Social policy: a major lever to reduce social health inequalities at birth

Social health inequalities (SHIs) result from the unequal distribution of resources, which allows people to have more or less control over their living circumstances and their health.1–3 Economic resources are particularly important as they can be ‘transformed’ into other types of resources. They determine access to material goods and influence social participation.4–7 These kind of resources have a considerable impact on the population’s health and on certain groups in particular, such as children.8–15 Through the redistribution of economic resources, social policies contribute to reduce the poverty of households and economic inequalities. They constitute a major lever to improve the population’s health and diminish SHI.7 15–21

One of the major recommendations to reduce SHI consists of improving children’s health as early as possible, especially through policies which increase families’ income and contribute to reducing the intergenerational transmission of poverty.19 20 22 It is indeed well established that poverty before and during pregnancy is associated with an increased risk of stress and poor health for the mother, which in turn affects the child’s health at birth (preterm birth, low birth weight). The family’s poverty and/or health problems at birth have a negative effect on all aspects of the child’s development (cognitive, health and behavioural) as well as on health in adulthood.12 23–26 Between rich countries, poverty and economic inequalities are the main factors that explain the differences in health inequalities in terms of infant mortality and adverse pregnancy outcomes.27 28

Although there is a consensus on the importance of social policy as a means to reduce income inequalities, to improve health and to reduce SHI from birth, few studies examine the way in which many components of these policies interact to influence household poverty and to accentuate or mitigate SHI in different contexts.11 29 30 The scarcity of such evaluative studies is due, in particular, to the complexity of the task: multicausality (inter-relation between many types of policies, influence of others contextual factors), impossibility to carry out experimental studies, long-term health effects, and so on.31 Comparing different contexts with significant social policy differences constitutes a relevant approach to analyse the effect of these policies on SHI.29 30 32 Indeed, variations in social policy between countries or regions provide an opportunity, based on natural experiments, to investigate the mechanisms by which these policies influence SHI in each context.11 29 33

The targeting effect of income support policies

The policies that contribute to improve the income of households present two features: (1) They are a conjunction of numerous measures from various social policies, pursuing a variety of goals.34–39 These measures can be grouped into two categories: those aimed at non-working households or low-income households (unemployment benefits, social assistance benefits, measures that facilitate access to services such as housing, transport or healthcare) and those aimed at helping families with children (mainly child benefits and childcare). These governmental aids are delivered in several forms, whose combination varies by the country: direct financial help, tax deductions, subsidies for services. (2) These measures are applied according to various selectivity and generosity criteria.39–43 The selection of the beneficiary households for each programme and the amount allocated depends on many criteria including most notably the following: income, employment status (full time or part time), household composition (single parents, number of children and their ages) and other specific criteria: disabilities, citizenship, residence status and so on. These targeting criteria vary according to the applicable measure and according to the country, which entails a differential impact of income support policies (ISPs) on the income of different types of families inside a country and across countries, which is called targeting effect. In other words, a single mother with two children, unemployed, could be entitled to more or less benefits depending on their country of residence. The same goes for other family configurations. The comparative evaluation of social policies should reveal this targeting effect and the differential impact on the poverty of households in different contexts.

The limits of usual studies on the links between social policy and SHI

This is a relatively recent research field. Most research consists of quantitative studies analysing the correlation between the characteristics of social protection systems and health indicators. They are for the most part based on welfare state typologies (categorising the social protection systems of industrialised countries) and analyse the correlation between these regimes, poverty rates and health indicators.30 32 44 The most influential classification, proposed by Esping-Andersen45 (later adapted by other authors30 44), distinguishes three regime types: liberal, conservative and social democrat. The latter is said to have the most state intervention: relatively generous universal benefits, a significant redistribution and a strong commitment towards equality.

These typologies are an interesting starting point to analyse the relation between different social protection systems and the population’s health. For example, several studies show that Scandinavian countries (social democrat systems) have better indicators in terms of infant mortality.30 Yet these typologies have some limitations.30 32 44 First, some results of studies based on these typologies are inconsistent. They vary especially according to the health indicators considered, the period and gender. Second, and most importantly, they are global macrosocial analyses which do not allow to decompose the support granted by the state into its many components nor to understand how these components interact to influence the poverty of households and contribute to SHI in different contexts. The regimes that they describe constitute ideal types which fail to reflect the diversity of the real state action. Even when states have the same social policy regime, they diverge in several aspects, such as its components, the selection process and the generosity.43 Even within a state, there can be important regional differences in the conception of social policy. Studies based on welfare state types fail to reveal such differences between states or regions and the potential impact they might have on different population groups. Their contribution is somewhat limited when it comes to identifying concrete political measures to reduce SHI. Moreover, these studies often analyse the effect of the welfare state on population health rather than SHI.

The need to study contexts in order to overcome these limitations

As some authors highlight, there is a need for studies which evaluate in detail specific policies in different contexts with different methods. In order to better analyse the influence of social policy on SHI, a comparative analysis must take into account the specificity of contexts.30 33 44 46 47 The comparison is more enlightening when it relates to specific policies instead of state policy as a whole and analyses its effect on specific sociodemographic groups. Such is the case with Burstrom et al’s study,33 which compares single mothers with mothers in a couple, living in three countries with different family policies regime, on several factors: sociodemographic characteristics, living conditions (poverty, unemployment), lifestyle (tobacco use) and self-reported health. This study is interesting for its explanatory perspective, showing single mothers as a particularly disadvantaged group in all countries. It also reveals important differences in the mother’s situation between countries, which reflect differences in family policies. The authors do point out a limitation in their study. A subgroup analysis was not carried out to account for the heterogeneity of single-parent households. Indeed, these households differ in particular according to the number of children. An analysis of the poverty level of single-parent households according to the number of children could reveal subgroup differences across countries. Moreover, present day social policy is not neutral, as it is influenced by historical and sociopolitical factors, which need to be highlighted as well.48 49

Our review of the literature shows that to better understand the mechanisms by which social policy contributes to reduce or increase SHI in different contexts, it is important to: (1) Clearly specify the object of the evaluation (types of policies, population groups). (2) Analyse the context of policy (historical, implementation). (3) Analyse the impact of policy on the poverty of different subgroups in each context. (4) Take into account contextual factors, other than poverty, which may interact with poverty and influence SHI in each context.

Research questions and hypotheses

This research aims to explore the link between ISPs, the poverty level of households receiving social assistance and unfavourable pregnancy outcomes, through a Brussels-Montreal comparison, in order to formulate hypotheses on the mechanisms by which these policies contribute to increase or decrease health inequalities at birth in both regions.

The choice of cases

Brussels and Montreal are two regions that present numerous similarities. The poverty50 51 and recent immigration rates are high in both cities, with approximately 50% of births being to immigrant families.52 53 The two regions are located in two industrialised countries with a comparable level of economic development and quality of healthcare. A public health insurance system guarantees access to healthcare, especially for the most precarious. Perinatal health indicators in the general population in Brussels and Montreal are also comparable.52 53

These two regions also present notable differences in the conception of ISPs. A first glance reveals significant differences and similarities in terms of child benefits54 55 and social assistance policy.56–58 In Quebec and in Brussels, child benefits are universal, with an extra supplement granted to some households according to various criteria. The amount is significantly greater in Quebec (where it is provided both by the province and the state), and it increases considerably for lower-income households. In Belgium, household income does not generally determine the amount of benefits. Supplements may be granted on the basis of income if the parent is receiving social assistance, is single, disabled or unemployed for at least 6 months. Child benefits increase with the child’s age and especially with their rank (third and beyond > second > first), whereas in Quebec they decrease from the first to the second child. Social assistance is relatively much more substantial in Belgium than in Quebec. Single-parent families and those with children receiving social assistance are entitled to extra benefits. Social assistance recipients receive further help through services (non-monetary or subsidies) such as priority access to social housing, extra healthcare reimbursements and free transport passes.

These differences between the two regions prompt the following initial remarks: (1) Single parenthood and number of children are two important dimensions when it comes to analysing income inequalities between households in the two regions. (2) In Quebec, child benefits (generous and income based) greatly contribute to increase the income of households receiving social assistance. The situation of households with children should be relatively more advantageous than that of households without children. This is consistent with the high levels of poverty among households without children in Quebec. (3) In Brussels, social assistance and non-monetary services should constitute an important part of the income of social assistance recipients. The fact that the basic child benefits are less generous than in Quebec might entail a relative disadvantage for households with children with respect to those without children.

However, this preliminary analysis must be nuanced by taking into account the increased child benefits for households on social assistance in Belgium, supplements based on single parenthood and the number of children, and the differential impact of services, so that the situation of various subgroups in both regions can be compared in detail. For the analysis of perinatal health inequalities, the differences observed between households in the two regions, according to the presence or not of children, make it relevant to compare the perinatal health indicators of children from primiparous mothers (firstborn) to those of children of multiparous mothers.

Methods and analysis

Two case studies will be carried out from a comparative perspective. Data collection and the comparative analysis will be done in four phases that will overlap over time.

Patient and public Involvement statement

Patients were not involved in this study.

Model family method limitations

The analysis is illustrative rather than exhaustive, as it treats theoretical rights. It does not encompass all the situations that households encounter. Indeed, the rights are not always automatically activated, either because the eligible persons do not request it for various reasons (no take-up), or because the available resources do not make it possible to satisfy all the requests. The two main measures analysed, social assistance and family allowance, are less concerned by this remark since, generally, the rights become active if the eligibility conditions are met. It is especially in terms of access to services, such as social housing, that significant differences could be found between different types of households within a region, and between the two regions. Do not take into account the difficulties of access to social housing for some households would lead to underestimation of the poverty of these households. The analysis will consider possible differences between the two regions in this respect.

Moreover, other income sources are not addressed by the model family method (family and community support, undeclared work). Also, ISPs often undergo changes (often minor ones) with potential impact on the available household income. Attention will be given to such changes in the period covered by the study. Any comparative effort, based on the model family method, requires equal hypotheses for each country. It is advised to compare few countries with a small number of case types so as to look at the families’ situations as exhaustively as possible.

Income-related health inequalities cannot be completely comparable

The Brussels income data are taken directly from the yearly household income whereas the Montreal data are ecological, in relation to the average income of the neighbourhood. This might entail some difficulties when attempting to compare very low-income groups and income-related health inequalities. Nevertheless, geographical data are allowed as proxies of the individual situation when the zones are defined in such a way as to have a high socioeconomic homogeneity. The more homogeneous a zone is in this respect, the more the data reflect the situation of a composite or average person. Using this kind of indicator is relevant in urban centres with high density populations and high homogeneity,65 80 81 as is the case with Montreal.

Other potential confounding factors not taken into account in the quantitative analysis of SHI

There are multiple perinatal health determinants.27 82 Factors which are not included in the multivariate models might diverge between the countries and contribute to the observed health inequalities. However, the analysis will take migration into account, one of the main factors influencing the relation between poverty and SHI. Based on the available literature, others factors, such as smoking during the pregnancy, will be compared in both regions. Attention will also be given to the aspects of the healthcare system, especially access to perinatal healthcare, which may differ between the two regions. Moreover, the comparative design of the study should ensure a comparability of both regions on several dimensions.