This BMJ Leader blog series has been produced in collaboration with the Health Equity Evidence Centre (HEEC). HEEC are dedicated to generating solid and reliable evidence about what works to address health and care inequalities. By adopting innovative methodologies, they efficiently map successful strategies for reducing health and care inequalities, and subsequently empower policymakers and practitioners to make evidence-informed decisions for all.
The blog below has been written by Dr John Ford, Director of the HEEC and gives an overview of the work of the centre.
In 1998, the UK government embarked on one of the world’s first cross-government strategies to reduce health inequalities. And it worked. The two main targets – to reduce inequalities in life expectancy and infant mortality – were achieved, marking one of the few examples of a large-scale health inequalities policy change that has been effective.
However, amazingly, we don’t know how and why it worked. Part of the challenge was that the strategy was wide-ranging encompassing Sure Start Centres, Health Action Zones, income tax changes, and programmes for “troubled families”. Much of the learning has been lost, leaving the current government who have set out to reduce regional health inequalities with little evidence of what to do.
Why do we still have so little understanding of how to address health inequalities?
Building the evidence base of what works to address inequalities is not straightforward. First, the evidence base is broad and disparate. Research is spread across multiple siloed academic disciplines, with a predominance of evidence describing the problem rather than identifying solutions. Health researchers in other areas have matured their evidence base over decades; we now have in-depth research on diabetes prevention, cancer treatment, and cardiac rehabilitation. Health inequalities evidence needs to be handled differently. It is not like clinical research where the causes of ill health under investigation are diseases or conditions amendable to drugs or procedures. Rather, the causes of ill health and their solutions in relation to health inequalities cut across multiple conditions, environmental factors, services and populations
Second, there are power imbalances. Research produced by academic institutions, especially quantitative data, often does not reflect the lived experiences of individuals. The process of data collection, curation, and analysis often loses the nuance and reality of people’s lives. The focus on statistical inference and concerns about outliers prioritises the average over those at the margins. And those living at the furthest margins are often poorly captured in health data as they face challenges accessing care. Additionally, we often do not consider evidence which is generated outside academic institutions. This is particularly important for disadvantaged communities, where evidence of lived experiences and what would help them is often distant from the expectations of academic publications.
Third, the evidence on health inequalities is biased by the research that is funded. Discrete interventions that can be easily implemented and described are more likely to receive funding than complex policy changes focusing on upstream structural reforms. This results in a large number of studies exploring small-scale interventions, which undoubtedly have their place, but at the expense of research into systemic changes such as funding and contractual reforms. Consequently, the evidence base contains far more studies on interventions that bolt onto existing services—such as staff training or translation services—rather than structural change that may have the largest impact, such as reallocating resources.
Fourth, health inequalities outcomes are often long-term and require large participant numbers to demonstrate measurable improvements. Early results from the national health inequalities strategy in the 2000s suggested it was not effective, but later data, which accounted for lag effects, revealed its positive impact. Designing and delivering studies with sufficient follow-up and participant numbers to show a closing of the gap is challenging for system-level changes. Therefore the evidence is awash with small scale studies with short follow-up.
Despite these challenges, more and more research articles exploring health inequalities are being published every year and our understanding is developing. Over the past 20 years, the number of studies focused on health inequalities has increased ten-fold, but the quality varies. Unpicking this growing body of mixed-quality evidence and drawing meaningful conclusions to inform policy and practice is becoming increasingly difficult.
Harnessing innovation in evidence synthesis
Successfully navigating the existing evidence is a key first step in building our understanding of what works to address health and care inequalities. Machine learning offers opportunities to make the process of identifying and synthesising evidence more efficient and effective. We use machine learning technology, EPPI Reviewer software, to create living evidence maps of what works to address health inequalities. The algorithm rapidly screens over 200 million articles every month and prioritises the most relevant ones. A researcher then manually reviews these articles to ensure they meet our eligibility criteria and labels them so they appear in our living evidence maps. Because this process is efficient, we can repeat it every month. As researchers, we then turn our living evidence maps into accessible evidence briefs for policy makers and practitioners.
We believe this is the future of how users will navigate large volumes of evidence. Traditional reviewing methods take time, effort, and quickly become outdated, especially with the burgeoning number of research articles. However, producing living evidence maps alone will not lead to the necessary shift in the evidence base – we need to think differently about how we use evidence. For many health inequalities topics, there is simply not enough evidence for what works due to the population group, health condition, or service. Traditional reviews may simply conclude that there is insufficient evidence and that more research is needed—an unhelpful outcome for policymakers and practitioners faced with decisions to make.
Instead, we use the principle of transferable evidence in our evidence briefs. For example, there is strong evidence that ensuring services are culturally competent improves access and outcomes for diverse communities. So, even if there isn’t a study examining culturally competent services for, for instance, Somali women seeking diabetes services, it is a principle that can still be applied.
If we are to make progress on health inequalities, building an understanding of what works is imperative. This requires the collective efforts of policymakers, practitioners, researchers, funders, publishers, and communities, while also embracing the opportunities presented by the latest technology. If we fail to build the evidence, we may find ourselves in a similar situation 25 years from now, regretting missed opportunities to learn the lessons of what works to address health inequalities.
In this blog series – a collaboration between HEEC and BMJ Leader – we dive deeper into how we apply our methods in practice, the lessons this generates for healthcare leaders, and ultimately, what works to address health and care inequalities. We hope you enjoy reading!
Declaration of interests
I have read and understood the BMJ Group policy on declaration of interests and declare the following interests: none.