#CriticalThinkingLive │ Enhancing healthcare equity: diversity in healthcare research

Clinical trials evaluate the safety and effectiveness of medicines. Ensuring that people from diverse backgrounds and geographies take part in them is key to advancing health equity.

Inclusive clinical research starts by addressing a question: how do we ensure that the design and other criteria align with the population we intend to treat?

Hosted on Tuesday 5 September and conducted by Tamsin Rose, our #CriticalThinkingLive with Dr Samit Hirawat, Executive Vice President, Chief Medical Officer and Head of Development at Bristol Myers Squibb (BMS), explored the company’s journey towards diversity in healthcare research. The conversation is summarised below.

Tamsin Rose: Why is it so important that the clinical research that we carry out and the clinical trials that we use for new medications are reflective of the diversity of our populations?

Dr Samit Hirawat: There is a clear mismatch between the diversity of patients with whom doctors work and the clinical research that is conducted.

Diversity in clinical trials and data is critically important to ensure that new drugs and treatments are safe and effective for all. If clinical trials only include certain demographics who are privileged enough to participate in the trials, the resulting data may not apply to more diverse patient populations.

TR: In your professional experience as a medical doctor, how does this lack of diversity impact the way we deliver care to patients?

SH: I have experienced this firsthand in my work in oncology, observing limitations when recommending treatments for Asian patients because the drugs had not been properly tested in Asian populations. I have looked at data from the perspective of diversity in gender, age, race and ethnicity to see how potential treatments may work differently in certain groups. This mismatch means that doctors cannot confidently prescribe treatments for underrepresented groups, since the data is not properly applicable to those diverse patient populations.

TR: Tell me about how you are leading change within BMS. How is the company ‘walking the talk’ on diversity?

SH: At BMS we have taken several steps to increase diversity in our clinical trials. In 2020, we made a public commitment to bring diversity to the forefront, establishing concrete objectives and targets. We have pledged to establish 25% of our clinical trial sites in underserved communities where minority populations live and to ensure that 18% of participants are from Black and African American, Hispanic and Latin communities. In 2021, we have achieved and surpassed these targets.

We have now made even more efforts as we are going to start collecting data on the diversity of our clinical investigators because it is generally true that patients are gravitating towards physicians who look like them, who talk like them.

To maintain trust with patients and society, the pharmaceutical industry must demonstrate a commitment to ethical practices, inclusion of diverse populations in clinical trials and access to medicines for all those who need them. By adopting the right policies and incentives, governments can help ensure the industry contributes to improving health outcomes and equity in healthcare. The role of the pharmaceutical industry goes beyond developing new drugs to generating and sharing knowledge that can benefit public health.

TR: I’m really struck by the number of times you’ve mentioned data and how important it is that the data we collect is accurate. Why is it so critical to make sure that our data sets are clean so that the artificial intelligence (AI) of the future delivers that health outcome you are looking for?

SH: AI and machine learning tools can play an important role in improving clinical trials and data collection, especially to help address issues of diversity and underrepresentation. AI can be used in several ways:

• To analyse existing datasets and impute missing data, which can help researchers understand how to proceed and where data gaps exist. This can provide a starting point for creating a more representative dataset for future use.
• To identify patterns and make predictions that can help with trial design and operation. AI tools can be trained on diverse training data to improve their accuracy for all populations.
• To assist in data collection by identifying variables that are not being captured properly, such as ethnicity data, or suggest improvements to data collection forms and processes.
• To standardise data reporting and analysis across trials to minimise bias, identifying inconsistencies and detecting issues that might be missed by human reviewers.

While AI will not solve the problems of underrepresentation and bias on its own, it can be a valuable tool when trained on robust and diverse datasets. Used appropriately with the right safeguards, AI has the potential to help clinical trials generate more accurate and representative data that benefits all patient populations.

TR: We are right in the middle of creating the European Health Data Space (EHDS), which is our first attempt to federate and pull together data that sits in silos all across Europe. What advice would you give in terms of constructing the databases? And when it comes to the revision of our pharmaceutical legislation, how do we encourage diversity to be respected?

SH: Getting these policies right will be important to ensure Europe remains attractive for the industry.

I believe the EHDS presents an important opportunity if implemented correctly. By enabling the sharing of health data across borders, it could lead to improved diagnostics, boost scientific research and help companies develop better-tailored medicines and services.

However, over the past few decades, Europe’s share of new treatments has declined compared to investments in the United States. The revision of the pharmaceutical legislation is a key chance to protect Europe’s life science sector for the future if done properly. One should be aware of the unintended consequences of reducing data protection periods for drugs from eight to six years, which could negatively impact investment in research and development (R&D).

I hope EU decision-makers, current ones and future ones, will use the next few years to future-proof the European regulatory framework, so that science-driven companies, such as BMS, see Europe as a competitive environment for investments in R&D and manufacturing.

TR: What would success look like, particularly in Europe? In the next 12 months, we are going to have elections at the European Parliament and a new College of Commissioners, how would the clinical trials landscape be more diverse and inclusive at the end of the next mandate?

SH: Ultimately, the test of success is whether clinical trial data leads to a reduction of health disparities by ensuring new drugs and technologies work for all patients, not just some. The impact on real-world health outcomes should be the primary measure, not just the number of participants from diverse backgrounds.

Some markers of success can include:

• The ability to confidently discuss efficacy and safety data with all patient groups, not just those represented in trials.
• Ensuring that people in all communities, regardless of socioeconomic status, have equal access to safe and effective treatments.
• Standardised ways of collecting, reporting and analysing data to minimise bias and missing information.
• Inclusion of data about diverse patient populations in drug labels and assessments to guide appropriate use.