Regional variation in patient selection and treatment for lower extremity vascular disease in the Vascular Quality Initiative.
Academic Article
Overview
abstract
OBJECTIVE: Prior studies on the cause and effect of surgical variation have been limited by utilization of administrative data. The Vascular Quality Initiative (VQI), a robust national clinical registry, provides anatomic and perioperative details allowing a more robust analysis of variation in surgical practice. METHODS: The VQI was used to identify all patients undergoing infrainguinal open bypass or endovascular intervention from 2009 to 2014. Asymptomatic patients were excluded. The 16 regional groups of the VQI were used to compare variation in patient selection, operative indication, technical approach, and process measures. χ2 analysis was used to assess for differences across regions where appropriate. RESULTS: A total of 52,373 interventions were included (31%). Of the 16,145 bypasses, 5% were performed for asymptomatic disease, 26% for claudication, 56% for chronic limb-threatening ischemia (CLI) (61% of these for tissue loss), and 13% for acute limb-threatening ischemia. Of the 35,338 endovascular procedures, 4% were for asymptomatic disease, 40% for claudication, 46% for CLI (73% tissue loss), and 12% for acute limb-threatening ischemia. Potentially unwarranted variation included proportion of prosthetic conduit for infrapopliteal bypass in claudication (13%-41%, median, 29%; P < .001), isolated tibial endovascular intervention for claudication (0.0%-5.0%, median, 3.0%; P < .001), discharge on antiplatelet and statin (bypass: 62%-84%; P < .001; endovascular: 63%-89%; P < .001), and ultrasound guidance for percutaneous access (claudication: range, 7%-60%; P < .001; CLI: 5%-65%; P < .001). Notable areas needing further research with significant variation include proportion of CLI vs claudication treated by bypass (38%-71%; P < .001) and endovascular intervention (28%-63%; P < .001), and use of closure devices in percutaneous access (claudication; 26%-76%; P < .001; CLI: 30%-78%; P < .001). CONCLUSIONS: Significant variation exists both in areas where evidence exists for best practice and, therefore, potentially unwarranted variation, and in areas of clinical ambiguity. Quality improvement efforts should be focused on reducing unwarranted variation. Further research should be directed at identifying best practice where no established guidelines and high variation exists.
