Analysis of a recent clinical trial finds that only 14% of recruitment sites were located in ZIP codes with the highest incidence of disease.
Ensuring that clinical trials are representative of real-world populations continues to be a strategic imperative for Clinical Development teams, even amid regulatory uncertainty. Closing the gap between study populations and real-world demographics is especially critical in therapeutic areas where underrepresented populations are most affected, such as metabolic dysfunction-associated steatohepatitis, or MASH — formerly known as nonalcoholic steatohepatitis (NASH). This chronic liver condition, which affects more than 14.9 million U.S. adults and is becoming more prevalent and has well-documented disparities in diagnosis: Nearly half of all MASH patients are of Hispanic/Latino ethnicity.
Recently, an FDA-approved treatment became available for patients with noncirrhotic MASH. However, the FDA concluded that differences in how the treatment worked across races could not be determined by the phase 3 trial due to the small number of enrolled participants from non-White groups.
Tracking the Impact of Screening Criteria
While inclusion and exclusion (I/E) criteria are essential to ensure scientific integrity and patient safety, they can also unintentionally filter out patients who reflect the real-world population and whose participation is crucial to understanding a treatment’s efficacy and safety across diverse populations.
To explore how I/E criteria can influence trial diversity, we modeled different scenarios based on criteria used in the MASH phase 3 trial, which included a MASH diagnosis, history of alcohol use, cirrhosis, comorbid liver conditions, thyroid disorders, certain lab thresholds (AST, ALT, eGFR, A1c, platelets), and GLP-1 use.
We compared the race and ethnicity (R&E) of several cohorts, including those with and without lab data, those with closed claims only to ensure the highest fidelity, and those with select thyroid disorders (an exclusion criterion) and without. This allowed us to isolate the cumulative and individual impacts of trial criteria on R&E representation.
Our baseline MASH population cohort of medically insured patients was 62.8% White, 18% Hispanic/Latino, 9.3% Black/African American, 5.1% Asian or Pacific Islander (AAPI), and 4.8% “Other.” Our analysis found that applying these criteria affected representation of patients of color in some groups more than others, a pattern that is only possible to detect by integrating highly curated specialty data with claims data.
Among our findings:
- The addition of lab criteria reduced the share of Hispanic/Latino patients — the ethnic minority group most impacted by MASH — by 5 percentage points. Diversity was negatively impacted overall by the addition of this criteria, declining 3 percentage points. Conversely, the share of Black patients increased by 4 percentage points
- Limiting the cohort to patients with closed claims decreased diversity overall, with the share of people of color declining by 2 percentage points
- Adding a thyroid-related exclusion to the cohort (criteria that was added to the real-world trial as an amendment) affected diversity only slightly. The Hispanic/Latino participant share increased by 1 percentage point, Black participant share increased by 2 percentage points, and AAPI and “Other” patients remained the same
- Overall, removing all three of these criteria (lab, closed claims only, and select thyroid disorders) increased representation of patients of color by 3 percentage points, from 37% to 40%. The Black patient share increased by 4 percentage points, and the Hispanic/Latino patient share remained about the same
Key Learnings
While I/E criteria are determined based on scientific, medical, and regulatory considerations, modeling the criteria enables sponsors to understand how it may affect real-world populations and, in turn, get a jump-start on strategizing how best to compensate for any negative impact.
For instance, it’s informative to know that the amendment that added the thyroid-related exclusion did not appear to negatively impact the distribution of screening-eligible patients within different racial and ethnic groups. Conversely, learning that some of the lab values could disproportionately impact Hispanic/Latino patients offers sponsors the foresight to proactively engage with sites that see high volumes of this population to offset the impact.
Using RWD — Including Demographic Insights — to Guide Site Selection
In addition to analyzing the impact of I/E recruitment on trial diversity, we also looked at how well U.S. recruitment sites aligned with the real-world distribution of patients with MASH.
- We identified the 100 ZIP codes that had the highest MASH patient volume. Of the 78 ZIP codes encompassed in the trial, only 14% were aligned
- Within the top 10 highest-density ZIP codes, there was only one trial site
- For example, ZIP code 90025 in south Los Angeles had the highest MASH patient volume but no exact trial site match — though California did host several trial locations
From Awareness to Action
Understanding how I/E criteria and site selection can affect clinical trial recruitment is the first step toward improving trial diversity. Our analysis demonstrates how using highly curated real-world data — including R&E insights — enables teams to model and flag criteria that may disproportionately exclude underrepresented groups and select sites that enable patients most impacted by the disease to participate in trials.
To see more articles like this, follow Komodo Health on X, LinkedIn, or YouTube, and visit Insights on our website.
