DOI: 10.1097/CCM.0000000000000061

,  

,  

PMID: 24534958

Issn Print: 0090-3493

Publication Date: 2014/03/01


Print

Volume-Outcome Relationships in Extracorporeal Membrane Oxygenation: Is Bigger Really Better?*

Graeme MacLaren; Sara K. Pasquali; Heidi J. Dalton

    loading  Checking for direct PDF access through Ovid

Excerpt

The association between larger volumes and improved performance has been accepted for years in various industries and has become an increasingly important topic in medicine. The notion that surgeons who more frequently perform a procedure achieve better outcomes than those with less experience has been grudgingly accepted for years (1), and several hundred publications clearly demonstrate a volume-outcome relationship for a host of medical procedures (2, 3). However, although the relationship is consistent, its magnitude varies and the underlying associations differ between surgical disciplines (2, 4). The strongest associations between higher volume and improved outcome are observed when the procedure evaluated is high risk or high complexity (5). Extracorporeal membrane oxygenation (ECMO) is a good example of such a procedure. Patients placed on ECMO are desperately ill and unstable, often making the initiation process difficult and fraught with potential complications. As the majority of programs have several clinicians who perform ECMO cannulation, experience with access techniques in a variety of patient populations may be diluted. Furthermore, placement of access cannulas is not the only facet in the care of ECMO patients where volume and experience may play a role. How well the ICU team has been trained; the quality of ancillary clinical staff, radiological and laboratory support, and the effectiveness of teamwork between multiple medical and nursing disciplines all may play a more important role in determining patient outcomes (6). In other words, when applied to complex interventions such as ECMO, center volume may not be a good surrogate for quality.
In this issue of Critical Care Medicine, Freeman et al (7) present an analysis of the volume-outcome relationship in 7,322 pediatric and neonatal ECMO patients from 40 hospitals participating in the Pediatric Health Information System. Centers were categorized as low, medium, or high volume based on the number of ECMO patients supported annually (< 20, 20–49, > 50 patients per year, respectively). After adjusting for possible confounders, they demonstrated that ECMO patients managed in low-volume centers had statistically higher mortality rates than in busier centers, with a cutoff for increased mortality noted at center volumes of less than 22 ECMO patients per year. This work complements the study by Karamlou et al (8), which noted a similar impact on survival for patients receiving ECMO for cardiac failure. Centers supporting less than 15 patients per year had reduced survival, despite risk adjustment. Higher volume centers have also been shown to use fewer resources. The randomized trial of ECMO for adult respiratory failure in the United Kingdom showed improved survival in patients treated at a single site, leading commentators to conclude that experience both with ECMO and conventional management provided superior care at high-volume centers (9).
On the basis of this evidence, should smaller centers stop doing ECMO and refer all patients to institutions with higher case volumes? On the surface, this may seem to be logical. In some parts of the world, the impact of volume and outcome in cardiac surgical programs has resulted in concentration of such procedures in large-volume centers. This is not the case in the United States, however, where having a functioning cardiac program is considered an essential element of hospital care in many small, medium, and large centers.
Although the data in the current report are compelling, there are several limitations. Freeman et al (7) used an administrative dataset based primarily on discharge coding and financial information to conduct their analysis; such datasets are limited by inherent difficulties with accurate case ascertainment and risk adjustment (10). Important variables that could not be assessed included patient complexity, severity of illness, pre-ECMO physiology, and institutional thresholds to initiate ECMO.
    loading  Loading Related Articles