Pediatric inpatient and intensive care specialists working outside of tertiary medical centers confront difficult clinical scenarios related to how best to care for extremely ill children who may or may not benefit from advanced medical technology, and these clinicians are often faced with limited local availability. Extracorporeal membrane oxygenation (ECMO) is a technology that is only available at a subset of tertiary care centers, and the decision to risk the transfer of a child for the potential benefit of ECMO is challenging. This article is aimed at addressing the main factors and ethical principles related to this decision-making: (1) whether ECMO is the standard of care, (2) clinical decision analysis of the risks and benefits, (3) informed consent and education of the parents and/or guardians, and (4) institutional leadership decision-making. A decisional framework is proposed that incorporates a thoughtful shared decision-making algorithm.
Physicians working in PICUs outside of tertiary care centers have many resources available to them. However, there may be additional therapies or interventions that are only offered at specific high-volume centers with certain subspecialties. As a result, these physicians may face situations in which they desire to escalate care without access to all potentially life-saving resources. Extracorporeal membrane oxygenation (ECMO), for example, is a highly advanced therapy that provides lung or heart and lung bypass that can be used as a rescue therapy for pediatric patients with severe life-threatening cardiopulmonary diseases, such as myocarditis and other acquired heart failure, pediatric acute respiratory distress syndrome (PARDS), refractory status asthmaticus, and refractory septic shock. ECMO is not available at many PICUs outside of a subset of tertiary care centers; it is highly specialized and resource-intensive, requiring a trained multidisciplinary team and a critical volume of patients to provide the therapy safely, reliably, and effectively. Data on the effectiveness of ECMO at centers that have specialized teams suggest that survival is enhanced, even with children who are moribund or those undergoing active cardiopulmonary resuscitation from a witnessed arrest.1,2
If a critically ill child is deemed to be eligible for ECMO, referring hospital PICUs may attempt to transfer these patients preemptively to a tertiary care center where this therapy is available. The questions of who, when, and whether to transfer from a referring hospital to a tertiary care hospital is often fraught with uncertainty. Fellowship training for pediatric critical care physicians almost invariably occurs in centers with ECMO; therefore, most are unprepared or untrained for this dilemma. Referring PICUs successfully care for the vast majority of critically ill children without the use of ECMO. It is clinically challenging for a physician to predict which small percentage of critically ill children could have significant clinical deterioration to an extent that would outstretch the limits of local resources. The decision about who and when to transfer is further complicated by the inherent risks of transporting a clinically worsening patient. In addition, transferring sick children out of the referring hospitals has the detrimental effect of decreasing local patient volume and decreasing the referring PICU’s staff experience and expertise in the care of the most critically ill children. A long-distance transfer also impairs the patient’s family’s ability to be close to home and forces them to leave care teams and support systems they trust. If transfers are initiated too frequently and too early, they could place unnecessary strain on the resources at accepting tertiary care centers, particularly if most transfers respond to conventional care and do not end up requiring specialized therapies. On the other hand, if a transfer is initiated too late, a patient who may have benefited from the use of a specialized therapy, such as ECMO, may die. As a result of these factors, physicians at referring PICUs face a conundrum when a patient, despite optimization and maximum support at the local level, has progressively worsened, usually quickly, to a severity of illness with a high risk of death and is a candidate for ECMO support but remains at a non-ECMO center.
Little consensus exists as to the best approach to managing this difficult situation. In fact, in the last 6 months in a local PICU at a regional hospital, 3 physicians asked themselves these questions when caring for 3 children with severe refractory hypoxic and hypercarbic respiratory failure. These children all required high ventilator support and would likely have been candidates for ECMO if they had been admitted to an ECMO center. Faced with these questions, the 3 caring, compassionate, and skilled clinicians all chose different paths. One chose to keep the child in the local PICU and continue conventional management; one chose to transfer the severely ill child by helicopter to a regional ECMO center, where the patient was placed onto ECMO; and the third physician mobilized the resources of adult physicians at the same institution to initiate ECMO in the local PICU (for the first time in this institution’s history) and then transferred the child on ECMO to a regional ECMO center. Each of these 3 options involves a high risk of death. The end result was 2 survivors (1 with morbidity) and 1 death. No option is clearly superior on the basis of risk/benefit analysis. Discussions with families were all different. All 3 clinicians still wonder if they did the right thing.
The purpose of this article is to provide an ethical analysis of the issues raised when considering ECMO for children with severe life-threatening cardiopulmonary diseases at a non-ECMO center. First, we will analyze the state of ECMO to determine the certainty or uncertainty of its role as standard of care in pediatrics and the implications of the technological imperative of using ECMO. Second, we will approach the ethical considerations from the perspective of the clinician who must make critical decisions in a highly stressed and time-sensitive situation. Third, we will seek to understand the obligations to the child’s family for informed consent in a truly life or death situation for their child. Finally, we will assess the administrative decision-making that is required by referring pediatric centers to devise the policies and procedures (on the basis of ethical principles) that will define a standard of care at these institutions. This will hopefully lead to an ethical framework to help these decisions and discussions with families to be rationally and ethically sound in the future.
ECMO As Standard of Care?
The question of whether ECMO should be available to all children who are severely sick, regardless of where they are, requires a careful analysis of the risks and benefits of the therapy. There is the risk of overestimating the effect of ECMO on the basis of the technological imperative in health care. This refers to the increased use of advanced technologies by physicians to “do everything possible” to extend the life of the individual, which may save them or may just prolong the inevitability of death. As Hofmann3 has outlined, however, “There is no technological imperative, but technology promotes a moral imperative; in particular, it promotes a moral imperative to proper assessment.”
Therefore, an attempt at a proper assessment is needed. When the focus is on ECMO in pediatric patients with PARDS, excluding neonates and adults, the published data are conflicting at best and have been the focus of a recent consensus guideline report.1 In this summary, the authors evaluated the literature and concluded that ECMO should be considered and that strong consideration for the quality of life of the child (due to morbidity) and the likelihood of benefit must be included. There are no controlled studies on the effectiveness of ECMO in PARDS; there are only case series. One clinical trial was stopped because of too-high survival in the control group.4 Case series have revealed a decrease in mortality at ECMO centers when using ECMO for PARDS, but others have revealed no change in mortality when propensity score–matched cohorts are compared.4,5 International databases on >2000 cases of ECMO for PARDS revealed an overall survival rate of 54% and revealed that complications of the therapy are common.6 A historical eligibility criterion for ECMO in PARDS has been the oxygenation index (OI), which factors in hypoxia and mechanical support. An OI >40 was considered to carry with it an 85% predicted mortality and thus warrant the risks of ECMO. However, the most recent analysis of the OI in PARDS revealed significant improvement in mortality if an OI > 40 without ECMO.7 In the 2020 “Surviving Sepsis Campaign International Guidelines for the Management of Septic Shock and Sepsis-Associated Organ Dysfunction in Children,” the authors suggested using “venovenous ECMO in children with sepsis-induced PARDS and refractory hypoxia (weak recommendation, very low quality of evidence)” and “venoarterial ECMO as a rescue therapy in children with septic shock only if refractory to all other treatments (weak recommendation, very low quality of evidence).”8 Refractory shock is poorly defined in the literature. In a subsequent meta-analysis, Ramanathan et al9 found that children with sepsis and refractory shock who were cannulated to ECMO had survival rates as high as 59% but noted several limitations with the study and call for more work to define the thresholds and timing for initiation of ECMO.
It is important to assess the feasibility and utility of creating a pediatric ECMO team at a referring PICU. ECMO requires a team that includes surgeons, respiratory therapists, nurses, perfusionists, intensivists, hematologists, or anticoagulation specialists and proper equipment. Those same consensus guidelines mentioned above state that competencies for all involved with ECMO should be required. The volume of ECMO cases has been shown to be significantly associated with mortality, with a vol of 22 cases per year being the inflection point of worsening survival at low-volume centers.10 Many referring PICUs may have 1 to 10 patients a year who are ECMO eligible. With so few patients and so many involved stakeholders in the process, competency is a limiting factor. Simulation programs have grown to augment live experiences for training, and ECMO simulation focuses on those providers routinely involved in providing robust care at ECMO centers (intensivists, respiratory therapists and/or specialized perfusionists, surgeons, etc). Simulation may be useful and necessary for team building and communication, maintenance of skills and/or competencies, and technical troubleshooting given the critical nature and complexity of a successful ECMO run. In one of the recent cases referred to earlier, adult specialists were used for pediatric ECMO, but each provider was outside the scope of their practice and was using techniques and equipment not specific for children. The added risk of initiating ECMO at a non-ECMO center has not been formally evaluated.
A proper assessment of the process of transferring critically ill children to another hospital must include an evaluation of the inherent risks of relocation. Prolonged time away from an ICU for an emergent transport, even with highly specialized pediatric transport teams, increases the chances of mortality and morbidity en route or soon afterward. Systemic review of intrahospital transport of critically ill children revealed an incidence of reported adverse events ranging widely from 0.11% to 75%.11 Patients who were sicker had higher rates of adverse events, with incidence of unplanned extubation rates as high as 3.5%. Hypotension during transport ranged from 3.3% to 37.5%, with vasoactive infusions being interrupted in up to 5% of patients. In a tertiary care center comparing critically ill children admitted via transport versus those admitted from their own emergency department, there was a significantly increased odds of mortality within 72 hours of admission in the transport group (odds ratio 2.18; 95% confidence interval 1.07–4.45).12,13 Transferring on ECMO has also been evaluated. On ECMO, specialized transport teams have successfully shown no increase in mortality, although up to 20% of patients have serious complications during the transport.14
Putting these facts together to help create an educated clinician at a referring PICU is not easy; there is lack of evidence-based strong recommendation for ECMO in patients with PARDS, and there are undoubtedly risks in transporting these patients as well. Therefore, the clinicians must make decisions on the basis of strong uncertainty, which is the crux of the dilemma. ECMO should be considered, as the experts recommend, but it is not standard of care. How clinicians hold uncertainty during that time of consideration, logically, ethically, and/or emotionally, while making decisions becomes the most significant component. The impact of the desire to do everything possible cannot be quantified when faced with a dying child.
Clinician Decision-making
Having established that ECMO should be considered for all 3 scenarios listed earlier, what should a clinician do with the uncertainty of the benefits of ECMO and risks of transport and ECMO-related complications? The sicker a child gets, the higher the benefit of ECMO may be, but the risks of transport also increase. Logically, and strictly using a risk/benefit analysis, the data do not support transferring to an ECMO center if the child has reached such a state of critical illness. Ethically, the clinician does not have an obligation to try treatments or therapies with uncertain benefits. Uncertainty, however, can be interpreted differently on the basis of one’s emotional state. “Uncertain” could mean that there may be a benefit, and it is thus worth a try, whereas “uncertain” in a different emotional state could mean taking a child and family through a complicated and risky process that likely does not outweigh the potential benefits. Deciding what one would want to do if it were one’s own child often becomes intertwined in the decision-making. When the prospect of losing a child becomes a likely reality for compassionate physicians, their decision-making can be clouded by emotional, and at times, brave attempts at doing everything. If a child dies locally without ECMO, the clinician must be assured that they have done everything within their means to save the child.
Providing local ECMO without a team specifically in place for pediatric patients introduces unknown benefits and risks that make the decision even murkier. However, transporting pediatric patients on ECMO is a possibility, although data are much sparser than data in adults. There are 2 methods of transporting patients on ECMO. Primary transport involves the receiving institution providing a transport team that cannulates the patient at the referring facility and then transports to the ECMO center. Secondary transport is when the patient is already supported with ECMO at the referring facility and is then transported to the ECMO center. In the absence of an experienced pediatric ECMO team at the referring center, secondary transport should be excluded. There is the potential that an adult-sized child (teenager) may be safely placed onto ECMO with a team of adult colleagues familiar with cardiac bypass in adults. Although complications may arise for patients transported on ECMO, mortality is similar to that for patients on ECMO who do not require transportation. In the absence of significant data specific to pediatric ECMO transportation, this option should be decided on a case-by-case basis in consultation with the receiving ECMO center.14,15
Under these circumstances, seeking a second or third opinion can provide valuable counsel to the clinician. Consultation with tertiary ECMO centers is warranted; these centers have criteria for ECMO eligibility, and if a patient is not an ECMO candidate on the basis of other factors (such as degree of multisystem organ failure or irreversible underlying pathophysiology), then no decision is needed and local care should continue. Consultation with PICU colleagues or leadership may provide direction when trying to make a sound decision. Emotions may be extremely high among the clinical team, and data suggest that this can significantly impact clinical decision-making.16 Decisions to stay or transfer for ECMO may be more objectively vetted with other colleagues to mitigate the emotional component of such highly charged decisions. Having collaborating clinicians, including one with less emotional ties, can only enhance the confidence of the recommendations to stay or transfer.
Parental Informed Consent
If we are employing a shared decision-making process, we accept that clinicians do not have unilateral decision-making authority over whether to offer ECMO, and we acknowledge that we must include parents in these discussions because they make medical decisions on behalf of their children. These decisions, however, have often not included parents in the conversation. Parents of critically ill children are assumed to have an extremely high degree of stress given the potential of losing a child. Parents seek trust in the medical team to fuel their hope, especially as their child gets sicker. Introducing uncertainty into the conversation about doing everything for their child is truthful yet worrisome for clinicians who feel a need to provide confidence in such a dire situation. Parents may not have an awareness of their own technological imperative biases or fears but would still be expected to provide educated, logical, and sound consent. Paternalism in decision-making, particularly with life-altering therapies, such as ECMO, is not justified. Paternalism in medicine has been routinely disregarded as ethically sound as the central role of autonomy has evolved. However, examples of justified paternalism still exist in ethical literature.17 Certitude of the right decision and what is clearly in the best interest of the medical decision-maker is required, however, and the obvious uncertainty of this dilemma has been outlined above.
The clinical team owes it to the parents to give them a chance to make a shared decision. As described in an article by Young et al,18 allocation of resource decisions in the ICU should be disclosed for reasons that promote transparency, professionalism, and effective health care delivery. Ethically, the rights of autonomy apply even in the most difficult clinical scenarios. Consent to stay despite a relatively close ECMO center or consent to transfer in critical condition must be educated and include an understanding of the risks and benefits. If the decision to go onto ECMO locally is employed, parents should be informed of the reasons why the PICU is considered a non-ECMO center and of who will be doing the procedures and why. It is also necessary to address the impacts of transferring to a different region away from home and transitioning to a new care team. Often these sick children have multiple providers who have developed a relationship with the family and child.
The clinician caring for a child severely ill enough to be considered a pediatric ECMO candidate has a moral obligation to include the parents in any decision. The clinician has the burden of providing an understanding of such uncertainty while maintaining confidence. Shared decisions will be stronger and build trust. Disagreements between the parents’ choice and the clinician’s choice must be honored and addressed to aid mutual understanding of the values driving each decision. Ultimately, parental authority must be honored.
Table 1 offers a framework to integrate the role of parental decision-making into the difficult process of informed consent for transporting a critically ill child for potential ECMO. Second opinions are imbedded into the model to help the clinician make appropriate recommendations if there are conflicts and when there is uncertainty. It is critical for the clinician to appreciate the nuances between the 3 categories in the table: may be beneficial, uncertain, and may be more harmful. Understanding these nuances will help the clinician to be aware of the degree to which their own interpretation of risk is influencing their discussion and recommendations with the parents or guardians.
Transfer Grid for Potential ECMO After Consultation With Parents
. | Parents Want ECMO . | Parents Uncertain . | Parents Do Not Want ECMO . |
---|---|---|---|
Clinician: ECMO may be beneficial | Transfer while seeking second opiniona,b | Stay while seeking second opiniona | Stay |
Clinician: ECMO uncertain | Transfer while seeking second opiniona,b | Stay while seeking second opiniona | Stay |
Clinician: ECMO more likely to be harmful than beneficial | Stay while seeking second opiniona | Stay | Stay |
. | Parents Want ECMO . | Parents Uncertain . | Parents Do Not Want ECMO . |
---|---|---|---|
Clinician: ECMO may be beneficial | Transfer while seeking second opiniona,b | Stay while seeking second opiniona | Stay |
Clinician: ECMO uncertain | Transfer while seeking second opiniona,b | Stay while seeking second opiniona | Stay |
Clinician: ECMO more likely to be harmful than beneficial | Stay while seeking second opiniona | Stay | Stay |
Includes risks associated with transport.
Other PICU intensivist or PICU leadership as second opinion.
If local ECMO policies, procedures, and personnel are in place, local ECMO cannulation is possible.
PICU and Pediatric Administrative Decision-making
The PICU and pediatric leadership team must be sure that it is appropriate to remain a non-ECMO center. On the basis of published data, if a PICU has the potential to cannulate ∼20 pediatric patients a year, that may be a time to introduce ECMO to the institution.10,11 In addition, it is the responsibility of the referring PICUs to have advanced medical support to provide all other modalities of proven intensive care therapies. To keep a child at the local PICU, all other ICU care must be equivalent so that the only variable is ECMO. Clinicians must be confident in their local team to provide the best care they can if they are to recommend staying. If other support modalities, such as ventilator use or vasopressors, are not optimized, then there are services in addition to ECMO that may prompt the transfer of a patient to a tertiary care center. Leadership must be vigilant about establishing highly functioning critical care teams. Currently, there is no ethical obligation to become an ECMO center because of the uncertainty of the effectiveness of the therapy.
Leadership and administration have biases toward keeping children in their local PICU. Caring for patients who are sick raises the caliber of care in a unit as a whole. Children who are sick rally a team to function at their best and provide significant incentive for all caregivers to come together for the sake of the child. Each good outcome increases the confidence and expertise of the team and raises the level of care for all patients. Each undesirable outcome provides a different type of incentive to learn, grow, and improve. Obviously, children who are sick require a large investment of staff, time in the PICU, multiple modalities of support, multiple consultants, and many billable hours and services.
It is imperative to have all institutional stakeholders involved in the discussions as to whether to offer ECMO locally, on a case-by-case basis, by a non–pediatric-trained adult cardiac bypass team. Emergency procedures under critical conditions require clearly defined inclusion criteria, such as patient age and size; well-delineated roles and responsibilities, with a distinct chain of command; and established efforts to create policies, grant pediatric privileges, and ensure safety. There is too much potential variability in practice and risk for poor outcomes and associated liability without preestablished criteria and policies in place before any ECMO cannulation.
Finally, each referring PICU should establish transport relationships with ECMO centers in advance of potential ECMO transfers to understand the skill level of the teams and the capabilities of their transport vehicles and to discuss the limits of transport eligibility.
Discussion
If a child, who may have survived on ECMO at a regional center, dies locally; if a patient dies during transport or soon afterward; or if a patient has serious complication and dies going onto ECMO at a non-ECMO center, can the clinician rest assured and/or look eye to eye to a parent and tell them that they have done the best they can? All 3 clinicians in the introduction were left wondering if the decisions they made was the best for their patients.
Is this situation similar to care in the 1930s, when the iron lung became available for the treatment of respiratory failure from polio in limited settings? How often were children or adults transferred (or considered for transfer) to an iron lung center when they developed severe polio? How many patients died during that era who may have been saved? How many patients made it to an iron lung but died anyway despite the new technology? This technology changed and improved over time so that mechanical ventilation now is undoubtedly the standard of care for respiratory failure. ECMO may evolve in that manner with time. Currently, ECMO support is only an option in medical establishments with a multitude of resources. Many developing countries do not have the luxury to worry about the potential for ECMO cannulation because they often suffer more acutely from a lack of basic mechanical ventilation support and do not have the option for ECMO. ECMO rescue therapy, therefore, should be appreciated as an option for a select few in the world.
This analysis was done to provide a framework for the clinicians at non-ECMO centers with capabilities to transfer to a tertiary care center that does have ECMO. We believe the clinicians at non-ECMO centers can make informed decisions and provide appropriate counsel if they appreciate the relevant facts and seek advice before deciding. Clinicians can make this final decision confidently when the following concepts are thoroughly explored.
The clinician should have an understanding of (1) the local PICU’s limitations (2) the risks and benefits of transport to an ECMO center, and (3) the risks and benefits of ECMO both locally and at a tertiary center with ECMO. The institutions much be consistent and have coherent strategies to (1) avoid variability in patient inclusion and exclusion criteria and (2) clearly delineate the multidisciplinary team’s and team members’ roles and responsibilities with associated privileges in place. All parties involved must seek to be honest and teach families the complexities of these decisions, so truly informed consent can be obtained.
ECMO centers, on the other hand, are in a unique position to (1) provide local and regional PICUs with an educational framework for early identification of potential ECMO candidates (eg, inclusion and exclusion criteria and monitoring parameters, such as OI and lactate levels) and (2) provide case-by-case consultation services to help the local or regional intensivists determine a patient’s eligibility for ECMO. During this consultation, these ECMO centers can draw on their unique experiences and proven outcomes, in combination with the patient specifics, to help the referring intensivists assess the risk/benefit of transfer. If a patient is deemed to be a candidate for ECMO and will transfer to an ECMO center, there should be a plan in place to return that patient to the local or regional PICU once the patient is on a clinically improving trajectory and never needed ECMO or has been successfully decannulated from ECMO and is stable for transport.
Fellowship programs for pediatric intensivists should consider adding decision analysis and risk/benefit analysis training into their curricula because many intensivists do not practice at tertiary centers and will face this and other scarce resource decisions during their career. This framework may not prevent compassionate clinicians from wondering if they did the best for their patients, but it hopefully provides a means of assuring all involved that the right questions have been asked, shared, and informed appropriately with families and that the decision is sound and consistent with the wishes of the child’s family.
FUNDING: No external funding.
Dr Bateman conceptualized the article and prepared the initial manuscript; Drs Johnson, Tiber, and Gauguet, Ms Fortier, and Dr Valentine provided significant manuscript content and concepts and contributed to critical reviews and revisions of the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
References
Competing Interests
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
Comments