analysing other populations in Spain (Bacigalupe et al., employed and housewives could be assigned to an occupational social 2016; Bartoll et al., 2015; Regidor, Barrio, Bravo, & De La Fuente, 2014; class although they were not in paid employment at the time the survey Urbanos-Garrido & Lopez-Valcarcel, 2015). Possibly, people might was conducted. This is because occupational social class was defined overstate their health perception when comparing themselves with according to current or most recent occupation. If never occupied, then others in worse conditions such as, for example, the unemployed. the occupational social class of the main breadwinner in the household Moreover, unemployment benefits can be claimed for up to 2 years. was assigned to the selected mother in our sample. This is the reason Thus, self-perceived health amongst the unemployed in 2011–2012 why many unemployed mothers and those mothers who are housewives might still not have worsened by as much as expected (Bartoll et al., could be assigned to one of the categories of occupational social class. 2015). However, this option was not possible for lone mothers-household The decrease in smoking for all groups of mothers suggests that new heads who were unemployed or housewives because they were the anti-smoking laws in Spain have had a positive effect. In addition, due main breadwinners in the household. to the economic crisis, more disadvantaged mothers may have experi- enced greater difficulty in being able to afford cigarettes and may have 4.2. Changes in the prevalence of health outcomes and health behaviours been forced to quit smoking (Bacigalupe et al., 2016). In fact, un- employed lone mothers from both manual and non-manual social During the economic crisis, the results suggest that self-perceived classes have shown the greatest decrease in the prevalence of smokers. 7 S. Trujillo-Alemán et al. SSM - Population Health 7 (2019) 100367 The prevalence of sleeping less than 6 hours/day seems to have couple and lone mothers could be explained by a socioeconomic per- increased in both couple and lone mothers from both the manual and spective, indicating that employment status might also play an im- non-manual social classes who were in paid employment. This could be portant role. Paid employment provides more opportunities to lone because balancing family and work forces working mothers to reduce mothers to become independent and live in their own household as they the time spent in personal care (which includes hours of sleep) in order are the main breadwinners. Likewise, paid employment could provide a to gain time to care for dependent children (Artazcoz, Borrell, Benach, buffer against the lack of a partner’s financial support. Cortès, & Rohlfs, 2004). Despite not showing an interaction between type of motherhood and survey year, some differences in inequalities between 2003–2004 4.3. Inequalities in health and health behaviours: Interactions between type and 2011–2012 deserve discussion. Inequalities in health and health of motherhood, occupational social class and employment status behaviours between lone and couple mothers indicate an increase amongst the manual social class and the employed. Because working This study reveals that inequalities in health and health behaviours conditions can worsen in times of economic crisis, inequalities are not between couple and lone mothers in Spain could exist for those be- simply a matter of being in paid employment or not, but are rather a longing to the manual social class. Results indicate that lone mothers matter of working conditions (Dziak et al., 2010). During the time from the manual social class could be at higher risk of all the health period studied, the implementation of various labour market reforms outcomes and health behaviours analysed in this study. However, it has increased job precariousness and has weakened worker rights does not seem that lone mothers amongst the non-manual social class (Bacigalupe et al., 2016; Bartoll et al., 2015; Eurofound, 2015). Wor- reported worse health outcomes, suggesting that occupational social sening of working conditions could affect lone mothers-household class might play an important role in the association between type of heads to a greater extent. For instance, part-time work may have be- motherhood, and health outcomes and health behaviours. Therefore, come more scarce, increasing the difficulties of lone mothers in bal- inequalities between lone and couple mothers might be a matter of ancing work and family life. This could negatively impact health, al- socioeconomic position and women’s economic ability to cope with though these effects would probably become evident in the long term. motherhood, especially lone motherhood. Lone mothers have been A possible explanation for the lack of statistically significant inter- shown to be more exposed to economic strain, which may be even action between type of motherhood and survey year in this study could greater amongst those from the manual social class, since lone mothers be the latency period in times of economic crisis. From the second from the non-manual social class probably have greater purchasing quarter of 2008, the Spanish economy entered a crisis. However, the power, preventing them from a higher risk of economic strain (Fritzell largest cutbacks in welfare state and public policy expenditure in Spain & Burström, 2006). started in 2010 (Bacigalupe et al., 2016). Thus, insufficient time may For mothers, having a partner living in the same household may have passed since the implementation of the cutbacks, and any changes provide affection, emotional support or help with childcare. However, in inequalities in health and health behaviours will probably be re- this support could be also be provided by the mothers’ relatives and vealed later. Indeed, it is important to distinguish between short- and friends. This is true for mothers living with a partner, as well as for long-term effects of an economic crisis on health (Bacigalupe et al., those who have separated from their partners or chosen lone mother- 2016). The possible lag between the effect of an economic crisis and hood. Family and friends may gradually help the lone mother to adapt changes in health outcomes is not obvious and it is not clear which to her new situation substituting the emotional or practical support no health outcomes will rapidly be affected by an economic crisis. In ad- longer provided by the partner. Therefore, social, community and fa- dition, for some time, the initial effects of economic crises could be mily networks play an important role in providing emotional support mitigated partly by unemployment benefits and partly by family net- and help with childcare. However, it is less likely that family and works (Regidor et al., 2014). In the Spanish socioeconomic and political friends share the financial burdens of raising children to the same ex- context, family support could have acted as a buffer, masking some of tent as a partner. Consequently, the greatest contribution of a partner the effects of the economic crisis. With women playing an important might be economic, which could explain why lone mothers from the role as caregivers and the low female participation in the labour manual social class showed worse results than couple mothers, while market, social protection is largely based on family support, in contrast lone mothers from the non-manual social class did not show worse to a more central role of the state (Artazcoz et al., 2014). Consequently, health outcomes. Lone mothers with economic resources could pay for despite the cutbacks in social policies, mothers in general may have domestic services and childcare needs, experiencing fewer difficulties in continued to receive family support, protecting especially lone mothers dealing with lone motherhood. Thus, lone mothers might enjoy the best from increased health risks (Berkman et al., 2015). part of being a mother, eschewing unpleasant tasks such as laundry or Despite analysing two fairly different moments in the Spanish so- cleaning. In contrast, many couple mothers might feel responsible for cioeconomic and political context, we found non-statistically significant all the housework in addition to childcare (Korpi, 2000), taking on their changes in inequalities in health and health behaviours. This seems to partner’s housework too, while lone mothers might not feel this pres- highlight that the higher exposure of lone mothers-household heads sure, since they do not cohabit with a partner. Another factor is the from the manual social class is not purely arbitrary or is only explained route into lone motherhood. Some mothers are lone mothers by choice, by the economic crisis, but is a structural process. As previously men- while a significant proportion are lone mothers as a result of separation, tioned, the value of a partner to a mother lies mainly in providing fi- divorce or widowhood. Lone mothers by choice may show different nancial support, which is less likely to be supplied by family or friends. characteristics and it is probable that they are more frequently from the Other types of support such as affection, emotional support or helping non-manual social class, enjoying social support and good health. with childcare may be supplied by social, community and family net- When employment status was analysed in addition to occupational works. The way in which lone mothers organise the upbringing and social class, patterns of inequalities remained similar. Amongst the care of their children goes beyond the nuclear and extended family, manual social class, lone mothers seem to have a higher risk of poor with the community and social networks in which they are involved health outcomes and health behaviours than couple mothers, regardless acquiring a greater role (Di Nella et al., 2014). Good social, community of whether they were employed, unemployed or housewives. Amongst and family networks could mitigate inequalities between couple and the non-manual social class in paid employment, lone motherhood lone mothers (Burström et al., 2010). However, a differential char- seems to confer a protective effect for health although non-statistically acteristic between manual and non-manual social classes and between significant. Therefore, amongst women in a more advantaged socio- employed and unemployed mothers, is the ability to pay for some economic position, lone motherhood seems not to have a negative ef- services, such as cleaning and childcare. If the welfare state considered fect. Once again, these results highlight that inequalities between the plurality of family models and redistributed wealth, other types of 8 S. Trujillo-Alemán et al. SSM - Population Health 7 (2019) 100367 motherhood not implying risks for health might be possible, as long as statistically significant results). Therefore, our results could indicate the minimum material needs were covered. Therefore, the socio- that, when the minimum material resources are covered, lone mothers economic and political context also plays an important role in gen- are not at higher risk of poor health. The greater vulnerability of lone erating inequalities between couple and lone mothers (Burström et al., mothers in comparison with couple mothers is partly because lone 2010). Lone mothers are at increased risk of social exclusion and some mothers have to shoulder the burden of childcare and being the factors that may help to explain their greater precariousness than breadwinner largely alone. Therefore, it seems that it is not a matter of couple mothers are the lack of family protection policies in the context having a partner or not, but rather of the lack of financial support of welfare regimes (Di Nella et al., 2014). available to women to raise their children, in a context where social Finally, other authors have pointed out that inequalities between protection policies for families and children are scarce, as is the case of couple and lone mothers could also be influenced by gender inequal- Spain. Despite some differences in the patterns of inequalities in health ities. The capitalist model of production organised around the sexual and health behaviours between couple and lone mothers over the study division of work underestimates domestic and care work, without period, we cannot confirm that inequalities changed between considering the importance of this work for the sustainability for the 2003–2004 and 2011–2012. Possible explanations include the Spanish human species (Di Nella et al., 2014). Lone motherhood illustrates the ‘traditional family policy model’ characterised by strong family soli- difficulties of balancing personal time, paid and unpaid work time, and darity and the latency period between the onset of the crisis and its family time. Time management policies are one of the central axes for effect on health outcomes and health behaviours. tackling gender inequality and the discrimination of women in society, as they would assign an adequate value to caregiving. It is essential to Acknowledgements recognise the importance of caregiving and domestic responsibilities (Almeda & Di Nella, 2011). We would like to acknowledge Gail Craigie who has reviewed the English version of the manuscript and edited it to eliminate possible 4.4. Limitations and strengths grammatical or spelling errors and to conform to correct scientific English. The cross-sectional design of this study does not allow us to draw conclusions on the causality of the associations found. However, con- Funding sidering that lone mothers are a socioeconomically disadvantaged group, it is likely that their excess risk of poor health outcomes and This work was partially supported by the SOPHIE Project - health behaviours is underestimated (Rousou et al., 2013). It is also Evaluating the Impact of Structural Policies on Health Inequalities and worth-mentioning that, when analysing self-reported data, there could their Social Determinants and Fostering Change [278173-2]; by grant be differences by socioeconomic factors in misreporting chronic con- from the Carlos III Institute of Health, Ministry of Economy and ditions or rating self-perceived health (Dowd, 2012; Mackenbach, Competitiveness (Spain) and European Union ERDF funds (European Looman, & van der Meer, 1996). In addition, it would have been in- Regional Development Fund) under the Health Strategy Action 2013- teresting to study inequalities in mental health between couple and lone 2016 with reference PI13/02292; and the program of social determi- mothers, but we were unable to do so because there is no information nants of health of the CIBER of Epidemiology and Public Health on this topic in the 2003–2004 data. Moreover, income level is an (CIBERESP). important factor in understanding inequalities between couple and lone mothers, but comparable data from the two waves of the Spanish Na- Note tional Health Survey were not available. Finally, the main limitation is probably the small sample size that, in addition to the performance of This article is part of Sara Trujillo-Alemán’s Ph.D. research, cur- different stratifications, became even smaller in each stratified group. rently being developed at the Department of Experimental and Health Because of this, results should be interpreted with caution since asso- Sciences of the Universitat Pompeu Fabra (Barcelona). ciations could be unstable. Despite this limitation, authors highlight the value of analysing inequalities in health between couple and lone mo- Declaration of interest thers within different socioeconomic groups, as, for instance, occupa- tional social class and employment status. This is the main strength of The authors declare they have no conflicts of interest. this study, which shows the importance of including socioeconomic factors when studying inequalities in health and health behaviours References between lone and couple mothers in Spain, as well as in other similar socioeconomic and political contexts, revealing that inequality might Almeda, E., & Di Nella, D. (2011). Monoparentalidad, género y bienestar [Single par- not depend simply on the presence or absence of a partner. To further enthood, gender and well-being]. In E. Almeda, & D. Di Nella (Vol. Eds.), (1st ed.). Bienestar, protección social y monoparentalidad: Las familias monoparentales a debate understand inequalities in health and health behaviours between [Welfare, social protection and single-parenthood: Single parent families for discussion]: 2, couple and lone mothers in Spain and other similar contexts, as well as (pp. 93–123). Barcelona: Copalqui Editorial. 〈http://www.ub.edu/tiifamo/wp- to address the impact of the economic crisis in these inequalities, content/uploads/2014/11/volumen2_cap5.pdf〉. Alvarez, G. G., & Ayas, N. T. (2004). The impact of daily sleep duration on health: A longitudinal data are necessary. review of the literature. Progress in Cardiovascular Nursing, 19(2), 56–59. https://doi. org/10.1111/j.0889-7204.2004.02422.x. 5. Conclusions Artazcoz, L., Borrell, C., Benach, J., Cortès, I., & Rohlfs, I. (2004). Women, family de- mands and health: The importance of employment status and socio-economic posi- tion. Social Science Medicine, 59(2), 263–274. https://doi.org/10.1016/j.socscimed. Inequalities in health and health behaviours between couple and 2003.10.029. lone mothers seem to exist in Spain amongst those belonging to the Artazcoz, L., Cortès, I., Puig-Barrachina, V., Benavides, F. G., Escribà-Aguir, V., & Borrell, C. (2014). Combining employment and family in Europe: The role of family policies manual social class. Lone mothers-household heads from the manual in health. The European Journal of Public Health, 24(4), 649–655. https://doi.org/10. social class are at higher risk of poor health and health behaviours in 1093/eurpub/ckt170. comparison to couple mothers (statistically significant results). Bacigalupe, A., & Escolar-Pujolar, A. (2014). The impact of economic crises on social However, amongst the non-manual social class, lone mothers-house- inequalities in health: What do we know so far? International Journal for Equity in Health, 13(1), 52. https://doi.org/10.1186/1475-9276-13-52. hold heads report better self-perceived health and are less likely to have Bacigalupe, A., Shahidi, F. V., Muntaner, C., Martín, U., & Borrell, C. (2016). Why is there at least one selected chronic condition, possibly due to a protective so much controversy regarding the population health impact of the great recession? effect of the higher income of the non-manual social class (non- Reflections on three case studies. International Journal of Health Services: Planning, 9 S. Trujillo-Alemán et al. SSM - Population Health 7 (2019) 100367 Administration, Evaluation, 46(1), 5–35. https://doi.org/10.1177/ Fritzell, S., Ringbäck Weitoft, G., Fritzell, J., & Burström, B. (2007). From macro to micro: 0020731415611634. The health of Swedish lone mothers during changing economic and social circum- Barros, A. J. D., & Hirakata, V. N. (2003). Alternatives for logistic regression in cross- stances. Social Science Medicine, 65(12), 2474–2488. https://doi.org/10.1016/j. sectional studies: An empirical comparison of models that directly estimate the socscimed.2007.06.031. prevalence ratio. BMC Medical Research Methodology, 3(1), 21. https://doi.org/10. Fritzell, S., Vannoni, F., Whitehead, M., Burström, B., Costa, G., Clayton, S., & Fritzell, J. 1186/1471-2288-3-21. (2012). Does non-employment contribute to the health disadvantage among lone Bartoll, X., Toffolutti, V., Malmusi, D., Palència, L., Borrell, C., & Suhrcke, M. (2015). mothers in Britain, Italy and Sweden? Synergy effects and the meaning of family Health and health behaviours before and during the Great Recession, overall and by policy. Health Place, 18(2), 199–208. https://doi.org/10.1016/j.healthplace.2011.09. socioeconomic status, using data from four repeated cross-sectional health surveys in 007. Spain (2001–2012). BMC Public Health. https://doi.org/10.1186/s12889-015- Gorber, S. C., Schofield-Hurwitz, S., Hardt, J., Levasseur, G., & Tremblay, M. (2009). The 2204-5. accuracy of self-reported smoking: A systematic review of the relationship between Berkman, L. F., Zheng, Y., Glymour, M. M., Avendano, M., Börsch-Supan, A., & Sabbath, self-reported and cotinine-assessed smoking status. Nicotine Tobacco Research: Official E. L. (2015). Mothering alone: Cross-national comparisons of later-life disability and Journal of the Society for Research on Nicotine and Tobacco, 11(1), 12–24. https://doi. health among women who were single mothers. Journal of Epidemiology Community org/10.1093/ntr/ntn010. Health, 69(9), 865–872. https://doi.org/10.1136/jech-2014-205149. Idler, E. L., & Benyamini, Y. (1997). Self-rated health and mortality: A review of twenty- Bull, T., & Mittelmark, M. B. (2009). Work life and mental wellbeing of single and non- seven community studies. Journal of Health and Social Behavior, 38(1), 21–37. single working mothers in Scandinavia. Scandinavian Journal of Public Health, 37(6), Instituto Nacional de Estadística (2017). Spanish National Health Survey. Retrieved 562–568. https://doi.org/10.1177/1403494809340494. November 13, 2017, from 〈https://www.msssi.gob.es/estadEstudios/estadisticas/ Burström, B., Diderichsen, F., Shouls, S., & Whitehead, M. (1999). Lone mothers in encuestaNacional/〉. Sweden: Trends in health and socioeconomic circumstances, 1979–1995. Journal of Instituto Nacional de Estadística (2018). INEbase. Retrieved December 18, 2018, from Epidemiology Community Health, 53(12), 750–756. https://doi.org/10.1136/jech.53. 〈https://www.ine.es/〉. 12.750. Korpi, W. (2000). Faces of inequality: Gender, class, and patterns of inequalities in dif- Burström, B., Whitehead, M., Clayton, S., Fritzell, S., Vannoni, F., & Costa, G. (2010). ferent types of welfare states. Social Politics, 7(2), 127–191. https://doi.org/10.1093/ Health inequalities between lone and couple mothers and policy under different sp/7.2.127. welfare regimes - the example of Italy, Sweden and Britain. Social Science Medicine, Lahelma, E., Arber, S., Kivelä, K., & Roos, E. (2002). Multiple roles and health among 70(6), 912–920. https://doi.org/10.1016/j.socscimed.2009.11.014. British and Finnish women: The influence of socioeconomic circumstances. Social Comisión para Reducir las Desigualdades Sociales en Salud en España [CRDSS-E] (2012). Science Medicine, 54, 727–740. https://doi.org/10.1016/S0277-9536(01)00105-8. Propuesta de políticas e intervenciones para reducir las desigualdades sociales en Mackenbach, J., Looman, C., & van der Meer, J. (1996). Differences in the misreporting of salud en España [A proposal of policies and interventions to reduce social inequalities chronic conditions, by level of education: the effect on inequalities in prevalence in health in Spain]. Gaceta Sanitaria, 26(2), 182–189. https://doi.org/10.1016/j. rates. American Journal of Public Health, 86(5), 706–711. gaceta.2011.07.024. Regidor, E., Barrio, G., Bravo, M. J., & De La Fuente, L. (2014). Has health in Spain been Di Nella, D., Almeda Samaranch, E., & Ortiz Monera, R. (2014). Perspectiva no declining since the economic crisis? Journal of Epidemiology Community Health, 68, androcéntrica en los estudios sobre familias monoparentales. Reflexiones e im- 280–282. https://doi.org/10.1136/jech-2013-202944. plicaciones metodológicas [Non-androcentric perspective in studies of one-parent Reynolds, J., Rueda, S., & Lahoz, M. (2007). Lone mothers in Spain. Policy context and families. Reflections and methodological implications]. Athenea Digital, 14(4), individual outcomes in terms of health and health-related behaviors. Tackling health in- 181–207. https://doi.org/10.5565/rev/athenea.1360. equalities in Europe: An integrated approach (EUROTHINE). Final Report. TheThe Diderichsen, F., Evans, T., & Whitehead, M. (2001). The social basis of disparities in Netherlands Netherlands: University Centre Medical Rotterdam. Departament of health. In T. Evans, M. Whitehead, & F. Diderichsen (Eds.). Challenging inequities in Public Health301–318. 〈http://ec.europa.eu/health/ph_projects/2003/action1/ health. From ethics to action (pp. 12–23). New York: Oxford University Press. docs/2003_1_16_frep_en.pdf〉. Domingo-Salvany, A., Bacigalupe, A., Carrasco, J. M., Espelt, A., Ferrando, J., & Borrell, Rius, C., Pérez, G., Martínez, J. M., Bares, M., Schiaffino, A., Girspert, R., & Fernández, E. C. (2013). Propuestas de clase social neoweberiana y neomarxista a partir de la (2004). An adaptation of Charlson comorbidity index predicted subsequent mortality Clasificación Nacional de Ocupaciones 2011 [Proposals for social class classification in a health survey. Journal of Clinical Epidemiology, 57, 403–408. https://doi.org/10. based on the Spanish National Classification of Occupations 2011 using neo- 1016/j.jclinepi.2003.09.016. Weberian and neo-Marxist]. Gaceta Sanitaria, 27(3), 263–272. https://doi.org/10. Roos, E., Burström, B., Saastamoinen, P., & Lahelma, E. (2005). A comparative study of 1016/j.gaceta.2012.12.009. the patterning of women's health by family status and employment status in Finland Domingo-Salvany, A., & Marcos Alonso, J. (1989). Propuesta de un indicador de la "clase and Sweden. Social Science Medicine, 60(11), 2443–2451. https://doi.org/10.1016/j. social" basado en la ocupación [Proposal of an indicator of «social class» based on the socscimed.2004.11.020. occupation]. Gaceta Sanitaria, 3(10), 320–326. https://doi.org/10.1016/S0213- Rousou, E., Kouta, C., Middleton, N., & Karanikola, M. (2013). Single mothers' self-as- 9111(89)70948-1. sessment of health: A systematic exploration of the literature. International Nursing Dowd, J. B. (2012). Whiners, deniers, and self-rated health: What are the implications for Review, 60(4), 425–434. https://doi.org/10.1111/inr.12044. measuring health inequalities? A commentary on Layes, et al. Social Science and Servicio Público de Empleo Estatal (2011). Informe del Mercado de Trabajo Estatal. Datos Medicine, 75(1), 10–13. https://doi.org/10.1016/j.socscimed.2012.01.036. 2010 Informe General, I〈http://www.sepe.es/indiceObservatorio/buscar.do?indice = Dziak, E., Janzen, B. L., & Muhajarine, N. (2010). Inequalities in the psychological well- 1&tipo = 1&periodo = anual&ambito = Nacional&tema = &idioma = es〉. being of employed, single and partnered mothers: The role of psychosocial work Sperlich, S., Maina, M. N., & Noeres, D. (2013). The effect of psychosocial stress on single quality and work-family conflict. International Journal for Equity in Health, 9, 6. mothers' smoking. BMC Public Health, 13, 1125. https://doi.org/10.1186/1471-2458- https://doi.org/10.1186/1475-9276-9-6. 13-1125. Espelt, A., Marí-Dell’Olmo, M., Penelo, E., & Bosque-Prous, M. (2016). Estimación de la Urbanos-Garrido, R. M., & Lopez-Valcarcel, B. G. (2015). The influence of the economic razón de prevalencia con distintos modelos de regresión: Ejemplo de un estudio in- crisis on the association between unemployment and health: An empirical analysis for ternacional en investigación de las adicciones. Adicciones, 29(2), 105–112. https:// Spain. The European Journal of Health Economics: HEPAC: Health Economics in doi.org/10.20882/adicciones.823. Prevention and Care, 16(2), 175–184. https://doi.org/10.1007/s10198-014-0563-y. Eurofound (2015). Spain: Figures show how the crisis has hit salaries. Retrieved Van de Velde, S., Bambra, C., Van der Bracht, K., Eikemo, T. A., & Bracke, P. (2014). December 18, 2018, from 〈https://www.eurofound.europa.eu/publications/article/ Keeping it in the family: The self-rated health of lone mothers in different European 2015/spain-figures-show-how-the-crisis-has-hit-salaries〉. welfare regimes. Sociology of Health Illness, 36(8), 1220–1242. https://doi.org/10. European Commission (2018). Eurostat database. Retrieved August 26, 2018, from 1111/1467-9566.12162. 〈http://ec.europa.eu/eurostat/data/database〉. Whitehead, M., Burström, B., & Diderichsen, F. (2000). Social policies and the pathways Fritzell, S., & Burström, B. (2006). Economic strain and self-rated health among lone and to inequalities in health: A comparative analysis of lone mothers in Britain and couple mothers in Sweden during the 1990s compared to the 1980s. Health Policy, Sweden. Social Science Medicine, 50(2), 255–270. https://doi.org/10.1177/ 79(2–3), 253–264. https://doi.org/10.1016/j.healthpol.2006.01.004. 070674371105600908. 10