Introduction
Poverty and social exclusion are two concepts that describe people with scarce resources to have a dignified life, and those who have been separated from society [
1]. Fighting poverty and social exclusion has always been a priority of the European Union (EU), which has typically measured these inequalities through the AROPE index (at risk of poverty or social exclusion). This index has been widely used in the Horizon 2020 programme [
2], and in the 2030 Agenda [
3], and it is composed of three sub-indicators: risk of poverty (based on household income), low work intensity (considering working hours) and severe material deprivation (such as not being able to afford certain goods or services). Meeting the conditions for at least one of the three sub-indicators implies being AROPE [
2,
4]. In 2018, Spain had one of the highest rates of AROPE in the European Union (EU) (26.1%) [
2,
5]. When assessing child poverty, the AROPE rate in Spain in 2018 differed depending on the type of family: children from two-parent families had a rate of 25.8%, while those from single-parent families presented a rate of 50% [
6].
Children and adolescents are marked by critical periods of development, and not achieving a certain skill in a certain moment might have lifelong implications, even when remedial actions were implemented at later stages [
7]. Socioeconomic inequalities may affect children’s development and mental health [
8], which can be assessed through internalizing and externalizing problems. Internalizing, or emotional, problems are inward-directed symptoms that bring about suffering in the child [
9]. They include anxiety, depression, somatic complaints and withdrawal [
10], and their prevalence is around 8.7–22.6% in Spanish adolescents [
11]. Several studies have shown adverse effects of economic hardship [
12,
13], low socioeconomic status (SES) or parental education level [
10,
14,
15] on internalizing problems [
16,
17]. Externalizing, or behavioural, problems describe outward-directed symptoms that, in addition to producing suffering in the child, also cause discomfort in other people [
9]. They comprise aggressive and oppositional behaviours, inattention/hyperactivity and emotion dysregulation [
9], and their prevalence is around 2.4–14.6% in Spanish adolescents [
11]. Poverty [
13,
16,
17] and low parental education level [
13,
15] were also associated with externalizing behaviours.
To understand how social inequalities affect mental health, Bronfenbrenner’s ecological systems theory (BEST), the family stress model (FSM) and the parental profile must be considered. According to the BEST [
18], a child is the centre of concentric spheres of influence. Variations in the furthest structural determinants can affect children through family-specific factors such as parenting practices [
14] or difficulties. Proximal social systems (family, school or community) can help promote the development of protective mechanisms that compensate the effect of unfavourable structural conditions [
19‐
23]. Interventions on these proximal factors could be more feasible in the short term [
24]. Stronger community, social and school networks have been positively related to better developmental outcomes [
20,
21], even in children from lower socioeconomic positions [
19,
20]. Assessing the role of these conditions is crucial to identify moderating factors. The importance of studying moderation (or effect modification) is clearly reflected in BEST, where community, social and school networks may temper or modulate the magnitude of the effect of socioeconomic strain on children’s mental health.
The FSM and the parental profile could describe a mediational pathway between economic strain and child’s mental health. The FSM posits that financial difficulties in the family generate stress that affects parenting practices, which may in turn influence child emotional and behavioural outcomes [
25‐
27]. Several studies in this line have described how higher stress [
17,
26,
28], maternal depression and harsh parenting [
29] mediated internalizing and externalizing problems. Parental profile encompasses knowledge (for example, about developmental stages in children), attitudes (such as father’s involvement), beliefs (like environmentalist outlook on development), and feelings (parental self-efficacy), about parenting. Finally, the effect of parenting knowledge on children’s mental health has been explained as follows: parents in situation of poverty or social exclusion are more likely to have less knowledge about child development [
30], and lower parenting self-efficacy (the confidence of doing well as parents) [
31]. This could result in a poor parental profile and therefore lower quality parent–child interactions, thereby increasing the risk of children having socioemotional problems [
32‐
34].
The INMA (INfancia y Medio Ambiente—Environment and Childhood) Study is a Spanish multicentre mother-and-child cohort [
35]. Its main purpose is to describe how environmental conditions affect children’s growth and development. Previous analyses with our data found a social gradient in child cognitive development at the age of 1–2 years [
36] and 5 years [
37,
38] when SES indicators such as parental social class, educational level or employment status were used.
This work provides several novelties with respect to previous studies. Firstly, in comparison to socioeconomic indicators such as education, employment and social class, AROPE may have greater sensitivity to detect children at extreme risk, as it provides a deeper understanding of multidimensional poverty or exclusion. Secondly, epidemiological work focuses on socioeconomic inequalities and their impact on mental health, but rarely emphasizes the family and social environment as a key factor. One of the main strengths of this study is the fact that it provides a more comprehensive approach to the poverty–family–mental health pathway.
The first aim is to determine whether the family risk of poverty and social exclusion, as measured by the AROPE indicators, is related to internalizing and externalizing problems in children aged 7–11 from two regions in Spain (Gipuzkoa and Valencia), with distinct SES levels [
2]. The second aim is to assess the dimensions of the family context that mediates or moderates the effect of poverty on children’s mental disorders. We hypothesize that: (a) children with a worse socioeconomic situation have a greater number of internalizing and externalizing problems, (b) parents’ stress caused by economic strain and the parental profile fostering child development act as a mediating pathway, and (c) the organization of the physical environment and social context acts as a moderating factor.
Discussion
We found that children from households at risk of poverty and exclusion and those with lower quality in the family context had higher scores for internalizing and externalizing problems. We also evaluated whether parental stress and parental profile were both mediators in the effect of poverty on children’s mental health, and if the physical environment and social context played a moderator role. Mediation analyses showed a direct and an indirect effect of risk of poverty and social exclusion on the outcomes, in both the simple and the simultaneous mediation, the latter demonstrating that both subscales can jointly mediate 42% of internalizing and 62% of externalizing problems. A moderation effect of the physical environment and social context was found for internalizing problems, with a positive relationship with the AROPE score in families with lower and middle quality on this subscale, while families with higher quality were not significantly affected by the AROPE. This fact suggests that a higher score on this subscale has a protective effect for poverty or social exclusion.
Several cohort studies have also explored behavioural outcomes in relation to family and community deprivation, such as the TRAILS Study [
47], the ALSPAC Study [
48] and the Millennium Cohort [
16], which have widely depicted how youths with internalizing and externalizing problems are more frequently born to parents with a low social class or income, unemployment and primary education. The INMA study previously described the effect of gender and socioeconomic inequities on child cognitive development [
37] and also analysed the factors associated with risk of poverty or social exclusion [
4]. As far as we know, this is the first work to describe the relationship between poverty, family context and children’s mental health in a Spanish population [
4].
Our work, as well as several other studies, has respected the layered structures to examine how socioeconomic hardship impacts children’s mental health. We found that both internalizing and externalizing problems were higher in more impoverished families and this relation was partially mediated by parental stress and parental profile. We did not find any studies with this mediation pattern, although some multi-level work has also described the effect of SES on mental health: one of them described how income inequality and family disruption were related to emotional problems [
12]. The other two found that family poverty, parental stress and authoritative parenting were associated with poorer mental health in children [
49,
50].
The parental stress and conflict subscale is composed of the factors of parental stress, frequency of and exposure to conflict, and conflict resolution. We found that greater AROPE was related to higher parental stress, and that this stress produced more risk of internalizing and externalizing problems. Three studies found trends compatible with our results [
17,
25,
26]. One of these publications examined externalizing problems at two time points during childhood and found that economic hardship and pressure led to emotional distress, couple conflict, harsh parenting and externalizing problems [
26]. We did not include some relational characteristics such as couple conflict or harsh parenting in our work, but our results point in the same direction as these findings. The second study tested to what extent two theories (family investment and family stress) explained the adverse relation between socioeconomic strain and externalizing problems. They found that the FSM was the pathway that best explained the relation between economic strain and adolescent delinquency, mediated by parents’ depression, caregiver conflict and parenting practices [
25]. The third study is from the Millennium Cohort Study, and found that permanent income had a protective effect for children’s mental health. When reports were made by parents, this relation was mediated by maternal distress, but this did not occur when children’s mental health was reported by teachers [
17]. This could represent a potential bias for the child’s psychopathology, and perhaps FSM is accountable only when both parental stress and the child’s behaviour are reported by parents. In addition to this problem, we must also keep in mind the potential reverse hypothesis: we argued that family stress was related to mental health problems, yet the child’s behaviour could be responsible for the parenting stress. A recent study found that in childhood, family stress was a predictor of externalizing, rather than the opposite, but in adolescence this relationship seems to be reciprocal [
51].
In our study, parental profile was composed of factors such as parental self-efficacy, parental knowledge regarding development stages, assertiveness, theories on an environmentalist outlook on development, and father’s involvement. Our results showed that families with greater risk of poverty or exclusion had a poorer parental profile. This knowledge, feelings and attitudes about parenting presented an association with both internalizing and externalizing problems. We did not find any research describing the whole poverty–parental profile–mental health axis, but several publications did find that better parenting knowledge [
30], self-efficacy [
33] or parental involvement [
52] reduced internalizing and externalizing problems.
Several studies have described the relation between parental stress and parental profile [
31,
33,
53]. We chose a correlational approach between them to respect the two-way relation of these interdependent parenting characteristics [
31]. We found a simultaneous mediation in both problems, but the magnitude and percentage of mediation was greater for externalizing problems. Parenting self-efficacy and parenting stress appear to be related [
31,
54] and may be modulated through positive parenting programmes, as a reduction in stress and an increase in parenting self-efficacy have been observed in the short term [
53], and an improvement in behaviour has been seen in the long term [
33]. Some studies supported the evidence of our findings, as stress may reduce self-efficacy [
54] and self-efficacy could predict parenting stress [
55]. An Australian study considered children’s outcomes, and its aim was not to assess mental health-related factors, but to observe which factors were related to parental self-efficacy. This could be partially predicted (37%) when employing parenting stress, parental education and child’s mental health as predictors [
31]. This study could be the most similar to ours, even when its hypothesis is reversed. They both have common factors, considering parental education (as a socioeconomic indicator), parenting stress, parental self-efficacy (as part of the parental profile) and the child’s mental health.
Organization of the Physical Environment and Social Context encompassed several related factors, such as quality of the physical environment, social support networks, promotion of child’s social relationships, and relations with the school. An Australian study described how community characteristics have a substantial impact on the child’s physical, mental and behavioural development, and more deprived areas have less appropriate neighbourhoods for children [
21]. For instance, in the physical domain adverse behavioural outcomes are related to greater distance from green spaces and higher population density [
21]. Another study examined the relations between socioeconomic characteristics and internalizing and externalizing problems, in a second factor: social support. This was inversely associated with both mental health problems in families and high socioeconomic status and low stress. This effect was not found for their low-status and high-stress counterparts, who reported greater behavioural problems in their children regardless of their social support [
56]. We did not explore our data stratifying by socioeconomic position to observe the social support. Conversely, we did stratify physical environment and social context in tertiles to observe the AROPE risk in each stratum. Despite this methodological difference, both approaches rely on the fact that social support and socioeconomic position mutually influence each other. A third factor considered was the child’s social relationships, whereby it appears that having and keeping a best friend at childhood reduces mental health problems [
57]. A final consideration in the Organization of the Physical Environment and Social Context is the interplay between school, friends and family. In particular, the relation of parents with the school in adverse environments is important. A recent study found that parental warmth and teacher support combined additively to reduce the effects of adversity in relation to internalizing problems [
58].
Several limitations should be considered: first, there is the problem of representativeness. Due to sample attrition, conclusions might not be generalizable to other regions. Second, to check the family stress model, parental profile and the social context mechanisms more accurately, it would be necessary to measure variables that could have been overlooked, such as parental mental health or alcohol intake at evaluation time [
17,
38]. Third, although we compared two cohorts that have proved to be substantially different, we did not find any interaction effect by cohort, and adding to the sample from other cohorts might help us to provide evidence that could be extrapolated to the general population in Spain. Fourth, income reports and family context scores could be biased, as many participants may have refused to answer when asked about their household’s income, and others could have masked family context answers for desirability. Fifth, the AROPE, HEFAS 7–11 and CBCL were reported at the same follow-up, a correlational rather than a causal relationship should be established [
38,
52]. Sixth, parental characteristics such as stress and mental health could be biasing children’s symptoms, as they tend to over-report behaviour problems [
17,
59]. Lastly, very few fathers answered the HEFAS 7–11 in comparison to mothers, so we could not stratify our analysis by respondent. However, no differences in subscales across respondents were observed, and results and significance did not change when we added the type of respondent, so simpler models were kept.
Our work also has several different strengths: first, we presented a new adaptation of the AROPE, to establish a continuous variable. This allowed us to increase the power of our analysis and provided richer information on the participating families and how much they are affected by poverty and exclusion. Second, we considered a full roster of covariates to improve the fit of our models and to control for potential confounders. Third, a strong measurement for family context, with good psychometric properties has been employed to describe the family characteristics. Fourth, to our knowledge, this is the first study to explore these characteristics as mediators between poverty and internalizing and externalizing problems, even though these factors have been interrelated in the literature. Lastly, the analysis was performed in two different cohorts with different social and cultural characteristics and with children ranging from 7 to 11 years of age: these facts endow our work with additional robustness.
By considering possible paths of intervention to ameliorate children’s symptoms, indirect and direct actions could be undertaken. Indirect interventions could consist in preventing economic inequities, eliminating the upstream causes of poverty itself, by economic compensation policies, such as providing a basic guaranteed income, implementing specific policies for single-parent families, reducing unemployment rates or increasing the minimum wage. These are proposals that are in line with the Spanish Government’s Strategy to fight poverty and exclusion [
60], which were included as part of the agreement for forming the coalition government [
61]. Proposals included in this agreement comprise the Minimum Vital Income, which was implemented in June 2020 [
62,
63]. Future research will have to unveil the effectiveness of the Subsistence Income as a compensation mechanism.
Conversely, direct interventions are more related to families and the immediate environment around the school. These proximal and family factors could become the main asset for preventing the negative impact of socioeconomic disadvantage on children’s mental health problems, as positive parenting and community strategies may be implemented to foster the child’s wellbeing. This is in line with Recommendation 19 (2006) of the Council of Ministers of Europe to member states [
64]. There is a need to invest in positive parenting programmes that can have a positive influence on children’s psychological development and indeed reduce the symptoms of internalizing and externalizing problems [
24]. These programmes may mobilize parents to ask for more playgrounds, green areas or services in their neighbourhood. Positive parenting programmes could also promote relations with educational and health services, which might help to identify youths at risk of mental health problems. Education and health systems must provide parents with developmental knowledge to improve their parental self-efficacy. Finally, parenting programmes should offer them tools to promote hope and stress management to foster parent–child interactions [
65]. In conclusion, preventing economic inequities by economic compensation policies such as the Subsistence Income, improving the neighbourhood and immediate environment around the school and social support, and promoting positive parenting programmes to strengthen parental self-efficacy could all improve mental health in childhood.