SCCM Resource Library-Association between Study Factors and Enrollment by Race/Ethnicity in Pediatric Acute Care Research

Video Transcription

Good morning. My name is Erin Paquette, and I'm a pediatric intensivist at the Ann and Robert H. Lurie  Children's Hospital of Chicago and assistant professor of pediatrics at the Northwestern University Feinberg School of Medicine.  I appreciate the opportunity to share my research with you this morning. Today I'll be discussing the association between study level factors and enrollment  by race and ethnicity in pediatric acute care research. While I very much appreciate the opportunity to share our research findings, I would like  to acknowledge and thank our entire research team for this work. The core question I have focused my research on is how to ensure that minority children  are proportionally represented in research. While children of minority race and ethnicity suffer greater morbidity and mortality from  critical illness, inadequate enrollment of diverse groups in research threatens study feasibility through potential for premature study termination, validity, and generalizability  as study findings may not be applicable to all diverse groups, and risks unfair distribution of research benefit and burden, potentially leading to exploitation of some groups who  are overrepresented in research or lack of benefit for those who are underrepresented. The prior literature demonstrates conflicting data on enrollment of children in research  with studies showing both over and under enrollment. Based on these findings and the previous literature, we hypothesize that acutely ill  children of minority race and ethnicity are underrepresented in some types of research and overrepresented in others.  We hypothesize that factors related to this study will be associated with differential enrollment by race and ethnicity.  To begin to answer this question, we retrospectively assessed participation across studies in the  Pediatric Emergency Care Applied Research Network and the Collaborative for Pediatric Critical Care Research Network with different study designs.  Nine studies comprised more than 64,000 patients with a mix of interventional and observational study designs and a variety of consent requirements.  To assess participation by race and ethnicity in different studies, we applied similar methods across nine studies in the two research networks. This graph summarizes that data.  We took the difference between the percentage of each race in the study population and the  percentage of each race in the research network's registry database to obtain a positive or  negative deflection from zero, where a positive deflection above the line represents overrepresentation,  where there is more of that race in the study than in the general population as reflected by the registry, and a negative deflection below the line represents underrepresentation,  where there are fewer of that race in the study than in the general population in the registry.  You can see that both over and underrepresentation by different races are seen across the network studies.  To begin to try to understand the mechanisms behind differences in enrollment, we examined several study-level factors.  Because we had only enrollment data available, we first conducted bivariate analyses to assess whether for each factor there was a higher or lower odds of having enrolled a Black,  Hispanic, or other race participant compared to a white participant in each study. The study-level factors we evaluated are shown on this slide.  We looked at consent mechanism, including three different types of consent, written consent, verbal consent, and a waiver of consent, as well as assent.  The type of study, whether it was interventional or more than minimal risk, or observational  minimal risk, the study procedures employed, including blood draws, genetic material collection,  administration of study drug, use of surveys or chart extraction, or other study interventions. We also evaluated the study population, whether in the PICU, ED, or both.  We'll show significant findings next. This table shows the odds of participants representing a certain race compared to white based on consent mechanism.  The top two rows represent formal consent mechanisms. The bottom rows represent more informal mechanisms.  Participants in studies with formal consent mechanisms have a lower odds of being non-Hispanic Black.  Participants in studies with less formal consent mechanisms have a higher odds of being non-Hispanic Black.  Similarly, when looking at Hispanic versus white and other versus white participants, there was a higher odds of enrolling a Hispanic or other participant versus white participant  with formal consent mechanisms and a lower odds of enrolling a Hispanic or other versus white participant when there was a less formal mechanism of consent.  When examining study type, there was a higher odds of having enrolled a non-Hispanic Black versus white participant in an observational study and a lower odds of having enrolled  a Hispanic or other participant versus white participant in an observational study. This table examines different associations between different study procedures and the  odds of having enrolled a Black versus white, Hispanic versus white, or other versus white participant, respectively.  Studies collecting genetic information using blood draws or administering study drugs had a lower odds of having enrolled a Black versus white participant.  While studies involving genetic testing and blood draws had a higher odds of having enrolled a Hispanic versus white participant and other versus white participant.  Meanwhile, studies involving survey administration, chart extraction, or other study examinations  and interventions had a higher odds of having enrolled a non-Hispanic Black participant  compared to a white participant, and studies involving survey administration or other examinations  and interventions had a lower odds of having enrolled a Hispanic versus white participant or an other versus white participant.  On multivariable modeling, in three separate models of Black versus white, Hispanic versus white, or other versus white, multiple associations remain significant.  On the y-axis here is odds of enrollment, and on the x-axis are study factors with odds shown by race and ethnicity for each study factor that was significant.  Written consent mechanism was associated with lower odds of having enrolled a Black versus white participant with an odds ratio less than one, while written consent was associated  with a higher odds of having enrolled a Hispanic versus white participant or an other versus white participant.  Observational studies had a higher odds of having enrolled an other versus white participant.  Studies involving blood draws had a higher odds of enrolling Black versus white participants,  while studies involving survey or chart extraction had a higher odds of enrolling other versus white participants.  All P values were significant to P of .006 or lower in these models. In conclusion, we've seen that enrollment in clinical trials is not randomly distributed by race and ethnicity.  Clinical study level factors are associated with differential enrollment by race and ethnicity.  Prospective evaluation of enrollment by race, ethnicity, and other sociodemographic characteristics  can help to better elucidate mechanisms or reasons underlying this differential enrollment. To better understand the interplay between factors, our next step is a prospective study  built on a mixed methods design. We will characterize children involved in critical care in emergency research studies by their risk for health disparities.  We will define a child at risk for health disparities according to the National Institute for Minority Health and Disparities risk factors for disparities, including being of  low income, non-English speaking, rural, and of minority race and ethnicity. Aim 1 of this study will prospectively characterize enrollment by CRHD status, leading to modeling  of enrollment in trials based on this status across approximately 26 studies enrolling in the pediatric ICU and emergency department at Lurie Children's Hospital.  Aim 2 will then elucidate qualitative insights into reasons for decisions and factors impacting decisions regarding enrollment from those who choose and those who refuse to enroll  in a trial or study. We will integrate Aim 1 and 2 data, and in Aim 3, we'll attempt to determine a prioritization  order for relevant factors impacting decisions about research participation in order to begin  to develop future interventions to improve the enrollment of diverse populations in pediatric emergency and critical care research.  Thank you for the opportunity again to present today. I'm happy to take any questions here or to continue the conversation via my email address or Twitter handle.

Video Summary

Dr. Erin Paquette, a pediatric intensivist at the Ann and Robert H. Lurie Children's Hospital of Chicago, discusses her research on the association between study level factors and enrollment by race and ethnicity in pediatric acute care research. The aim of the research is to ensure that minority children are proportionally represented in research to avoid disparities in healthcare. The study found that enrollment in clinical trials is not randomly distributed by race and ethnicity, and there are various factors that contribute to differential enrollment. The findings highlight the need for prospective evaluation and interventions to improve enrollment of diverse populations in pediatric research. Further research will involve a mixed methods approach to better understand the reasons for enrollment decisions and develop interventions to address the issue.

Asset Subtitle

Pediatrics, Quality and Patient Safety, Research, 2022

Asset Caption

INTRODUCTION: The pediatric literature shows under- and over- representation of diverse groups in research based on participant factors such as race/ethnicity (R/E). R/E is a social construct acting as a proxy for social factors underlying such associations. We hypothesize that study factors correlate with differential enrollment of Black v White participants (BP v WP) and Hispanic v White participants (HP v WP). METHODS: Retrospective review of enrollment in 2 pediatric trials networks. Study factors included: consent mechanism (CM) (written/assent/verbal/waiver); study type (ST) (interventional/observational); study procedures (SP) (blood draw/genetic information/study drug/survey/chart review/other); recruitment site (RS) (ED/PICU). Models accounting for clustering of participants in studies estimated odds ratios (OR, 95% confidence interval) for enrolling BP v WP and HP v WP by study factor. Multivariable analysis included significant factors. RESULTS: 63434 enrolled participants were evaluated in 8 studies. Factors associated with having enrolled a BP v WP (p < 0.0001 unless stated) were CM (written OR 0.70, 0.62-0.79; assent OR 0.65, 0.57-0.75; verbal OR 1.48, 1.27-1.72; waiver OR 1.39; 1.24-1.57); ST (observational OR 1.46, 1.21-1.75) & SP (Blood draw OR 0.72, 0.64-0.81; genetic information OR 0.73, 0.65-0.82; study drug OR 0.75, 0.65-0.86; survey OR 1.38, 1.23-1.55; chart review OR 1.53, 1.33-1.77; other OR 1.41, 1.25-1.59) & RS (PICU OR 0.69, 0.54-0.71). On multivariable analysis, CM (written OR 0.724, 0.653-0.80) and SP (blood draw OR 1.21, 1.09-1.36, p = 0.0006) remained significant. Factors associated with having enrolled a HP v WP included CM (written OR 2.36, 1.855-3.003); waiver OR 0.4; 0.35-0.45); ST (observational OR 0.69, 0.50-0.95, p = 0.02) & SP (blood draw OR 2.52, 2.23-2.86; genetic information OR 2.60, 2.4-2.82; survey OR 0.40, 0.35-0.47; other OR 0.43, 0.33-0.55). On multivariable analysis, CM (written OR 1.771, 1.60-1.96) remained significant. CONCLUSION: Study factors correlate with differential enrollment by R/E in pediatric multisite studies. This may reflect participant preferences, implicit bias or systemic racism. Prospective work could clarify the relationship between social and study-level factors affecting enrollment and identify opportunities to mitigate bias and systemic racism.

Meta Tag

Content Type Presentation

Knowledge Area Pediatrics

Knowledge Area Quality and Patient Safety

Knowledge Area Research

Knowledge Level Intermediate

Knowledge Level Advanced

Membership Level Select

Tag Pediatrics

Tag Evidence Based Medicine

Tag Outcomes Research

Year 2022

Keywords

pediatric acute care research

enrollment by race and ethnicity

disparities in healthcare

clinical trials

diverse populations

	Society of Critical Care Medicine 500 Midway Drive Mount Prospect, IL 60056 USA Phone: +1 847 827-6888 Fax: +1 847 439-7226 Email: support@sccm.org	Contact Us About SCCM Newsroom Advertising & Sponsorship DONATE	MySCCM LearnICU Patients & Families Surviving Sepsis Campaign Critical Care Societies Collaborative	GET OUR NEWSLETTER
© Society of Critical Care Medicine. All rights reserved. \| Privacy Statement \| Terms & Conditions The Society of Critical Care Medicine, SCCM, and Critical Care Congress are registered trademarks of the Society of Critical Care Medicine.