Introduction

Among the most enduring and consistent observations in intensive care unit (ICU) organization is the finding that ICUs that care for a high volume of patients experience improved risk-adjusted outcomes. This finding, known as the “volume-outcome relationship,” was first robustly demonstrated in mechanically ventilated patients in 2006 [1] and has been re-demonstrated in multiple other ICU cohorts since [2]. Yet despite the growing evidence in support of volume-outcome relationships, clinicians and policy makers have struggled to translate these relationships into improved outcomes for ICU patients. This article will review recent research findings related to the ICU volume-outcome relationship, with a particular focus on data suggesting how we might harness these relationships to improve the quality of care in all ICUs.

Potential mechanisms of the volume-outcome relationship

A major unknown in the volume-outcome literature is the mechanism underlying the observed relationships. In the ICU there are three potential mechanisms [3]. First is “selective referral,” by which patients are referred to high-quality providers, thus driving an increase in volume. Under this mechanism, high quality leads to high volume, not the other way around. Second is “practice-makes perfect,” by which clinical experience causes providers to improve over time. Under this mechanism, high volume leads to high quality. Third is the notion that there are organizational structures common at high volume hospitals that are associated with outcome, such as high-intensity physician staffing, multidisciplinary care teams, or protocol use. Under this mechanism, we would observe no volume outcome relationships if we could properly control for these factors.

Although the exact mechanism of the volume-outcome relationship remains uncertain, several recent studies help shed light on the issue. In the US, Iwashyna et al. [4] examined the transfer destination of patients with acute myocardial infarction admitted to hospitals not capable of coronary revascularization and transferred to capable hospitals. They found that within a transfer distance of 100 miles, only 13.1 % of patients were transferred to the best hospital in the area, and in 38.6 % of cases patients bypassed a higher quality hospital en route to their destination hospitals. This suggests that physicians are not good at selecting high-quality transfer destinations for critically ill patients, making “selective referral” an unlikely mechanism for the volume-outcome relationship.

Other recent studies demonstrate that clinicians tend to be more adherent to evidence-based practices at high-volume hospitals. In one study, patients with acute respiratory distress syndrome (ARDS) were more likely to be ventilated with a lung-protective ventilation strategy, and less likely to be ventilated with injurious tidal volumes, in high-volume, academic hospitals with trained intensivists [5]. In a randomized trial of extracorporeal membrane oxygenation (ECMO) in the UK, patients with severe ARDS transferred to a large regional referral center were also more likely to receive lung-protective ventilation, even if they did not receive ECMO [6]. Together, these studies provide preliminary data that either clinical experience or organizational factors explain the volume-outcome relationship. Future research is necessary to further delineate these two potential mechanisms.

Lessons from negative volume-outcome studies

Another recent advance in our understanding of the volume-outcome relationship came with the publication of two high-impact papers demonstrating an absence of a volume-outcome relationship in key patient groups [7, 8]. The first study was performed in 119 hospitals in the Veterans Health Administration network, a large government-run health system in the US [7]. Despite substantial variation in volume, higher volume hospitals experienced no improvements in risk-adjusted mortality among mechanically ventilated patients (odds ratio for each 50-patient increase: 0.98, 95 % CI 0.87–1.10). The second study was performed in 170 general critical care units participating in the UK’s Intensive Care National Audit and Research Centre’s case mix program [8]. This study, which focused on patients admitted with severe sepsis, also found no improvements in mortality at higher volume hospitals (odds ratio comparing the highest and lowest volume quartiles for mechanically ventilated septic patients: 0.92, 95 % CI 0.79–1.08).

These studies are important because they show that some health systems are able to overcome the volume-outcome relationship and provide high-quality care at low volume hospitals. Both the US Veterans Affairs and the UK health systems make use of routine audit and feedback of quality data to participating hospitals and routinely share quality improvement strategies across hospitals. It is possible that these efforts resulted in quality improvements at low-volume hospitals. These results should be interpreted with caution, since the volume range was relatively narrow in both of these studies, and volumes were generally low compared to other studies [2]. Nonetheless the studies provide the early insight that consistent high quality at small hospitals is possible.

Outreach strategies in low-volume ICUs

One response to the volume-outcome relationship is to regionalize ICU care by systematically transferring high-risk patients to regional centers of excellence. ICU regionalization is actively being explored by several groups [9]. Another response is to use community outreach to improve the quality of care in small volume hospitals without transferring patients. A recent study to explore this strategy used videoconferencing-based education and guideline dissemination to improve the quality of care at six community hospitals in Ontario, Canada [10]. In a cluster-randomized trial design, the authors found that evidence-based care practices increased significantly compared to control hospitals (summary odds ratio 2.79, 95 % CI 1.0–7.74), with the greatest increase in semi-recumbent positioning to prevent pneumonia and precautions to prevent catheter-related bloodstream infections. This study demonstrates that use of enabling technologies for community outreach may be an efficient way to improve outcomes in small hospitals. Similar results were recently reported using more traditional quality improvement approaches based upon education, such as the Surviving Sepsis Campaign [11].

Conclusions

Volume-outcome relationships have great potential to provide insight into novel ways to improve the quality of critical care. Recent advances suggest that the mechanism of the volume-outcome relationship is through better implementation of evidence-based practice, through either clinical experience or adherence to care protocols and other organizational factors. Some health systems have overcome the volume-outcome relationship in the ICU, indicating that community outreach, through either tele-health or other novel strategies, is a potentially valuable tool for improving critical care outcomes in small hospitals.