- Email: [email protected]
Readmissions after thoracic endovascular aortic repair
From the Society for Clinical Vascular Surgery
Readmissions after thoracic endovascular aortic repair Bindu Kalesan, PhD, MPH,a Thomas W. Cheng, MS,b Alik Farber, MD,b Yi Zuo, MPH,a Jeffrey A. Kalish, MD,b Douglas W. Jones, MD, MPH,b and Jeffrey J. Siracuse, MD,b Boston, Mass
ABSTRACT Objective: The care of patients undergoing thoracic endovascular aortic repair (TEVAR) can be resource intensive, which can be driven by readmissions. Our objective was to characterize index readmissions at 30, 90, and 180 days after TEVAR. Methods: A retrospective analysis of the Nationwide Readmissions Database was performed for patients who underwent TEVAR in 2013. Multivariable analysis identified independent predictors for index readmission at 30, 90, and 180 days. Results: There were 4045 TEVARs performed for descending thoracic aortic dissection (37.7%), nonruptured aneurysm (56%), and ruptured aneurysm (6.3%). There were 419 (11.1%) index readmissions at 30 days, 895 (23.6%) at 90 days, and 1131 (29.8%) at 180 days. The most frequent reason for index readmission was heart related at 30 days (15.5%) and aorta related at 90 days (18%) and 180 days (19.6%). Reinterventions were performed at 6.4%, 9.5%, and 9.7% of 30-, 90-, and 180-day readmissions, respectively. The majority of these included additional endovascular stent graft placement (51.9% of reinterventions at 30 days, 67.7% at 90 days, and 65.9% at 180 days). In multivariable analysis, 30-day index readmission was associated with initial ruptured presentation (odds ratio [OR], 1.48; 95% confidence interval [CI], 1.1-3.5; P ¼ .023) and diagnosis-related group (DRG) severity grades of moderate (OR, 1.42; 95% CI, 0.74-2.73), major (OR, 2.47; 95% CI, 1.28-4.74), and extreme (OR, 1.60; 95% CI, 0.76-3.36; P ¼ .009). Index readmission at 90 days was independently associated with initial ruptured presentation (OR, 1.88; 95% CI, 1.18-3.01; P ¼ .008), urgent/emergent TEVAR (OR, 1.41; 95% CI, 1.08-1.85; P ¼ .014), and DRG severity grades of moderate (OR, 1.53; 95% CI, 0.95-2.47), major (OR, 2.27; 95% CI, 1.39-3.7), and extreme (OR, 2.45; 95% CI, 1.43-4.18; P ¼ .002). Finally, at 180 days, initial ruptured presentation (OR, 1.66; 95% CI, 1.05-2.62; P ¼ .029), urgent/ emergent TEVAR (OR, 1.37; 95% CI, 1.08-1.79; P ¼ .013), and DRG severity grades of moderate (OR, 1.55; 95% CI, 1.01-2.38), major (OR, 2.15; 95% CI, 1.38-3.33), and extreme (OR, 2.39; 95% CI, 1.47-3.89; P ¼ .002) were, again, independently associated with index readmission. Conclusions: A large portion of patients treated with TEVAR were readmitted most commonly for heart-related reasons at 30 days and aorta-related reasons at 90 and 180 days. TEVAR performed to treat initial aortic rupture and greater DRG severity grade were independently associated with an index readmission at 30, 90, and 180 days. Urgent/emergent TEVAR was independently associated with an index readmission at 90 and 180 days. These factors are important to consider in using readmissions as a quality measure. (J Vasc Surg 2018;-:1-11.)
As our health care delivery system transitions from feefor-service to value-based care, hospital readmission, a surrogate for both quality of care and resource utilization, has become an ever important metric.1,2 A number of studies have demonstrated that patients undergoing From the Center for Clinical Translational Epidemiology and Comparative Effectiveness Research, Preventive Medicine and Epidemiology, Department of Medicine, Boston University School of Medicinea; and the Division of Vascular and Endovascular Surgery, Boston University, Boston Medical Center.b Author conflict of interest: B.K. and Y.Z. are funded by the Evans Research Foundation. Presented as a plenary oral presentation at the Forty-fifth Annual Symposium of the Society for Clinical Vascular Surgery, Lake Buena Vista, Fla, March 18-22, 2017. Additional material for this article may be found online at www.jvascsurg.org. Correspondence: Jeffrey J. Siracuse, MD, Department of Surgery, Boston University School of Medicine, 88 E Newton St, Ste C520, Boston, MA 02118 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2018 by the Society for Vascular Surgery. Published by Elsevier Inc. https://doi.org/10.1016/j.jvs.2017.12.035
vascular surgery procedures have 30-day readmission rates as high as 28.5%.2-6 In particular, patients with Medicare undergoing vascular surgery have a 30-day readmission rate of 23.9%, which is highest among other surgical specialties.2 The Centers for Medicare & Medicaid Services introduced the Hospital Readmissions Reduction Program in 2012 to reduce readmissions by penalizing centers with higher than expected readmission rates by withholding up to 1% of Medicare reimbursement. Multiple independent predictors of readmission have been identified across common vascular surgery procedures, including lower extremity bypass, carotid endarterectomy, open abdominal aortic aneurysm repair, and endovascular abdominal aortic aneurysm repair.7-10 However, there are limited data on readmissions for patients undergoing thoracic endovascular aortic repair (TEVAR). As TEVAR becomes a standard in treating patients with descending thoracic aortic disease, it is paramount to characterize predictors for readmissions after TEVAR in this cohort of high-risk patients.11-16 Therefore, our objectives were to describe readmission and reintervention rates after TEVAR both to evaluate reasons 1
2
Journal of Vascular Surgery
Kalesan et al
---
for index readmissions and to identify factors that are independently associated with index readmission at 30, 90, and 180 days.
ARTICLE HIGHLIGHTS d
METHODS Database. The Nationwide Readmissions Database (NRD) is an inpatient database designed by the Agency for Healthcare Research and Quality (AHRQ) as part of the Healthcare Cost and Utilization Project to identify areas for improving patient care, specifically hospital readmissions.17 The NRD includes all hospitalizations including readmissions within a state except for those at rehabilitation and long-term acute care hospitals.17 This study used the 2013 NRD, which consists of 21 states and captures 49.3% of the U.S. resident population and 49.1% of all hospitalizations within the United States in 2013.17 Selection of patients. Patients were first identified if they were older than 18 years and if they had a primary or secondary diagnosis for dissection of the thoracic aorta, dissection of the thoracoabdominal aorta, ruptured thoracic aneurysm, thoracic aneurysm without rupture, and any aortic aneurysm without rupture using the respective International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes: 441.01, 441.03, 441.1, 441.2, and 441.9. Patients with a primary or secondary procedure code for TEVAR (ICD-9: 39.73) were then included. All available diagnosis and procedure codes in the NRD database were evaluated. Patients who underwent open surgical repair were excluded with the following primary or secondary ICD-9 procedure code: resection of aorta or thoracic vessel with anastomosis (38.34, 38.35) or with replacement (38.44, 38.45) and repair of a vessel with a patch (39.57, 39.58). Patients who underwent proximal debranching or iliac access were identified using the respective primary or secondary ICD-9 procedure codes: 39.22 and 39.29. Patients were then classified into three indications at initial presentation: dissection, nonruptured aneurysm, and ruptured aneurysm. The indication was determined by the primary diagnosis code or by the secondary diagnosis code that was closest to the primary diagnosis code by numerical order. Sample groups and data collection. The resulting cohort of patients was categorized into two groups: those with an index readmission and those not readmitted after the TEVAR procedure at 30, 90, and 180 days. Demographics, comorbidities, indication at initial presentation, Elixhauser score, and hospital characteristics were collected at each time point. The Elixhauser score is a single numeric comorbidity score used in primarily administrative databases to assess the patient’s disease burden.18,19 The Elixhauser score consists of 30 comorbidity measures, such as hypertension, congestive heart failure, chronic pulmonary disease, obesity, and renal
2018
d
d
Type of Research: Retrospective analysis of prospectively collected data of the Nationwide Readmissions Database Take Home Message: In this study of 4045 thoracic endovascular aortic repair (TEVAR) patients enrolled in the Nationwide Readmissions Database, patients were most frequently readmitted for cardiac reasons at 30 days and aorta-related reasons at 90 and 180 days. TEVAR for aneurysm rupture predicted readmission at 30, 90, and 180 days. Recommendation: The authors suggest considering readmissions for ruptured aneurysms and for cardiac and aorta-related reasons as quality measures of TEVAR.
failure, that may have an impact on outcomes, such as length of stay, hospital charges, and in-hospital mortality.18,19 The Elixhauser score can also be implemented to evaluate risk of readmission.20 Hospital characteristics collected included hospital bed size (large, medium, and small). A hospital bed size defined by the AHRQ depends on the region (Northeast, Midwest, Southern, and Western) and location and teaching status of the hospital (rural, urban, and nonteaching, and urban and teaching).17 An urban teaching hospital with >425 beds in the Northeast region, 375 beds in the Midwest region, 450 beds in the Southern region, and 325 beds in the Western region was considered a large bed size.17 Patients who were readmitted at 30 days were included in the later readmission time points. Also, patients who died during follow-up were not included in the analysis as demonstrated in previous studies that analyzed readmission for resource utilization.5,21 Outcome variables collected at 30, 90, and 180 days were survival, readmission reasons, and reinterventions. Readmission reasons after the index TEVAR were identified by the primary ICD-9-CM diagnosis code and categorized by similarity. These categories included cardiac, aortic, infectious, gastrointestinal, other, renal failure, pulmonary, wound complication, peripheral ischemia, altered mental status, hypertension, musculoskeletal, trauma, bleeding, venous thromboembolism, and stroke. The aortic category was subcategorized to those with a reintervention. Reinterventions for TEVAR and open repair were identified using the previously described primary or secondary ICD-9 procedure codes. The Boston University School of Medicine Institutional Review Board approved this study, and informed consent was waived. Statistical analysis. Weighted numbers and percentages were used for all statistical analysis as recommended by the AHRQ.17 As the 2013 NRD does not include
Journal of Vascular Surgery Volume
-,
Number
Kalesan et al
3
-
all states, this method of weighing allows national statistics to be evaluated on the basis of the participating states. Univariate analysis performed between index readmission and those not readmitted at 30-, 90-, and 180-day time points included c2 test and survey linear regression model as appropriate. Univariate analyses that included c2 test and survey linear regression model were performed as appropriate for those who survived and those who died after the TEVAR procedure at 30-, 90-, and 180-day time points. All patient and hospital covariates were considered for possible confounders and underwent a variance inflation factor test to evaluate for multicollinearity. Covariates with a variance inflation factor >7 were excluded from the multivariable model. Multivariable survey logistic regression was used to identify factors that independently associated with index readmission at 30, 90, and 180 days. The software used for statistical analysis was Stata Statistical Software, release 14 (StataCorp LP, College Station, Tex). Statistical significance was set at P < .05.
RESULTS There were 4045 patients who underwent TEVAR in 2013 from a total of 82,518 patients identified with a primary or secondary diagnosis of thoracic aortic dissection and aneurysm (Fig). Comorbidities of the overall patient population are listed in Table I. The two most prevalent comorbidities were hypertension (83.4%) and peripheral vascular disorders (42.9%). The mean Elixhauser comorbidity score was 4.29. Overall, TEVARs divided by indication at initial presentation were thoracic aneurysm without rupture (56%), ruptured thoracic aneurysm (6.3%), and descending dissection in the thoracic aorta (37.7%; Table II). The mean age was 67.7 6 0.4 years; 60.8% were male, and 51.5% of TEVARs were urgent/emergent. Patients who were younger than 65 years represented 35.4% of the study population, and 36.8% were not on Medicare. A large portion (89.7%) of patients presented to a hospital within the state in which they resided. Insurance was largely (84.6%) either private insurance or Medicare. Proximal debranching and iliac access were performed in 14.2% and 1.3% of cases, respectively. The most common hospitals performing TEVARs were hospitals with a large bed size (85.3%) and teaching hospitals in the metropolitan area (80.9%). At 30 days, there were 419 (10.4%) index readmissions (Table II). A major diagnosis-related group (DRG) severity grade (55.3% vs 39.8%; P < .001) and the Elixhauser score (4.7 vs 4.3; P < .003) were significantly higher in proportion for those readmitted compared with those not readmitted at 30 days. Indication for TEVAR and urgent/ emergent TEVARs were not statistically different between those readmitted and those not readmitted at 30 days. Overall, the length of stay was 9.5 6 0.3 days. The length of stay for those readmitted at 30 days and
those not readmitted at 30 days was 8.2 6 0.5 days and 9.6 6 0.4 days, respectively (P ¼ .009). The three most frequent reasons for the index readmission at 30 days were cardiac (15.5%), aorta related (14.8%), and nonaorta infections (12.3%; Table III). Nonaorta-related reasons accounted for 85.2% of index readmissions at 30 days. During an index readmission at 30 days, there were 27 (0.7% of index TEVARs) reinterventions (Table IV). Additional TEVARs were performed in 51.9% of cases and an open intervention in 37% of cases at 30 days. In multivariable analysis, DRG severity grades of moderate (odds ratio [OR], 1.42; 95% confidence interval [CI], 0.74-2.73), major (OR, 2.47; 95% CI, 1.28-4.74), and extreme (OR, 1.60; 95% CI, 0.76-3.36) were independently associated with index readmission at 30 days (P ¼. 009; Table V). Also, initial presentation with ruptured thoracic aortic aneurysm (OR, 1.48; 95% CI, 1.1-3.5; P ¼ .023) was independently associated with index readmission at 30 days. The patients’ demographic characteristics, such as female sex, location of residence, type of insurance, and income status, were not independent predictors for readmission at 30 days. Requirement for iliac access or proximal debranching and hospital urban teaching status were also not independent predictors for readmission at 30 days. A summary of covariates independently associated with an index readmission at 30, 90, and 180 days is described in Table VI. At 90 days, there were 895 (22.1%) index readmissions (Table VII). Patients readmitted at 90 days were significantly older (69.7 6 0.8 years vs 67.2 6 0.5 years; P ¼ .02). Age >60 years (79.8% vs 72.7%; P ¼ .008), patients residing in the same state as the hospital (94.6% vs 88.3%; P ¼ .001), a major DRG severity grade (49.1% vs 39.2%; P < .004), and a higher Elixhauser score (4.7 vs 4.2; P < .0003) were significantly higher in proportion for those readmitted compared with those not readmitted at 90 days. A smaller proportion of patients younger than 65 years were readmitted at 90 days (29.5% vs 37%; P ¼ .018), and a higher proportion of patients on Medicare insurance were readmitted at 90 days (69.9% vs 61.3%; P ¼ .01). Patients who originally presented with ruptured thoracic aortic aneurysms (9.3% vs 5.5%; P ¼ .025) and those with urgent/emergent TEVARs (58.4% vs 49.5%; P ¼ .004) were also significantly higher in proportion for those readmitted at 90 days. The three most frequent reasons for readmission at 90 days were aorta related (18%), heart related (15.3%), and nonaorta infections (12.6%; Table III). Nonaorta-related reasons accounted for 82% of index readmissions at 90 days. During an index readmission at 90 days, there were 96 (2.4% of index TEVARs) reinterventions (Table IV), with additional TEVAR performed in 67.7% of cases and open intervention in 29.2% of cases. In multivariable analysis, covariates that were independently associated with index readmission at 90 days were again the DRG severity grades of moderate (OR, 1.53; 95% CI, 0.95-2.47),
4
Journal of Vascular Surgery
Kalesan et al
---
2018
Total number of patients greater than 18 years hospitalized in 2013 = 13,172,268 (unwt); 29,936,410 (wt) Number of patients without a primary or secondary diagnosis for aortic dissection or aneurysm = 13,136,070 (unwt); 29,853,892 (wt) Number of patients with a primary or secondary diagnosis for aortic dissection or aneurysm = 36,198 (unwt); 82,518 (wt) Number of patients without a primary or secondary procedure code for TEVAR in 2013 = 34,323 (unwt); 78,169 (wt) Number of patients with a primary or secondary procedure code for TEVAR in 2013= 1,875 (unwt); 4,349 (wt) Number of primary or secondary procedure code for open repair in 2013 = 94 (unwt); 221 (wt) Number of All TEVARs in 2013= 1,781 (unwt); 4,128 (wt) Number of Index TEVARs in 2013 = 1,745 (unwt); 4,045 (wt) In-hospital mortality= 119 (unwt); 256 (wt) Number of surviving index TEVARs in 2013= 1,626 (unwt); 3,789 (wt) Total number of readmissions in 2013 = 520 (unwt); 1,219 (wt)
Fig. Flow chart for study population. TEVAR, Thoracic endovascular aortic repair; unwt, unweighted; wt, weighted.
major (OR, 2.27; 95% CI, 1.39-3.7), and extreme (OR, 2.45; 95% CI, 1.43-4.18; P ¼ .002). Patients with an initial presentation of ruptured thoracic aortic aneurysms (OR, 1.88; 95% CI, 1.18-3.01; P ¼ .008) and urgent/emergent TEVARs (OR, 1.41; 95% CI, 1.08-1.85; P ¼ .014) were again independently associated with index readmission at 90 days. The patients’ demographics, requirement for iliac access or proximal debranching, and hospital urban teaching status were again not independent predictors for readmission at 90 days. At 180 days, there were 1131 (28%) index readmissions (Table VIII). Patients readmitted at 180 days were again significantly older (69.1 6 0.7 years vs 67.2 6 0.5 years; P ¼ .019). Age >60 years (78.9% vs 72.5%; P ¼ .01), patients residing in the same state as the hospital (94.8% vs 87.1%; P < .0001), a major DRG severity grade (48.3% vs 38.7%; P ¼ .004), and a higher Elixhauser score (4.7 vs 4.2; P < .0001) were again significantly higher in proportion for those readmitted compared with those not readmitted at 180 days. Again, patients readmitted at 180 days were significantly in higher proportion for urgent/emergent TEVARs (57.2% vs 49.3%; P ¼ .005). At 180 days, the three most frequent reasons for readmission were aorta related (19.6%), heart related (13.8%), and nonaorta infections (12.6%; Table III). Nonaorta-related reasons accounted for 80.4% of index readmissions at
180 days. There were 129 (3.2% of index TEVARs) reinterventions during an index readmission at 180 days (Table IV). Additional TEVARs were performed in 65.9% of cases and an open intervention in 31.8% of cases. In multivariable analysis, the DRG severity grades of moderate (OR, 1.55; 95% CI, 1.01-2.38), major (OR, 2.15; 95% CI, 1.38-3.33), and extreme (OR, 2.39; 95% CI, 1.47-3.89; P ¼ .002) were again independently associated with index readmission at 180 days (Table V). Patients with an initial presentation of ruptured thoracic aortic aneurysm (OR, 1.66; 95% CI, 1.05-2.62; P ¼ .029) and urgent/emergent TEVARs (OR, 1.37; 95% CI, 1.08-1.79; P ¼ .013) were again independently associated with index readmission at 180 days. The patients’ demographics, requirement for iliac access or proximal debranching, and hospital urban teaching status were, again, not independent predictors for readmission at 180 days. The mortality rate at 30 days was 6.1%; at 90 days, 7.6%; and at 180 days, 8.4%. Univariate analyses for mortality at 30 days, 90 days, and 180 days are depicted in Supplementary Tables I,II, and III, respectively (online only).
DISCUSSION Our study demonstrated that 10.4% of patients undergoing TEVAR were readmitted at 30 days, 22.1% at 90 days, and 28% at 180 days. Although the most
Journal of Vascular Surgery Volume
-,
Number
Kalesan et al
5
-
Table I. Comorbidities of the patients Comorbidities Alcohol abuse
135 (3.3)
Chronic blood loss anemia
36 (.9)
Chronic pulmonary disease
1127 (27.9)
Coagulopathy
500 (12.4)
Congestive heart failure Deficiency anemias
55 (1.4) 781 (19.3)
Depression
405 (10)
Diabetes (uncomplicated)
605 (14.9)
Drug abuse
103 (2.6)
Fluid and electrolyte disorder
1294 (32)
Hypertension
3373 (83.4)
Hypothyroidism
387 (9.6)
Liver disease
64 (1.6)
Lymphoma
33 (.8)
Metastatic cancer
30 (.7)
Obesity
546 (13.5)
Other neurologic disorders
238 (5.9)
Paralysis
197 (4.9)
Peripheral vascular disorders Psychoses
1735 (42.9) 97 (2.4)
Pulmonary circulation disorder
17 (.4)
Solid tumor without metastasis
88 (2.2)
Renal failure
766 (18.9)
Rheumatoid arthritis
134 (3.3)
Valvular disease
46 (1.1)
Weight loss
277 (6.8)
Elixhauser comorbidity score
4.29 (1.89)
Values are weighted frequencies and weighted percentages using survey analysis. All data presented are number (%), except for the Elixhauser comorbidity score, which is mean (95% confidence interval [CI]).
frequent primary reason for index readmission was heart related at 30 days and aorta related at 90 and 180 days, the majority of reasons for readmission were not aorta or heart related. These included nonaorta infections, renal failure, and pulmonary conditions. The length of stay for those who were readmitted and those not readmitted at 30 days was significantly higher by 1.4 days for those not readmitted. Patients who originally presented with ruptured thoracic aortic aneurysm and those with a DRG severity grade of moderate, major, or extreme were independently associated with an index readmission at 30, 90, and 180 days. Urgent/emergent TEVAR was independently associated with an index readmission at 90 and 180 days. These findings confirmed that despite the successful treatment by TEVAR of patients with descending aortic dissection or aneurysm, improvement is needed to reduce readmissions, particularly for patients with ruptured thoracic aortic aneurysms. In addition, these time points at 30, 90, and 180 days are important as they demonstrate perioperative outcomes
after TEVAR, global hospital payment, and long-term outcomes, respectively. Our study’s readmission rate was similar to or lower than that in the literature. Jones et al15 reported a 30-day readmission rate of 21% for patients with type B aortic dissections treated with TEVAR between 2000 and 2010, and this rate was higher than our 30-day readmission rate for TEVARs performed in 2013. Another study analyzing the use of TEVAR from 2000 to 2007 for nonruptured thoracic aortic aneurysm reported a 30-day readmission rate of 22% during the period of 2000 to 2001 and 16.5% during 2006 to 2007.22 This trend toward lower readmission rates may speak to improved operator techniques and stent technology over time. However, these studies used Medicare data and investigated readmission rates only at 30 days, whereas our study’s database included all-payer inpatient visits and long-term readmission data at 90 and 180 days in addition to 30-day readmission data. The all-payer inpatient visits in the NRD include a large portion of patients (36.8%) who were not captured in the Medicare database as the NRD is hospital based, whereas the Medicare database is insurance based. Our study also provides follow-up data on TEVAR performed for patients younger than 65 years who may not be present in Medicare database studies because of the general Medicare age restriction. In addition, the NRD is different from the Vascular Quality Initiative and the National Surgical Quality Improvement Program databases; the Vascular Quality Initiative does not have any readmission data, and the National Surgical Quality Improvement Program database has only 30-day readmission data. Our findings provide improved generalizability to the entire population treated with TEVAR. In our study, the most frequent reason for readmission was heart related (15.5%) at 30 days and aorta related at 90 and 180 days (18% and 19.6%, respectively). Our 30day rate for heart-related readmission was comparable to that of another study in which 18.5% of 30-day readmissions after elective endovascular aneurysm repair for nonruptured abdominal aortic aneurysm were heart related.23 At 90 days, our aorta-related readmission finding was comparable to that of another study reporting that 18.5% of readmissions after TEVAR were aneurysm related.14 As aorta-related readmissions are the most frequent reason for readmission at 90 days, this may affect how hospital global payments are issued at 90 days. At 180 days, aorta-related readmission remained the most frequent reason for readmission, which suggests the extent of impact that performing TEVAR has on reimbursement for certain hospital systems, particularly accountable care organizations. However, the magnitude of impact on reimbursement is unclear, and it is outside the scope of our study. Also, it is difficult to assess as the health care system is changing drastically in the near future with the transition from a fee-for-service to
6
Journal of Vascular Surgery
Kalesan et al
---
2018
Table II. Baseline patient and hospital characteristics for 30-day index readmission Characteristics
Overall (N ¼ 4045)
Readmitted at 30 days (n ¼ 419)
Not readmitted at 30 days (n ¼ 3626)
P value
Patient Age, years, mean (SE)
67.7 (0.4)
.94
Age $60 years
3004 (74.3)
309 (73.8)
2695 (74.3)
.89
Age <65 years
1431 (35.4)
147 (35.1)
1283 (35.4)
.95
2459 (60.8)
231 (55)
2228 (61.5)
.15
Male sex
67.7 (0.4)
67.6 (1.4)
Location of patient’s residence Central metropolitan ($1 million)
.45 997 (24.7)
118 (28.1)
880 (24.3)
Fringe metropolitan ($1 million)
916 (22.7)
105 (25.1)
811 (22.4)
Metropolitan (250,000-999,999)
932 (23.1)
81 (19.2)
852 (23.6)
Micropolitan (<25,000)
1189 (29.5)
116 (27.6)
1074 (29.7)
3422 (84.6)
354 (84.9)
3068 (84.6)
621 (15.4)
63 (15.1)
558 (15.4)
2557 (63.2)
277 (66.1)
2280 (62.9)
Insurance Private/Medicare Medicaid/other
.91
Insurance Medicare Not Medicare
.41 1488 (36.8)
142 (33.9)
1346 (37.1)
Lowest annual income (<$38,000)
1045 (26.1)
169 (26.3)
876 (26.1)
.98
Patient resides in same state as hospital
3628 (89.7)
396 (94.4)
3232 (89.1)
.16
464 (11.5)
25 (5.9)
440 (12.1)
97 (23.1)
997 (27.5)
DRG severity Minor
.001
Moderate
1094 (27)
Major
1674 (41.4)
232 (55.3)
1442 (39.8)
814 (20.1)
66 (15.7)
748 (20.6)
4.3 (0)
4.7 (0.2)
4.3 (0.1)
2263 (56)
Extreme Elixhauser score, mean (SE)
.003
Indication at initial presentation Aneurysms without rupture Aneurysms with rupture Dissections
224 (53.4)
2039 (56.2)
255 (6.3)
40 (9.6)
214 (5.9)
.48
1527 (37.7)
155 (36.9)
1372 (37.8)
.81
576 (14.2)
57 (13.5)
519 (14.3)
.72
52 (1.3)
6 (1.3)
46 (1.3)
.94
150 (3.7)
22 (5.2)
128 (3.5)
.057
Concomitant procedures Proximal debranching Iliac access Hospital Bed size Small Medium Large
.50 50 (11.8)
394 (10.9)
3451 (85.3)
443 (11)
348 (82.9)
3104 (85.6)
722 (17.9)
87 (20.8)
635 (17.5)
Teaching status Metropolitan nonteaching Metropolitan teaching Not metropolitan Urgent/emergent procedure
.58 3274 (80.9)
327 (78.1)
48 (1.2)
5 (1.2)
44 (1.2)
226 (53.8)
1855 (51.2)
2080 (51.5)
2947 (81.3) .58
DRG, Diagnosis-related group; SE, standard error. All values are weighted frequency and percentages. The number is the total sample size for the respective cohort. Values are reported as number (%) unless otherwise indicated.
value-based health care. In addition, these aorta-related readmissions could be related to planned procedures as depicted by Glebova et al,9 who concluded that planned readmissions after vascular surgery were more frequent
compared with those after other surgical procedures. Unfortunately, given limitations of our database, our study cannot explain whether planned procedures represented a majority of readmissions.
Journal of Vascular Surgery Volume
-,
Number
Kalesan et al
7
-
Table III. Primary reasons for index readmission at 30 days, 90 days, and 180 days Category
30 Days, No. (%)
90 Days, No. (%)
180 Days, No. (%)
Heart related
65 (15.5)
137 (15.3)
156 (13.8)
Aorta related
222 (19.6)
62 (14.8)
161 (18)
Reinterventionsa
27 (6.4)
85 (9.5)
110 (9.7)
Nonaorta infections
52 (12.3)
113 (12.6)
142 (12.6)
Gastrointestinal
38 (9.1)
81 (9.1)
116 (10.3)
Other
41 (9.8)
108 (12.1)
138 (12.2)
Renal failure
30 (7.2)
40 (4.5)
47 (4.2)
Pulmonary
28 (6.7)
63 (7)
66 (5.8)
Wound complication
22 (5.3)
32 (3.6)
32 (2.8)
Peripheral ischemia
17 (4.1)
23 (2.6)
38 (3.3)
Altered mental status
13 (3.1)
21 (2.3)
23 (2)
Hypertension
11 (2.7)
14 (1.5)
19 (1.7)
10 (2.4)
22 (2.4)
28 (2.5)
9 (2.2)
22 (2.5)
36 (3.2)
Musculoskeletal Trauma Bleeding
9 (2.1)
12 (1.3)
14 (1.2)
Venous thromboembolism
8 (1.9)
24 (2.7)
24 (2.1)
Stroke
6 (1.4)
23 (2.6)
31 (2.7)
Frequencies for each reason for readmission are not mutually exclusive between the 30-, 90-, and 180-day time points. a Reinterventions is a subcategory that describes the number of reinterventions that occurred during an aorta-related readmission, and the percentage is in respect to the number of index readmissions.
Table IV. Reinterventions after thoracic endovascular aortic repair (TEVAR) Type of reintervention during index readmission
30 Days, No. (%)
90 Days, No. (%)
180 Days, No. (%)
Additional TEVAR
14 (51.9)
65 (67.7)
85 (65.9)
Open intervention
10 (37)
28 (29.2)
41 (31.8)
Additional TEVAR and open intervention Total
3 (11.1)
3 (3.1)
3 (2.3)
27 (100)
96 (100)
129 (100)
Frequencies for reinterventions were not mutually exclusive between the 30-, 90-, and 180-day time points. All values and percentages are weighted. The percentages are percentages of total number of reinterventions at each respective time point.
The rates for other primary reasons for readmission that constituted the majority of readmission in our study, particularly infection and renal failure, were echoed by multiple studies.5,6,22-24 In our study, 12.3% of 30-day readmissions were related to infections, which was similar to the reported range of 10.3% to 11.9% for infections associated with 30-day readmissions for common vascular surgery procedures, including lower extremity bypass, carotid endarterectomy, and abdominal aortic procedures.6,23 This similarity continued for patients readmitted for renal failure as our study demonstrated that 7% of 30-day readmissions were related to renal failure, and Greenblatt et al24 reported that 6% of Medicare patients were readmitted for renal failure at 30 days after an elective endovascular aneurysm repair for nonruptured abdominal aortic aneurysm. This indicates that patients undergoing TEVAR were readmitted for various primary reasons other than heart- or aorta-related reasons. In our study, the length of stay for those not readmitted at 30 days was significantly longer compared with those readmitted at 30 days. As current hospital practices are
to decrease length of stay, there is concern for higher readmission rate as patients may not be optimized for an early discharge. However, our study demonstrates that early discharge did not adversely affect readmission rate. In our multivariable analysis, urgent/emergent TEVAR was not independently associated with 30-day readmission. This was echoed in a study analyzing the association between urgent/emergent TEVARs for nonruptured thoracic aortic aneurysm and 30-day readmission in Medicare patients (OR, 0.95; 95% CI, 0.72-1.27; P ¼ .75).22 This suggests that although urgent/emergent TEVARs may be relatively safe during the perioperative phase, the urgency of TEVAR should be considered in using readmission for reimbursement past 30 days as our study demonstrated urgent/emergent TEVARs to be independently associated with an index readmission at 90 and 180 days. The limitations of this study include the inherent use of a retrospective database. A national database is subject to selection bias for performing TEVAR. Also, although the ICD-9 codes used in this study have been validated by another study, there are potential errors in coding
8
Journal of Vascular Surgery
Kalesan et al
---
2018
Table V. Multivariable analysis of predictors for index readmission 30 Days Covariate
90 Days
P value
OR (95% CI)
180 Days
P value
OR (95% CI)
OR (95% CI)
P value
Patient Female sex
1.16 (0.84-1.6)
.36
Location of patient’s residence Central metropolitan ($1 million)
.99
1.02 (0.67-1.56)
Metropolitan (250,000-999,999)
0.89 (0.57-1.38)
Micropolitan (<25,000)
0.84 (0.54-1.29)
Reference .98 (0.71-1.36)
.51 1.05 (0.77-1.42) 0.92 (0.67-1.25)
0.78 (0.56-1.09)
0.83 (0.61-1.13)
.85
Private/Medicare
.11
.3
Reference
Reference
Reference
0.96 (0.61-1.51)
0.75 (0.53-1.07)
.85 (0.62-1.16)
Low income quartile (<$38,000)
0.93 (0.64-1.36)
.72
DRG severity Minor
0.84 (0.63-1.12)
.25
.009
0.97 (0.75-1.26)
.002
Reference
Reference
Reference
1.42 (0.74-2.73)
1.53 (0.95-2.47)
1.55 (1.01-2.38)
Major
2.47 (1.28-4.74)
2.27 (1.39-3.7)
2.15 (1.38-3.33)
Extreme
1.60 (0.76-3.36)
2.45 (1.43-4.18)
2.39 (1.47-3.89)
1.12 (0.79-1.61)
.51
1.41 (1.08-1.85)
.014
Aneurysms without rupture
1.28 (0.87-1.88)
Aneurysms with rupture
1.48 (1.1-3.5)
.21
1.33 (1-1.78)
.023
1.88 (1.18-3.01)
Proximal debranching
1.06 (0.68-1.66)
Iliac access
1.48 (0.42-5.25)
.78
1.32 (0.68-1.66)
.55
1.65 (0.63-4.34)
.83 .002
Moderate
Urgent/emergent procedure
.88
Reference
0.94 (0.67-1.3)
Insurance
1.02 (0.81-1.28)
.48
Reference
Fringe metropolitan ($1 million)
Medicaid/other
1 (0.78-1.28)
.8
1.37 (1.08-1.79)
.013
.052
1.26 (0.87-1.88)
.091
.008
1.66 (1.05-2.62)
.029
.097
1.35 (1-1.84)
.052
.31
1.83 (0.72-4.63)
.2
Indication at initial presentation
Procedure
Hospital Teaching status Metropolitan nonteaching
.93
.63
.21
Reference
Reference
Reference
Metropolitan teaching
0.93 (0.63-1.37)
1.09 (0.80-1.48)
1.07 (0.81-1.41)
Not metropolitan
0.93 (0.2-4.44)
1.66 (0.54-5.1)
2.53 (0.91-7.04)
CI, Confidence interval; DRG, diagnosis-related group; OR, odds ratio. Boldface entries indicate statistical significance (P < .05).
Table VI. Summary of significant factors associated with index readmission Covariate
Readmitted at 30 days
Readmitted at 90 days
Readmitted at 180 days
þ
þ
þ
þþ
þþ
þþ
þ
þþ
þþ
Initial presentation of ruptured aneurysm
þ
þ
þ
Urgent/emergent procedure
0
þ
þ
DRG severity grade Moderate Major Extreme
DRG, Diagnosis-related group. A þ and þþ represent an odds ratio (OR) of 1 to 2 and 2 to 3, respectively. A 0 represents an OR that is not statistically significant.
and the limited anatomic detail associated with codes.15 One limitation is how the indications were categorized. As the indication was identified by numerical order for diagnosis for dissection, nonruptured aneurysm, and ruptured aneurysm, patients who presented with a dissection and then ruptured during the initial hospitalization could have been categorized as dissection.
Although this is a limitation, all diagnosis codes for each patient including any dissection, nonruptured aneurysms, and ruptured aneurysms were placed into the multivariable models for readmission at 30, 90, and 180 days. Our study is further limited by how the ICD-9 diagnosis codes for readmissions in the database are captured as they were on admission rather than at
Journal of Vascular Surgery Volume
-,
Number
Kalesan et al
9
-
Table VII. Baseline patient and hospital characteristics for 90-day index readmission Characteristics
Overall (N ¼ 4045)
Readmitted at 90 days (n ¼ 895)
67.7 (0.4)
69.4 (0.8)
Not readmitted at 90 days (n ¼ 3150)
P value
Patient Age, years, mean (SE)
67.2 (0.5)
.02
Age $60 years
3004 (74.3)
714 (79.8)
2290 (72.7)
.008
Age <65 years
1431 (35.4)
264 (29.5)
1166 (37)
.018
2459 (60.8)
523 (58.5)
1936 (61.4)
.36
997 (24.7)
239 (26.7)
758 (24.1)
Male sex Location of patient’s residence Central metropolitan ($1 million)
.64
Fringe metropolitan ($1 million)
916 (22.7)
213 (23.7)
704 (22.4)
Metropolitan (250,000-999,999)
932 (23.1)
196 (21.9)
736 (23.4)
Micropolitan (<25,000)
1189 (29.5)
247 (27.6)
943 (30)
3422 (84.6)
785 (87.9)
2637 (83.7)
621 (15.4)
108 (12.1)
513 (16.3)
2557 (63.2)
626 (69.9)
1931 (61.3)
Insurance Private/Medicare Medicaid/other
.062
Insurance Medicare Not Medicare
.01 1488 (36.8)
269 (30.1)
1219 (38.7)
Lowest annual income (<$38,000)
1045 (26.1)
209 (23.5)
837 (26.9)
Patient resides in same state as hospital
3628 (89.7)
846 (94.6)
2782 (88.3)
464 (11.5)
60 (6.7)
404 (12.8)
DRG severity Minor
.21 .001 .004
Moderate
1094 (27)
220 (24.6)
874 (27.7)
Major
1674 (41.4)
439 (49.1)
1235 (39.2)
814 (20.1)
176 (19.7)
638 (20.2)
4.3 (0)
4.7 (0.1)
4.2 (0.1)
2263 (56.0)
489 (54.6)
1774 (56.3)
255 (6.3)
83 (9.3)
172 (5.5)
1527 (37.7)
323 (36.1)
1204 (38.2)
.46
576 (14.2)
142 (15.8)
434 (13.8)
.37
52 (1.3)
13 (1.5)
38 (1.2)
.64
150 (3.7)
31 (3.5)
119 (3.8)
Extreme Elixhauser score, mean (SE)
.0003
Indication at initial presentation Aneurysms without rupture Aneurysms with rupture Dissections
.59 .025
Concomitant procedures Proximal debranching Iliac access Hospital Bed size Small Medium Large
.97 443 (11.0)
99 (11.1)
3451 (85.3)
765 (85.4)
2687 (85.3)
344 (10.9)
722 (17.9)
158 (17.7)
564 (17.9)
3274 (80.9)
724 (80.9)
48 (1.2)
13 (1.4)
36 (1.1)
2081 (51.5)
523 (58.4)
1557 (49.5)
Teaching status Metropolitan nonteaching Metropolitan teaching Not metropolitan Urgent/emergent procedure
.90 2551 (81) .004
DRG, Diagnosis-related group; SE, standard error. All values are weighted frequency and percentages. The number is the total sample size for the respective cohort. Values are reported as number (%) unless otherwise indicated.
discharge. Another limitation of the study is that the NRD captures only readmissions within the state and does not allow interstate data to be collected. However, because a large majority (89.7%) of patients were
readmitted at the same state, the impact of patients’ being readmitted at another state may be relatively low. Moreover, our study is limited by staged TEVAR procedures, which could inflate the reintervention and
10
Journal of Vascular Surgery
Kalesan et al
---
2018
Table VIII. Baseline patient and hospital characteristics for 180-day index readmission Characteristics
Overall (N ¼ 4045)
Readmitted at 180 days (n ¼ 1131)
Not readmitted at 180 days (n ¼ 2914)
P value
Patient Age, years, mean (SE)
67.7 (0.4)
69.1 (0.7)
67.2 (0.5)
.019
Age $60 years
3004 (74.3)
892 (78.9)
2112 (72.5)
.01
Age <65 years
1431 (35.4)
363 (32.1)
1068 (36.6)
2459 (60.8)
669 (59.2)
1790 (61.4)
997 (24.7)
286 (25.3)
711 (24.5)
Male sex Location of patient’s residence Central metropolitan ($1 million)
.74
Fringe metropolitan ($1 million)
916 (22.7)
271 (24)
645 (22.2)
Metropolitan (250,000-999,999)
932 (23.1)
241 (21.3)
692 (23.8)
Micropolitan (<25,000)
1189 (29.5)
333 (29.4)
856 (29.5)
3422 (84.6)
981 (86.9)
2441 (83.8)
621 (15.4)
147 (13.1)
474 (16.2)
2557 (63.2)
764 (67.6)
1793 (61.5)
Insurance Private/Medicare Medicaid/other
.13
Insurance Medicare Not Medicare
.068 1488 (36.8)
366 (32.4)
1122 (38.5)
Lowest annual income (<$38,000)
1045 (26.1)
287 (25.6)
759 (26.4)
Patient resides in same state as hospital
3628 (89.7)
1072 (94.8)
2556 (87.1)
464 (11.5)
79 (7)
385 (13.2)
DRG severity Minor
.45
.74 .0001 .004
Moderate
1094 (27)
287 (25.4)
807 (27.7)
Major
1674 (41.4)
546 (48.3)
1128 (38.7)
814 (20.1)
218 (19.3)
596 (20.4)
4.3 (0)
4.7 (0.1)
4.2 (0.1)
<.0001
2263 (56)
.69
Extreme Elixhauser score, mean (SE) Indication at initial presentation Aneurysms without rupture Aneurysms with rupture Dissections
623 (55.1)
1641 (56.3)
255 (6.3)
95 (8.4)
160 (5.5)
.06
1527 (37.7)
413 (36.5)
1114 (38.2)
.53
576 (14.2)
177 (15.6)
399 (13.7)
.3
52 (1.3)
18 (1.6)
34 (1.2)
150 (3.7)
38 (3.3)
113 (3.9)
Procedure Proximal debranching Iliac access
.44
Hospital Bed size Small Medium Large
.32 121 (10.7)
322 (11)
3451 (85.3)
443 (11)
971 (85.9)
2480 (85.1)
722 (17.9)
194 (17.1)
529 (18.1)
Teaching status Metropolitan nonteaching Metropolitan teaching Not metropolitan Urgent/emergent procedure
.16 3274 (80.9) 48 (1.2) 2080 (51.5)
915 (80.9) 22 (2) 647 (57.2)
2360 (81) 26 (.9) 1433 (49.3)
.005
DRG, Diagnosis-related group; SE, standard error. All values are weighted frequency and percentages. The number is the total sample size for the respective cohort. Values are reported as number (%) unless otherwise indicated.
readmission rates because we cannot delineate planned and unplanned readmissions. Furthermore, TEVAR device information is unavailable in this database, which limits this study’s ability to characterize whether a
correlation exists between device and rate of readmission or reintervention. However, given these limitations, our study demonstrates contemporary results for the follow-up period after TEVAR.
Journal of Vascular Surgery Volume
-,
Number
Kalesan et al
11
-
CONCLUSIONS Our study demonstrated that a large proportion of patients are readmitted at 30, 90, and 180 days, with the most frequent reason for index readmission heart related at 30 days and aorta related at 90 and 180 days. However, the majority of readmissions were for nonheart and nonaorta-related reasons. Patients with an initial presentation of ruptured thoracic aortic aneurysms and urgent/emergent TEVARs in particular are more likely to be readmitted, and these factors should be accounted for in analyzing quality of care measures such as readmission rate.
AUTHOR CONTRIBUTIONS Conception and design: BK, JS Analysis and interpretation: BK, TC, AF, YZ, JK, DJ, JS Data collection: BK, TC, YZ, JS Writing the article: TC, JS Critical revision of the article: BK, TC, AF, YZ, JK, DJ, JS Final approval of the article: BK, TC, AF, YZ, JK, DJ, JS Statistical analysis: BK, YZ Obtained funding: BK Overall responsibility: JS
REFERENCES 1. Schroeder SA, Frist W. Phasing out fee-for-service payment. N Engl J Med 2013;368:2029-32. 2. Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare fee-for-service program. N Engl J Med 2009;360:1418-28. 3. Gonzalez AA, Girotti ME, Shih T, Wakefield TW, Dimick JB. Reliability of hospital readmission rates in vascular surgery. J Vasc Surg 2014;59:1638-43. 4. Wiseman JT, Guzman AM, Fernandes-Taylor S, Engelbert TL, Saunders RS, Kent KC. General and vascular surgery readmissions: a systematic review. J Am Coll Surg 2014;219: 552-69.e2. 5. Siracuse JJ, Shah NK, Peacock MR, Tahhan G, Kalish JA, Rybin D, et al. Thirty-day and 90-day hospital readmission after outpatient upper extremity hemodialysis access creation. J Vasc Surg 2017;65:1376-82. 6. Tahhan G, Farber A, Shah NK, Krafcik BM, Sachs TE, Kalish JA, et al. Characterization of planned and unplanned 30-day readmissions following vascular surgical procedures. Vasc Endovascular Surg 2017;51:17-22. 7. Duwayri Y, Goss J, Knechtle W, Veeraswamy RK, Arya S, Rajani RR, et al. The readmission event after vascular surgery: causes and costs. Ann Vasc Surg 2016;36:7-12. 8. Engelbert TL, Fernandes-Taylor S, Gupta PK, Kent KC, Matsumura J. Clinical characteristics associated with readmission among patients undergoing vascular surgery. J Vasc Surg 2014;59:1349-55. 9. Glebova NO, Bronsert M, Hicks CW, Nehler MR, Henderson WG. Contributions of planned readmissions and patient comorbidities to high readmission rates in vascular surgery patients. J Vasc Surg 2014;63:746-55.e2.
10. Glebova NO, Bronsert M, Hammermeister KE, Nehler MR, Gibula DR, Malas MB, et al. Drivers of readmissions in vascular surgery patients. J Vasc Surg 2016;64:185-94.e3. 11. Fanelli F, Dake MD. Standard of practice for the endovascular treatment of thoracic aortic aneurysms and type B dissections. Cardiovasc Intervent Radiol 2009;32:849-60. 12. Patel HJ, Williams DM, Upchurch GR, Dasika NL, Deeb GM. A comparative analysis of open and endovascular repair for the ruptured descending thoracic aorta. J Vasc Surg 2009;50:1265-70. 13. Goodney PP, Travis L, Lucas FL, Fillinger MF, Goodman DC, Cronenwett JL, et al. Survival after open versus endovascular thoracic aortic aneurysm repair in an observational study of the medicare population. Circulation 2011;124:2661-9. 14. Karimi A, Walker KL, Martin TD, Hess PJ, Klodell CT, Feezor RJ, et al. Midterm cost and effectiveness of thoracic endovascular aortic repair versus open repair. Ann Thorac Surg 2012;93:473-9. 15. Jones DW, Goodney PP, Nolan BW, Brooke BS, Fillinger MF, Powell RJ, et al. National trends in utilization, mortality, and survival after repair of type B aortic dissection in the Medicare population. J Vasc Surg 2014;60:11-9.e1. 16. Zimmerman KP, Oderich G, Pochettino A, Hanson KT, Habermann EB, Bower TC, et al. Improving mortality trends for hospitalization of aortic dissection in the National Inpatient Sample. J Vasc Surg 2016;64:606-15.e1. 17. NRD overview. Healthcare Cost and Utilization Project (HCUP). Agency for Healthcare Research and Quality, Rockville, Md. Available at: https://www.hcup-us.ahrq.gov/ nrdoverview.jsp. Accessed August 29, 2017. 18. Elixhauser A, Steiner C, Harris DR, Coffey RM. comorbidity measures for use with administrative data. Med Care 1998;36:8-27. 19. van Walraven C, Austin PC, Jennings A, Quan H, Forster AJ. A modification of the Elixhauser comorbidity measures into a point system for hospital death using administrative data. Med Care 2009;47:626-33. 20. Moore BJ, White S, Washington R, Coenen N, Elixhauser A. Identifying increased risk of readmission and in-hospital mortality using hospital administrative data. Med Care 2017;55:698-705. 21. Krafcik BM, Komshian S, Lu K, Roberts L, Farber A, Kalish JA, et al. Short and long-term readmission rates after infrainguinal bypass in a safety net hospital are higher than expected. J Vasc Surg 2017;66:1786-91. 22. Brooke BS, Goodney PP, Powell RJ, Fillinger MF, Travis LL, Goodman DC, et al. Early discharge does not increase readmission or mortality after high-risk vascular surgery. J Vasc Surg 2013;57:734-40. 23. Vogel TR, Symons RG, Flum DR. Longitudinal outcomes after endovascular repair of abdominal aortic aneurysms. Vasc Endovascular Surg 2008;42:412-9. 24. Greenblatt DY, Greenberg CC, Kind AJ, Havlena JA, Mell MW, Nelson MT, et al. Causes and implications of readmission after abdominal aortic aneurysm repair. Ann Surg 2012;256: 595-605.
Submitted Sep 15, 2017; accepted Dec 5, 2017.
Additional material for this article may be found online at www.jvascsurg.org.
11.e1
Journal of Vascular Surgery
Kalesan et al
---
2018
Supplementary Table I (online only). Baseline patient and hospital characteristics for 30-day mortality Characteristics
Overall (N ¼ 4045)
Dead at 30 days (n ¼ 249)
Alive at 30-days (n ¼ 3796)
P value
Patient Age, years, mean (SE)
67.7 (0.4)
69.0 (1.7)
67.6 (0.5)
Age, $60 years
3004 (74.3)
185 (74.4)
2819 (74.3)
.97
Male sex
2459 (60.8)
154 (61.8)
2305 (60.7)
.85
997 (24.7)
90 (36.1)
908 (24)
Location of patient’s residence Central metropolitan ($1 million)
.032
Fringe metropolitan ($1 million)
916 (22.7)
37 (14.7)
879 (23.2)
Metropolitan (250,000-999,999)
932 (23.1)
73 (29.3)
859 (22.7)
Micropolitan (<25,000)
1189 (29.5)
50 (19.9)
1140 (30.1)
3422 (84.6)
200 (80.4)
3222 (84.9)
Insurance Private/Medicare Medicaid/other
.28 621 (15.4)
49 (19.6)
572 (15.1)
Lowest annual income (<$38,000)
1045 (26.1)
68 (27.9)
978 (26)
Patient resides in same state as hospital
3628 (89.7)
235 (94.3)
3394 (89.4)
464 (11.5)
0
464 (12.2) 1089 (28.7)
1674 (41.4)
56 (22.6)
1617 (42.6)
814 (20.1)
188 (75.7)
626 (16.5)
5 (0.2)
4.3 (0.1)
.002
2263 (56)
71 (28.6)
2192 (57.7)
<.0001
255 (6.3)
51 (20.5)
204 (5.4)
<.0001
1527 (37.7)
127 (50.9)
1400 (36.9)
.014
576 (14.2)
42 (16.9)
534 (14.1)
.49
52 (1.3)
11 (4.5)
41 (1.1)
150 (3.7)
12 (4.7)
139 (3.7)
1094 (27)
Major Elixhauser score, mean (SE)
.20
4 (1.7)
Moderate Extreme
.66 <.0001
DRG severity Minor
.43
4.3 (0)
Indication at initial presentation Aneurysms without rupture Aneurysms with rupture Dissections Concomitant procedures Proximal debranching Iliac access
.003
Hospital Bed size Small Medium Large
.86 28 (11)
416 (11)
3451 (85.3)
443 (11)
210 (84.3)
3242 (85.4)
722 (17.9)
62 (25.1)
660 (17.4)
3274 (80.9)
186 (74.9)
3088 (81.3)
48 (1.2)
0
48 (1.3)
219 (88.1)
1861 (49.1)
Teaching status Metropolitan nonteaching Metropolitan teaching Not metropolitan Urgent/emergent procedure
.25
2080 (51.5)
<.0001
DRG, Diagnosis-related group; SE, standard error. All values are weighted frequency and percentages. The number is the total sample size for the respective cohort. Values are reported as number (%) unless otherwise indicated.
Journal of Vascular Surgery Volume
-,
Number
Kalesan et al
11.e2
-
Supplementary Table II (online only). Baseline patient and hospital characteristics for 90-day mortality Characteristics
Overall (N ¼ 4045)
Dead at 90 days (n ¼ 309)
Alive at 90 days (n ¼ 3736)
P value
Patient Age, years, mean (SE)
67.7 (0.4)
69.9 (1.4)
Age, $60 years
3004 (74.3)
234 (77)
2766 (74)
67.5 (0.5)
.53
Male sex
2459 (60.8)
191 (61.7)
2268 (60.7)
.84
997 (24.7)
109 (35.3)
888 (23.8) 865 (23.2)
Location of patient’s residence Central metropolitan ($1 million)
.012
Fringe metropolitan ($1 million)
916 (22.7)
51 (16.5)
Metro (250,000-999,999)
932 (23.1)
93 (30.1)
839 (22.5)
Micropolitan (<25,000)
1189 (29.5)
56 (18.1)
1134 (30.4)
3422 (84.6)
246 (79.6)
3176 (85.1)
Insurance Private/Medicare Medicaid/other
.17 621 (15.4)
63 (20.4)
558 (14.9)
Lowest annual income (<$38,000)
1045 (26.1)
73 (24.3)
972 (26.3)
Patient resides in same state as hospital
3628 (89.7)
291 (94.1)
3337 (89.3)
.61 .16 <.0001
DRG severity Minor
.13
464 (11.5)
4 (1.2)
460 (12.3)
17 (5.4)
1077 (28.8)
Moderate
1094 (27)
Major
1674 (41.4)
68 (21.9)
814 (20.1)
221 (71.6)
593 (15.9)
4.3 (0)
5 (0.2)
4.2 (0.1)
2263 (56)
107 (34.7)
2156 (57.7)
<.0001
255 (6.3)
59 (19.2)
196 (5.2)
<.0001
1527 (37.7)
142 (46.1)
1385 (37.1)
.062
576 (14.2)
52 (16.9)
524 (14)
.44
52 (1.3)
11 (3.6)
40 (1.1)
.016
150 (3.7)
14 (4.6)
136 (3.6)
Extreme Elixhauser score, mean (SE)
1606 (43) .0005
Indication at initial presentation Aneurysms without rupture Aneurysms with rupture Dissections Concomitant procedures Proximal debranching Iliac access Hospital Bed size Small Medium Large
.77 38 (12.4)
405 (10.8)
3451 (85.3)
443 (11)
256 (82.9)
3195 (85.5)
722 (17.9)
72 (23.2)
651 (17.4)
3274 (80.9)
237 (76.8)
48 (1.2)
0
48 (1.3)
266 (86)
1815 (48.6)
Teaching status Metropolitan nonteaching Metropolitan teaching Not metropolitan Urgent/emergent procedure
.32
2080 (51.5)
3037 (81.3) <.0001
DRG, Diagnosis-related group; SE, standard error. All values are weighted frequency and percentages. The number is the total sample size for the respective cohort. Values are reported as number (%) unless otherwise indicated.
11.e3
Journal of Vascular Surgery
Kalesan et al
---
2018
Supplementary Table III (online only). Baseline patient and hospital characteristics for 180-day mortality Characteristics
Overall (N ¼ 4045)
Dead at 180 days (n ¼ 341)
Alive at 180 days (n ¼ 3704)
P value .0v65
Patient Age, years, mean (SE)
67.7 (0.4)
70.2 (1.4)
67.5 (0.5)
Age $60 years
3004 (74.3)
268 (78.6)
2736 (73.9)
.27
Male sex
2459 (60.8)
210 (61.5)
2249 (60.7)
.88
997 (24.7)
114 (33.5)
883 (23.9)
Location of patient’s residence Central metropolitan ($1 million)
.02
Fringe metropolitan ($1 million)
916 (22.7)
57 (16.7)
859 (23.3)
Metro (250,000-999,999)
932 (23.1)
103 (30.1)
830 (22.5)
Micropolitan (<25,000)
1189 (29.5)
67 (19.7)
1122 (30.4)
3422 (84.6)
278 (81.5)
3144 (84.9)
Insurance Private/Medicare Medicaid/other
.37 621 (15.4)
63 (18.5)
558 (15.1)
Lowest annual income (<$38,000)
1045 (26.1)
80 (23.8)
966 (26.4)
Patient resides in same state as hospital
3628 (89.7)
323 (94.7)
3305 (89.2)
Minor
464 (11.5)
Moderate
1094 (27)
Major
1674 (41.4)
Extreme Elixhauser score, mean (SE)
.1 <.0001
DRG severity
814 (20.1)
4 (1.1)
460 (12.4)
19 (5.5)
1075 (29)
82 (24.1)
1592 (43)
236 (69.3)
578 (15.6)
4.3 (0)
5.1 (0.2)
4.2 (0.1)
2263 (56)
118 (34.5)
2146 (57.9)
<.0001
255 (6.3)
67 (19.7)
188 (5.1)
<.0001
1527 (37.7)
156 (45.8)
1371 (37)
.071
576 (14.2)
66 (19.4)
510 (13.8)
.13
52 (1.3)
13 (3.8)
39 (1)
150 (3.7)
14 (4.2)
136 (3.7)
.0001
Indication at initial presentation Aneurysms without rupture Aneurysms with rupture Dissections Concomitant procedures Proximal debranching Iliac access
.006
Hospital Bed size Small Medium Large
.94 38 (11.3)
405 (10.9)
3451 (85.3)
443 (11)
288 (84.5)
3163 (85.4)
722 (17.9)
76 (22.3)
646 (17.4)
3274 (80.9)
265 (77.7)
3010 (81.2)
48 (1.2)
0
48 (1.3)
295 (86.6)
1785 (48.3)
Teaching status Metropolitan nonteaching Metropolitan teaching Not metropolitan Urgent/emergent procedure
.35
2080 (51.5)
<.0001
DRG, Diagnosis-related group; SE, standard error. All values are weighted frequency and percentages. The number is the total sample size for the respective cohort. Values are reported as number (%) unless otherwise indicated.
Readmissions after thoracic endovascular aortic repair Bindu Kalesan, PhD, MPH,a Thomas W. Cheng, MS,b Alik Farber, MD,b Yi Zuo, MPH,a Jeffrey A. Kalish, MD,b Douglas W. Jones, MD, MPH,b and Jeffrey J. Siracuse, MD,b Boston, Mass
ABSTRACT Objective: The care of patients undergoing thoracic endovascular aortic repair (TEVAR) can be resource intensive, which can be driven by readmissions. Our objective was to characterize index readmissions at 30, 90, and 180 days after TEVAR. Methods: A retrospective analysis of the Nationwide Readmissions Database was performed for patients who underwent TEVAR in 2013. Multivariable analysis identified independent predictors for index readmission at 30, 90, and 180 days. Results: There were 4045 TEVARs performed for descending thoracic aortic dissection (37.7%), nonruptured aneurysm (56%), and ruptured aneurysm (6.3%). There were 419 (11.1%) index readmissions at 30 days, 895 (23.6%) at 90 days, and 1131 (29.8%) at 180 days. The most frequent reason for index readmission was heart related at 30 days (15.5%) and aorta related at 90 days (18%) and 180 days (19.6%). Reinterventions were performed at 6.4%, 9.5%, and 9.7% of 30-, 90-, and 180-day readmissions, respectively. The majority of these included additional endovascular stent graft placement (51.9% of reinterventions at 30 days, 67.7% at 90 days, and 65.9% at 180 days). In multivariable analysis, 30-day index readmission was associated with initial ruptured presentation (odds ratio [OR], 1.48; 95% confidence interval [CI], 1.1-3.5; P ¼ .023) and diagnosis-related group (DRG) severity grades of moderate (OR, 1.42; 95% CI, 0.74-2.73), major (OR, 2.47; 95% CI, 1.28-4.74), and extreme (OR, 1.60; 95% CI, 0.76-3.36; P ¼ .009). Index readmission at 90 days was independently associated with initial ruptured presentation (OR, 1.88; 95% CI, 1.18-3.01; P ¼ .008), urgent/emergent TEVAR (OR, 1.41; 95% CI, 1.08-1.85; P ¼ .014), and DRG severity grades of moderate (OR, 1.53; 95% CI, 0.95-2.47), major (OR, 2.27; 95% CI, 1.39-3.7), and extreme (OR, 2.45; 95% CI, 1.43-4.18; P ¼ .002). Finally, at 180 days, initial ruptured presentation (OR, 1.66; 95% CI, 1.05-2.62; P ¼ .029), urgent/ emergent TEVAR (OR, 1.37; 95% CI, 1.08-1.79; P ¼ .013), and DRG severity grades of moderate (OR, 1.55; 95% CI, 1.01-2.38), major (OR, 2.15; 95% CI, 1.38-3.33), and extreme (OR, 2.39; 95% CI, 1.47-3.89; P ¼ .002) were, again, independently associated with index readmission. Conclusions: A large portion of patients treated with TEVAR were readmitted most commonly for heart-related reasons at 30 days and aorta-related reasons at 90 and 180 days. TEVAR performed to treat initial aortic rupture and greater DRG severity grade were independently associated with an index readmission at 30, 90, and 180 days. Urgent/emergent TEVAR was independently associated with an index readmission at 90 and 180 days. These factors are important to consider in using readmissions as a quality measure. (J Vasc Surg 2018;-:1-11.)
As our health care delivery system transitions from feefor-service to value-based care, hospital readmission, a surrogate for both quality of care and resource utilization, has become an ever important metric.1,2 A number of studies have demonstrated that patients undergoing From the Center for Clinical Translational Epidemiology and Comparative Effectiveness Research, Preventive Medicine and Epidemiology, Department of Medicine, Boston University School of Medicinea; and the Division of Vascular and Endovascular Surgery, Boston University, Boston Medical Center.b Author conflict of interest: B.K. and Y.Z. are funded by the Evans Research Foundation. Presented as a plenary oral presentation at the Forty-fifth Annual Symposium of the Society for Clinical Vascular Surgery, Lake Buena Vista, Fla, March 18-22, 2017. Additional material for this article may be found online at www.jvascsurg.org. Correspondence: Jeffrey J. Siracuse, MD, Department of Surgery, Boston University School of Medicine, 88 E Newton St, Ste C520, Boston, MA 02118 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2018 by the Society for Vascular Surgery. Published by Elsevier Inc. https://doi.org/10.1016/j.jvs.2017.12.035
vascular surgery procedures have 30-day readmission rates as high as 28.5%.2-6 In particular, patients with Medicare undergoing vascular surgery have a 30-day readmission rate of 23.9%, which is highest among other surgical specialties.2 The Centers for Medicare & Medicaid Services introduced the Hospital Readmissions Reduction Program in 2012 to reduce readmissions by penalizing centers with higher than expected readmission rates by withholding up to 1% of Medicare reimbursement. Multiple independent predictors of readmission have been identified across common vascular surgery procedures, including lower extremity bypass, carotid endarterectomy, open abdominal aortic aneurysm repair, and endovascular abdominal aortic aneurysm repair.7-10 However, there are limited data on readmissions for patients undergoing thoracic endovascular aortic repair (TEVAR). As TEVAR becomes a standard in treating patients with descending thoracic aortic disease, it is paramount to characterize predictors for readmissions after TEVAR in this cohort of high-risk patients.11-16 Therefore, our objectives were to describe readmission and reintervention rates after TEVAR both to evaluate reasons 1
2
Journal of Vascular Surgery
Kalesan et al
---
for index readmissions and to identify factors that are independently associated with index readmission at 30, 90, and 180 days.
ARTICLE HIGHLIGHTS d
METHODS Database. The Nationwide Readmissions Database (NRD) is an inpatient database designed by the Agency for Healthcare Research and Quality (AHRQ) as part of the Healthcare Cost and Utilization Project to identify areas for improving patient care, specifically hospital readmissions.17 The NRD includes all hospitalizations including readmissions within a state except for those at rehabilitation and long-term acute care hospitals.17 This study used the 2013 NRD, which consists of 21 states and captures 49.3% of the U.S. resident population and 49.1% of all hospitalizations within the United States in 2013.17 Selection of patients. Patients were first identified if they were older than 18 years and if they had a primary or secondary diagnosis for dissection of the thoracic aorta, dissection of the thoracoabdominal aorta, ruptured thoracic aneurysm, thoracic aneurysm without rupture, and any aortic aneurysm without rupture using the respective International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes: 441.01, 441.03, 441.1, 441.2, and 441.9. Patients with a primary or secondary procedure code for TEVAR (ICD-9: 39.73) were then included. All available diagnosis and procedure codes in the NRD database were evaluated. Patients who underwent open surgical repair were excluded with the following primary or secondary ICD-9 procedure code: resection of aorta or thoracic vessel with anastomosis (38.34, 38.35) or with replacement (38.44, 38.45) and repair of a vessel with a patch (39.57, 39.58). Patients who underwent proximal debranching or iliac access were identified using the respective primary or secondary ICD-9 procedure codes: 39.22 and 39.29. Patients were then classified into three indications at initial presentation: dissection, nonruptured aneurysm, and ruptured aneurysm. The indication was determined by the primary diagnosis code or by the secondary diagnosis code that was closest to the primary diagnosis code by numerical order. Sample groups and data collection. The resulting cohort of patients was categorized into two groups: those with an index readmission and those not readmitted after the TEVAR procedure at 30, 90, and 180 days. Demographics, comorbidities, indication at initial presentation, Elixhauser score, and hospital characteristics were collected at each time point. The Elixhauser score is a single numeric comorbidity score used in primarily administrative databases to assess the patient’s disease burden.18,19 The Elixhauser score consists of 30 comorbidity measures, such as hypertension, congestive heart failure, chronic pulmonary disease, obesity, and renal
2018
d
d
Type of Research: Retrospective analysis of prospectively collected data of the Nationwide Readmissions Database Take Home Message: In this study of 4045 thoracic endovascular aortic repair (TEVAR) patients enrolled in the Nationwide Readmissions Database, patients were most frequently readmitted for cardiac reasons at 30 days and aorta-related reasons at 90 and 180 days. TEVAR for aneurysm rupture predicted readmission at 30, 90, and 180 days. Recommendation: The authors suggest considering readmissions for ruptured aneurysms and for cardiac and aorta-related reasons as quality measures of TEVAR.
failure, that may have an impact on outcomes, such as length of stay, hospital charges, and in-hospital mortality.18,19 The Elixhauser score can also be implemented to evaluate risk of readmission.20 Hospital characteristics collected included hospital bed size (large, medium, and small). A hospital bed size defined by the AHRQ depends on the region (Northeast, Midwest, Southern, and Western) and location and teaching status of the hospital (rural, urban, and nonteaching, and urban and teaching).17 An urban teaching hospital with >425 beds in the Northeast region, 375 beds in the Midwest region, 450 beds in the Southern region, and 325 beds in the Western region was considered a large bed size.17 Patients who were readmitted at 30 days were included in the later readmission time points. Also, patients who died during follow-up were not included in the analysis as demonstrated in previous studies that analyzed readmission for resource utilization.5,21 Outcome variables collected at 30, 90, and 180 days were survival, readmission reasons, and reinterventions. Readmission reasons after the index TEVAR were identified by the primary ICD-9-CM diagnosis code and categorized by similarity. These categories included cardiac, aortic, infectious, gastrointestinal, other, renal failure, pulmonary, wound complication, peripheral ischemia, altered mental status, hypertension, musculoskeletal, trauma, bleeding, venous thromboembolism, and stroke. The aortic category was subcategorized to those with a reintervention. Reinterventions for TEVAR and open repair were identified using the previously described primary or secondary ICD-9 procedure codes. The Boston University School of Medicine Institutional Review Board approved this study, and informed consent was waived. Statistical analysis. Weighted numbers and percentages were used for all statistical analysis as recommended by the AHRQ.17 As the 2013 NRD does not include
Journal of Vascular Surgery Volume
-,
Number
Kalesan et al
3
-
all states, this method of weighing allows national statistics to be evaluated on the basis of the participating states. Univariate analysis performed between index readmission and those not readmitted at 30-, 90-, and 180-day time points included c2 test and survey linear regression model as appropriate. Univariate analyses that included c2 test and survey linear regression model were performed as appropriate for those who survived and those who died after the TEVAR procedure at 30-, 90-, and 180-day time points. All patient and hospital covariates were considered for possible confounders and underwent a variance inflation factor test to evaluate for multicollinearity. Covariates with a variance inflation factor >7 were excluded from the multivariable model. Multivariable survey logistic regression was used to identify factors that independently associated with index readmission at 30, 90, and 180 days. The software used for statistical analysis was Stata Statistical Software, release 14 (StataCorp LP, College Station, Tex). Statistical significance was set at P < .05.
RESULTS There were 4045 patients who underwent TEVAR in 2013 from a total of 82,518 patients identified with a primary or secondary diagnosis of thoracic aortic dissection and aneurysm (Fig). Comorbidities of the overall patient population are listed in Table I. The two most prevalent comorbidities were hypertension (83.4%) and peripheral vascular disorders (42.9%). The mean Elixhauser comorbidity score was 4.29. Overall, TEVARs divided by indication at initial presentation were thoracic aneurysm without rupture (56%), ruptured thoracic aneurysm (6.3%), and descending dissection in the thoracic aorta (37.7%; Table II). The mean age was 67.7 6 0.4 years; 60.8% were male, and 51.5% of TEVARs were urgent/emergent. Patients who were younger than 65 years represented 35.4% of the study population, and 36.8% were not on Medicare. A large portion (89.7%) of patients presented to a hospital within the state in which they resided. Insurance was largely (84.6%) either private insurance or Medicare. Proximal debranching and iliac access were performed in 14.2% and 1.3% of cases, respectively. The most common hospitals performing TEVARs were hospitals with a large bed size (85.3%) and teaching hospitals in the metropolitan area (80.9%). At 30 days, there were 419 (10.4%) index readmissions (Table II). A major diagnosis-related group (DRG) severity grade (55.3% vs 39.8%; P < .001) and the Elixhauser score (4.7 vs 4.3; P < .003) were significantly higher in proportion for those readmitted compared with those not readmitted at 30 days. Indication for TEVAR and urgent/ emergent TEVARs were not statistically different between those readmitted and those not readmitted at 30 days. Overall, the length of stay was 9.5 6 0.3 days. The length of stay for those readmitted at 30 days and
those not readmitted at 30 days was 8.2 6 0.5 days and 9.6 6 0.4 days, respectively (P ¼ .009). The three most frequent reasons for the index readmission at 30 days were cardiac (15.5%), aorta related (14.8%), and nonaorta infections (12.3%; Table III). Nonaorta-related reasons accounted for 85.2% of index readmissions at 30 days. During an index readmission at 30 days, there were 27 (0.7% of index TEVARs) reinterventions (Table IV). Additional TEVARs were performed in 51.9% of cases and an open intervention in 37% of cases at 30 days. In multivariable analysis, DRG severity grades of moderate (odds ratio [OR], 1.42; 95% confidence interval [CI], 0.74-2.73), major (OR, 2.47; 95% CI, 1.28-4.74), and extreme (OR, 1.60; 95% CI, 0.76-3.36) were independently associated with index readmission at 30 days (P ¼. 009; Table V). Also, initial presentation with ruptured thoracic aortic aneurysm (OR, 1.48; 95% CI, 1.1-3.5; P ¼ .023) was independently associated with index readmission at 30 days. The patients’ demographic characteristics, such as female sex, location of residence, type of insurance, and income status, were not independent predictors for readmission at 30 days. Requirement for iliac access or proximal debranching and hospital urban teaching status were also not independent predictors for readmission at 30 days. A summary of covariates independently associated with an index readmission at 30, 90, and 180 days is described in Table VI. At 90 days, there were 895 (22.1%) index readmissions (Table VII). Patients readmitted at 90 days were significantly older (69.7 6 0.8 years vs 67.2 6 0.5 years; P ¼ .02). Age >60 years (79.8% vs 72.7%; P ¼ .008), patients residing in the same state as the hospital (94.6% vs 88.3%; P ¼ .001), a major DRG severity grade (49.1% vs 39.2%; P < .004), and a higher Elixhauser score (4.7 vs 4.2; P < .0003) were significantly higher in proportion for those readmitted compared with those not readmitted at 90 days. A smaller proportion of patients younger than 65 years were readmitted at 90 days (29.5% vs 37%; P ¼ .018), and a higher proportion of patients on Medicare insurance were readmitted at 90 days (69.9% vs 61.3%; P ¼ .01). Patients who originally presented with ruptured thoracic aortic aneurysms (9.3% vs 5.5%; P ¼ .025) and those with urgent/emergent TEVARs (58.4% vs 49.5%; P ¼ .004) were also significantly higher in proportion for those readmitted at 90 days. The three most frequent reasons for readmission at 90 days were aorta related (18%), heart related (15.3%), and nonaorta infections (12.6%; Table III). Nonaorta-related reasons accounted for 82% of index readmissions at 90 days. During an index readmission at 90 days, there were 96 (2.4% of index TEVARs) reinterventions (Table IV), with additional TEVAR performed in 67.7% of cases and open intervention in 29.2% of cases. In multivariable analysis, covariates that were independently associated with index readmission at 90 days were again the DRG severity grades of moderate (OR, 1.53; 95% CI, 0.95-2.47),
4
Journal of Vascular Surgery
Kalesan et al
---
2018
Total number of patients greater than 18 years hospitalized in 2013 = 13,172,268 (unwt); 29,936,410 (wt) Number of patients without a primary or secondary diagnosis for aortic dissection or aneurysm = 13,136,070 (unwt); 29,853,892 (wt) Number of patients with a primary or secondary diagnosis for aortic dissection or aneurysm = 36,198 (unwt); 82,518 (wt) Number of patients without a primary or secondary procedure code for TEVAR in 2013 = 34,323 (unwt); 78,169 (wt) Number of patients with a primary or secondary procedure code for TEVAR in 2013= 1,875 (unwt); 4,349 (wt) Number of primary or secondary procedure code for open repair in 2013 = 94 (unwt); 221 (wt) Number of All TEVARs in 2013= 1,781 (unwt); 4,128 (wt) Number of Index TEVARs in 2013 = 1,745 (unwt); 4,045 (wt) In-hospital mortality= 119 (unwt); 256 (wt) Number of surviving index TEVARs in 2013= 1,626 (unwt); 3,789 (wt) Total number of readmissions in 2013 = 520 (unwt); 1,219 (wt)
Fig. Flow chart for study population. TEVAR, Thoracic endovascular aortic repair; unwt, unweighted; wt, weighted.
major (OR, 2.27; 95% CI, 1.39-3.7), and extreme (OR, 2.45; 95% CI, 1.43-4.18; P ¼ .002). Patients with an initial presentation of ruptured thoracic aortic aneurysms (OR, 1.88; 95% CI, 1.18-3.01; P ¼ .008) and urgent/emergent TEVARs (OR, 1.41; 95% CI, 1.08-1.85; P ¼ .014) were again independently associated with index readmission at 90 days. The patients’ demographics, requirement for iliac access or proximal debranching, and hospital urban teaching status were again not independent predictors for readmission at 90 days. At 180 days, there were 1131 (28%) index readmissions (Table VIII). Patients readmitted at 180 days were again significantly older (69.1 6 0.7 years vs 67.2 6 0.5 years; P ¼ .019). Age >60 years (78.9% vs 72.5%; P ¼ .01), patients residing in the same state as the hospital (94.8% vs 87.1%; P < .0001), a major DRG severity grade (48.3% vs 38.7%; P ¼ .004), and a higher Elixhauser score (4.7 vs 4.2; P < .0001) were again significantly higher in proportion for those readmitted compared with those not readmitted at 180 days. Again, patients readmitted at 180 days were significantly in higher proportion for urgent/emergent TEVARs (57.2% vs 49.3%; P ¼ .005). At 180 days, the three most frequent reasons for readmission were aorta related (19.6%), heart related (13.8%), and nonaorta infections (12.6%; Table III). Nonaorta-related reasons accounted for 80.4% of index readmissions at
180 days. There were 129 (3.2% of index TEVARs) reinterventions during an index readmission at 180 days (Table IV). Additional TEVARs were performed in 65.9% of cases and an open intervention in 31.8% of cases. In multivariable analysis, the DRG severity grades of moderate (OR, 1.55; 95% CI, 1.01-2.38), major (OR, 2.15; 95% CI, 1.38-3.33), and extreme (OR, 2.39; 95% CI, 1.47-3.89; P ¼ .002) were again independently associated with index readmission at 180 days (Table V). Patients with an initial presentation of ruptured thoracic aortic aneurysm (OR, 1.66; 95% CI, 1.05-2.62; P ¼ .029) and urgent/emergent TEVARs (OR, 1.37; 95% CI, 1.08-1.79; P ¼ .013) were again independently associated with index readmission at 180 days. The patients’ demographics, requirement for iliac access or proximal debranching, and hospital urban teaching status were, again, not independent predictors for readmission at 180 days. The mortality rate at 30 days was 6.1%; at 90 days, 7.6%; and at 180 days, 8.4%. Univariate analyses for mortality at 30 days, 90 days, and 180 days are depicted in Supplementary Tables I,II, and III, respectively (online only).
DISCUSSION Our study demonstrated that 10.4% of patients undergoing TEVAR were readmitted at 30 days, 22.1% at 90 days, and 28% at 180 days. Although the most
Journal of Vascular Surgery Volume
-,
Number
Kalesan et al
5
-
Table I. Comorbidities of the patients Comorbidities Alcohol abuse
135 (3.3)
Chronic blood loss anemia
36 (.9)
Chronic pulmonary disease
1127 (27.9)
Coagulopathy
500 (12.4)
Congestive heart failure Deficiency anemias
55 (1.4) 781 (19.3)
Depression
405 (10)
Diabetes (uncomplicated)
605 (14.9)
Drug abuse
103 (2.6)
Fluid and electrolyte disorder
1294 (32)
Hypertension
3373 (83.4)
Hypothyroidism
387 (9.6)
Liver disease
64 (1.6)
Lymphoma
33 (.8)
Metastatic cancer
30 (.7)
Obesity
546 (13.5)
Other neurologic disorders
238 (5.9)
Paralysis
197 (4.9)
Peripheral vascular disorders Psychoses
1735 (42.9) 97 (2.4)
Pulmonary circulation disorder
17 (.4)
Solid tumor without metastasis
88 (2.2)
Renal failure
766 (18.9)
Rheumatoid arthritis
134 (3.3)
Valvular disease
46 (1.1)
Weight loss
277 (6.8)
Elixhauser comorbidity score
4.29 (1.89)
Values are weighted frequencies and weighted percentages using survey analysis. All data presented are number (%), except for the Elixhauser comorbidity score, which is mean (95% confidence interval [CI]).
frequent primary reason for index readmission was heart related at 30 days and aorta related at 90 and 180 days, the majority of reasons for readmission were not aorta or heart related. These included nonaorta infections, renal failure, and pulmonary conditions. The length of stay for those who were readmitted and those not readmitted at 30 days was significantly higher by 1.4 days for those not readmitted. Patients who originally presented with ruptured thoracic aortic aneurysm and those with a DRG severity grade of moderate, major, or extreme were independently associated with an index readmission at 30, 90, and 180 days. Urgent/emergent TEVAR was independently associated with an index readmission at 90 and 180 days. These findings confirmed that despite the successful treatment by TEVAR of patients with descending aortic dissection or aneurysm, improvement is needed to reduce readmissions, particularly for patients with ruptured thoracic aortic aneurysms. In addition, these time points at 30, 90, and 180 days are important as they demonstrate perioperative outcomes
after TEVAR, global hospital payment, and long-term outcomes, respectively. Our study’s readmission rate was similar to or lower than that in the literature. Jones et al15 reported a 30-day readmission rate of 21% for patients with type B aortic dissections treated with TEVAR between 2000 and 2010, and this rate was higher than our 30-day readmission rate for TEVARs performed in 2013. Another study analyzing the use of TEVAR from 2000 to 2007 for nonruptured thoracic aortic aneurysm reported a 30-day readmission rate of 22% during the period of 2000 to 2001 and 16.5% during 2006 to 2007.22 This trend toward lower readmission rates may speak to improved operator techniques and stent technology over time. However, these studies used Medicare data and investigated readmission rates only at 30 days, whereas our study’s database included all-payer inpatient visits and long-term readmission data at 90 and 180 days in addition to 30-day readmission data. The all-payer inpatient visits in the NRD include a large portion of patients (36.8%) who were not captured in the Medicare database as the NRD is hospital based, whereas the Medicare database is insurance based. Our study also provides follow-up data on TEVAR performed for patients younger than 65 years who may not be present in Medicare database studies because of the general Medicare age restriction. In addition, the NRD is different from the Vascular Quality Initiative and the National Surgical Quality Improvement Program databases; the Vascular Quality Initiative does not have any readmission data, and the National Surgical Quality Improvement Program database has only 30-day readmission data. Our findings provide improved generalizability to the entire population treated with TEVAR. In our study, the most frequent reason for readmission was heart related (15.5%) at 30 days and aorta related at 90 and 180 days (18% and 19.6%, respectively). Our 30day rate for heart-related readmission was comparable to that of another study in which 18.5% of 30-day readmissions after elective endovascular aneurysm repair for nonruptured abdominal aortic aneurysm were heart related.23 At 90 days, our aorta-related readmission finding was comparable to that of another study reporting that 18.5% of readmissions after TEVAR were aneurysm related.14 As aorta-related readmissions are the most frequent reason for readmission at 90 days, this may affect how hospital global payments are issued at 90 days. At 180 days, aorta-related readmission remained the most frequent reason for readmission, which suggests the extent of impact that performing TEVAR has on reimbursement for certain hospital systems, particularly accountable care organizations. However, the magnitude of impact on reimbursement is unclear, and it is outside the scope of our study. Also, it is difficult to assess as the health care system is changing drastically in the near future with the transition from a fee-for-service to
6
Journal of Vascular Surgery
Kalesan et al
---
2018
Table II. Baseline patient and hospital characteristics for 30-day index readmission Characteristics
Overall (N ¼ 4045)
Readmitted at 30 days (n ¼ 419)
Not readmitted at 30 days (n ¼ 3626)
P value
Patient Age, years, mean (SE)
67.7 (0.4)
.94
Age $60 years
3004 (74.3)
309 (73.8)
2695 (74.3)
.89
Age <65 years
1431 (35.4)
147 (35.1)
1283 (35.4)
.95
2459 (60.8)
231 (55)
2228 (61.5)
.15
Male sex
67.7 (0.4)
67.6 (1.4)
Location of patient’s residence Central metropolitan ($1 million)
.45 997 (24.7)
118 (28.1)
880 (24.3)
Fringe metropolitan ($1 million)
916 (22.7)
105 (25.1)
811 (22.4)
Metropolitan (250,000-999,999)
932 (23.1)
81 (19.2)
852 (23.6)
Micropolitan (<25,000)
1189 (29.5)
116 (27.6)
1074 (29.7)
3422 (84.6)
354 (84.9)
3068 (84.6)
621 (15.4)
63 (15.1)
558 (15.4)
2557 (63.2)
277 (66.1)
2280 (62.9)
Insurance Private/Medicare Medicaid/other
.91
Insurance Medicare Not Medicare
.41 1488 (36.8)
142 (33.9)
1346 (37.1)
Lowest annual income (<$38,000)
1045 (26.1)
169 (26.3)
876 (26.1)
.98
Patient resides in same state as hospital
3628 (89.7)
396 (94.4)
3232 (89.1)
.16
464 (11.5)
25 (5.9)
440 (12.1)
97 (23.1)
997 (27.5)
DRG severity Minor
.001
Moderate
1094 (27)
Major
1674 (41.4)
232 (55.3)
1442 (39.8)
814 (20.1)
66 (15.7)
748 (20.6)
4.3 (0)
4.7 (0.2)
4.3 (0.1)
2263 (56)
Extreme Elixhauser score, mean (SE)
.003
Indication at initial presentation Aneurysms without rupture Aneurysms with rupture Dissections
224 (53.4)
2039 (56.2)
255 (6.3)
40 (9.6)
214 (5.9)
.48
1527 (37.7)
155 (36.9)
1372 (37.8)
.81
576 (14.2)
57 (13.5)
519 (14.3)
.72
52 (1.3)
6 (1.3)
46 (1.3)
.94
150 (3.7)
22 (5.2)
128 (3.5)
.057
Concomitant procedures Proximal debranching Iliac access Hospital Bed size Small Medium Large
.50 50 (11.8)
394 (10.9)
3451 (85.3)
443 (11)
348 (82.9)
3104 (85.6)
722 (17.9)
87 (20.8)
635 (17.5)
Teaching status Metropolitan nonteaching Metropolitan teaching Not metropolitan Urgent/emergent procedure
.58 3274 (80.9)
327 (78.1)
48 (1.2)
5 (1.2)
44 (1.2)
226 (53.8)
1855 (51.2)
2080 (51.5)
2947 (81.3) .58
DRG, Diagnosis-related group; SE, standard error. All values are weighted frequency and percentages. The number is the total sample size for the respective cohort. Values are reported as number (%) unless otherwise indicated.
value-based health care. In addition, these aorta-related readmissions could be related to planned procedures as depicted by Glebova et al,9 who concluded that planned readmissions after vascular surgery were more frequent
compared with those after other surgical procedures. Unfortunately, given limitations of our database, our study cannot explain whether planned procedures represented a majority of readmissions.
Journal of Vascular Surgery Volume
-,
Number
Kalesan et al
7
-
Table III. Primary reasons for index readmission at 30 days, 90 days, and 180 days Category
30 Days, No. (%)
90 Days, No. (%)
180 Days, No. (%)
Heart related
65 (15.5)
137 (15.3)
156 (13.8)
Aorta related
222 (19.6)
62 (14.8)
161 (18)
Reinterventionsa
27 (6.4)
85 (9.5)
110 (9.7)
Nonaorta infections
52 (12.3)
113 (12.6)
142 (12.6)
Gastrointestinal
38 (9.1)
81 (9.1)
116 (10.3)
Other
41 (9.8)
108 (12.1)
138 (12.2)
Renal failure
30 (7.2)
40 (4.5)
47 (4.2)
Pulmonary
28 (6.7)
63 (7)
66 (5.8)
Wound complication
22 (5.3)
32 (3.6)
32 (2.8)
Peripheral ischemia
17 (4.1)
23 (2.6)
38 (3.3)
Altered mental status
13 (3.1)
21 (2.3)
23 (2)
Hypertension
11 (2.7)
14 (1.5)
19 (1.7)
10 (2.4)
22 (2.4)
28 (2.5)
9 (2.2)
22 (2.5)
36 (3.2)
Musculoskeletal Trauma Bleeding
9 (2.1)
12 (1.3)
14 (1.2)
Venous thromboembolism
8 (1.9)
24 (2.7)
24 (2.1)
Stroke
6 (1.4)
23 (2.6)
31 (2.7)
Frequencies for each reason for readmission are not mutually exclusive between the 30-, 90-, and 180-day time points. a Reinterventions is a subcategory that describes the number of reinterventions that occurred during an aorta-related readmission, and the percentage is in respect to the number of index readmissions.
Table IV. Reinterventions after thoracic endovascular aortic repair (TEVAR) Type of reintervention during index readmission
30 Days, No. (%)
90 Days, No. (%)
180 Days, No. (%)
Additional TEVAR
14 (51.9)
65 (67.7)
85 (65.9)
Open intervention
10 (37)
28 (29.2)
41 (31.8)
Additional TEVAR and open intervention Total
3 (11.1)
3 (3.1)
3 (2.3)
27 (100)
96 (100)
129 (100)
Frequencies for reinterventions were not mutually exclusive between the 30-, 90-, and 180-day time points. All values and percentages are weighted. The percentages are percentages of total number of reinterventions at each respective time point.
The rates for other primary reasons for readmission that constituted the majority of readmission in our study, particularly infection and renal failure, were echoed by multiple studies.5,6,22-24 In our study, 12.3% of 30-day readmissions were related to infections, which was similar to the reported range of 10.3% to 11.9% for infections associated with 30-day readmissions for common vascular surgery procedures, including lower extremity bypass, carotid endarterectomy, and abdominal aortic procedures.6,23 This similarity continued for patients readmitted for renal failure as our study demonstrated that 7% of 30-day readmissions were related to renal failure, and Greenblatt et al24 reported that 6% of Medicare patients were readmitted for renal failure at 30 days after an elective endovascular aneurysm repair for nonruptured abdominal aortic aneurysm. This indicates that patients undergoing TEVAR were readmitted for various primary reasons other than heart- or aorta-related reasons. In our study, the length of stay for those not readmitted at 30 days was significantly longer compared with those readmitted at 30 days. As current hospital practices are
to decrease length of stay, there is concern for higher readmission rate as patients may not be optimized for an early discharge. However, our study demonstrates that early discharge did not adversely affect readmission rate. In our multivariable analysis, urgent/emergent TEVAR was not independently associated with 30-day readmission. This was echoed in a study analyzing the association between urgent/emergent TEVARs for nonruptured thoracic aortic aneurysm and 30-day readmission in Medicare patients (OR, 0.95; 95% CI, 0.72-1.27; P ¼ .75).22 This suggests that although urgent/emergent TEVARs may be relatively safe during the perioperative phase, the urgency of TEVAR should be considered in using readmission for reimbursement past 30 days as our study demonstrated urgent/emergent TEVARs to be independently associated with an index readmission at 90 and 180 days. The limitations of this study include the inherent use of a retrospective database. A national database is subject to selection bias for performing TEVAR. Also, although the ICD-9 codes used in this study have been validated by another study, there are potential errors in coding
8
Journal of Vascular Surgery
Kalesan et al
---
2018
Table V. Multivariable analysis of predictors for index readmission 30 Days Covariate
90 Days
P value
OR (95% CI)
180 Days
P value
OR (95% CI)
OR (95% CI)
P value
Patient Female sex
1.16 (0.84-1.6)
.36
Location of patient’s residence Central metropolitan ($1 million)
.99
1.02 (0.67-1.56)
Metropolitan (250,000-999,999)
0.89 (0.57-1.38)
Micropolitan (<25,000)
0.84 (0.54-1.29)
Reference .98 (0.71-1.36)
.51 1.05 (0.77-1.42) 0.92 (0.67-1.25)
0.78 (0.56-1.09)
0.83 (0.61-1.13)
.85
Private/Medicare
.11
.3
Reference
Reference
Reference
0.96 (0.61-1.51)
0.75 (0.53-1.07)
.85 (0.62-1.16)
Low income quartile (<$38,000)
0.93 (0.64-1.36)
.72
DRG severity Minor
0.84 (0.63-1.12)
.25
.009
0.97 (0.75-1.26)
.002
Reference
Reference
Reference
1.42 (0.74-2.73)
1.53 (0.95-2.47)
1.55 (1.01-2.38)
Major
2.47 (1.28-4.74)
2.27 (1.39-3.7)
2.15 (1.38-3.33)
Extreme
1.60 (0.76-3.36)
2.45 (1.43-4.18)
2.39 (1.47-3.89)
1.12 (0.79-1.61)
.51
1.41 (1.08-1.85)
.014
Aneurysms without rupture
1.28 (0.87-1.88)
Aneurysms with rupture
1.48 (1.1-3.5)
.21
1.33 (1-1.78)
.023
1.88 (1.18-3.01)
Proximal debranching
1.06 (0.68-1.66)
Iliac access
1.48 (0.42-5.25)
.78
1.32 (0.68-1.66)
.55
1.65 (0.63-4.34)
.83 .002
Moderate
Urgent/emergent procedure
.88
Reference
0.94 (0.67-1.3)
Insurance
1.02 (0.81-1.28)
.48
Reference
Fringe metropolitan ($1 million)
Medicaid/other
1 (0.78-1.28)
.8
1.37 (1.08-1.79)
.013
.052
1.26 (0.87-1.88)
.091
.008
1.66 (1.05-2.62)
.029
.097
1.35 (1-1.84)
.052
.31
1.83 (0.72-4.63)
.2
Indication at initial presentation
Procedure
Hospital Teaching status Metropolitan nonteaching
.93
.63
.21
Reference
Reference
Reference
Metropolitan teaching
0.93 (0.63-1.37)
1.09 (0.80-1.48)
1.07 (0.81-1.41)
Not metropolitan
0.93 (0.2-4.44)
1.66 (0.54-5.1)
2.53 (0.91-7.04)
CI, Confidence interval; DRG, diagnosis-related group; OR, odds ratio. Boldface entries indicate statistical significance (P < .05).
Table VI. Summary of significant factors associated with index readmission Covariate
Readmitted at 30 days
Readmitted at 90 days
Readmitted at 180 days
þ
þ
þ
þþ
þþ
þþ
þ
þþ
þþ
Initial presentation of ruptured aneurysm
þ
þ
þ
Urgent/emergent procedure
0
þ
þ
DRG severity grade Moderate Major Extreme
DRG, Diagnosis-related group. A þ and þþ represent an odds ratio (OR) of 1 to 2 and 2 to 3, respectively. A 0 represents an OR that is not statistically significant.
and the limited anatomic detail associated with codes.15 One limitation is how the indications were categorized. As the indication was identified by numerical order for diagnosis for dissection, nonruptured aneurysm, and ruptured aneurysm, patients who presented with a dissection and then ruptured during the initial hospitalization could have been categorized as dissection.
Although this is a limitation, all diagnosis codes for each patient including any dissection, nonruptured aneurysms, and ruptured aneurysms were placed into the multivariable models for readmission at 30, 90, and 180 days. Our study is further limited by how the ICD-9 diagnosis codes for readmissions in the database are captured as they were on admission rather than at
Journal of Vascular Surgery Volume
-,
Number
Kalesan et al
9
-
Table VII. Baseline patient and hospital characteristics for 90-day index readmission Characteristics
Overall (N ¼ 4045)
Readmitted at 90 days (n ¼ 895)
67.7 (0.4)
69.4 (0.8)
Not readmitted at 90 days (n ¼ 3150)
P value
Patient Age, years, mean (SE)
67.2 (0.5)
.02
Age $60 years
3004 (74.3)
714 (79.8)
2290 (72.7)
.008
Age <65 years
1431 (35.4)
264 (29.5)
1166 (37)
.018
2459 (60.8)
523 (58.5)
1936 (61.4)
.36
997 (24.7)
239 (26.7)
758 (24.1)
Male sex Location of patient’s residence Central metropolitan ($1 million)
.64
Fringe metropolitan ($1 million)
916 (22.7)
213 (23.7)
704 (22.4)
Metropolitan (250,000-999,999)
932 (23.1)
196 (21.9)
736 (23.4)
Micropolitan (<25,000)
1189 (29.5)
247 (27.6)
943 (30)
3422 (84.6)
785 (87.9)
2637 (83.7)
621 (15.4)
108 (12.1)
513 (16.3)
2557 (63.2)
626 (69.9)
1931 (61.3)
Insurance Private/Medicare Medicaid/other
.062
Insurance Medicare Not Medicare
.01 1488 (36.8)
269 (30.1)
1219 (38.7)
Lowest annual income (<$38,000)
1045 (26.1)
209 (23.5)
837 (26.9)
Patient resides in same state as hospital
3628 (89.7)
846 (94.6)
2782 (88.3)
464 (11.5)
60 (6.7)
404 (12.8)
DRG severity Minor
.21 .001 .004
Moderate
1094 (27)
220 (24.6)
874 (27.7)
Major
1674 (41.4)
439 (49.1)
1235 (39.2)
814 (20.1)
176 (19.7)
638 (20.2)
4.3 (0)
4.7 (0.1)
4.2 (0.1)
2263 (56.0)
489 (54.6)
1774 (56.3)
255 (6.3)
83 (9.3)
172 (5.5)
1527 (37.7)
323 (36.1)
1204 (38.2)
.46
576 (14.2)
142 (15.8)
434 (13.8)
.37
52 (1.3)
13 (1.5)
38 (1.2)
.64
150 (3.7)
31 (3.5)
119 (3.8)
Extreme Elixhauser score, mean (SE)
.0003
Indication at initial presentation Aneurysms without rupture Aneurysms with rupture Dissections
.59 .025
Concomitant procedures Proximal debranching Iliac access Hospital Bed size Small Medium Large
.97 443 (11.0)
99 (11.1)
3451 (85.3)
765 (85.4)
2687 (85.3)
344 (10.9)
722 (17.9)
158 (17.7)
564 (17.9)
3274 (80.9)
724 (80.9)
48 (1.2)
13 (1.4)
36 (1.1)
2081 (51.5)
523 (58.4)
1557 (49.5)
Teaching status Metropolitan nonteaching Metropolitan teaching Not metropolitan Urgent/emergent procedure
.90 2551 (81) .004
DRG, Diagnosis-related group; SE, standard error. All values are weighted frequency and percentages. The number is the total sample size for the respective cohort. Values are reported as number (%) unless otherwise indicated.
discharge. Another limitation of the study is that the NRD captures only readmissions within the state and does not allow interstate data to be collected. However, because a large majority (89.7%) of patients were
readmitted at the same state, the impact of patients’ being readmitted at another state may be relatively low. Moreover, our study is limited by staged TEVAR procedures, which could inflate the reintervention and
10
Journal of Vascular Surgery
Kalesan et al
---
2018
Table VIII. Baseline patient and hospital characteristics for 180-day index readmission Characteristics
Overall (N ¼ 4045)
Readmitted at 180 days (n ¼ 1131)
Not readmitted at 180 days (n ¼ 2914)
P value
Patient Age, years, mean (SE)
67.7 (0.4)
69.1 (0.7)
67.2 (0.5)
.019
Age $60 years
3004 (74.3)
892 (78.9)
2112 (72.5)
.01
Age <65 years
1431 (35.4)
363 (32.1)
1068 (36.6)
2459 (60.8)
669 (59.2)
1790 (61.4)
997 (24.7)
286 (25.3)
711 (24.5)
Male sex Location of patient’s residence Central metropolitan ($1 million)
.74
Fringe metropolitan ($1 million)
916 (22.7)
271 (24)
645 (22.2)
Metropolitan (250,000-999,999)
932 (23.1)
241 (21.3)
692 (23.8)
Micropolitan (<25,000)
1189 (29.5)
333 (29.4)
856 (29.5)
3422 (84.6)
981 (86.9)
2441 (83.8)
621 (15.4)
147 (13.1)
474 (16.2)
2557 (63.2)
764 (67.6)
1793 (61.5)
Insurance Private/Medicare Medicaid/other
.13
Insurance Medicare Not Medicare
.068 1488 (36.8)
366 (32.4)
1122 (38.5)
Lowest annual income (<$38,000)
1045 (26.1)
287 (25.6)
759 (26.4)
Patient resides in same state as hospital
3628 (89.7)
1072 (94.8)
2556 (87.1)
464 (11.5)
79 (7)
385 (13.2)
DRG severity Minor
.45
.74 .0001 .004
Moderate
1094 (27)
287 (25.4)
807 (27.7)
Major
1674 (41.4)
546 (48.3)
1128 (38.7)
814 (20.1)
218 (19.3)
596 (20.4)
4.3 (0)
4.7 (0.1)
4.2 (0.1)
<.0001
2263 (56)
.69
Extreme Elixhauser score, mean (SE) Indication at initial presentation Aneurysms without rupture Aneurysms with rupture Dissections
623 (55.1)
1641 (56.3)
255 (6.3)
95 (8.4)
160 (5.5)
.06
1527 (37.7)
413 (36.5)
1114 (38.2)
.53
576 (14.2)
177 (15.6)
399 (13.7)
.3
52 (1.3)
18 (1.6)
34 (1.2)
150 (3.7)
38 (3.3)
113 (3.9)
Procedure Proximal debranching Iliac access
.44
Hospital Bed size Small Medium Large
.32 121 (10.7)
322 (11)
3451 (85.3)
443 (11)
971 (85.9)
2480 (85.1)
722 (17.9)
194 (17.1)
529 (18.1)
Teaching status Metropolitan nonteaching Metropolitan teaching Not metropolitan Urgent/emergent procedure
.16 3274 (80.9) 48 (1.2) 2080 (51.5)
915 (80.9) 22 (2) 647 (57.2)
2360 (81) 26 (.9) 1433 (49.3)
.005
DRG, Diagnosis-related group; SE, standard error. All values are weighted frequency and percentages. The number is the total sample size for the respective cohort. Values are reported as number (%) unless otherwise indicated.
readmission rates because we cannot delineate planned and unplanned readmissions. Furthermore, TEVAR device information is unavailable in this database, which limits this study’s ability to characterize whether a
correlation exists between device and rate of readmission or reintervention. However, given these limitations, our study demonstrates contemporary results for the follow-up period after TEVAR.
Journal of Vascular Surgery Volume
-,
Number
Kalesan et al
11
-
CONCLUSIONS Our study demonstrated that a large proportion of patients are readmitted at 30, 90, and 180 days, with the most frequent reason for index readmission heart related at 30 days and aorta related at 90 and 180 days. However, the majority of readmissions were for nonheart and nonaorta-related reasons. Patients with an initial presentation of ruptured thoracic aortic aneurysms and urgent/emergent TEVARs in particular are more likely to be readmitted, and these factors should be accounted for in analyzing quality of care measures such as readmission rate.
AUTHOR CONTRIBUTIONS Conception and design: BK, JS Analysis and interpretation: BK, TC, AF, YZ, JK, DJ, JS Data collection: BK, TC, YZ, JS Writing the article: TC, JS Critical revision of the article: BK, TC, AF, YZ, JK, DJ, JS Final approval of the article: BK, TC, AF, YZ, JK, DJ, JS Statistical analysis: BK, YZ Obtained funding: BK Overall responsibility: JS
REFERENCES 1. Schroeder SA, Frist W. Phasing out fee-for-service payment. N Engl J Med 2013;368:2029-32. 2. Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare fee-for-service program. N Engl J Med 2009;360:1418-28. 3. Gonzalez AA, Girotti ME, Shih T, Wakefield TW, Dimick JB. Reliability of hospital readmission rates in vascular surgery. J Vasc Surg 2014;59:1638-43. 4. Wiseman JT, Guzman AM, Fernandes-Taylor S, Engelbert TL, Saunders RS, Kent KC. General and vascular surgery readmissions: a systematic review. J Am Coll Surg 2014;219: 552-69.e2. 5. Siracuse JJ, Shah NK, Peacock MR, Tahhan G, Kalish JA, Rybin D, et al. Thirty-day and 90-day hospital readmission after outpatient upper extremity hemodialysis access creation. J Vasc Surg 2017;65:1376-82. 6. Tahhan G, Farber A, Shah NK, Krafcik BM, Sachs TE, Kalish JA, et al. Characterization of planned and unplanned 30-day readmissions following vascular surgical procedures. Vasc Endovascular Surg 2017;51:17-22. 7. Duwayri Y, Goss J, Knechtle W, Veeraswamy RK, Arya S, Rajani RR, et al. The readmission event after vascular surgery: causes and costs. Ann Vasc Surg 2016;36:7-12. 8. Engelbert TL, Fernandes-Taylor S, Gupta PK, Kent KC, Matsumura J. Clinical characteristics associated with readmission among patients undergoing vascular surgery. J Vasc Surg 2014;59:1349-55. 9. Glebova NO, Bronsert M, Hicks CW, Nehler MR, Henderson WG. Contributions of planned readmissions and patient comorbidities to high readmission rates in vascular surgery patients. J Vasc Surg 2014;63:746-55.e2.
10. Glebova NO, Bronsert M, Hammermeister KE, Nehler MR, Gibula DR, Malas MB, et al. Drivers of readmissions in vascular surgery patients. J Vasc Surg 2016;64:185-94.e3. 11. Fanelli F, Dake MD. Standard of practice for the endovascular treatment of thoracic aortic aneurysms and type B dissections. Cardiovasc Intervent Radiol 2009;32:849-60. 12. Patel HJ, Williams DM, Upchurch GR, Dasika NL, Deeb GM. A comparative analysis of open and endovascular repair for the ruptured descending thoracic aorta. J Vasc Surg 2009;50:1265-70. 13. Goodney PP, Travis L, Lucas FL, Fillinger MF, Goodman DC, Cronenwett JL, et al. Survival after open versus endovascular thoracic aortic aneurysm repair in an observational study of the medicare population. Circulation 2011;124:2661-9. 14. Karimi A, Walker KL, Martin TD, Hess PJ, Klodell CT, Feezor RJ, et al. Midterm cost and effectiveness of thoracic endovascular aortic repair versus open repair. Ann Thorac Surg 2012;93:473-9. 15. Jones DW, Goodney PP, Nolan BW, Brooke BS, Fillinger MF, Powell RJ, et al. National trends in utilization, mortality, and survival after repair of type B aortic dissection in the Medicare population. J Vasc Surg 2014;60:11-9.e1. 16. Zimmerman KP, Oderich G, Pochettino A, Hanson KT, Habermann EB, Bower TC, et al. Improving mortality trends for hospitalization of aortic dissection in the National Inpatient Sample. J Vasc Surg 2016;64:606-15.e1. 17. NRD overview. Healthcare Cost and Utilization Project (HCUP). Agency for Healthcare Research and Quality, Rockville, Md. Available at: https://www.hcup-us.ahrq.gov/ nrdoverview.jsp. Accessed August 29, 2017. 18. Elixhauser A, Steiner C, Harris DR, Coffey RM. comorbidity measures for use with administrative data. Med Care 1998;36:8-27. 19. van Walraven C, Austin PC, Jennings A, Quan H, Forster AJ. A modification of the Elixhauser comorbidity measures into a point system for hospital death using administrative data. Med Care 2009;47:626-33. 20. Moore BJ, White S, Washington R, Coenen N, Elixhauser A. Identifying increased risk of readmission and in-hospital mortality using hospital administrative data. Med Care 2017;55:698-705. 21. Krafcik BM, Komshian S, Lu K, Roberts L, Farber A, Kalish JA, et al. Short and long-term readmission rates after infrainguinal bypass in a safety net hospital are higher than expected. J Vasc Surg 2017;66:1786-91. 22. Brooke BS, Goodney PP, Powell RJ, Fillinger MF, Travis LL, Goodman DC, et al. Early discharge does not increase readmission or mortality after high-risk vascular surgery. J Vasc Surg 2013;57:734-40. 23. Vogel TR, Symons RG, Flum DR. Longitudinal outcomes after endovascular repair of abdominal aortic aneurysms. Vasc Endovascular Surg 2008;42:412-9. 24. Greenblatt DY, Greenberg CC, Kind AJ, Havlena JA, Mell MW, Nelson MT, et al. Causes and implications of readmission after abdominal aortic aneurysm repair. Ann Surg 2012;256: 595-605.
Submitted Sep 15, 2017; accepted Dec 5, 2017.
Additional material for this article may be found online at www.jvascsurg.org.
11.e1
Journal of Vascular Surgery
Kalesan et al
---
2018
Supplementary Table I (online only). Baseline patient and hospital characteristics for 30-day mortality Characteristics
Overall (N ¼ 4045)
Dead at 30 days (n ¼ 249)
Alive at 30-days (n ¼ 3796)
P value
Patient Age, years, mean (SE)
67.7 (0.4)
69.0 (1.7)
67.6 (0.5)
Age, $60 years
3004 (74.3)
185 (74.4)
2819 (74.3)
.97
Male sex
2459 (60.8)
154 (61.8)
2305 (60.7)
.85
997 (24.7)
90 (36.1)
908 (24)
Location of patient’s residence Central metropolitan ($1 million)
.032
Fringe metropolitan ($1 million)
916 (22.7)
37 (14.7)
879 (23.2)
Metropolitan (250,000-999,999)
932 (23.1)
73 (29.3)
859 (22.7)
Micropolitan (<25,000)
1189 (29.5)
50 (19.9)
1140 (30.1)
3422 (84.6)
200 (80.4)
3222 (84.9)
Insurance Private/Medicare Medicaid/other
.28 621 (15.4)
49 (19.6)
572 (15.1)
Lowest annual income (<$38,000)
1045 (26.1)
68 (27.9)
978 (26)
Patient resides in same state as hospital
3628 (89.7)
235 (94.3)
3394 (89.4)
464 (11.5)
0
464 (12.2) 1089 (28.7)
1674 (41.4)
56 (22.6)
1617 (42.6)
814 (20.1)
188 (75.7)
626 (16.5)
5 (0.2)
4.3 (0.1)
.002
2263 (56)
71 (28.6)
2192 (57.7)
<.0001
255 (6.3)
51 (20.5)
204 (5.4)
<.0001
1527 (37.7)
127 (50.9)
1400 (36.9)
.014
576 (14.2)
42 (16.9)
534 (14.1)
.49
52 (1.3)
11 (4.5)
41 (1.1)
150 (3.7)
12 (4.7)
139 (3.7)
1094 (27)
Major Elixhauser score, mean (SE)
.20
4 (1.7)
Moderate Extreme
.66 <.0001
DRG severity Minor
.43
4.3 (0)
Indication at initial presentation Aneurysms without rupture Aneurysms with rupture Dissections Concomitant procedures Proximal debranching Iliac access
.003
Hospital Bed size Small Medium Large
.86 28 (11)
416 (11)
3451 (85.3)
443 (11)
210 (84.3)
3242 (85.4)
722 (17.9)
62 (25.1)
660 (17.4)
3274 (80.9)
186 (74.9)
3088 (81.3)
48 (1.2)
0
48 (1.3)
219 (88.1)
1861 (49.1)
Teaching status Metropolitan nonteaching Metropolitan teaching Not metropolitan Urgent/emergent procedure
.25
2080 (51.5)
<.0001
DRG, Diagnosis-related group; SE, standard error. All values are weighted frequency and percentages. The number is the total sample size for the respective cohort. Values are reported as number (%) unless otherwise indicated.
Journal of Vascular Surgery Volume
-,
Number
Kalesan et al
11.e2
-
Supplementary Table II (online only). Baseline patient and hospital characteristics for 90-day mortality Characteristics
Overall (N ¼ 4045)
Dead at 90 days (n ¼ 309)
Alive at 90 days (n ¼ 3736)
P value
Patient Age, years, mean (SE)
67.7 (0.4)
69.9 (1.4)
Age, $60 years
3004 (74.3)
234 (77)
2766 (74)
67.5 (0.5)
.53
Male sex
2459 (60.8)
191 (61.7)
2268 (60.7)
.84
997 (24.7)
109 (35.3)
888 (23.8) 865 (23.2)
Location of patient’s residence Central metropolitan ($1 million)
.012
Fringe metropolitan ($1 million)
916 (22.7)
51 (16.5)
Metro (250,000-999,999)
932 (23.1)
93 (30.1)
839 (22.5)
Micropolitan (<25,000)
1189 (29.5)
56 (18.1)
1134 (30.4)
3422 (84.6)
246 (79.6)
3176 (85.1)
Insurance Private/Medicare Medicaid/other
.17 621 (15.4)
63 (20.4)
558 (14.9)
Lowest annual income (<$38,000)
1045 (26.1)
73 (24.3)
972 (26.3)
Patient resides in same state as hospital
3628 (89.7)
291 (94.1)
3337 (89.3)
.61 .16 <.0001
DRG severity Minor
.13
464 (11.5)
4 (1.2)
460 (12.3)
17 (5.4)
1077 (28.8)
Moderate
1094 (27)
Major
1674 (41.4)
68 (21.9)
814 (20.1)
221 (71.6)
593 (15.9)
4.3 (0)
5 (0.2)
4.2 (0.1)
2263 (56)
107 (34.7)
2156 (57.7)
<.0001
255 (6.3)
59 (19.2)
196 (5.2)
<.0001
1527 (37.7)
142 (46.1)
1385 (37.1)
.062
576 (14.2)
52 (16.9)
524 (14)
.44
52 (1.3)
11 (3.6)
40 (1.1)
.016
150 (3.7)
14 (4.6)
136 (3.6)
Extreme Elixhauser score, mean (SE)
1606 (43) .0005
Indication at initial presentation Aneurysms without rupture Aneurysms with rupture Dissections Concomitant procedures Proximal debranching Iliac access Hospital Bed size Small Medium Large
.77 38 (12.4)
405 (10.8)
3451 (85.3)
443 (11)
256 (82.9)
3195 (85.5)
722 (17.9)
72 (23.2)
651 (17.4)
3274 (80.9)
237 (76.8)
48 (1.2)
0
48 (1.3)
266 (86)
1815 (48.6)
Teaching status Metropolitan nonteaching Metropolitan teaching Not metropolitan Urgent/emergent procedure
.32
2080 (51.5)
3037 (81.3) <.0001
DRG, Diagnosis-related group; SE, standard error. All values are weighted frequency and percentages. The number is the total sample size for the respective cohort. Values are reported as number (%) unless otherwise indicated.
11.e3
Journal of Vascular Surgery
Kalesan et al
---
2018
Supplementary Table III (online only). Baseline patient and hospital characteristics for 180-day mortality Characteristics
Overall (N ¼ 4045)
Dead at 180 days (n ¼ 341)
Alive at 180 days (n ¼ 3704)
P value .0v65
Patient Age, years, mean (SE)
67.7 (0.4)
70.2 (1.4)
67.5 (0.5)
Age $60 years
3004 (74.3)
268 (78.6)
2736 (73.9)
.27
Male sex
2459 (60.8)
210 (61.5)
2249 (60.7)
.88
997 (24.7)
114 (33.5)
883 (23.9)
Location of patient’s residence Central metropolitan ($1 million)
.02
Fringe metropolitan ($1 million)
916 (22.7)
57 (16.7)
859 (23.3)
Metro (250,000-999,999)
932 (23.1)
103 (30.1)
830 (22.5)
Micropolitan (<25,000)
1189 (29.5)
67 (19.7)
1122 (30.4)
3422 (84.6)
278 (81.5)
3144 (84.9)
Insurance Private/Medicare Medicaid/other
.37 621 (15.4)
63 (18.5)
558 (15.1)
Lowest annual income (<$38,000)
1045 (26.1)
80 (23.8)
966 (26.4)
Patient resides in same state as hospital
3628 (89.7)
323 (94.7)
3305 (89.2)
Minor
464 (11.5)
Moderate
1094 (27)
Major
1674 (41.4)
Extreme Elixhauser score, mean (SE)
.1 <.0001
DRG severity
814 (20.1)
4 (1.1)
460 (12.4)
19 (5.5)
1075 (29)
82 (24.1)
1592 (43)
236 (69.3)
578 (15.6)
4.3 (0)
5.1 (0.2)
4.2 (0.1)
2263 (56)
118 (34.5)
2146 (57.9)
<.0001
255 (6.3)
67 (19.7)
188 (5.1)
<.0001
1527 (37.7)
156 (45.8)
1371 (37)
.071
576 (14.2)
66 (19.4)
510 (13.8)
.13
52 (1.3)
13 (3.8)
39 (1)
150 (3.7)
14 (4.2)
136 (3.7)
.0001
Indication at initial presentation Aneurysms without rupture Aneurysms with rupture Dissections Concomitant procedures Proximal debranching Iliac access
.006
Hospital Bed size Small Medium Large
.94 38 (11.3)
405 (10.9)
3451 (85.3)
443 (11)
288 (84.5)
3163 (85.4)
722 (17.9)
76 (22.3)
646 (17.4)
3274 (80.9)
265 (77.7)
3010 (81.2)
48 (1.2)
0
48 (1.3)
295 (86.6)
1785 (48.3)
Teaching status Metropolitan nonteaching Metropolitan teaching Not metropolitan Urgent/emergent procedure
.35
2080 (51.5)
<.0001
DRG, Diagnosis-related group; SE, standard error. All values are weighted frequency and percentages. The number is the total sample size for the respective cohort. Values are reported as number (%) unless otherwise indicated.