Impact of provider characteristics on use of endovenous ablation procedures in Medicare beneficiaries.

OBJECTIVE: The objective of this study was to assess the association between provider characteristics and intensity of endovenous therapy (EVT) utilization in the Medicare population. METHODS: The Medicare Provider Utilization and Payment Data Public Use Files (2012-2014) were queried to construct a database of providers performing EVT using laser or radiofrequency ablation techniques for treatment of lower extremity venous reflux. A utilization index (UI; EVT procedure per patient treated per year) was calculated for each provider, and median services per county were determined. Provider specialty, geographic region, and site of service (facility vs outpatient) were determined for each patient. Multivariate regression analysis was used to identify provider characteristics associated with a UI above the 75th percentile. RESULTS: There were 6599 providers who performed EVT in 405,232 Medicare beneficiaries during the study period. Intensity of EVT use by providers was assessed by the calculated UI, the average number of EVT procedures performed in treated patients per year (range, 1-4). Vascular surgeons had the lowest UI among all provider specialties (1.32). By multivariate analysis, the likelihood of a provider's UI being >1.8 (top 25%) was associated with provider training in a field other than surgery, cardiology, or radiology (odds ratio [OR], 3.35; 2.74-4.09); services performed in an outpatient setting (OR, 2.62; 1.97-3.47); and providers who perform high annual volume of EVT (OR, 8.68; 7.59-9.91). A high annual volume provider was defined as one whose EVT volume was ≥75th percentile nationally. CONCLUSIONS: There is great variation in intensity of vein ablation procedures performed on Medicare beneficiaries by geographic location and provider specialty. High-volume providers and those with a specialty not traditionally associated with the management of lower extremity chronic venous disease are more likely to perform more EVT procedures per patient.

Characterizing the evolution of perioperative outcomes and costs of endovascular abdominal aortic aneurysm repair.

OBJECTIVE: The purpose of this study is to characterize the evolution in perioperative outcomes and costs of endovascular abdominal aortic aneurysm (AAA) repair (EVAR) by detailing changes in adjusted outcomes and costs over time. METHODS: National Inpatient Sample (2000-2011) data were used to evaluate patient characteristics, outcomes, and perioperative costs for elective EVAR performed for intact AAA. Outcomes were adjusted for patient demographics and comorbidities, and hospital factors by multivariate analysis. Costs were calculated from hospital cost to charge ratio files and adjusted to 2011 dollars. RESULTS: From 2000 to 2011, 185,249 patients underwent elective EVAR for intact AAA. The absolute rates of in-hospital major morbidity, mortality, and procedural costs all decreased significantly over time (P < .0001). The prevalence of major comorbidities in patients undergoing EVAR, including obesity, diabetes, and dyslipidemia, all increased significantly over time. After adjusting for multiple demographics, comorbidities, and hospital-level factors, recent outcomes of EVAR (2009-2011) remain superior to the early experience (2000-2002) with respect to mortality and major complications. CONCLUSIONS: From 2000-2011, the perioperative outcomes of EVAR improved significantly despite a higher prevalence of comorbidities among patients undergoing repair. Concurrently, procedure-associated costs declined. Advanced technology is often implicated in escalating healthcare spending, and the value of novel techniques is often questioned. These findings highlight that, in the case of EVAR, procedural outcomes have improved while the initial costs of repair have declined over time. EVAR offers an interesting example for stakeholders to consider in the era of cost-containment pressures and criticism of nascent, expensive technology in healthcare.

Impact of "Do Not Resuscitate" Status on the Outcome of Major Vascular Surgical Procedures.

BACKGROUND: Patients with Do Not Resuscitate (DNR) orders may still be offered surgery that aims to prolong or improve quality of life. The widely accepted approach of "required reconsideration" mandates that patients and surgeons discuss perioperative risks and expected outcomes in the context of the patient's values and preferences. However, surgical outcomes in this patient population have not been well-defined. The objectives of this study are to assess outcomes in DNR patients undergoing major vascular procedures, and develop an evidence basis for informed, shared decision-making. METHODS: Patients undergoing common major vascular procedures were identified in the 2007-2010 National Surgical Quality Improvement Project databases. DNR patients were defined as those with an active DNR order within 30 days before surgery. Demographics, comorbidities, procedural details, and complications were compared with those without DNR orders. To isolate the impact of DNR status, multivariate regression and 1:1 propensity score matching were used to compare outcomes between DNR patients and a non-DNR cohort of comparably high-risk patients. RESULTS: Of 110,279 patients undergoing major vascular surgery, 1,565 (1.4%) had active DNR orders 30 days preceding surgery. DNR patients were more likely to be functionally dependent (69% vs. 15%; P < 0.0001), over 80 years of age (53% vs. 20%; P < 0.001), and suffer from a variety of cardiac, pulmonary, and systemic comorbidities. The most common procedures in DNR patients were major amputation (38.4%), lower extremity bypass (20%), and peripheral thromboembolectomy (11.7%). Unadjusted 30-day mortality was significantly higher among DNR patients (21% vs. 3.4%; P < 0.001). After 1:1 propensity score matching, with the 2 cohorts differing only with respect to DNR status, perioperative mortality remained significantly higher among DNR patients (21% vs. 13%; P < 0.01). There was a trend toward reduced cardiopulmonary resuscitation in patients with recent DNR (1.7% vs. 2.6%; P = 0.07). CONCLUSIONS: DNR patients are at high risk for major complications and mortality after vascular surgery procedures. Compared with a matched cohort of "high-risk" non-DNR patients, those with DNR orders suffered equivalent rates of postoperative morbidity, but markedly increased mortality. This suggests that DNR status, independent of comorbidities and perioperative complications, may increase the risk of "failure to rescue." These findings have implications not only for risk adjustment, but also provide an evidence basis for shared decision-making in challenging circumstances.

Variability in anesthetic considerations for arteriovenous fistula creation.

INTRODUCTION: Anesthetic options for arteriovenous fistula (AVF) creation include regional anesthesia (RA), general anesthesia (GA) and local anesthetic for select cases. In addition to the benefits of avoiding GA in high-risk patients, recent studies suggest that RA may increase perioperative venous dilation and improve maturation. Our objective was to assess perioperative outcomes of AVF creation with respect to anesthetic modality and identify patient-level factors associated with variation in contemporary anesthetic selection. METHODS: National Surgical Quality Improvement Project (NSQIP) data (2007-2010) were accessed to identify patients undergoing AVF creation. Univariate analysis and multivariate logistic regression were performed to assess the relationships among patient characteristics, anesthesia modality and outcome. RESULTS: Of 1,540 patients undergoing new upper extremity AVF creation, 52% were male and 81% were younger than 75 years. Anesthesia distribution was GA in 85.2%, local/monitored anesthetic care (MAC) in 2.9% and RA in 11.9% of cases. By multivariate analysis, independent predictors of RA were dyspnea at rest (hazard ratio [HR] 2.3, 95% confidence interval [CI] 1.1-4.9), age >75 (HR 1.6, 95% CI 1.1-2.3) and teaching hospital status as indicated by housestaff involvement (HR 3.7, 95% CI 2.5-5.5). RA was associated with higher total operative time, duration of anesthesia, length of time in operating room and duration of anesthesia start until surgery start (p<0.01). There were no differences between perioperative complications or mortality among anesthetic modalities, although all deaths occurred in the GA group. DISCUSSIONS: Despite recent reports highlighting potential benefits of RA for AVF creation, GA was surprisingly used in the vast majority of cases in the United States. The only comorbidities associated with preferential RA use were advanced age and dyspnea at rest. Practice environment may influence anesthetic selection for these cases, as a nonteaching environment was associated with GA use. The trend seen here toward higher mortality in GA and the potential perioperative benefits of RA for the access should encourage more widespread use of RA in practice for this high-risk patient population.

The Comprehensive Risk Assessment for Bypass (CRAB) facilitates efficient perioperative risk assessment for patients with critical limb ischemia.

OBJECTIVE: Specific perioperative risk assessment models have been developed for bariatric, pancreatic, and colorectal surgery. A similar instrument, specific for patients with critical limb ischemia (CLI), could improve patient-centered clinical decision making. We describe a novel tool to predict 30-day major morbidity and mortality (M&M) after bypass surgery for CLI. METHODS: Data for 4985 individuals from the 2007 to 2009 National Surgical Quality Improvement Program were used to develop and internally validate the model. Outcome measures included mortality, major morbidity, and a composite end point (M&M). M&M included mortality and the most severe postoperative morbidities that were highly associated with death (eg, sepsis and major cardiopulmonary complications). More than 30 preoperative factors were tested for association with 30-day mortality, major morbidity, and M&M. Significant predictors in multivariate models were assigned integer values (points), which were added to calculate a patient's Comprehensive Risk Assessment For Bypass (CRAB) score. Performance was assessed (C-index) across all outcome measures and compared with other general tools (American Society of Anesthesiologists class, Surgical Risk Scale) and existing CLI-specific survival prediction models (Finnvasc score, Edifoligide for the Prevention of Infrainguinal Vein Graft Failure [PREVENT III] score) on a distinct validation sample (n = 1620). RESULTS: In the derivation data set (n = 3275), the 30-day mortality rate was 2.9%. The rate of any major morbidity was 19.1%. The composite end point M&M occurred in 10.1%. Significant predictors of M&M by multivariate analysis included age >75 years, prior amputation or revascularization, tissue loss, dialysis dependence, severe cardiac disease, emergency operation, and functional dependence. Applied to a distinct validation sample of 1620 patients, higher CRAB scores were significantly associated with higher rates of mortality, all major morbidities, and M&M (P < .0001). Comparison with other models by assessment of area under the receiver-operating characteristic curve revealed the CRAB was a more accurate predictor of mortality, all major morbidity, and M&M. CONCLUSIONS: The CRAB is a CLI-specific, risk assessment instrument derived from multi-institutional American College of Surgeons-National Surgical Quality Improvement Program surgical outcomes data that out-performs existing prognostic risk indices in the prediction of clinically significant adverse events after bypass surgery. Use of the CRAB as a risk assessment tool provides an evidence basis for patient-centered clinical decision making and may have a role in identifying patients at higher risk for surgical revascularization in whom an endovascular approach is preferable.

Clinical, demographic, and medicolegal factors associated with geographic variation in inferior vena cava filter utilization: an interstate analysis.

OBJECTIVES: Geographic variability exists in the use of IVC filters (IVCF). We hypothesized that variation in IVCF use is incompletely explained by variation in the prevalence of deep-vein thrombosis (DVT) and pulmonary embolism (PE) and may result from different practice patterns regarding prophylactic IVCF use. We characterize geographic variation in IVCF use at the state level and evaluate its association with clinical factors, patient demographics, and the medicolegal environment. METHODS: Healthcare Cost and Utilization Project State Inpatient Database records were accessed to identify 230,445 IVCFs placed from 2006 to 2008 in 33 states. Similar queries were performed for DVT and PE. Additional state data were obtained from public sources. Analyses included descriptive statistics, Spearman Correlation (SC), Wilcoxon rank-sum test, and characterization of variability. RESULTS: Overall, IVCF use correlated with the prevalence of DVT (SC = 0.89, P < .01). States on the East coast have significantly greater rates of IVCF use per 100K (mean ± SD = 41.2 ± 16.7 vs 27.8 ± 11.1, P < .05) and greater rates of IVCF per DVT (20.2 ± 4.5% vs 15.2 ± 2.9%; P < .005), despite similar rates of DVT per 100K (198.1 ± 51.2 vs 177.7 ± 46.7, P = NS) compared with all other states. Overall, states with the greatest rate of IVCF per DVT were (in descending order): Rhode Island, New Jersey, Florida, New York, and West Virginia. Rates of detected PE per 100K in these states were not significantly different from all other states (95.6 ± 16.6 vs 90.4 ± 16.1, P = NS). In these states, a greater percentage of IVCF recipients were older than 85 (15.3% vs 11.8%; P < .01); fewer were pediatric (0.3% vs 0.7%; P < .05) or aged 45 to 64 (26.1% vs 32.4%; P < .001). There were no differences in patient sex, race, insurance type, hospital size, or teaching status. States with high rates of IVCF per DVT were noted to have significantly greater rates of paid malpractice claims per 100K (4.9 ± 2.51 vs 1.1 ± 0.8; P = .001), and annual general surgeon liability insurance premiums ($78,630 ± 34,822 vs $43,989 ± 17,794; P < .05). CONCLUSION: Variation in IVCF use is incompletely explained by clinical factors. High rates of IVCF per DVT in some states may represent increased use of prophylactic IVCF in states with litigious medicolegal environments.