Real world evidence can help in tackling health inequalities

A focus on:

Real world evidence (RWE) can contribute to and inform decisions regarding the approval and use of therapeutics and medicines. This has the potential to increase our ability to tackle health inequalities in diverse populations, but there are issues around the availability and quality of the real world data upon which RWE relies.

Our vision:

Addressing barriers to using real world data, such as representation, inclusion, diversity and accessibility, will help build the RWE needed to tackle health inequalities.

Looking to the future:

Identify and address gaps in availability and quality of real world data, encourage sharing of existing data, and promote inclusion of diverse communities and populations in designing studies that better model the real world.

Dr Ben Bray

Principal

Challenges in using real world evidence to address health inequalities

Real world evidence (RWE) describes information and scientific insights derived from real world data, and is increasingly being used to help develop new medicines. Real world data typically is more representative of patients in the real world compared to the sometimes highly selective patient cohorts (often younger, fitter and less medically complex) who are recruited to clinical trials. Indeed, the issue of lack of diversity in clinical trials has recently drawn much criticism and there are now many initiatives across the pharmaceutical industry and academia to improve the diversity of the patients recruited to clinical trials.

RWE is now recognised by medicine regulators such as the Food and Drug Administration (FDA) and European Medicines Agency (EMA) as being a valuable source of insights and data to complement clinical trials in establishing the effectiveness and safety of new medicines. For payers and other bodies such as the National Institute for Health and Care Excellence (NICE) who are responsible for evaluating the value of new medicines and making decisions about reimbursement, RWE can help to address areas of uncertainty about the cost-effectiveness of new medicines and the size of the benefit that they will bring to patients.

The amount and quality of real world health data available for research is varied and growing. It includes both data collected specifically for observational research and also data generated for other reasons but which can potentially be repurposed to answer scientific research questions.

Frequently used sources of data used for real world evidence include:

electronic health records
health insurance claims
data used to administer or manage healthcare services
observational research cohorts
patient generated data from apps and digital devices

In theory, RWE is well suited to identifying and understanding the impact of health inequalities, since it captures the messy reality of healthcare as delivered in real life, and can help to identify disparities in access to treatments or in the quality of care that patients receive. It is also often very large, with some real world data sources including information on millions of patients. This is helpful when trying to understand the health of marginalised groups in society who are small in number but may experience large disparities in health outcomes.

One of the key considerations in RWE is that the types of research questions which can be answered and the accuracy and validity of the results is highly dependent on the quality and nature of the source of the underlying real world data. This is especially true when the data was generated primarily for uses other than research, such as in the case of electronic health records or insurance claims databases, where an understanding of how, where and why the data was originally generated is critical in interpreting the data correctly.

RWE is widely used[as part of the data and evidence that payers and Health Technology Appraisal (HTA) bodies use to determine the potential value of a new medicine. For health equity to be considered as part of this assessment, appropriate data and evidence need to be included in these submissions to establish and define the scale of unmet health need. Gaps in what is available and possible with real world data can make this challenging.

Lack of diversity in real world data may impact on early research efforts to identify new medicines and bring them out of the laboratory into testing in clinical trials. For example, there are hundreds of gene therapies currently being developed which aim to correct for specific genetic abnormalities which cause disease. Lack of access to genetic data from diverse populations could result in gene therapies being developed which are less useful for people in some groups. Similarly, gaps in knowledge about epidemiology, population health and burden of illness in people living in data-poor countries could mean that less research attention is focused on developing treatments for the most pressing health challenges in these countries.

Lack of diversity in real world data may also impact the post-authorisation surveillance of new medicines, where RWE is often used to identify any long-term safety risks of medicines. For example, risk minimisation measures are often taken to make sure that the potential risks of medicines are managed and, without appropriate data, it may be difficult to identify if these are as effective as they should be in all groups, such as non-native language speakers who may have difficulty in understanding drug safety warnings.

Real world evidence could be used to tackle health inequalities if five key data issues are addressed

Although RWE holds much promise in helping to shine a light on the scale and impact of health inequalities and supporting the development of new medicines to contribute to tackling health inequalities, in practice there are many challenges.

We have identified solutions to five key barriers to using real world data as part of efforts to identify and tackle health inequalities. These range from global scale inequities in real world data availability to source-specific challenges where data is useful and relevant but inaccessible. Removing these barriers would go a long way towards reducing health inequalities locally and globally.

1. Represent whole populations

The largest scale challenge is the global inequality in the availability of real world data. The data sources used to drive much of the RWE to support the development of new medicines are almost exclusively from high income countries in North America, Europe and Asia. The vast majority of people living in low or middle income countries (the majority of the world’s population) are not included in any large scale real world data sources, and live in ‘data poor countries’. Without representation in data widely used for health research, the health challenges of people living in countries with the poorest health outcomes remain largely invisible.

2. Include more people with a range of characteristics

The second main barrier is exclusion of people with certain characteristics from existing real world data sources. This particularly affects people who already experience their own barriers in access to healthcare and the lack of real world health data is often a consequence of this. For example, refugees and migrants, travellers, homeless people and those without insurance coverage (in countries with insurance-based health systems) are often not included in even otherwise large and comprehensive real world data sources.

3. Include variables so that subgroups can be identified

Being included in a data source does not automatically mean that individuals with characteristics relevant to health equality can be identified. For example, even if people of diverse ethnicities are included in a real world data source, unless ethnicity status is recorded, then it is not possible to identify health disparities between people of different ethnicity. Similarly, even when relevant variables are potentially available, the data itself may be of poor quality with a high proportion of missing data or errors. This is a commonly encountered problem and may relate to a perception that collecting data specifically relevant to diversity or health equity is a ‘nice to have’ and less important than other data.

4. Measure relevant health measures and outcomes

A harder to define but important barrier is the relevance of the data included in real world data sources to the health needs of specific groups in society. Not everyone thinks about health in the same way and some health outcomes are more relevant to some people than others. For example, being able to find employment or living independently may be particularly important health outcomes for people with learning disabilities, but these outcomes are often not available in real world health data sources. As a result, RWE generated using this data might not address research questions which are important or relevant for all.

5. Make sure that real world data is simple to access and use

This challenge is in theory the most tractable to short term improvement: access to data. There are now vast quantities of real world health data of all types and sizes being generated every day, and many large and well established real world data sources are available for research. Much of this data is highly relevant for understanding health inequalities. But even if data relevant to health equity is collected, it does not mean that it is available or published, and in practice there are often significant barriers to access. For example, large disease registries and surveillance systems for cancer or cardiovascular disease may have data relevant to health equity included in their data collections, but may not publish data or summary statistics broken down by dimensions of health equity (e.g. cancer incidence in people of different ethnicity).

What does good look like: Genomics England

There are some efforts underway to recognise and tackle diversity in real world data and evidence. For example, Genomics England is a very large research initiative in the UK which is collecting very detailed genetic data on people with rare diseases or cancer, and has already helped advance the scientific understanding of many types of diseases. They have recently launched a Diverse Data initiative, aiming to increase the amount of genetic data available for research from people from under-served communities.

Calls to action to improve availability and use of real world data

The amount and type of real world data continues to grow, and there is huge potential for its use in RWE to make a major contribution to identifying and addressing health inequalities. How can we do this?

1) Identify inequalities gaps in the availability and quality of existing real world data sources.

2) Highlight the issue of “data poverty” and the implications this has for people not represented in the data sources often used for health research.

3) Research funders and study sponsors should make sure that they include health inequalities relevant data when setting up new studies and data collections.

4) Research funders should support the development of scientific capacity to collect and analyse real world health data in data low and middle-income countries.

5) Involve people from different backgrounds, in prioritising research questions and in designing real world studies, particularly marginalised populations with high unmet health needs.

6) Make better use of existing data by encouraging data custodians (e.g. registries, real world health databases) to publish and health inequalities data that is collected but not currently reported.