My name is Jenny Kingsley. I am at Children's Hospital Los Angeles, where I am a pediatric intensivist and an assistant professor at the Keck School of Medicine at the University of Southern California. I also am faculty in the Center for Bioethics there. I have no financial disclosures, sadly. And I'm here to talk about equity and ECMO. And I'm hoping to bring some of the concepts that my amazing panelists have already talked about to fuller discussion in the next 15 minutes. Before we go further, I think it's important that we define health equity so we are all on the same page. So I define health equity as the concept that everyone has a fair and just opportunity to be as healthy as possible. It requires removing obstacles to health, such as poverty, discrimination, and their consequences, including powerlessness and lack of access to good health care, quality education, and housing. I'm hoping to spend the next 15 minutes highlighting what we know exists in terms of disparities and ECMO utilization in pediatrics. And I want to explore how some of our biases might play into these disparities. And then I'm going to hopefully end with some high-level suggestions for practical or pragmatic solutions, recognizing that I also don't have the answers. Welcome to ethics. ECMO decisions are complex, as you've already heard. There's high mortality and morbidity. ECMO is instituted emergently when death is the alternative. And its use is increasing. Indications are broadening. And management complexity is increasing. There's no standard of care for candidacy, initiation, or discontinuation, as these two lovely people have discussed. Outcomes are uncertain. And in the setting of all of that complexity and uncertainty, there is disparity. And I want to show you what we know about those disparities today. Let's start with a really high-level examination of ECMO utilization and the social determinants of health. There's been a lot of research about disparity in pediatric ECMO utilization and outcomes. The majority of those show that there's no statistical difference by elements of social determinants of health and ECMO utilization. Yet there's still some signal to suggest that disparity exists, because we have reason to believe that those studies may lack the scientific rigor that we really need to see a signal. This is a figure from Moynihan et al. They did a scoping review describing the results of studies examining the associations between social determinants of health and ECMO utilization. To orient you to the figure, the size of the circle or the bubble is proportionate to the number of studies showing that association with the legend on the left side of your screen. So about 55% of the studies show that there's no statistical signal or no statistical difference between social determinants of health and ECMO utilization. But when you drill down, the vast majority of studies have inconsistent presence, strength, and directions of associations. Analyses were imprecise, and there was inconsistently applied confounder adjustments. With the remaining of studies that did show a statistical significance, statistically significant difference, you guys try saying that when your mouth is dry, 84% found adverse outcomes or lower utilization in underrepresented, diverse, or under-resourced populations. So there is some evidence out there to suggest that this is a very real fact that we're seeing. What underlies that? We don't know the answer. It's not just enough to describe disparity. We need to understand the mechanistic drivers that lead to disparities in ECMO utilization by social determinants of health. And as Mithya mentioned, there's many different layers of bias and structural discrimination that all interplay to create this disparity. And the vast majority of studies don't dive deeper to examine the mechanistic drivers. In the Moynihan et al. scoping review, just 17% of papers that described disparities by ECMO utilization went to that deeper level. We know that there are societal factors. There's structural discrimination. We know that geography plays a huge role in access to care, which we'll talk about next. And then we know that there's individual factors. Things like race, ethnicity, racism, immigration status, the language that you speak, your insurance status, all play a role in your ability to get care and may play a role in your ability to get on ECMO. So let's talk about geographic variation. Access to ECMO is neither equal nor equitable. 16 million children, so somewhere between a fourth and a fifth of the pediatric population of the United States, live greater than 60 miles from a pediatric ECLS center. When you drill down on that data, there's even larger discrepancies between children that live in a rural setting versus an urban setting. Nearly half of children that live in a rural setting live greater than 60 miles from an ECLS center. That number drops to just 17% in children living in an urban center. What you see on this slide is a map of the United States. And each dot, whether red or blue, is a census tract. Red is greater than 61 miles away from a ECMO center, and then the blue is within 20 miles of an ECMO center. So there's a little bit of a gap there, but why is this a problem? When children have inequitable or unequal access to care, it creates a de facto deference to the first come, first serve policy, which seems like it would be a pretty explicit criteria for getting on ECMO. But in fact, it further disadvantages people that already have trouble getting their foot in the door. And what this means is that kids may show up sicker, or they may show up after all the other ECMO circuits are used. It also means that people who are already on ECMO have ownership of that ECMO circuit, and we don't have triage policies that are implemented or discontinuation policies that are implemented to allow us to redistribute those resources. Lack of equitable access may have important and significant downstream effects on patient level ECLS utilization and on outcomes, and may in part be a driver of the disparities that we saw on previous slides. When we look at even more of a micro level, there's huge variability in decision making on an individual level, both on an individual institution level and on a provider-to-provider. This is a figure from a cross-sectional survey of ECMO program directors inquiring about candidacy status of patients based off of a given vignette. The left side of the screen is a brief synopsis of the vignettes that the program directors received. And generally, there was a lot more information in the study, but each case was designed such that the degree of cardiopulmonary compromise could justify ECMO cannulation. But past medical history, comorbidities, and surrounding circumstances differ so that the study was really designed to help us understand what's motivating our decision making when it comes to cannulation and candidacy. And what's really interesting is that when participants were asked to explain the factors that were animating their decision making, they looked beyond factors of severity of illness, but instead named values, things like quality of life, neurologic status. And those values were used both to justify offering ECMO and used to not offer ECMO. So I pulled just one, but each vignette has examples of this in the paper. This is a patient who had an electrocution, a prolonged arrest, and had an uncertain neurologic status that was hemodynamically unstable in the PICU. One respondent said, the most important factor in my decision is the uncertain prognosis from the acute injury. So I prefer to buy time with ECMO to properly assess the patient's response. Another respondent used the same uncertainty and neurologic status to say that this is a contraindication for ECMO. While some variability is to be expected and is even normal, the degree of subjectivity that we see in this paper, but also I would argue in all of our day-to-day practice, suggests or supports that ECMO utilization is susceptible to bias. Where does this come up in our day-to-day? I think a lot of it is couched in terms of quality of life determinations. Clinicians, and that's problematic because clinicians and parents have different estimations of quality of life. So what do we know about what we as clinicians think about quality of life? We tend to focus on hardship. We tend to focus on the potential or experience of suffering, which is important. And we should pay attention to our concerns. But we also need to make sure that we are being able to explicitly define what suffering means and looks like. Suffering is often invoked to refer to very different experiences and unconsciously is used to smuggle value judgments about quality of life into decision making. We need to make sure that we're not using the term suffering without clarity about what our specific concerns are, that we aren't exaggerating our claims of suffering, and that that claim from suffering is not shoehorning us into a decision to not offer ECMO. We also tend to focus on our patients' health care utilization. This comes up with patients with high degrees of medical complexity. Two-thirds of residents admit to feeling frustrated caring for children who cannot be cured or function independently. Seventy percent of those residents questioned the aggressiveness of our treatments. How do we marry that experience and feeling and judgment with the experience and feeling of families who love and care for these children? As clinicians, we tend to use composite outcomes to try to find prognoses for our patients, which may lead us to make decision making based on a diagnosis like trisomy 13 or 18 and miss the nuances and the particular circumstances of the particular patient in front of us. Finally, we fall prey to the disability hierarchy. The disability hierarchy is the idea that some impairments are positioned as worse or more severe than others. We know that clinicians rank severe neurologic injury as a major factor that's important for quality of life, which not all families agree with. Parents, on the other hand, focus on gains. They focus on the strength of their relationship with their child. They focus on knowing their child best. And they have a deep understanding and gratitude for their child's life. They focus on meaning making and are motivated by being a good parent. We've heard that the majority of clinicians view decisions around ECMO as being not a shared decision, but as one that we own. We also heard an argument for more shared decision making in these two panelists. I worry that if clinicians only own decision making and we have such disparate values, that we're never going to connect and be able to get to that actual ideal state of shared decision making. How does this show up in conversations and candidacy determination? I worry that we create false objectivity, that we rely on terms like hopelessly ill or that patient has a poor prognosis, which may not be as objective as we think it is. Instead, it's a clinician's personal set of values that we might be expressing about what a life worth living looks like or what a life worth saving looks like. And while we should have opinions about what is right for our patient, we need to make sure that we're not conveying those values as objective fact, but in fact, recognizing the positionality that we bring to the conversation. We risk generalizing a clinician's expertise from medical to moral without recognizing that transition. So where do we go from here? We know there's disparity and we know we all have bias. As Nithya said, we really need candidacy decision making and ECMO decision making that's explicit, consistent, clear, and transparent for clinicians and families that appeal to established ethical guidelines. And I'm not here to suggest that we should abdicate decision making to the values of parents only. I think doing so would actually be a dereliction of our responsibility to make a strong medical recommendation. But we have to do that in the context of our individual patient. With market population heterogeneity, subjectivity and perceived benefits and burdens and what's an acceptable outcome with a limited ability to prognosticate accurately, it's not surprising we don't have a one size fits all guideline. We're never going to get there. But distributive and procedural justice require more transparency and more accountability. I borrowed this phrase from Dr. Moynihan to suggest that we need to move from discrimination to discretion. We need a transparent process-based approach like the one Nithya described, where we're giving personalized recommendations that continues throughout the course of a patient's stay in our ICU or hospital. It needs to be re-evaluated constantly and with the support of a team. We also need accountability metrics, something like a dashboard where we can clearly see how our decision-making is changing based off of patient population. I know that this doesn't exist. And I know that this would be really challenging to do. Our data isn't precise. ECMO is hard to research. And we don't really register who we say isn't a candidate. So we don't have the data. And this is a huge blind spot for us. But I would suggest that it would be a place to start. With that, I'm going to close. Thank you guys so much. Thank you.

health equity

ECMO utilization

social determinants

clinical biases

ethical guidelines