Addressing the diversity issue in clinical trials

The clinical development industry has a diversity problem. With constant challenges in recruitment and participation, certain ethnic groups are consistently under-represented in clinical trials around the world. But these groups still fall ill, and they still need access to the latest therapies that are safe, efficacious and accessible. We also know that due to genetic, social and economic factors, certain patient demographics may be disproportionately affected by some health conditions or more prone to adverse reactions.

Yet, due to a variety of factors including recruitment bias, the location of recruiting sites and historical issues of mistrust, clinical trial participants are not always representative of the patient population. This means the industry is unable to develop effective therapies that benefit the entire population receiving drugs.

Recently, the clinical development community has begun to address diversity in clinical trials. The European Medicines Agency (EMA) updated its Clinical Trials Regulation to ensure sponsors justify any non-representative procedures in January this year; this was followed by the US Food and Drug Administration (FDA), which issued new guidance on meaningful representation of racial and ethnic groups in April. To ensure these measures have the desired impact on trial participation, a patient-centric approach is needed. By harnessing the advanced technological capabilities now available, greater insights can be taken from data. There is also pending US legislation aimed at increasing diversity in clinical trials.

Diversity matters
With any clinical trial, safety is paramount. A biased or inadequate assessment of adverse reactions jeopardises regulatory approval. Understanding where the benefit-risk lies and which groups are most – or least – likely to tolerate a therapy allows healthcare professionals (HCPs) to make more informed decisions when it comes to prescribing medical treatments. The safety and efficacy of any treatment varies depending on a range of factors.

Many of these are obvious, and when recruiting for a clinical trial, patient sex, race, ethnicity and age must all be considered. If not, developers may have to bear the cost – and patients the risk – of any adverse reactions or lack of efficacy identified after a drug has gone to market.

Racial-based variations in treatment responses are not, in and of themselves, enough to derail a clinical trial. In many cases, variations could be seen as an opportunity, with some therapies being more effective in certain racial groups. One such example is the heart failure drug BiDil (isosorbide dinitrate/hydralazine), which was approved by the FDA in 2005. BiDil was the first ever drug recommended for use only in a single racial group after it was found to reduce deaths in African- American heart failure patients by 43%. This was an important finding and provided an alternative treatment option for a group that is statistically less likely to respond to the traditional therapies intended to control blood pressure such as beta- blockers, angiotensin-converting-enzyme (ACE) inhibitors and angiotensin receptor-blocking agents.

While some therapies are better tolerated by certain races or ethnic groups, other factors must also be considered. For example, the antiplatelet drug Clopidogrel is often used to reduce a person’s risk of heart disease, heart attacks or stroke.When in the body, the drug is converted to its active metabolite by a group of enzymes called cytochrome P450 (CYP). The difference in the genes responsible for CYP enzyme production varies across ethnicities. For example, South Asians are more likely to have a variation in the gene responsible for producing CYP enzymes, which causes them to metabolise Clopidogrel more slowly. In such patients, HCPs may consider the use of other antiplatelet agents in the class.

By identifying variations such as these during the clinical trial stage, we can work towards developing the best treatment strategies for all groups with a particular condition – saving money and improving patient experiences.

The current state of trial participation
With the new regulations from the FDA and EMA, the current lack of standards in reporting of trial participant demographics is going to become an increasing challenge. Representation cannot be reliably measured or compared if there is no consistency in metrics across government bodies, census data, healthcare providers, industry regulators and pharmaceutical companies.

Without clear metrics on race and ethnicity when designing clinical trials, it is much harder to identify clinically meaningful variations that could result in trials failing or requiring increased amendments. But race is not currently being considered as a standard element of protocols in the trial process. For example, a recent analysis by Phesi showed that, between 2007 and 2022, just 26% (460/1,777) of clinical trials with UK participation reported on racial and/or ethnic data – suggesting that almost three-quarters of trials did not see ethnicity as a primary concern.

Of the 460 UK-recruiting trials that did report on ethnicity, 49% of these reported no Asian participation and 43% reported no black participation. To have entire racial groups excluded from clinical research is clearly an issue and with the new guidance from the EMA, many developers may see delays in approvals as a result. Furthermore, 5.9% patients in these trials identified as ‘mixed race’ – a descriptor that does not align with census data, or provide further context to ethnic history, making it difficult to identify trends or race-associated variations.

The UK is not unique in these representation failings. According to Phesi data, 42% of cancer clinical trials do not include a single black patient. This is in spite of the US census reporting that 18.7% of the US population identifies as Hispanic or Latin American and 12.1% of the US population identifies as black or African American.

These disparities penetrate throughout the entire population and clinical trials often fail to encompass the diversity in the population. Further analysis has found that Asian, Hispanic and Latino, Native American and Alaska Native, Native Hawaiian and other Pacific Islander patient sub- populations were all significantly and consistently under-represented in clinical trials during the past decade. This means almost a third of the US population is not being represented adequately in clinical trials.

Trial diversity doesn’t only vary based on location, but also by disease indication. In 2016, clinical trials for cardiovascular therapies had less than 3% inclusion for black patients, despite this group being more likely to experience severe cardiovascular disease. Yet, trials focusing on psychiatric diseases have an over-representation of black and African-American patients – with 24.2% of patients participating in trials for psychiatric diseases being black or African American.

Given the inconsistencies in patient recruitment based on race, it is key that trials are better designed. Selection of adequate investigator sites and countries is needed at the planning stage to ensure adequate representation of racial groups in patients recruited into the trials. Population and disease occurrence data are vital tools for trial sponsors, allowing them to avoid oversights such as these and ensure that trial participant groups are an accurate representation of the patient population.

Data-driven design
So, what can the clinical development community do to prioritise diversity in clinical trials? And can we improve the clinical development process – and patient outcomes – by doing so? The answer to these questions lies in the intelligent mining and analysis of data. Thanks to the experience of previous clinical trials, a wealth of useful data is available that can be applied to optimise trial designs going forward. By knowing the demographics of patients at different recruitment sites and knowing what is needed to truly represent the patient population, diversity can be incorporated into protocol design at the earliest stage. By employing a data-driven approach when selecting recruitment sites, biopharmaceutical companies can reliably ensure that the right mix of patients are recruited to a trial from the beginning – putting patients first and reducing delays at the approval stage of the development process.

In addition to optimising protocol design, data can be leveraged to greater sophistication – and can even be used to fill recruitment gaps. By using large volumes of data collated from similar or identical trials, a ‘digital’ trial arm can be produced, eliminating the need for control groups in certain circumstances and reducing patient burden. These digital arms can also be used in place of real-world patients when recruitment challenges arise, plugging the diversity gap in ongoing trials. In many disease indications, generating digital trial arms is already possible thanks to a comprehensive and diverse existing dataset that represents all races. However, in rarer disease indications, sufficient, detailed and consistent information about different ethnic groups is still needed.

Truly representative clinical trials will have benefits across the scope of clinical research, informing future strategies that will lead to more efficient, cost-effective and equitable clinical trials. This can only be achieved by properly collecting, analysing and applying data on patient demographics. By making representation in clinical research a priority, not only will trust in the clinical research process be restored in key groups, but the clinical development process will become more efficient, more inclusive, diverse and equitable – ultimately leading to better therapies that benefit more patients, making it to market faster.