Background

This report describes an attempt to look at how much health promotion and public health research relating to young people has tackled health inequalities, and in what ways it has done this. Health inequalities are recognised as an important problem nationally and internationally. There is policy interest in improving the health of the most disadvantaged, reducing the gap between the most and least disadvantaged, and reducing gradients across the whole population. Health inequalities arise from variations in social, economic and environmental influences along the life course. Health promotion, particularly when it uses social and structural interventions developed by multi-disciplinary teams working with young people, not merely for them, has the potential to reduce health inequalities among young people immediately, and in their later lives.

Inequalities research draws on a range of evidence. Observational studies describe the magnitude and severity of health problems and of inequalities. The findings of these studies can guide the targeting of interventions to reduce disadvantage or inequalities. Interventions evaluated using controlled trials or some other evaluation design can answer questions about effective ways of reducing inequalities. Research reporting people’s views and experiences adds valuable qualitative data to observational and intervention evaluations.

Research questions and methods

The review described in this report had two aims: to provide a descriptive map of the extent to which health promotion and public health intervention research has focused on inequalities in young people’s health; and to look in more detail at the methods used by researchers for defining and measuring inequalities. We defined ‘young people’ as those aged from 11 to 21. Addressing inequalities requires methods for including diverse populations in research and discriminating between them. Methods for the latter are addressed in this report; methods for the former are addressed in another study conducted at the same time.

What research has been done? The systematic map

The first aim of the work was to identify and describe the available research evidence in order to answer the following questions:

• How much research activity in health promotion and public health has addressed health inequalities among young people?
• What types of research have looked at gaps or gradients in health status?
• How much of this research specifically relates to socially disadvantaged and vulnerable young people (those considered at risk in various ways; those with social or complex needs; those living in marginalised communities; looked-after young people; young parents; school non-attendees)?
• How much of the research in this area addresses the impact of structural interventions?

We answered these questions by looking at a group of studies derived from two sources: records of health promotion and public health research included in commercial databases which express an explicit focus on inequalities (the ‘inequalities research dataset’); and studies included in BiblioMap, a register rich in intervention evaluations as it contains the Trials Register of Promoting Health Interventions (TRoPHI), a web-based database of randomised controlled trials and controlled trials (non-randomised) of public health and health promotion interventions hosted by the EPPI Centre (the ‘intervention research dataset’).

Criteria for studies to form part of the map included comparing the health status or two or more demographic groups and having a publication date of 1996 or later. These criteria were first applied to titles and abstracts. Potentially relevant studies were subsequently re-screened using the full paper. We used a coding scheme to classify studies. The codes enabled us to describe the literature at a broad level, including different study types, settings, topics, populations and outcomes. We also coded studies according to whether they presented an analysis of inequalities in terms of gaps, gradients or both.

How has the research been done? The methods study

The second aim of this work was to look at the methods that researchers have used to define and assess health inequalities in outcome evaluations of interventions. We asked how existing research has:

• described populations and measures of disadvantage;
• described health status and difference between populations;
• incorporated the experiences and opinions of young people into the development and evaluation of interventions; and
• recruited and retained young people participating in intervention evaluations.

To answer these questions we used a different set of studies: intervention evaluations included in previous EPPI Centre systematic reviews in areas where young people experience health inequalities (healthy eating, mental health, physical health, teenage pregnancy and teenage parent support) (Harden et al. 2001, Harden et al. 2006, Rees et al. 2001, Shepherd et al. 2001).

A key issue for researchers in the area of health inequalities is how to define and assess socio-economic position (SEP). In examining this, we used the classificatory framework PROGRESS (Place of residence, Race /ethnicity, Occupation, Gender, Religion, Education, Socio-economic status and Social capital) (Evans and Brown 2003). We expanded PROGRESS to form PROGRESS-Plus (Kavanagh et al. 2008) with the addition of the three further variables of age, disability and sexual orientation, which are also implicated in health inequalities and addressed by discrimination legislation to support equal opportunities and human rights, (Department of Health 2007, US Department of Health and Human Services 2000) and specific vulnerable or excluded groups.

Results

What research has been done? The systematic map

Our systematic map of the research included studies across a wide range of health promotion and public health topics, settings and study designs. There were a total of 191 studies: 154 were found in the inequalities research dataset and 43 in the intervention research dataset, with 6 common to both. The most common health status measured in the inequalities research was physical health (Body Mass Index (BMI), disease) and, in the intervention research, health-related behaviours. Most of the studies were conducted in the USA (55% in the inequalities and 72% in the intervention research datasets).

Most of the studies used observational designs. We found only 46 outcome evaluations of health promotion and public health interventions that addressed health inequalities by comparing distinct groups; only 12 of these studies evaluated structural interventions or environmental modifications, and 6 evaluated interventions at the level of social networks. Most studies sampled broad populations rather than well-defined disadvantaged groups.

The most common difference examined in the research was gender (56% of the inequalities and 81% of the intervention research datasets), followed by ethnicity (56% and 35%) and SES-relevant differences (55% and 21%). Almost half the studies included comparisons relevant to SES. A wide range of methods were used to measure socio-economic status, including single measures such as occupational class, parental education and income, and multiple measures comprising combinations of these. Ten studies used nine different composite measures of SES.

The bulk of the studies in the inequalities dataset contained data relevant to both gaps and gradients (51% compared with 23% of the studies in the intervention dataset). In the inequalities dataset, 25% of the studies focused on gaps only (mostly gender comparison) and 23% on gradients. The comparable figures for the intervention research dataset were 67% gaps studies (again, mostly focused on gender) and 9% gradients studies.

How has the research been done? The methods study

Most of the outcome evaluations of interventions we reviewed in depth recruited young people through schools or agencies such as social services. In many studies the recruitment methods used would have excluded the most disadvantaged. A notable feature of these studies was that, although most (n=21 of 28) gave figures for attrition, only about half of these (n=11) reported on the socio-demographic characteristics of participants who dropped out.

Few studies involved young people or their parents actively in choosing research priorities or intervention aims (n=2), or developing interventions (n=3). More elicited their views as research participants for the needs assessment (n=4) or process evaluation (n=7).

Most studies did not explicitly aim to measure or reduce inequalities. Half the studies (n=15) provided subgroup data, but this was not always analysed appropriately.

Recommendations and implications

There is an extensive research literature to inform policy interest about the causes of health inequalities and how they develop along the life course. There is less research directly addressing how to reach the policy goal of reducing inequalities. We sought a literature about health promotion, inequalities and young people’s health, but found instead largely disconnected literatures of: observational studies addressing inequalities; discourses about inclusive research and public involvement for tackling inequalities; evaluations of health promotion (but not of its effects on inequalities); and methodologies for assessing health inequalities that had been applied almost exclusively to observational studies. Knowledge from these different literatures needs to be joined up in order to build an evidence base to support effective health promotion for young people that reduces, or at least does not increase, health inequalities. There are implications throughout the research pathway for: research priorities, the research community, study designs, methods for data collection and analysis, reporting and publishing.

Research priorities

Systematically mapping the literature and critiquing health promotion intervention studies in terms of inequalities research methods exposed a number of gaps in the evidence base. There is a need:

• to investigate appropriate research methods for the evaluation of interventions intended to reduce inequalities. This includes questions of when to conduct subgroup analysis and with which groups.
• to conduct high quality outcome evaluations of interventions which compare outcomes between different PROGRESS-Plus groups, especially SES comparisons. There is a particular need to conduct such evaluations in the UK. There is also a particular need to conduct such research with vulnerable groups.
• for rigorous evaluations of the effects of structural and social support interventions which earlier reviews have highlighted as having potential for reducing inequalities.
• for process evaluations which can provide information on the implementation of interventions and their acceptability to young people and their families.

These research needs should be considered alongside the findings of the priority setting exercises undertaken by the Cochrane Equity Field and the Cochrane Health Promotion and Public Health Field that are to be discussed at the Cochrane Colloquium in October 2008.

Research community

To address the effects of interventions on health inequalities, the research community, when funding, designing, reporting or publishing research, or reviewing its science or ethics, needs to draw on its knowledge and experience of:

• developing interventions for reducing inequalities, such as structural interventions or social support;
• experimental study designs for evaluating social interventions;
• measures of health inequalities;
• working with disadvantaged or marginalised groups in order to develop and implement inclusive recruitment strategies and minimise attrition;
• working in partnership with potential recipients in guiding public health research which addresses their health needs. 

When considering individual studies, and in the course of wider debate, it is beneficial to convene mixed groups, in terms of academic disciplines and professional and personal roles, in order to exchange such knowledge and experience.

Study design

Investigating inequalities requires study designs to take into account dimensions of inequalities pertinent to the focus of study, and whether the study aims to generate hypotheses or draw conclusions about the effects of interventions on inequalities.

• Much of the literature about investigating inequalities argues for attending to dimensions of inequality other than SES both for their own sake and in order to describe their interactions with SES as determinants of health; similar thoughtful approaches are needed to inform intervention studies taking into account health inequalities.
• The overlapping discourses of social determinants of health and equality of opportunity can inform the choice of population descriptors in studies of health inequalities. PROGRESS-Plus is a practical tool spanning these discourses for characterising populations when planning and reporting primary research, including sampling frames, recruitment and data collection, and analysis (including attrition). Data need not be collected for every dimension of PROGRESS-Plus; rather it provides a framework from which to choose appropriate dimensions for investigation.
• Primary studies drawing conclusions about the effects of interventions on inequalities need to be sufficiently large to allow subgroup analyses sensitive to statistically significant differences.
• Smaller studies may explore potential differences with subgroup analyses for the purpose of hypotheses generation. Hypotheses may be tested subsequently in sufficiently powered primary studies or in systematic reviews with statistical meta-analyses.

Data collection

The choice (or dearth) of socio-demographic data reported in some intervention studies suggests that this literature largely fails to draw on lessons learnt from the wider inequalities literature.

• It is essential to determine whether or not a study has an explicit focus on inequalities baseline socio-demographic data which is collected, reported and linked to outcomes data, in order to expose instances of interventions inadvertently increasing inequalities.
• Studies of inequalities in socio-economic status need to employ measures of SES that are meaningful and feasible for particular populations and settings. For instance, young people may offer more accurate descriptions of housing conditions than of parental occupations.
• There is a need to assess the validity of composite measures of socio-economic status for young people.

Analysis

Analytical methods need to be theoretically and technically sound:

• The rationale for comparing different socio-economic groups should be clear.
• In particular, there is a need for better theorising about the role of gender and ethnicity in addressing inequalities through subgroup analyses.
• Where subgroup analyses are used to investigate inequalities, they should be pre-specified and have an appropriate rationale. Subgroup analyses which are under-powered to investigate differential effects in evaluations of interventions should only be used for hypothesis generation.
• Attrition data should be reported and linked to socio-demographic data.
• There is a need to explore the potential of area measures for experimental designs, particularly the value of using area measures: as background information comparable to official national statistics; as inclusion criteria; and as analytical factors for understanding the effects of interventions and effect modifiers.

Reporting and publishing

To build an evidence base, individual studies need to contribute comprehensive reporting of socio-economic data, either in journals or in publicly accessible reports. Funders and publishers are well placed to encourage:

• better and fuller reporting of socio-demographic data of participants in primary research, if necessary through publicly accessible electronic appendices.
• reporting comparisons of health status between different socio-economic groups in abstracts.

This accumulative evidence base would be more accessible through bibliographic databases where structured abstracts and keywords include terms relating to health inequalities.

This report should be cited as: Oliver S, Kavanagh J, Caird J, Lorenc T, Oliver K, Harden A, Thomas J, Greaves A and Oakley A (2008) Health promotion, inequalities and young people’s health: a systematic review of research. London: EPPI Centre, Social Science Research Unit, Institute of Education, University of London.