The Underrepresentation of Black and Hispanic Men in Clinical Trials Compared With Their Share of Prostate Cancer Incidence: Journal Club – Christopher Wallis & Zachary Klaassen

January 19, 2023

In this UroToday Journal Club Christopher Wallis and Zachary Klaassen highlight the JAMA oncology publication Disparities in the Inclusion of Racial and Ethnic Minority Groups and Older Adults in Prostate Cancer Clinical Trials: A Meta-analysis.  Mandates have been put in place to emphasize both the importance of including racial and ethnic minorities in clinical trials and the reporting of the demographics of these trial populations to allow assessment of their inclusion. Adherence to these mandates is somewhat unclear. In this analysis, the authors performed an Ovid MEDLINE search from inception until February 2021, to identify data on trial characteristics in oncology clinical trials, reporting and representation according to race, ethnicity, and age to evaluate disparities in the inclusion of racial and ethnic minority groups and older adults across PCa clinical trials.


Christopher J.D. Wallis, MD, Ph.D., Assistant Professor in the Division of Urology at the University of Toronto.

Zachary Klaassen, MD, MSc, Urologic Oncologist, Assistant Professor Surgery/Urology at the Medical College of Georgia at Augusta University, Georgia Cancer Center

Read the Full Video Transcript

Chris Wallis: Hello, and thank you for joining us for this UroToday Journal Club discussion. Today, we're talking about a recent publication entitled, Disparities in the Inclusion of Racial and Ethnic Minority Groups and Older Adults in Prostate Cancer Trials, A Meta-Analysis.

I'm Chris Wallis, an assistant professor in the Division of Urology at the University of Toronto. With me today is Zach Klaassen an assistant professor in the Division of Urology at the Medical College of Georgia.
You can see here, the citation for this recent publication in JAMA Oncology.

Disparities in prostate cancer care and outcomes have been long recognized, and disparities, based on race in the incidence and outcomes, are among the greatest of all cancers. In the case of prostate cancer, Black men are approximately twice as likely to be diagnosed with prostate cancer, are substantially more likely to have recurrent disease or metastatic disease, and are two to three times more likely than non-Black men to die of prostate cancer. Additionally, older men of racial and ethnic minorities have particularly worse outcomes, more so than younger men.

When we consider representation in clinical trials, mandates have been put in place to emphasize both the importance of including racial and ethnic minorities in clinical trials, and the reporting of the demographics of these trial populations, to allow assessment of their inclusion. However, adherence to these mandates, which have been passed through Congress, is somewhat unclear. Precise quantification of enrollment disparities is important, because it has implications for policy, clinical decision-making and future trial design.

So in this analysis, the authors performed an Ovid MEDLINE search from inception until February 2021, to identify oncology clinical trials. Two reviewers extracted data on trial characteristics, and reporting and representation according to race, ethnicity, and age, in keeping with the FDA position statement.

In terms of the selection of trials for the assessment of race and ethnicity, they included trials reporting race or ethnicity, they were recruited exclusively from within the United States. In the context of age, they included trials regardless of region, as long as they reported the number of patients who are older than 65 years or younger age 65 years. The trial proportion was calculated as the number of participants in any given category among all enrolled participants in that trial.

The authors performed an assessment of population-based prostate cancer incidence for each race and ethnicity group using SEER 21, encompassing the time period of 2000 to 2018, and for each age category using the Global Burden of Disease database.

They calculated enrollment incidence ratios to assess disparities in the enrollment of patients to randomized controlled trials among each subgroup. This was defined as the ratio between the proportion of patients in any given subgroup within the context of the trial, and the overall prostate cancer incidence at the population level.

In terms of statistical analysis, the office performed meta-analysis using log-transformed enrollment incidence ratios from each trial, and pooled leads using the inverse variance waiting approach using random effects.

Meta-regression was performed to assess the association between trial characteristics and enrollment disparities, with a random effects meta-regression of study enrollment incidence ratios. They also assessed the trends in enrollment disparities between 1990 and 2020 using meta-regression.

They performed a number of post-hoc sensitivity analyses, assessing international trials, US-based trials, and trials that were of agents leading to an FDA approval.

I'm going to hand you over to Zach, to walk us through the results of this meta-analysis.

Zach Klaassen: Thanks so much, Chris. So we can see in this figure that there was 3407 studies that were assessed for eligibility. Ultimately, 286 studies were included in this study, including 45 trials with exclusive US recruitment included in the racial ethnic disparity meta-analysis, and 49 trials included in the age disparity meta-analysis.

This is a summary of the trial characteristics. As I mentioned, there was 286 trials in total. 44% of these were Phase II, 36 were Phase III. In terms of arms of the study, 91.6 were two arms studies, and 8.4% were greater than or equal to three arm studies. You can see here that, in terms of the size of the trial, 37.1% of the trials were less than 100 patients. 40.9% were 100 to 500 patients, and 22% were greater than 500 patients. With regards to the clinical setting, 15% were localized disease, 3.8% in the biochemical recurrence disease space, 21.3% in the metastatic castrate sensitive, and 56.3% with mCRPC.

In terms of the class of treatment regimen, the most common was anti-androgens at 49.3%, and chemotherapy at 36.7%.
Looking at the primary endpoint on the right side of this panel, overall survival was the primary endpoint in 20.3% of the trials, 16.8% for progression-free survival. Positive trials were present in 32.2% of cases, and negative trials in 39.5% of the included studies.

In terms of the funding source, industry was the most common at 58%, non-industry at 35%. Majority of these trials are multi-center trials at 85.7%.

In terms of regional recruitment, US only 22.7%, international with US recruitment at 21.7%, and excluding the US at 40.2%.
This figure looks at the trends in reporting of race, ethnicity, and age in the US prostate cancer trials, and this is stratified by brackets of years. So we can see on the X axis, from 1990 to '95 all the way up to 2016 to 2020. The blue columns are percent of US only trials reporting race. The orange columns are the percent of US only trials reporting ethnicity. And the gray columns are percent of trials reporting age proportions.

Couple of interesting points here. We can see that for ethnicity, this has only started to be reported in roughly 2001 onward, with a general decline in the reporting of ethnicity. When we look at the percent of trials reporting age, we see that it's relatively uncommon to report age, with the only in the most common or contemporary trials reporting up to 40% of age breakdown. When we look at the percent of US trials reporting race in the blue columns, we see a general increase in reporting of race, up to roughly 85%, as of the most contemporary trials.

This table looks at the participation of racial and ethnic subgroups compared with US population estimates. And this is a very nice table, because it shows the enrollment incidence ratio. So enrollment incidence ratio of one means that the number of trial participants is expected, based on the US overall population proportion. So less than one is under-represented, and over one is over-represented.

We can see here, that African American and Asian Pacific Islanders are under-represented, with the African American and Black patients at an enrollment incidence ratio of 0.7, and an Asian Pacific Islander of 0.48, with white patients being adequately represented at an enrollment incidence ratio of 1.0. With regards to ethnicity, Hispanic men are also under-represented, with an enrollment incidence of 0.62, and non-Hispanic men are appropriately represented in enrollment incidence ratios 1.04.

This is further elaborated in these figures looking at enrollment incidence ratios of subgroups by race. We can see here again, under-representation less than enrollment incidence ratio of one, over-representation of more than one. For Asian and Pacific Islander 0.48, and Black men 0.70, both under-represented, with white men being adequately represented with an enrollment incidence ratio of 1.00.

This figure looks at the observed and expected numbers of under-enrolled minority patients. We can see observed a number of participants in these trials in yellow, and the expected based on the population proportion in purple. And we see that for all of Black, Hispanic, and Asian Pacific Islander men, these were all under-represented based on what would be expected in the US population.

Again, going back to this figure which you saw previously, this is enrollment incident ratios for subgroups by ethnicity. We see non-Hispanic men appropriately represented, with an enrollment incidence ratio of 1.04, and Hispanic men being under-represented, with an enrollment incidence ratio of 0.62.

A similar figure looking at age, we see that there's very little discrepancy based on age. Older adults are adequately represented at enrollment incidence ratio of 1.0, and younger adults at a ratio of 0.94.

This table looks at the significant predictors of enrollment, looking at incidence ratio after multivariable metaregression. This is broken down by Black, Hispanic, Asian-Pacific Islanders, and older adults. And so, if we look at the Black men, we see that for larger trials, they were less likely to be enrolled, and for those exclusive US recruitment versus international recruitment, they were more likely to be enrolled. We do not see any discrepancies based on these breakdowns for Hispanic men. However, for Asian Pacific Islanders, we see that those trials that had a primary endpoint of overall survival, Asian and Pacific Islander men were less likely to be enrolled. And for older adults, we see that increasing trial size of more than 100 participants was associated with being more likely to be enrolled in these prostate cancer clinical trials.

This figure shows the trends in enrollment disparity in prostate cancer trials conducted between 1989 and 2000. And we can see here, this is broken down again by the previous allocation we just talked about in the previous table. For Black men, we see a general decline in participation from 1990 to 2020, but we see a progressive incline in participation for Asian or Pacific Islander men, as well as Hispanic men. And we see a relative stability in participation over the years for older adults, as we can see in Figure D.

This final table looks at the racial ethnic representation across trials with different levels of US recruitment, and across trials leading to FDA drug approvals. If we work from left to right in this table, among US only trials, as we've already mentioned, they have African American Black representation at enrollment incidence ratio of 0.70, Asian Pacific Islander 0.48, Hispanic 0.62, and non-Hispanic 1.04.

When we include US only plus international trials, this becomes worse for African-American men, down to an enrollment incidence ratio of 0.40. Again, decreasing for Hispanic representation, 0.54, and holding stable for non-Hispanic representation of 1.04. For international trials including in the US, we see again, a further discrepancy for African-American or Black men enrollment incidence ratio of 0.26, and again, worsening for Hispanic representation of 0.49.

And finally, for FDA approval, we see poor enrollment from African-American or Black representation, enrollment incidence ratio of 0.26, and also poor for Hispanic representation, 0.38. With improvement up to enrollment incidence ratio of 2.03 for Asian Pacific Islander representation, and appropriate representation for non-Hispanic men, enrollment incidence ratio of 1.06.
So by way of discussion, the results of this meta-analysis suggests that there is persistent under-representation of racial and ethnic minority groups in prostate cancer trials. And this is seen even after adjusting for population estimates of prostate cancer incidences.

Representation of Black patients has consistently remained low, and is significantly worse in larger trials. And less than 10% of trials report clinical outcomes by race and ethnicity, indications of the collective failure to account for differential health outcomes in marginalized patients.

It's important for journal editors to routinely ask for racial and ethnic demographic data and outcomes data reported by race and ethnicity in the trials they publish.

And finally, the FDA has recommended that clinical trial sponsors develop race and ethnicity diversity plans for inclusion of minorities. And compliance with these recommendations during submission of investigational new drug applications is very crucial.

So in conclusion, the results of this meta-analysis suggest that Black and Hispanic representation in prostate cancer clinical trials are significantly less than expected, based on cancer incidence.

Secondly, Black participation has remained consistently low during the past two decades.

And third, an actionable framework involving legislation, institutional support, and journal policy for data transparency is required to tackle enrollment inequity in future clinical trials.

Thank you very much for your attention, and we hope you enjoyed this UroToday Journal Club discussion of the recently published JAMA Oncology paper.