Prolonged Length of Stay as Quality Measure in Asymptomatic and Symptomatic Patients Undergoing Carotid Endarterectomy
Emma S. Zwanenburg, BS1, Alexander B. Pothof, MD, MS2, Thomas F. O'Donnell, MD2, Sarah E. Deery, MD, MPH2, Gert J. de Borst, MD3, Marc L. Schermerhorn, MD2
1UVA, Amsterdam, Netherlands, 2BIDMC, Boston, MA, USA, 3UMC Utrecht, Utrecht, Netherlands
OBJECTIVE: CMS recently introduced Merit-based Incentive Payment System (MIPS) scores to compare physicians, which includes their performance on ‘prolonged length of stay (LOS)’, i.e. LOS > 2 days, plus any failed discharge home in asymptomatic carotid endarterectomy (CEA) patients. A low MIPS score can significantly reduce a physician’s Medicare reimbursement. However, the clinical implications of prolonged LOS are largely unknown.
METHODS: We identified CEAs from the Targeted-Vascular NSQIP (2011-2015) and stratified patients based on their preprocedural neurological symptom. We compared 30-day outcomes between patients with and without prolonged LOS, and evaluated prolonged LOS’s ability to capture complications among asymptomatic and symptomatic patients, with 30-day ipsilateral stroke as our primary endpoint. We additionally assessed predictors of prolonged LOS in complication-free patients.
RESULTS: Of 16,739 CEA patients, 9,784 (58%) were asymptomatic. Of the 6,955 symptomatic patients; 1,216 presented with an ocular TIA; 2,635 with a hemispheric TIA; and 3,104 with a stroke. Among asymptomatic patients, 12.6% experienced prolonged LOS, compared to 13.5%; 19.4%; and 33% among patients with a preprocedural ocular TIA, hemispheric TIA, and stroke, respectively. Patients with prolonged LOS experienced higher rates of stroke than those without prolonged LOS (asymptomatic: 6.7% vs. 0.5%; ocular TIA: 4.8% vs. 0.9%; hemispheric TIA: 8.4% vs. 1.1%; stroke: 8.9% vs. 1.7%, all P < .01). Of all strokes among asymptomatic patients, 66% occurred in patients with prolonged LOS, compared to 47%; 64%; 72% among patients with a procedural ocular TIA, hemispheric TIA, and stroke, respectively (Table I). But, CNIs and TIAs were poorly captured by prolonged LOS. However importantly, among patients with no recorded complication in VQI, 14% did experience prolonged LOS and these patients experienced higher 30-day mortality rates than patients without prolonged LOS, with ambulatory status, congestive heart failure, and general anesthesia as the main predictors of prolonged LOS in patients without a recorded complication (Table II).
CONCLUSIONS: Prolonged LOS identifies a patient population with high rates of perioperative complications. However, a large proportion of complications-free patients experience prolonged LOS. The identified predictors of prolonged LOS among complication-free patients could guide quality improvement efforts to eliminate prolonged LOS among complication free patients. 
Back to 2018 Program
