Interventions to tackle health inequalities in cardiovascular risks for socioeconomically disadvantaged populations: a rapid review

Abstract Introduction Cardiovascular disease (CVD) has shown significant health inequalities for people with low socioeconomic status associated with more risk factors. This review was to synthesize interventions that targeted CVD risks and outcomes among socioeconomically disadvantaged populations and to understand the impact associated with these interventions. Sources of data Cochrane CENTRAL, MEDLINE, Embase, PsycINFO and CINAHL were searched for records published in the last decade using a systematic search strategy, complemented by screening the reference lists and citation indexes. Nineteen studies were included and a narrative synthesis with the effect direction plot was undertaken in which studies, interventions, participants and outcomes were examined according to the intervention type focusing on behaviours, lifestyle, education, medication and monitoring. Areas of agreement No universal definition of disadvantaged socioeconomic status was used with common factors relating to racial/ethnic minorities, low income and low or no health insurance. Mixed effects of interventions were reported on clinical outcomes including weight, body mass index, blood pressure, glycated haemoglobin and cholesterol. Areas of controversy Inconsistent effect was reported due to a large variety of settings, participants and intervention components although they are considered necessary to address the complex health needs of socioeconomically disadvantaged populations. Growing points There is inadequate evidence to determine whether any of the intervention types are effective in optimising lipids management for socioeconomically disadvantaged populations. Areas timely for developing research Research is needed with mixed evidence using real world evaluation and lived experience combined with health economic evaluation, on both mental and physical health outcomes.


Introduction
Hyperlipidaemia is characterized by elevated levels of lipids caused by acquired and genetic disorders.It is a chronic progressive disease associated with the development of cardiovascular disease (CVD), a leading cause of mortality resulting in nearly 18 million deaths annually, representing 32% of all deaths worldwide. 1CVD is caused by thrombosis or atherosclerosis restricting blood flow and is commonly presented as coronary heart disease (including angina and myocardial infarction), stroke, transient ischaemic attack and peripheral arterial disease.It is estimated that CVD could cost £9 billion in healthcare per year. 2 CVD risk can be reduced by modifying blood lipid profile targeting total cholesterol, non high density lipoprotein cholesterol and triglyceride level.Both national and international guidelines recommend assessment and management strategies including blood tests, statin treatment, modification of other risk factors such as smoking and obesity, management of secondary causes of dyslipidaemia and outcomes monitoring. 3Evidence also supports the effectiveness of lipid lowering therapies 4 and lifestyle modifications 5 ,6 in preventing CVD in adults.Consequently, there has been a decrease in overall CVD incidence over the last three decades with a stable mortality-to-incidence ratio worldwide. 7hilst the National Health Service (NHS) Long Term Plan has set up CVD ambitions for the next 10 years targeting atrial fibrillation, blood pressure and cholesterol, 8 CVD has shown significant health inequalities for people with low socioeconomic status associated with less access to care and more risk factors.People in the most deprived areas in UK were four times more likely to die prematurely due to CVD than those in the most affluent areas from 2017 to 2019. 9Also, high blood pressure is 30% more likely to be detected in the most deprived areas which presents the biggest single risk factor for heart attack and stroke. 9This could be attributed to a range of biological, behavioural and psychosocial risk factors that are more prevalent in disadvantaged individuals. 10The COVID-19 pandemic has further amplified the problems experienced as lockdowns, quarantines, and closure of some supporting services have all disrupted care and exacerbated health inequalities in CVD.This may result in a further considerable increase in CVD incidence, particularly with acute pathologies such as stroke, acute coronary syndrome and cardiogenic shock among individuals with lower socioeconomic status and vulnerable elderly populations. 11nterventions should target specific risk factors associated with low socioeconomic status when aiming to improve health outcomes.However most trials and evidence have not been adequately powered to engage people with low socioeconomic status in detecting effects 5 in improving CVD events, 12 mortality, 13 hypertension, 14 diabetes incidence, 15 metabolic syndrome, 16 diet 17 and physical activity 18,19 as well as reporting intervention harms. 20,21As such, there is a limited evidence base for interventions targeting socioeconomic disadvantage.In practice, the NHS Health Check launched in 2009 was designed to enable early detection of stroke, kidney disease, heart disease, type 2 diabetes or dementia amongst adults in UK aged 40 to 74. 22 However there has been poor engagement reported with the most disadvantaged groups with a higher risk of developing CVD. 23This led to a call for action to increase awareness and uptake from Public Health England in 2014, yet data suggests that less than half of the socioeconomically disadvantaged populations attended and received follow up support. 24Implications for both research and practice highlight that there is a need to investigate optimised interventions tailored for the characteristics and needs of individuals with low socioeconomic status.With limited but emerging research set up tackling health inequalities, there is no synthesis of current literature of interventions targeting CVD risks and outcomes for socioeconomically disadvantaged populations, hence this review.
The aim of this review was, therefore, to synthesize interventions that targeted CVD risks and outcomes among socioeconomically disadvantaged populations and to understand the impact associated with these interventions.

Methods
This review was undertaken and reported following the Cochrane Rapid Reviews Guidance 25 and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. 26It has been registered with PROSPERO (registration number CRD42022348881).The protocol has been published 27 elsewhere.

Type of studies
This review focused on empirical studies published in peer-reviewed scientific journals, within the last 10 years (to mirror the NHS long term plan) and in the English language.To ensure a degree of commonality in the health system as well as socioeconomic and demographic content, studies were included only if they were conducted in Organization for Economic Co-operation and Development (OECD) countries. 28

Type of participants
Studied were included if they involved adults with common CVD comorbidities who were from disadvantaged socioeconomic backgrounds (income, education, social class, deprivation, poverty or an areabased proxy for deprivation derived from place of residence).Comorbidities were referred to as conditions that can increase the risk of developing CVD including hypertension, diabetes mellitus, chronic kidney disease (CKD) and dyslipidaemia. 29

Type of interventions
Multifaceted interventions were included due to the need for the intervention to improve multiple factors associated with low socioeconomic status.

Type of outcome measures
There is no universal recommendation for the core outcomes sets in studies on CVD prevention, 30,31 studies were included regardless of outcomes measured or reported for health outcomes.This may include vascular related outcomes, cognitive and functional outcomes, lifestyle, medical risk factors, cardioprotective medications and patient reported outcome measures.Any measures of professionals', patients' and/or families' knowledge, attitudes or satisfaction were also included.

Data sources and search strategy
Detailed search strategies for Cochrane CENTRAL, MEDLINE, Embase, PsycINFO and CINAHL were developed by YF refined by NM and validated by an information specialist.Boolean operators were also used to maximize the retrieval of relevant records (supplemental material 1).The searches were conducted on May 26, 2022.Identified citations were exported to Endnote v20 for initial deduplication before being exported to Rayyan for title and abstract screening.This was conducted by NM and a random 10% of citations were independently screened by YF and KT.Full texts were retrieved and screened where citations appeared to meet the eligibility or where a decision to exclude could not be made on the information provided.Reference lists and citation indexes of relevant articles were scrutinized.Any discrepancies were resolved by discussion amongst the research team.

Data extraction
A data extraction sheet was developed and further piloted with two retrieved studies including the author's last name, publication date, location and setting, study design, the aim of the study, a brief description of the intervention, inclusion and exclusion criteria, method of recruitment, outcome measures, participant characteristics (number, gender, age and ethnicity) and primary findings.Where a study appears to have multiple citations, all information from multiple citations was used.

Quality assessment
Quality appraisal of included studies was performed using modified versions of the Critical Appraisal Skills Programme (CASP) tool.Any discrepancies were resolved by discussion amongst the research team.The decision was made to include all papers in data extraction irrespective of methodological quality to provide a complete overview of the existing literature.

Data analysis
Due to the level of heterogeneity of study settings, participants, intervention components and outcomes reported, a narrative synthesis was undertaken by YF, GY and NW, validated by KT to focus on the intervention components, reported effects and mechanisms leading to the outcomes.Interventions and outcomes were grouped according to the design and elements of the intervention and the effect size and 95% confidence interval reported.The effect direction plot table was made to support the synthesis and visualization of effect direction data according to the outcomes reported across the included studies indicating the impact on health outcomes, together with shades to represent study quality. 32

Study selection
A total of 24 136 records were yielded from the initial search and nine records were further identified from citation chaining.This resulted in 16 812 records after 7333 duplicates were removed.Following the screening of titles and abstracts, 76 studies were retrieved in full text and 56 were further excluded with common exclusion mainly due to lack of description of disadvantaged socioeconomic status (n = 20), intervention (n = 15) or outcomes (n = 9).A total of 20 citations were included with two 33,34 reporting the same study (Fig. 1).

Risk of bias
Quality assessment was completed for each included study.Of 13 RCTs, 10 33,35,39,41,43,44,47,48,50,52 were rated as low risk of bias and three 40,46,49 as medium risk of bias largely due to the inadequate description of the blinding process, the precision of the estimated effect and potential harms and costs.Two cohort studies 38 ,45 were rated as low risk of bias, two 37,42 medium risk of bias and one 36 high risk of bias.

Intervention and reported effects
Of 10 studies that reported the change in weight, six observed a decrease and three observed no difference.The conclusion could not be drawn in one study where no sufficient information was reported. 36Six of the 10 studies that reported the change in BMI observed a decrease, three observed no difference and one comprised insufficient information 36 to be asssessed.The majority of the studies that reported BP, total cholesterol, LDL and HDL observed no difference at follow up.A reduction was observed in all studies that reported HbA1C except for no difference in one study. 44Reported outcomes are presented in Table 2.

Behavioural interventions
Four studies 35,47,48,50 tested behavioural interventions for the management of weight, diabetes, hypertension and CKD.Two tested a behavioural weight management programme comprising cognitivebehavioural interventions, self-monitoring, dietary and physical activity advice and skills training.One compared with the usual care in populations with low literacy and limited access to health services, 35 and the other compared with group-based advice and support on diet and physical activity from the practice nurse with people living with high levels of social deprivation. 50Both studies reported lowered weight (−1.07, 95% confidence interval (CI) −1.94 to −0.22; −1.9, 95% CI -3.7 to −0.1) and BMI (0.41, 95% CI -0.73 to −0.09; −0.7, 95% CI -1.3 to 0.0) but no differences were identified in blood pressure.
One study compared usual care in people in low income, Spanish-speaking Latinos with type 2 diabetes with the special intervention comprising integrated medical and behavioural co-located visits, group-based diabetes self-management education sessions and care coordination. 47HbA1C was lowered (−0.32, 95%CI −0.49 to −0.15), but cholesterol remained unchanged.
In a 2×2 study, 48 patients with CKD from the safety-net primary care clinics received access to the CKD registry with feedback or a self-management programme or both.However no difference was observed in SBP in any of the intervention groups.

Lifestyle interventions
Five RCTs 33 ,39,40,43,44 and two cohort 36 ,37 studies investigated the effect of lifestyle interventions comprising physical activity, dietary intake support, education and skill building on metabolic syndrome, 36 CVD 37,39,40 and weight management. 33,43,44Results were unable to be synthesised in one 36 of the cohort studies rated a high risk of bias due to the lack of 95% CI or P-value reported.In the other cohort study, 37 participants with lower socio-economic status and less access to health care were provided with access to physical activity, dietary intake activity and heart-healthy education.Participants at the follow up experienced a weight reduction (P < 0.001, neither mean difference (MD) nor 95% CI reported) and BMI (P < 0.001), but their systolic and diastolic BP remained unchanged.
Three of the five RCTs reported the effects on weight at follow up compared with the usual care in low income populations.Two studies 33 ,43 reported a greater reduction (−4.51 kg, 95% CI −6.01 to −3.02; −2.5 lb, 95% CI −4.25 to −0.75) and one 44 reported weight as unchanged.
→ (phase2) (%,phase1) (%,phase2) : no change * For RCTs, this relates to the reported difference between intervention and control arms at follow up; For cohorts, this relates to the reported difference between baseline and follow up.* * RCTs where usual care was not used as control arm a = low risk of bias; b = some concerns; c = high risk of bias Four 39 ,40,43,44 RCTs reported the effects on BMI at follow up compared with CVD risk screening plus education in women who were un-or underinsured in two studies 39,40 and with usual care in low income populations in another two studies. 43 ,44 lowered BMI (P = 0.03, neither MD nor 95%CI reported; −0.46, 95%CI −0.76 to −0.14) at follow up was observed in two studies 40,43 which remained unchanged in the other two studies. 39,44tudies 33,39,40,44 which investigated the effects on both SBP and DBP reported no difference in BP between the intervention and the control at follow up.
Two RCTs reported the effects on HbA1C in low income populations 43 ,44 .Compared with usual care, a reduction (−0.07, 95%CI −0.10 to −0.04) was observed in one study 43 whereas no difference in the other study. 44our 39 ,40,43,44 studies reported the effects on cholesterol in women who were un-or underinsured 39,40 or low income populations. 33 ,44No differences were observed in total cholesterol and LDL at follow up.There was likely increased total cholesterol (P = 0.02, neither MD nor 95%CI reported) and LDL (P < 0.01, neither MD nor 95%CI reported) at 12 weeks, but they remained unchanged at 12 months. 40An increased HDL (4.6, 95%CI 2.9 to 6.3) was reported in one study, 33 but it was unchanged in the other study. 44

Education-based interventions
Five studies 38,42,45,49,51 including four cohorts 38 ,42,45,51 and one RCT, 49 investigated education-based interventions combining information sessions, coaching sessions and motivational interviewing for lifestyle recommendations.Particpants included people with diabetes, 38,45,49 CVD 42,51 and hypertension, and hyperlipidemia 38 and who were identified as either low income, 38,45,49 older women with histories of substance abuse and homelessness 42 and ethnic minorities. 51f the three studies 42,45,49 that reported the effects of the intervention on participants' weight change, one 49 observed a greater reduction in pounds (lbs) (P = 0.044, neither MD nor 95%CI reported) compared to the usual care at 6 months and the other two studies 42,45 identified no difference.
The outcome of cholesterol varied across studies that reported total cholesterol, LDL and HDL.Total cholesterol was lowered (P < 0.001, neither MD nor 95%CI reported) in one study 38 but unchanged in the other study. 42LDL was lowered (P = 0.04, neither MD nor 95%CI reported 38 ; −7.2, 95%CI −4.7 to −9.7 51 ) in two studies but unchanged in one study. 45DL was increased (1.2, 95%CI 2.1 to 0.3) in one study but unchanged in the other study. 38

Medication-based interventions
One RCT investigated the effects of medicationbased interventions on people living with hypertension who had low income. 41It compared interventions comprising an algorithm of antihypertensive medication adjustments, higher antihypertensive medication if needed and weekly telephone health coaches with patients receiving all but without antihypertensive medication. 41Both groups had a reduction in SBP and DBP, but no difference between them.However, when data from the two groups were combined, SBP significantly decreased by 21.8 mmHg between baseline and 6 months (P < 0.001).This suggested that health coaching itself was associated with improved blood pressure.

Monitoring-based interventions
Two RCTs studied the effects of monitor of BP and glucose amongst people with limited health insurance 52 and those from underserved areas. 46A greater reduction in SBP was found in both those who had low literacy with high information needs (−8.8, 95%CI −14.2 to −3.4), 52 and those who had the lowest income level (−4.23,P = 0.019, no 95%CI reported). 46A lower HbA1C but unchanged LDL was reported (−0.5, P ≤ 001, no 95%CI reported) in those with the lowest income level. 46

Discussion
This review synthesised interventions targeting CVD risks and outcomes and the effects reported on clinical outcomes for socioeconomically disadvantaged populations.A total of 19 studies with mixed quality of evidence were included resulting in five types of multifaceted interventions that were based on behavioural change, lifestyle, education, medication and monitoring.Mixed effects were reported for clinical measures that include weight, BMI, BP, HbA1C and cholesterol with inadequate evidence to determine whether any of the intervention types are effective in optimising lipids management for socioeconomically disadvantaged populations.
There was a limited definition of socioeconomically disadvantaged populations observed across the included studies.Although this review was set up to include populations with low levels of income, education, social class, deprivation, poverty, or an area-based proxy for deprivation, studies included commonly recruited participants according to their income, ethnicity or health insurance.None related to educational attainment or employment status, which are also important makers of socioeconomic status associated with CVD outcomes, particularly in high-income countries.There is a strong association between education and health literacy, which is likely found to be low in those who experience increased all cause mortality 53 or with low or no compliance with their medications. 54In another study, the unemployed population showed an increased risk of CVD events than the retired cohort after controlling for demographic factors, 55 indicating that job loss could lead to the negative effects of unemployment.Given the dynamic changes between these factors and CVD risks in one's life, multiple markers of socioeconomic status should be used in research and practice in predicting CVD risks or outcomes.
None of the included studies measured mental health except one which reported the unchanged mental health status, as part of the 12-item Short Form Health Survey, being unchanged compared with usual care. 48There is evidence increasingly suggesting that psychological factors affect socioeconomically disadvantaged populations and their CVD outcomes.Individuals with low income who experienced stress and depression reported over 45 and 30% higher risk of developing CVD and all cause mortality respectively after controlling for demographic, clinical and behavioural factors. 56imilarly, a higher risk of CHD mortality was also reported in another study amongst those with both low socioeconomic status and psychological distress. 57The inequalities in risks may highlight inadequate resources to address psychological distress events and related health behaviours including physical inactivity and smoking, at both patient and practice levels.Future interventions and research should also evaluate the impact of interventions on mental health wellbeing targeting CVD in socioeconomically disadvantaged populations.This review showed the inconsistent impact of multifaceted interventions on lipid management for socioeconomically disadvantaged populations, given the wide variety of settings, intervention components, approaches and targeted populations.Although interventions for CVD in general populations seem effective, difficulty in concluding the effectiveness of any interventions for vulnerable groups has been reported by the previous literature. 58his issue with large variety has also been suggested in the past literature as a barrier to determining the effectiveness of the interventions for socioeconomically disadvantaged groups, 59 however it is required and almost necessary to address extra care needs of these specific patient groups. 60Reviews with less broad eligibility focusing on more specific cohorts with subgroups analysis may be valuable to detect breakdown effects.It is also worth noting that most of the interventions examined in this review were adapted to specific research settings requiring extra resources, for example, materials being translated and contents being simplified for readability.This raises a question on the sustainability of scale up implementation within health systems.Future research using health economic evaluation is needed to confirm the cost-effectiveness.
Although this review had no limitation on study design when searching records in the databases, only studies with RCTs and cohort designs were included.Future studies are needed using qualitative or mixed methods to reflect lived experience and describe barriers and challenges in intervention delivery and implementation in real world settings.

Limitations
This review was limited by the fact that only studies undertaken in OECD countries published in English were included.This was to ensure the similarity of healthcare systems and socioeconomic and demographic structure, therefore the findings may be less generalisable for socioeconomically disadvantaged populations in low-or middleincome countries, where research reporting CVD in socioeconomically disadvantaged populations is limited often with conflicting results. 61,62In addition, potential theoretical bias may exist given the intervention was delivered or facilitated by either clinicians or researchers which may result in a placebo effect.

Conclusion
This review synthesised 19 studies presenting five types of intervention type focusing on behaviours, lifestyle, education, medication and monitoring.Definition of disadvantaged socioeconomic populations was inconsistently used to describe mainly relating to racial/ethnic minorities, low income and limited or no health insurance.There is inadequate evidence to determine whether any of the intervention types are effective in optimising lipids management for socioeconomically disadvantaged populations, due to a large variety of settings, participants and intervention components although they are considered necessary to address the complex health needs of socioeconomically disadvantaged populations in practice.Future research is needed with multi-factor defined populations using mixed evidence using real world evaluation and lived experience combined with health economic evaluation, on both mental and physical health outcomes.

:
positive health impact; : negative health impact;

Table 1
Characteristics of included studies (n = 19)

Table 1 Continued Citation Country; setting; SES definition Design; intervention length; Follow up Components of intervention (I) vs Control (C)
The Fat and Fibre Questionnaire; FVS: All Day Fruit and Vegetable Screener; CHAMPS: Community Healthy Activities Model Program for Seniors

Table 2
Effect direction plot for reported outcomes of included studies (n =19)