Theory adjudication
Theory 1 – the beneficial effect of the welfare state on child and maternal health especially in LMECs through progressive social policies (Table
3).
There was a total of 25 studies assessing the relationship between the welfare state and child and/or maternal health outcomes. Twenty-two studies (88%) were assessed as positive, i.e. demonstrating evidence that a more generous welfare state led to superior health outcomes. This proportion was comparable to the wider systematic review by Barnish et al. [
11] where the corresponding figure was 79 out of 102 (77%). Welfare state provision may be conceptualised using a range of classification schemes. One influential work is The Three Worlds of Welfare Capitalism [
44]. This broadly categorises welfare states into neoliberal regimes, social democratic regimes and residual regimes, the latter being characterised by family-first support and a minimalist welfare regime. Of these 22 studies, 21 [
27,
36,
38‐
40,
45‐
60] remained classified as positive when only child and maternal health outcomes were considered. In the study by Klomp and de Haan [
61], the outcomes were based on 16 national health indicators that cover child and maternal health aspects, although the analytic model is written up in a way that does not permit the separable effect of the welfare state variables on the child and maternal health outcomes to be distinguished.
Out of the initial 25 studies, 21 studies reported a positive association between more generous welfare state provision and more positive child and maternal health outcomes, and were subjected to more detailed theory adjudication. Among these 21 studies, 12 assessed infant mortality outcomes [
39,
40,
45,
46,
48‐
50,
52‐
54,
56,
57], five assessed under-five mortality [
40,
45,
49,
51,
60], four studies assessed each of maternal mortality [
27,
45,
47,
56], low birth weight [
40,
47‐
49] and child mortality [
27,
36,
53,
55,
60], three studies assessed child poverty [
36,
53,
56] and one study assessed each of child wellbeing [
58], fertility [
59] and neonatal mortality [
38]. Therefore, both child and maternal health were assessed. Among these 21 studies, there was greater focus on developed countries than LMECs. There were five studies that focused exclusively on LMECs [
38,
45,
51,
53,
57], while six studies considered both developed countries and LMECs [
27,
40,
54‐
56,
60], and the remaining ten studies included only developed countries. Studies differed with regard to how they assessed the welfare state – either in terms of welfare state typology classification or welfare state financial expenditure – but the differential effect on outcomes was small. Compared to the systematic review, a higher proportion of studies included in this realist review used an expenditure-based approach. Studies on LMECs all considered welfare state in terms of health expenditure and/or its impact on healthcare coverage. Studies on developed countries, or those that included both developed countries and LMECs varied in terms of the use of regime or expenditure-based methods. Studies by Wu and Chiang [
60] on developed countries and LMECs and by Chung and Muntaner [
49] considered social services expenditures rather than solely health expenditures. Studies using regime classifications frequently cite Esping-Andersen [
44] as an important influence, but exact classification schemes do differ. For example, Karim et al. [
54] assessing both developed countries and LMECs used a six-way classification scheme: Scandinavian, Anglo-Saxon, Bismarckian, Southern, Eastern European and East Asian” which contrasts for example with a simpler three-way system: Liberal, Conservative and Social Democratic used by Bambra [
46] or four-way system: Social Democratic, Christian Democratic, Liberal and Wage Earner used by Chung and Muntaner [
48], both in developed countries.
Across studies, there was consistent evidence to support the proposed mechanism being more generous welfare policies to improve the social conditions of the deprived, although it is noted that studies did not consistently provide sufficient granularity of information on welfare state exposures to determine exactly which policies were effective. It was clear across studies that greater expenditure on welfare state measures (in studies assessing welfare state financial expenditure) and approaches to welfare state provision corresponding to the social-democratic classification in Esping-Anderson [
44] predicted the best outcomes. A revised version of theory 1 is presented (Table
3).
Theory 2 – the beneficial effect of left-of-centre political tradition on child and maternal health outcomes especially in LMECS through a greater focus on progressive policies (Table
3).
There was a total of eight studies assessing the relationship between political tradition and child and/or maternal health outcomes. All eight studies (100%) were assessed as positive, i.e. demonstrating that left-of-centre political tradition led to superior health outcomes. This proportion was higher than in the wider systematic review by Barnish et al. [
11] where the corresponding figure was 15 out of 17 (88%). Of these eight studies, seven [
49,
62‐
67] remained classified as positive when only child and maternal health outcomes were considered. In the study by London and Williams [
68], the child and maternal health aspects of the composite index of 41 indicators of domestic well-being were not presented separately.
Among these seven studies, six assessed infant mortality [
49,
62,
63,
65‐
67], two assessed each of child mortality [
62,
64] and low birth weight [
49,
64] and one assessed under five mortality [
49]. These studies did not assess maternal health. Contrary to expectation, the evidence base was not concentrated largely in LMECs, but instead rather largely in developed countries. Lena and London [
62] assessed periphery and non-core nations, Moon and Dixon [
63] assessed a range of countries without restriction by level of economic development, while the remaining four studies were conducted exclusively in OECD or otherwise wealthy countries. As anticipated, the evidence base focused on different aspects of child mortality, although two studies also assessed low birth weight – another aspect of debility.
Regarding mechanisms, the beneficial effect of left-of-centre political tradition operating via an increased focus on the welfare state was supported by all six studies. For example, Chung and Muntaner [
49] concluded that “strong political will that advocates for more egalitarian welfare policies, including public medical services, is important in maintaining and improving the nation’s health”. Lena and London [
62] demonstrate that political systems exert an influence on health and well-being independent of national and international economic factors. This adds further weight to the argument that it is political will that makes the difference rather than solely affluence – since it determines how a country’s resources are deployed, and whether or not it is in ways that benefit population health. Moon and Dixon [
63] explore a slightly different aspect of the same conceptual phenomenon and find that in left-wing countries state strength promotes welfare performance, while in right-wing countries it inhibits the provision of basic needs, again emphasising the importance of political will. Moreover, political will may also be related to political capacity [
69,
70], which is the ability to exert reflexive policy learning when navigating a complex policy environment that is laden with multiple interests from multiple stakeholders. Muntaner et al. [
64] also emphasise the role of ideological outlook in promoting a strong welfare state. Work by Navarro and Shi [
67] and Navarro et al. [
65,
66] furthermore emphasises the role of levels of public expenditures and health care benefits coverage and degree of redistributional focus. More specifically, they emphasise additionally public support of services to children as an important predictor of health status, and an important mechanism by which the positive effect of left-of-centre political tradition on child health outcomes may operate. None of the seven studies assessed maternal health. A revised version of theory 2 is presented (Table
3).
Theory 3 – beneficial effect of democracy on child and maternal health outcomes especially in LMECs by promoting empowerment and provision (Table
3).
There was a total of 28 studies assessing the relationship between democracy and child and/or maternal health outcomes. Twenty studies (71%) were assessed as positive, i.e. demonstrating that greater democracy led to superior health outcomes. This proportion was comparable with the wider systematic review by Barnish et al. [
11] where the corresponding figure was 34 out of 44 (77%). Of these 20 studies, 16 [37–38, 58–59, 67–78) remained positive when only child and maternal health outcomes were considered, while one study [
71] became inconclusive and three studies [
68,
72,
73] did not present the analysis in a way that enabled the impact on child and maternal health to be assessed separately. Studies tended to include a large range of countries and this range often included the USA: a potential caveat regarding the beneficial effect of democracy relates to the USA where there is a high level of democracy, but a highly neoliberal health system in which a high proportion of the population lack insurance coverage. Among these 16 studies, 14 assessed infant mortality [
42,
43,
62,
63,
74‐
83], three assessed each of maternal mortality [
42,
75,
79] and child mortality [
62,
84,
85], two assessed fertility [
79,
84] and the same single study [
79] assessed each of receipt of prenatal care, skilled birth attendance, under five underweight, anaemia during pregnancy, haemoglobin levels during reproductive age, tetanus immunization among pregnant women and birth rate among women aged 15–19. The evidence is clear across studies that increased democracy is beneficial for both child and maternal health. However, studies relevant to this theme tended to provide too little information to allow a specific theory to be supported with regard to the mechanism by which democracy achieves a benefit for child and maternal health.
Theory 4 – negative short-term effect of the introduction of capitalist democracy in communist autocracies on child and maternal health outcomes through the erosion of state networks (Table
3).
There was a total of 28 studies assessing the relationship between democracy and child and/or maternal health outcomes. One study (4%) was assessed as negative, i.e. demonstrating evidence that greater democracy led to inferior health outcomes [
41]. This proportion was comparable to the wider systematic review by Barnish et al. [
11] where the corresponding figure was two out of 45 (5%).
However, this study [
41] was no longer assessed as negative when only child and maternal health outcomes were considered. Instead, infant mortality rates were shown to follow the global trend, i.e. fell as income rose, with increased income resulting from greater democratisation in the context of the transition from Communism in Central and Eastern Europe. Therefore, theory four was not supported given there was no evidence that increased democracy resulted in inferior child and/or maternal health outcomes.
By means of comparison, in the entire set of studies and outcomes from the systematic review, there is evidence to suggest that the introduction of democracy may have a negative effect on health by disrupting existing systems that underpin health at least in the short term. Gauri and Khaleghian [
86] found that increased democracy predicted reduced diphtheria, tetanus, pertussis and measles vaccine coverage in 2018 LMECs with the erosion of state networks of compliance likely to be a key factor. Meanwhile, Adeyi et al. [
41] found that increased democracy in the Central and Eastern European transition context exerted a negative effect on the probability of dying between 15 and 65 years. The authors explain that reduced life expectancy at birth results from rising income level being associated with an increased probability of death between the ages of 15 and 65, stating that in this context “the wealthier the society, the less healthy is its population, particularly for its males” [
41]. Systems factors were likely to play a key role, and it is also important to note the temporal effect.
Theory 5 – beneficial effect of globalisation on child and maternal health outcomes especially in LMECs through increased prosperity benefitting public services (Table
3).
There was a total of 13 studies that assessed the relationship between globalisation and child and/or maternal health outcomes. Five studies (38%) were assessed as positive, i.e. demonstrating evidence that greater globalisation led to superior health outcomes. This proportion was higher than in the wider systematic review by Barnish et al. [
11] where the corresponding figure was seven out of 28 (25%). Of these five studies, all [
83‐
87] remained classified as positive when only child and maternal health outcomes were considered.
Among these five studies, all assessed infant mortality and one study each assessed child mortality [
86] and under five mortality [
87]. Therefore, all outcomes related beneficially to globalisation were related to different aspects of child mortality, and none assessed maternal health. All five studies assessed a wide range of countries with a range of different world system roles and were not restricted to LMECs. The evidence does not appear to support the suggestion that the positive effects of globalisation are largely limited to LMECs. Gerring and Thacker [
88] explicitly consider the direct effects of globalisation rather than those indirectly routed via economic growth. This study found that openness to imports and long-term membership in neoliberally oriented international trade bodies are associated with lower rates of infant mortality at a global level when the analysis controls for general economic performance, as measured by GDP per capita. However, the authors acknowledge an inability to determine causal mechanisms and any theoretical claims in their paper are recognised as largely speculative. Whereas Owen and Wu [
89] and Moore et al. [
90] focus on economic aspects of globalisation, Mukherjee and Krieckhaus [
91] and Martens et al. [
87] exemplify important social and political aspects of globalisation playing a role in the benefit of globalisation for population health.
Across studies, there appears little consensus on how the potential positive impact of globalisation on population health may operate and no specific theory can be supported at this current time. This situation is comparable across the full range of systematic review outcomes. Additionally, authors also acknowledge that neoliberal policies create winners and losers and increase health inequalities, so care needs to be taken in the interpretation of evidence regarding the potential benefits of globalisation, especially when no single theory gains much empirical support.
Theory 6 – negative effect of globalisation on child health especially in LMECs by generating a commercially-driven obesogenic environment (Table
3).
There was a total of 13 studies assessing the relationship between globalisation and child and/or maternal health outcomes. Five studies (38%) were assessed as negative, i.e. demonstrating evidence that greater globalisation led to inferior health outcomes. This proportion was lower than in the wider systematic review by Barnish et al. [
11] where the corresponding figure was 14 out of 28. Of these five studies, all [
82,
92‐
95] remained classified as negative when only child and maternal health outcomes were considered, although it should be noted that in Fan & Le’au [
92], the effect was only found for neonatal rather than infant mortality.
Among these five studies, two each assessed infant mortality [
82,
94] and neonatal mortality [
92,
95] and one each assessed child mortality [
92] and youth smoking [
93]. None assessed maternal health. Whereas obesity was a common outcome among studies in the wider systematic review, in line with the frequent conceptualisation of the negative impact of globalization and trade liberalisation in terms of an obesogenic environment, obesity was not a measured outcome in any of the studies specifically in the child and maternal health context. While Fan & Le’au [
92] found support for a negative impact of increased globalisation on neonatal mortality, obesity and overweight were only assessed in an adult population. Therefore, theory 6, focusing on obesity, was not supported by the evidence. However, some of the broader conceptual points behind this theory found support in the evidence. In the context of youth smoking, Maynard [
93] provides support for the idea that increasing international trade dependency and membership in neoliberally oriented international trade bodies leads to increased youth smoking by increasing the power of multinational companies in the LMEC context resulting in greater availability and promotion of tobacco and weaker tobacco control policies. All the five studies considered exclusively the LMEC context, which is coherent with a theory proposing that the negative effects of globalisation on health largely operate in an LMEC context. Economic aspects of globalisation were dominant, also being the focus of Shandra et al. [
82], Shen and Williamson [
94] and Shen and Williamson [
95], which linked aspects of increased trade dependency to child and infant mortality. This conceptualisation is referred to in the literature as dependency theory. However, Fan and Le’au [
92] show that social and political aspects of globalisation may also play a role. Across studies, a revised version of theory 6 (see Table
3) gains substantial support, with economic factors clearly dominating the evidence base.