BackgroundEthnic health inequalities have long been recorded in Britain. Although socioeconomic deprivation and discrimination are important underlying factors, differential benefit from public health and health-care interventions contribute to these inequalities. Progress is hampered by a scarcity of evidence on what works, creating uncertainty about how to design and deliver interventions so as to mitigate disadvantage. Understanding the links between ethnicity, health care, and health presents substantial challenges in ethnic categorisation, comparability across diversity, elucidating causal pathways, and dangers of stereotyping and stigmatisation. Furthermore, sociologically driven research examining links between ethnicity and health remains largely descriptive. Therefore, that standard systematic reviews and national guidance generally produce limited recommendations for intervention in this area is not surprising. We need more sophisticated and critical approaches to evidence synthesis. We examine the potential of realist synthesis, a theory-driven approach to reviewing research evidence on complex social interventions that aims to show why and under what conditions particular interventions do or do not work. The approach is flexible and iterative, seeking to explicate underlying interventional theories and operational contexts, and often engaging policy or practice decision makers in the process. Such syntheses thereby produce theoretical claims alongside summaries of empirical evidence. MethodsWe undertook a narrative assessment of selected existing studies that used realist synthesis. Starting with a synthesis we had done previously on interventions to enhance minority ethnic women's breast-screening uptake, plus the wider literature on research practice in ethnicity and health, we developed a thematic framework, identifying potential strengths and weaknesses of the approach conceptually and operationally. We then undertook a pragmatic search to identify at least four studies that addressed ethnicity and health and used realist synthesis. The intention was not to be exhaustive, but to include sufficient studies to identify important common issues. The thematic framework loosely guided our exploration of how the realist synthesis performed across the studies and identified strengths and weaknesses. FindingsHow the realist synthesis approach was used across studies varied, in part suggesting different interpretations of the method itself, but also showing pragmatic decisions in the face of available material. In particular, researchers sometimes struggled to elucidate underlying interventional theories and found disappointing detail on context in many studies. Nevertheless, the approach had important potential benefits. For instance, the process of unpacking interventions to understand their logic meant that taken-for-granted understandings of the causes of ethnic inequality were explicitly examined. Similarly, the focus on identifying mechanisms usefully diverted attention away from fixed ethnic categories (and worries about their comparability) towards more general processes of exclusion and inclusion. The method's explicit consideration of context was helpful in overcoming the tendency to view the contingent nature of ethnic identity as an obstacle to synthesis. Also, more generally, the inclusion of diverse published work to identify and confirm emerging theory produced richer or unanticipated findings, or both. The resulting insights thereby had the potential to shift policy and practice intended to reduce inequalities in useful new directions. InterpretationRealist synthesis holds promise for improvement of the evidence base informing action on ethnic health inequalities. The pros and cons identified deserve attention by researchers undertaking syntheses or primary research, as well as by policy makers and practitioners who seek evidence to inform interventional strategies. Our findings are based on limited experience and, as further realist syntheses appear, revisiting and comparing these methods with others that are in development in this area, such as qualitative meta-synthesis, will be useful. FundingSS is a Senior Research Fellow funded by the NIHR School for Public Health Research.
Read full abstract
Feb 25, 2014
Geert Cornelis + 4
Open Access