
[Viktoria Tom/Adobe Stock]
The need for greater LGBTQ+ clinical trial inclusion
For decades, the LGBTQ+ community has faced significant barriers to equitable healthcare. Long-standing hurdles include higher rates of discrimination in healthcare visits and a greater likelihood to experience negative healthcare provider interactions. Those struggles also hinder the development of treatments and interventions that effectively address the unique health needs of LGBTQ+ individuals.

Mokash Sharma
“We see that items like race and ethnicity are drivers of inequity. There’s been a lot of work that we’ve done and reported on there,” said Mokash Sharma, SVP, Head of Development Operations at BMS. “In terms of sexual orientation, gender identity and intersex status (SOGIIS), there are drivers of different responses or potentially different impacts of treatment.”
A data-driven approach to closing the gap
Bristol Myers Squibb (BMS) aims to support diversity and inclusivity in clinical trials across multiple dimensions. This overarching goal aligns with BMS’s core business objectives and mission to bring more medicines to more patients faster, Sharma said.
BMS has pledged that by 2025, all new U.S. clinical trial study protocols incorporating electronic Clinical Outcome Assessments (eCOAs) will include SOGIIS data. The company has already surpassed 60% of this target.
Factors such as hormone replacement therapy (HRT), PrEP use and HIV status, for instance, are more prevalent in the LGBTQ+ community. “Our desire is that our trials should be as representative of the population we’re trying to treat as possible,” Sharma said. “We don’t want patients to be excluded just because of a different status, race, ethnicity, veteran status, or any other sorts of barriers to the benefits of clinical research and what it brings in terms of treatment options. We’re taking action to eliminate those barriers one by one.”
How BMS collects SOGIIS data
BMS’s initiative to collect SOGIIS data begins in the U.S., where the company has integrated this information into its electronic Clinical Outcome Assessment (eCOA) process. “In most of our studies now, certainly phase 2 and phase 3 studies, we implement eCOA,” Sharma said. “It’s basically a tablet or smartphone where we ask patients to enter data on their condition.”
At the end of these assessments, patients encounter dedicated pages where they can voluntarily share their sexual orientation, gender identity, and intersex status. BMS prioritizes patient choice, allowing individuals to decline to answer if they prefer.
Using AI to dismantle barriers
BMS is also using AI to ensure its clinical trial protocols are as inclusive as possible. The system analyzes a range of clinical data – publicly available datasets, commercially acquired datasets, and BMS’s own trial data – to pinpoint potential barriers to participation.
One way this plays out is in identifying and addressing exclusion criteria that disproportionately impact specific groups. For example, the AI system might flag hormone replacement therapy (HRT) status as a potential area where patients could be inadvertently excluded. “What we use the AI for is to ask: How many patients are actually undergoing HRT therapy of various kinds, and are those truly necessary to be excluded depending on the mechanism of our drug?” Sharma asked.
The AI system can suggest adjustments to inclusion/exclusion criteria that could significantly increase participation without compromising the scientific integrity of the trial. “We look at a lot of blood markers where we have a limit for inclusion or exclusion. What the AI engine helped us to do is to say that if we just tweaked by 5%, we would actually include maybe twice as many African American patients as we currently have.”
This same data-driven approach is being applied to enhance inclusivity for LGBTQ+ participants. While AI systems themselves require careful development to mitigate bias, they offer an unparalleled ability to analyze and uncover potential areas of bias across a range of factors — from race and ethnicity to SOGIIS status — that might otherwise go unnoticed by human researchers. “That’s where I think AI and large language models can help us to unravel what’s noise and what’s really a signal here,” Sharma said.
Promising early results and a roadmap for the future
The initial results from BMS’s pilot studies have been encouraging. Sharma reports, “We piloted this last year, and we were encouraged by the pilot studies. About 63% of eligible trials are collecting this data. It was quite encouraging.”
As data from the program starts to accumulate, BMS will analyze it to understand representation levels, identify disparities and develop targeted interventions to address those disparities. “Our game plan is to look at sources of inequity and expand our efforts. We quantify because without data, we don’t know what we’re talking about,” Sharma said. “We set targets to break down those barriers to the fullest extent possible. Race and ethnicity were the first steps. Now we add the lens of SOGIIS, and then we look to see how we can expand globally.”
The initiative aligns with BMS’s greater commitment to diversity and inclusion. “It certainly makes me proud that not only are we doing this stuff, but quite often we’re leading and driving the industry in this area,” Sharma said. “It’s something that we do believe in… It’s not an add-on activity. Our job is to make this business as usual.”
Filed Under: clinical trials, Drug Discovery, machine learning and AI, Regulatory affairs