Impact of hospital transfer to hubs on outcomes of cardiogenic shock in the real world

Aims Cardiogenic shock (CS) is associated with significant mortality, and there is a movement towards regional 'hub-and-spoke' triage systems to coordinate care and resources. Limited data exist on outcomes of patients treated at CS transfer hubs. Methods and results Cardiogenic shock hospitalizations were obtained from the Nationwide Readmissions Database 2010-2014. Centres receiving any interhospital transfers with CS in a given year were classified as CS transfer 'hubs'; those without transfers were classified as 'spokes.' In-hospital mortality was compared among three cohorts: (A) direct admissions to spokes, (B) direct admissions to hubs, and (C) interhospital transfer to hubs. Among hospitals treating CS, 70.6% were classified as spokes and 29.4% as hubs. A total of 130 656 (31.7%) hospitalizations with CS were direct admission to spokes, 253 234 (61.4%) were direct admissions to hubs, and 28 777 (7.0%) were transfer to hubs. CS mortality was 47.8% at spoke hospitals and was lower at hub hospitals, both for directly admitted (39.3%, P < 0.01) and transferred (33.4%, P < 0.01) patients. Hospitalizations at hubs had higher procedural frequency (including coronary artery bypass graft, right heart catheterization, mechanical circulatory support), greater length of stay, and greater costs. On multivariable analysis, direct admission to CS hubs [odds ratio (OR) 0.86, 95% confidence interval (CI) 0.84-0.89, P < 0.01] and transfer to hubs (OR 0.72, 95% CI 0.69-0.76, P < 0.01) were both associated with lower mortality. Conclusion While acknowledging the limited ability of the Nationwide Readmissions Database to classify CS severity on presentation, treatment of CS at transfer hubs was associated with significantly lower mortality within this large real-world sample.