Variations in the utilization of endovascular aneurysm repair reflect population risk factors and disease prevalence

Abstract

Objective: Highly variable utilization rates for a diverse group of surgical procedures are commonly ascribed to physician practice patterns rather than clinical considerations. A previous investigation by our group showed that variations in the rates of carotid endarterectomy (CEA) actually reflected regional risk factors for atherosclerosis, not physician density or other socio-economic drivers. In this study, we examine the use of endovascular abdominal aortic aneurysm repair (EVAR) over six years to test our hypothesis that the utilization of innovative vascular procedures by vascular surgeons more closely reflects disease prevalence and consistent clinical judgment than non-medical factors.

Methods: The Nationwide Inpatient Samples and State Inpatient Databases (2001-2006) were accessed to document the number and type of aneurysm repairs (EVAR versus open). Multiple metrics pertaining to clinical risk factors, socioeconomic status, access to care, provider distribution, and local healthcare capacity were quantitated for each state. We performed bivariate analysis, Pearson (PC) or Spearman (SC) correlations, and multiple regression modeling.

Results: The total number of aneurysms repaired has not changed significantly (from 45,828 in 2001 to 45,111 in 2006). Over the same interval, the number of open AAA repair nationwide decreased by 48% while the number of AAA repaired endovascularly increased by 105%. In 2005, the utilization rate of EVAR among 29 states ranged widely from 39.3% to 69.9%. Use of EVAR was highest in states with higher incidences of aneurysms (PC = 0.43, P < .05), greater number of deaths from heart disease (PC = 0.42, P < .05), greater number of diabetes discharges (PC = 0.48, P < .005), higher number of carotid stenosis discharges (PC = 0.40, P < .05), and higher number of chronic obstructive pulmonary disorder (COPD) discharges (SC = 0.43, P < .05). Regional malpractice pressure, specifically the number of paid claims and mean malpractice premium, both exhibited positive correlations with the EVAR rate. The number of physicians, vascular surgeons, hospital beds, teaching hospitals, or trauma centers did not predict high utilization of EVAR nor did the other socio-economic indices tested.

Conclusion: While there was substantial regional variation in the use of EVAR, utilization of the less morbid procedure was well correlated with higher risk populations (number of diabetic patients and deaths secondary to heart disease). Contrary to other studies of regional discrepancies in the utilization of some surgical procedures, it appears that the utilization of EVAR was not associated with physician distribution, socioeconomics, or other non-medical factors.