Who gets ECMO and who does not? Clinicians say guidelines are not driving the decision

JAMA Network

Extracorporeal membrane oxygenation - ECMO - is one of the most resource-intensive interventions in modern medicine. The machine takes over the work of the lungs, drawing blood out of the body, oxygenating it through an artificial membrane, and returning it to the circulation. For patients with severe respiratory failure who are dying despite maximum ventilator support, ECMO can be the difference between survival and death.

But ECMO is scarce. The machines are expensive. The circuits require specialized perfusionists to manage. The patients need round-the-clock intensive care nursing at ratios that strain even well-resourced hospitals. Not every patient who might benefit from ECMO can receive it, and not every center that has the equipment has the personnel to run it safely.

This scarcity creates a high-stakes allocation problem. Who gets offered ECMO, and on what basis? A new qualitative study published in JAMA Network Open suggests that the answer is less systematic than most people would assume.

Clinical intuition over checklists

Derek Soled and colleagues at Brigham and Women's Hospital conducted in-depth interviews with clinicians involved in ECMO candidacy decisions for patients with severe respiratory failure requiring venovenous ECMO - the configuration used for lung support, as opposed to venoarterial ECMO, which supports the heart. The interviews explored how clinicians actually make these decisions in practice, what factors they weigh, and how they handle uncertainty.

The findings paint a picture of decision-making driven primarily by individual clinical judgment rather than by standardized, objective guidelines. Clinicians described assessing patient suitability based on their experience, their reading of the clinical picture, and their sense of whether a particular patient was likely to benefit. Formal scoring systems and published criteria, while acknowledged, did not appear to function as the primary decision-making tools in most cases.

This is not unusual in critical care medicine. Many life-or-death decisions - when to intubate, when to withdraw life-sustaining treatment, when to pursue aggressive surgical intervention - rely on clinical judgment informed by experience rather than algorithmic protocols. But ECMO is different from most interventions in one critical respect: it is rationed. When judgment rather than criteria drives allocation of a scarce resource, the risk of inconsistency and inequity increases.

The variability problem

If two clinicians at the same hospital assess the same patient differently - one believing the patient is a good ECMO candidate and the other believing they are not - the patient's survival may depend on which clinician happens to be on call. If clinicians at different hospitals apply different thresholds for candidacy, then geography becomes a determinant of access to life-saving technology.

The study identifies this variability as a core concern. Without shared, objective criteria for ECMO candidacy, allocation becomes inherently inconsistent. A patient who would receive ECMO at one center might be declined at another, not because the second center lacks equipment but because its clinicians apply different judgments about suitability.

This variability can also intersect with known disparities in critical care. Research in other areas of medicine has shown that subjective clinical decisions are vulnerable to implicit biases related to race, socioeconomic status, age, and perceived social worth. Whether such biases influence ECMO allocation has not been extensively studied, but the reliance on subjective judgment rather than objective criteria creates the conditions under which bias can operate.

Why standardization is harder than it sounds

The absence of rigid guidelines is not simply an oversight. ECMO candidacy involves genuinely complex clinical assessments. The ideal ECMO patient has severe but potentially reversible respiratory failure, adequate vascular anatomy for cannulation, and no conditions that would make anticoagulation (required during ECMO) prohibitively dangerous. But each of these criteria involves judgment calls.

How reversible is reversible enough? A patient with severe pneumonia may recover lung function; a patient with end-stage pulmonary fibrosis almost certainly will not. Between those extremes lies a wide range of conditions where the probability of recovery is uncertain. Clinicians must estimate that probability based on incomplete information, and reasonable clinicians will disagree.

Age is another contested factor. Some centers set explicit age cutoffs for ECMO candidacy; others evaluate older patients on a case-by-case basis. Body mass index, pre-existing organ dysfunction, duration of mechanical ventilation before ECMO initiation, and the presence of other infections all factor into decisions in ways that are difficult to reduce to a simple checklist.

The clinicians interviewed in this study were not acting carelessly. They were navigating genuine clinical complexity with imperfect tools. The problem is systemic rather than individual: the field has not developed the consensus criteria that would allow standardization without sacrificing the nuance that individual cases demand.

What the study can and cannot tell us

As a qualitative study, this research describes how clinicians think about ECMO decisions. It does not quantify how often those decisions differ, does not track patient outcomes by decision pathway, and does not measure whether specific patient populations are systematically disadvantaged by current practices.

The study focused on venovenous ECMO for respiratory failure, and its findings may not generalize to venoarterial ECMO decisions, which involve different clinical considerations and different patient populations. The clinicians interviewed came from institutions experienced enough in ECMO to participate in candidacy decisions regularly; decision-making at smaller or less experienced centers may look different.

Qualitative research also captures what people say about their decision-making, which may differ from what they actually do. Clinicians may understate the role of guidelines or overstate the role of clinical judgment in retrospective interviews. Observational studies that track real-time decision-making would complement these findings.

Toward criteria that balance rigor and reality

The study's implicit argument is that the ECMO field needs to move toward more standardized candidacy criteria - not rigid algorithms that eliminate clinical judgment, but frameworks that constrain the range of acceptable variation and make the basis for decisions more transparent and auditable.

Other areas of transplant and critical care medicine have developed allocation frameworks that balance clinical judgment with standardized criteria. Organ transplant allocation, while imperfect, uses scoring systems that prioritize patients based on measurable factors. Triage protocols in mass casualty events apply explicit criteria to distribute scarce resources. ECMO could benefit from similar frameworks adapted to its clinical specifics.

The stakes are high enough to warrant the effort. For a patient with severe respiratory failure, the decision about whether to offer ECMO is the decision about whether to offer the last remaining chance at survival. Making that decision more consistent, more transparent, and more equitable is not just an academic exercise. It is a matter of who lives and who does not.