Table 2.

Summary of the published evidence contributing to each principle.

Principle	First author (year)	No. of included reviews	Domains covered	Evidence
Healthy-by-default and easy to use initiatives	Thomson (2018) [17]	29 SRs (150 studies)	Tobacco, alcohol, nutrition, reproductive health, infectious disease control, the environment, workplace regulations	•2 studies found USA food stamp (subsidy) programme had positive impacts on foetal survival and weight gain during pregnancy of low-income populations. •9 studies 10–20% increased intake of targeted foods or nutrients of participants in food subsidy programme •4 studies of taxes on unhealthy foods and drink showed positive equity effects on diet outcomes •1 SR found significant drop in casualties in the more deprived areas, compared to the less deprived areas from speed limit interventions. •1 study found reduced absolute inequalities in dental caries between the most affluent and least affluent areas associated with intervention that provided fluoridated toothpaste and daily toothbrushing supervision for 5-year-olds. •2 studies found evidence that fiscal incentive schemes (maternity allowance, childcare benefits) may decrease inequalities in vaccination rates.
	Eyles (2012) [50]	32 studies	Nutrition and diet	•11 out of 14 studies reporting impacts by SES found pro-health and pro-equity outcomes for food taxes and subsidies (although many note that taxes would be regressive with more financial burden on low-income individuals).
	McGill (2015) [60]	36 studies	Nutrition and diet	•10 of 18 “price” interventions were likely to reduce inequalities by improving healthy eating outcomes more for individuals of low SES, particularly when interventions were a combination of taxes and subsidies with all 6 respective studies reducing inequalities. •4 of 6 “place” interventions reduced inequalities and none widened them. •8 of 19 “person” (individual-based information and education) interventions widened inequalities.
	Cauchi (2016) [25]	63 SRs	Childhood obesity	•48 studies with positive outcomes reported the following effective environmental strategies: improving overall school food environment (nutrition standards, reformulating school lunches, removing vending machines/banning sale of sugar sweetened beverages/snacks high in fat, sugar, or salt), purchasing new PE/sports equipment, daily formal physical activity sessions, providing free or low-cost fruit, making playgrounds available for physical activity after school hours, providing free/low-cost water, providing healthy breakfasts at school, substituting sweetened beverages, reducing screen time at home.
	Beauchamp (2014) [54]	14 studies	Obesity	•5 of 6 interventions with a positive equity impact included structural changes to support behaviour change, 5 had a wide reach (3 community-based and 2 school-based), and all were multi-year in duration. •4 of 5 interventions with no beneficial impact among lower SES groups had low structural changes and 1 had moderate amounts of structural change, 3 were very short term (2–10 weeks), and 4 were based solely on information delivery.
	Durand (2014) [55]	19 studies	Shared decision-making	•5 of 7 studies differentiating outcome by disadvantage/literacy levels reduced disparities in knowledge, decisional conflict, uncertainty and treatment preferences suggesting SDM interventions could narrow health disparities by promoting skills/resources needed to engage in SDM.
	Moore (2015) [61]	20 studies	Universal school-based interventions on health behaviours	•Of 4 education-based interventions, 1 widened inequalities and 3 had a neutral effect. •Of 4 environmental interventions, 1 reduced inequalities and 3 had a neutral effect. •Interventions combining education and environmental change had mixed results.
	Carey (2019) [20]	6 studies	Personalisation schemes	•Accessing and benefiting from schemes based on personalisation requires high levels of skills and resources at the individual level. •Identified factors associated with better outcomes in personalisation schemes were higher levels of economic, cultural, social, and symbolic capital in the forms of education, being employed, having capable networks and support, knowledge and skills in navigating complex systems, household income, knowledge of where to access information and the capacity to self-manage individual budgets.
	Cairns (2015) [56]	18 studies	Obesity	•0 of 11 counselling or advice-based interventions reduced inequalities in obesity.
	Craike (2018) [28]	17 SRs	Physical activity	•1 SR found that 2 of 4 universal policies showed a positive equity impact on children's physical activity levels: provincial school physical education policy requiring students to take physical education to graduate from secondary school and a children's fitness tax credit.
	Haby (2016) [24]	15 SRs, 7 economic evaluations	Agriculture, food, nutrition	•1 SR reported on health inequality impact found reduction in health inequalities from balancing taxes on unhealthy foods with subsidies on healthy food.
Long-term, multi-sector action	Gibson (2011) [15]	5 SRs (130 studies)	Housing and neighbourhood conditions	•1 SR (72 studies) found highest efficacy in interventions aimed at multiple pathways (rehousing and changes to: indoor equipment or furniture; respondents' knowledge or behaviour; community norms or collective behaviour; housing policy or regulatory practices, and health practitioners' behaviour) and which are ecological (target multiple levels (i.e. individuals, households, housing and neighbourhoods)).
	Craike (2018) [28]	17 SRs	Physical activity	•3 reviews on children found that physical activity interventions, particularly those that were school-based and multicomponent were likely to be effective. Common elements of successful policy-focused interventions included enhancements to physical education, additional physical activity opportunities, school self-assessments, and education about physical activity. •1 SR on all age groups found intensive interventions are most likely to reduce socio economic status inequalities in physical activity.
	McGill (2015) [60]	36 studies	Nutrition and diet	•4 out 6 place-based interventions demonstrated to reduce inequalities were implemented in a range of settings including schools, workplaces, and communities/neighbourhoods.
	Naik (2019) [18]	62 (umbrella, meta-analyses, & narrative)	Macroeconomic determinants	•High quality SR showed evidence of pro-equity impact from taxing tobacco and moderate quality SR found mixed, but mostly positive impact on reductions in preterm births among mothers with low education and black mothers. Supported by findings of 4 other lower quality reviews. •3 reviews (low quality) found some association between unemployment insurance and reduced inequalities and better health outcomes. •4 reviews (moderate to low quality) on gendered health inequalities found positive equity impacts from the dual-earner policy model and welfare conditions reducing job precarity. •2 reviews (moderate quality) found pro-equity impacts of occupational health and safety regulations such as preventing toxin exposures.
	Simpson (2021) [53]	38 studies	Social security policy and mental health	•14 of 21 studies on expansionary policies (increased benefit amount or access) improved mental health; 4 studies evaluated inequalities of which 2 reduced inequalities and 2 had no impact. •11 of 17 studies on contractionary policies (decreased benefit amount or access) worsened mental health; 10 evaluated inequalities which widened in 3, narrowed in 2, and had mixed or no effects in 5.
	Macintyre (2020) [21]	15 SRs (1720 studies)	Adolescent health	•Evidence for market regulation impact in SR on youth smoking found 7 (of 38) studies showed positive impact on inequalities, 16 showed neutral effects, 12 negative impact, 4 mixed and 1 unclear. Taxation/increasing the price of cigarettes had the most evidence for positive equity impact.
Locally designed focus	Cauchi (2016) [25]	63 SRs	Childhood obesity	•Environmental interventions had beneficial equity impacts (ES: 0.09 [0.16, 0.02]). •Community-based interventions of any type & parental involvement resulted in small but consistently positive ES ranging from 0.094 [p = <0.001] to 0.151 [0.334, 0.031].
	Craike (2018) [28]	17 SRs	Physical activity	•1 SR on interventions with pre-schoolers: 6 of 11 included studies showed a significant effect; all 3 community-based interventions were effective. •9 SRs on adults found factors associated with higher effectiveness were: the involvement of the community in the design and implementation of interventions; developing community infrastructure to sustain effective interventions; interventions delivered through personal contact; and tailored interventions. •1 SR on all age groups found community settings were the most effective intervention setting for socioeconomically disadvantaged groups.
	Crocker-Buque (2016) [52]	41 studies	Immunisation	•16 studies on multicomponent locally designed interventions demonstrated higher efficacy from improving immunisation in children and adolescents in the short term for ethnically diverse, low-income populations.
	Pierron (2018) [29]	21 SRs	Supporting parenting	•1 SR found increased effectiveness from diversifying approaches shared between state, school, and neighbourhood organisations and varying intervention to local context and different cultures/societies. •2 SRs reported on necessity of integrating the entire network related to parenting (environment, professionals, organisations, social contexts, etc.).
	Thomson (2019) [30]	15 SRs (157 studies)	Community pharmacy-delivered interventions	•17 studies found increased vaccination rates among people who had missed vaccination the previous year or otherwise wouldn't have accessed vaccination services with pharmacy-delivery and that of those delivered a third of the vaccinations took place outside traditional working hours documenting the increased accessibility provided by community pharmacy networks. •1 study found increased breast and cervical cancer screening uptake among low- and moderate-income women.
Targeting disadvantaged communities	Moore (2015) [61]	20 studies	Universal school-based interventions on childhood health behaviours	•10 of 20 universal interventions had a neutral impact on inequalities. •6 of 20 universal interventions widened inequalities.
	Thomson (2018) [17]	29 SRs (150 studies)	tobacco, alcohol, nutrition, reproductive health, infectious disease control, the environment, workplace regulations	•3 studies documented a widening of socio-economic inequalities from mass media intervention for pre-conception folic acid use from the national campaign (which persisted for 3 years), but not in the local campaign. The studies showed worsening health inequality effects in terms of folate uptake by education level, and the prevalence of neural tube defects by ethnicity. •1 SR found that the Expanded Food and Nutrition Education Program (EFNEP) – a federal community outreach programme targeted at low-income families – increased fruit and vegetable consumption and had a positive effect on health inequalities. •2 studies found interventions targeted toward disadvantaged groups increased screening rates – particularly amongst lower socio-economic groups. •4 studies found positive effects of ‘reminder and recall’ systems when targeted at disadvantaged groups, but that universal systems had no effect on reducing inequalities in vaccine uptake rates. 7 studies found a combination of targeted and universal immunisations improved health outcomes for indigenous populations. •1 study found complex interventions targeted interventions were effective in encouraging child- hood vaccination when specifically targeted at lower SES groups of younger children.
	Cairns (2015) [56]	18 SRs	Obesity	•2 RCTs (strong/moderate quality) demonstrated reduced inequalities in physical activity interventions targeted at low-income workers. •1 observational study (moderate quality) showed increased inequalities from a universally delivered workplace physical activity intervention.
	Bird (2018) [27]	17 SRs	Built and natural environment	•1 SR found provision of affordable and diverse housing was found to be associated with higher or increased physical activity, primarily walking and perceived safety among those from low-income groups. •9 SRs reported that provision of affordable housing to vulnerable individuals with specific needs (those living with intellectual disability, substance users, individuals experiencing homelessness, and those living with a chronic condition) was associated higher or improved social, behavioural, physical and mental health-related outcomes.
	Gibson (2011) [15]	5 SRs (130 studies)	Housing and neighbourhood conditions	•30 studies found warmth and energy efficiency interventions had the clearest positive impacts on health. Interventions that reported the largest effects were targeted at vulnerable groups, including those with existing health conditions and the elderly.
	Durand (2014) [55]	19 studies	Shared decision-making	•3 studies suggested that despite knowledge levels being lower in disadvantaged groups pre-intervention, disparities between groups tended to disappear post-intervention, particularly when the intervention was adapted to disadvantaged groups' needs (e.g. low literacy).
Matching of resources to need	Barr (2017) [62]		NHS resource allocation	•Between 2001 and 2011 the increase in NHS resources to deprived areas accounted for a reduction in the gap between deprived and affluent areas in male mortality amenable to healthcare of 35 deaths per 100 000 population (95% confidence interval 27 to 42) and female mortality of 16 deaths per 100 000 (10–21). This explained 85% of the total reduction of absolute inequality in mortality amenable to healthcare during this time. •Each additional £10 m of resources allocated to deprived areas was associated with a reduction in 4 deaths in males per 100 000 (3.1–4.9) and 1.8 deaths in females per 100 000 (1.1–2.4).
	Barr (2014) [63]		UK Health Inequalities Strategy	•During the strategy the gap in life expectancy for men reduced by 0.91 months each year (0.54–1.27 months) and for women by 0.50 months each year (0.15–0.86 months) compared to increasing inequalities before and after strategy implementation. •By 2012 the gap in male life expectancy was 1.2 years smaller (95% confidence interval 0.8–1.5 years smaller) and the gap in female life expectancy was 0.6 years smaller (0.3–1.0 years smaller) than it would have been if the trends in inequalities before the strategy had continued.

SR = systematic reviews.