SCCM Resource Library-Designing Clinical Trials for Pediatric ARDS: Hopes and Hurdles

Video Transcription

Good morning. My name is Martha Curley, and I'd like to thank the organizing committee for giving  me this opportunity to present Designing Clinical Trials for PARS, Hurdles and Hopes.  I bring greetings from the University of Pennsylvania and Children's Hospital of Philadelphia, where I hold the Colcitt Chair in Pediatric Nursing Science.  The objective of this presentation is to discuss how to design clinical trials for PARS, including identification of the right population and novel study designs.  Essentially, I'm going to talk about the challenges in conducting PARS research and perhaps present some known and future strategies to overcome these challenges.  For disclosures, I do receive research support from the NHLBI and NICHD, which I am very much appreciative of. So what are we trying to do here?  We have a common desire to build evidence and clinical scholarship, evidence that ensures that our care is evidence-based and that changes in our approach to care are driven by data  and systematically evaluated. We also have a desire to build clinical scholarship to ensure the relevance of our research, mentoring  of junior investigators and research methods, and the growth of our subspecialty. That noted, we've taken a black eye.  It's been very difficult to generate the data that our vulnerable patients need and that our families deserve.  There have been major gaps in adequately powered RCTs that really make a difference, that really  make a difference in the shortest period of time that can be funded and a sufficient number of trialists also working together to build programs of research to fill these gaps.  Our challenges are many. One of the challenges is that our patient population in the PICU straddles the larger NICU in the adult population.  We're dealing with individuals who are evolving people. They are dependent upon their parents and guardians. They vary by age, size, development, maturation, and disease prevalence.  ATS consensus statement noted way back in 2005 that the pediatric respiratory system must grow and develop even when faced with injury and repair.  In this context, it is not surprising that the mechanism and consequences of disease are likely to be different in children.  These challenges also present in operational definitions. A lot of fine work has been proposed through the PILIC guidelines that were published in  2015 that got the clinical trial list on the same page, operationally defining what PARDS is if it's in children.  We are in the process of redoing those guidelines that hopefully will be agreed upon shortly. Challenges in PARDS research also have to do with the challenges of timing of PARDS.  For example, there are differences in whether or not a patient presents with PARDS within four days of endotracheal tube intubation or whether or not it occurs after four days.  In trying to get patients into clinical trials in an early phase, we have to come up with mechanisms in order to support that.  For example, continuous screening, assuring that physicians who are caring for these patients attest to the treatments.  Once a patient meets criteria, consenting them, then getting them into the trial quickly. But kids don't always read this prescriptive delineation of study entry.  In the prospect clinical trial in which myself, Ira, and Martin are the co-PIs, we've developed  an alternative consent process to try to get kids into clinical trials as soon as they meet the criteria, consenting early so that when criteria are met, for example, in the  middle of the night or on a weekend, that the patients can be immediately randomized and placed on protocol. We're also challenged with the consent and assent process in pediatrics.  Clinical consent, legal guardians, the rates vary across enrolling sites and studies. We have to identify who has legal guardianship for the patient, and they may not be physically  or emotionally available when they meet criteria, when their children meet criteria for study. As soon as the children are able, they need to be able to also provide assent.  The age of assent varies according to the child's cognitive capacity. And then uniquely in pediatrics, what happens when we enroll the 16, 17-year-old and follow them over time?  And what happens when they become a newly adult and consent has to occur after age 18? Ruth LeBay and colleagues wrote this paper describing our experience in asking parents  for permission to enroll their child into clinical trial best practices that we've learned over time.  Another challenge is the simple heterogeneity of the patient population, heterogeneity for age, size, development, organ maturation, and also disease clusters by age.  What produces PARs in a younger population changes over time.  In designing trials, we have to consider how much to narrow the inclusion criteria, how to stratify or control for age, and then have to set up for nonlinear effects.  For example, whether or not an effect would be developmentally sensitive so that you may lose the ability to determine effect if there was a use-shaped or a different type of impact  for your intervention in the different age groups, and then really also how to put those age groups together.  Further challenges that have to do with heterogeneity are related to the care and the practice variability that occurs from PICU to PICU to PICU.  There's variation in practice by PICU, and we're plagued in pediatrics by early adopters of interventions in PEAS critical care without pediatric-specific research.  We also have heterogeneity in what we consider to be minimum data sets, or even within those  data sets, using the same operational definitions from study to study to study, and then trying  to make cross-study comparisons by comparing Table 1, which varies by most interventional studies that we have in PEAS critical care.  And this, of which also needs to be mentioned, that because you design a trial and implement a trial, we have to report treatment fidelity, the extent to which the patients actually  receive the intervention through quality monitoring and through auditing.  We also have challenges in determining what primary outcome we should study within a PARDS  research.  In addition to survival, which is always critically important, which is low, thank goodness, in PARDS, there's now a movement towards long-term quality of life and functional  status outcome metrics, because these metrics have been articulated to be critically important in children.  That said, it's very difficult to measure a single construct over the developmental span that we see in pediatrics.  And here's an example of a quality of life measure that varies with six different instruments over our population of concern.  And then we are faced also with whether or not we should ask parents for their proxy  determinations, or when we can actually ask the children themselves, typically after eight years of age and when they're cognitively capable.  But our groups are wide from a developmental perspective, but we must move beyond mortality as our primary outcome variable in PARDS research.  These trials take time and money, and here are my beautiful grandchildren, and my oldest granddaughter, Claire, is pictured here.  And by the time we moved Restore, which I was involved with as a research idea in 2002, and published the trial in 2015, she was just a figment of our imagination and now a much  older child. We went from a research idea to a pilot trial to requesting funding for a larger multi-site clinical trial, running the clinical trial, then finally publishing in 2015.  So it takes a lot of time and money and focus to deliver on some of our trials.  The least of which is the recruitment challenges. Trying to get patients identified early and then get them into our trials is very difficult.  We have, in reality, few potential patients when we really apply our criteria. It can be considered, as compared to the adult population, a rare disease.  We have lots of seasonal variation, as you start to see here, but really the impact of COVID over the past two years really will change the face of clinical trials.  We're still struggling with, this is the prospect enrollment, of trying to get our patients enrolled and recovering from PARDS. And it's going to take some time for us to recover.  Challenges, when you have so few patients, we are always struggling on whether or not we should compete or collaborate.  And our stance is always an abundance mentality to collaborate and look for any way that we can possibly do many clinical trials at the same time.  As long as there's different primary outcomes, we have to assess the potential interaction effect on the primary outcome with several interventional trials.  Our capacity to statistically evaluate effect modification, whether or not we could balance subject assignment, adequately attribute any adverse events that may be there, and  evaluate how complex the interventions are. And can the bedside practitioner safely implement two complex interventions into the same patient population?  And then what to do with parent and family burden with multiple consent processes.  But given all these challenges, I'd like us to focus on what we can do differently and how we can do science differently. And I definitely believe that times, they are changing.  First of which is over the last 20 years, really, we've built networks to find colleagues  who mutually can help us with our PARDS research and completing our PARDS research from single site to multi-site clinical trials.  Our colleagues and our collegial base has grown over time with the Polisi network.  We now have over 90 PICUs coming together, trying to work together to collectively answer our complex questions.  And this is primarily in the United States, but now Polisi is international. And we also are forming a relationship with similar structures that are existing from  an international perspective. The Canadian Clinical Trials Network, the European Society for Pediatric and Neonatal  Intensive Care, our ANZIC colleagues, and our Asian PAC-MAN colleagues. So together, there's no question that collectively we cannot answer together.  We also are advantaged in that we are now starting at a different place because of our consensus conference, which demands an enormous amount of time.  But now with evidence-based processes, systematic reviews, evidence-based appraisals, we now  have stellar PLIC guidelines, PLIC 1 and 2, that really will help shape our PARDS research in the future.  Added to that, the TACSI recommendations, the brand new PANDEM recommendations on pain and agitation, delirium and withdrawal, early ambulation, and also, as I mentioned briefly,  two works coming together. Number one, to operationally define critical elements that must be studied in PARDS research,  led by Shan Ward, and then also a separate initiative on clinical trial design. Both these groups are in formation and hopefully will deliver soon.  We also can start at a better place because we now can rely on big data methodologies  to give us the data that we need to pristinely identify what the starting frequency, occurrence,  prevalence is of PARDS using electronic medical records, data warehouses that are run within  institutions, or, for example, Peter Lawson's work up in Toronto, auto-capturing data over time, primarily in the PEDS cardiac world, but also in moving out to the PEDS critical  care world. And then also using natural data sets to fix our power analyses from the FIS data set,  insurance claim data, but the potential of big data in informing PARDS research in the future is limitless.  We also, by design, have learned to transition from one question to the next. For example, transitioning our organizational structures over time so that when you start  and end a study, you don't start and end the enormous amount of work that's required  to launch these clinical trials.  So, for example, going from the restored trial to the half-pint trial to now the prospect clinical trial, we've learned from previous research, and we've learned to build the  infrastructure support to pay it forward.  Also, by design, we've learned that these clinical trials deserve many answers when they are put together.  Instead of just answering the primary parent clinical trial, we're now very familiar with adding more clinical questions in that are critically important to the field of PARDS  by adding in ancillary trials to the parent trial. And here is an example within prospect.  We are now working with three sets of investigators looking at endotypes in PARDS, looking at  outcomes evaluation, comparing prospect to ECMO. And we're also looking at now the impact of the microbiome and nutrition research in PARDS.  So this is done by design, just not answering one question, but looking for your colleagues to come and to propose additional research that can be layered onto the parent work.  So congratulations to these investigators in their own R01s linking onto the parent R01.  Alternative research designs is where we form our relationships with our interprofessional statistical colleagues to help us figure this out.  And this is a little bit of a joke slide, but using these alternative research designs  will help us creatively, in an innovative way, pull together information that's being generated in the field and applying it to pediatrics.  For example, well-structured factorial designs.  And we've demonstrated in pediatrics, we can enroll huge numbers of patients, so we can do factorial studies.  A Bayesian approach, linking estimates of treatment effects in children from values in the adult and the neonatal world to enhance our statistical power and to detect pediatric  specific effects.  Adaptive modeling.  Now, instead of waiting for these 1,000, 2,000 patient studies to answer the question and figure out what the impact of our studies will be, adaptive design uses prospectively  planned modification of one or more specified aspects of the study design and hypothesis based on analysis of data from subjects in the current study.  The data analysis are performed at planned interim looks with blinded and unblinded  methodologies that can occur with and without formal statistical hypothesis testing.  This type of adaptive modeling is being used in the prospect clinical trial, which is a two-by-two factorial response adaptive randomized control clinical trial of supine and prone  positioning in PARDS. Consecutive patients with PARDS will be randomized to one of four groups, stratified by age,  less than one, one to seven, and older plus eight, and also stratified by direct and indirect lung injury.  So here, what we do is we enroll, as currently designed, up to 400 patients, 100 patients in each of these four arms.  And then we update our randomization probabilities based upon the data that we see at 400. And what we'll do is that we will increase allocation to well-performing arms and decrease  allocation to poorly performing arms. So after 400 subjects, the randomization probabilities will change and will push newly  enrolled patients to well-performing arms away from poorly performing arms.  And that, I think, helps when you're sitting at the bedside consenting families for the clinical trial.  To decrease heterogeneity within prospect, we're enrolling early. So we are specifically gearing up this rapid, continuous screening process, getting patients  into the study early to decrease the heterogeneity associated with early versus late PARDS. And we're also increasing generalizability by enrolling internationally.  Currently, we've got 50. We're going to move to about 65, 70 international sites using the networks that I mentioned  earlier to work together to answer these complex questions that we have within prospect in our three ancillary trials.  Coming soon, I think, given where we are with PARDS and what the previous speakers, I hope, would have said within our presentation this morning, I hope we're going to move towards  unpacking the PARDS syndrome in and by itself. We need smarter trials targeting therapies that are designed for a specific patient subgroup  rather than all patients as if they were a homeogenous group.  We need to move towards genetic and biomarker-based risk stratification using either point of  care testing or other methodologies to really transform patient selection and get patients into trials earlier.  An excellent example of this is the recently published Lancet respiratory paper that Mary  Dahmer and Heidi Flory and others published that identified phenotypes in pediatric patients  with PARDS using a technique that was first described by Carolyn Kalfi and her colleagues using latent class analysis.  So looking at PARDS and looking at the different inflammatory pattern that these patients are  presenting with, and hopefully by using this work, paying it forward, getting more strategic and a precision way patients into these clinical trials.  Another methodology is really looking at differences, heterogeneity differences, using under-the-skin and over-the-skin variables.  And this was recently, a paper that was recently published by Tessie October and colleagues about how we can look at race and ethnicity and SES within our PARDS clinical trials.  Upon closing, I'd like to say that the future is bright and thank you for your time.  And specifically, I hope the next time we can get together, it won't be just with a talking head on a slide show, but hopefully we can get together and truly discuss opportunities  to advantage our PARDS population in the future. Thank you.

Video Summary

In this presentation, Martha Curley discusses the challenges and hopes in designing clinical trials for pediatric acute respiratory distress syndrome (PARDS). She highlights the unique challenges in conducting PARDS research, including the diverse patient population, operational definitions, and timing of PARDS. Curley emphasizes the importance of collaboration and building networks to collectively answer complex questions in PARDS research. She also discusses the use of big data methodologies, transitioning from one study to the next, and incorporating ancillary trials to address multiple research questions. Curley suggests alternative research designs such as factorial designs and adaptive modeling to enhance statistical power and detect pediatric-specific effects. She concludes by emphasizing the need for smarter trials targeting specific patient subgroups and the potential of genetic and biomarker-based risk stratification in patient selection. Overall, Curley is optimistic about the future of PARDS research and the opportunities to improve outcomes for pediatric patients.

Asset Subtitle

Pulmonary, Research, Pediatrics, 2022

Asset Caption

Pediatric acute respiratory distress syndrome (PARDS) is a manifestation of severe lung injury with a mortality rate of up to 35%. Despite this significance of PARDS in critically ill mechanically ventilated children, respiratory management remains largely supportive, with no data to support one approach over another. This session will highlight advances in the understanding of PARDS by relating experimental models to clinical experience and discussing recent advances in respiratory monitoring and management of PARDS and the identification of biomarkers that can be used for disease stratification and prognostication. Speakers will also discuss how to design clinical trials of PARDS, delivering much-needed scientific knowledge

Meta Tag

Content Type Presentation

Knowledge Area Pulmonary

Knowledge Area Research

Knowledge Area Pediatrics

Knowledge Level Advanced

Membership Level Select

Tag Acute Respiratory Distress Syndrome ARDS

Tag Clinical Research Design

Tag Pediatrics

Year 2022

Keywords

Martha Curley

clinical trials

pediatric acute respiratory distress syndrome

PARDS

challenges

hopes

	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.