Neurointerventional procedural volume per hospital in United States: implications for comprehensive stroke center designation.

BACKGROUND AND PURPOSE Availability of neurointerventional procedures is recommended as a necessary component of a comprehensive stroke center by various regulatory guidelines that also emphasize adequate procedural volumes. We studied the volumes of neurointerventional procedures performed in various hospitals across the United States with subsequent comparisons with rates of minimum procedural volumes recommended by various professional bodies or used in clinical trials to ensure adequate operator experience. METHODS We reviewed the Nationwide Inpatient Sample database in the United States for the years 2005 to 2008. Using International Classification of Disease-Clinical Modification, 9th revision, and Medicare severity diagnosis-related group codes, we identified among hospitals that admit stroke patients those that met the minimum criteria for overall and individual procedural volumes specified in various guidelines. We then compared the characteristics between the high-volume hospitals that performed at least 100 cervicocerebral angiograms and met ≥1 other procedural criterion (n=79) and low-volume hospitals (n=958). RESULTS Proportions of hospitals that met individual procedural volume criteria over the 4-year period according to procedure were: cervicocerebral angiography (7.0%-7.8%); endovascular acute ischemic stroke treatments (0.4%-2.6%); carotid angioplasty/stent placement (3.0%-5.3%); intracranial angioplasty/stent placement (0.3%-1.3%); and aneurysm embolization (1.3%-2.6%). There were significant trends for increasing numbers of all the endovascular procedures except intracranial angioplasty/stent placement over the course of 4 years. The high-volume hospitals were more likely to be urban teaching hospitals (70.9% versus 13.1%; P<0.001), had larger bed size (79.7% versus 26.9%; P<0.001), and had significantly higher rates of hemorrhagic stroke admissions and lower rates of transient ischemic attack admissions. Urban teaching location/status (OR, 8.92; CI, 4.3-18.2; P<0.001) and large bed size (OR, 4.40; CI, 2.0-9.5; P<0.001) remained as independent predictors of a high-volume hospital when adjusted for age, gender, risk factors, and stroke subtype. CONCLUSIONS There are very few hospitals in the United States that meet all the neurointerventional procedural volume criteria for all endovascular procedures recommended to ensure adequate operator experience. Our results support the creation of specialized regional centers for ensuring adequate procedural volume within treating hospitals.