Abstract
Objective:
Underinsured patients can experience worse preoperative medical optimization. We aimed to determine whether insurance status was associated with carotid endarterectomy (CEA) urgency and postoperative outcomes.
Methods:
We analyzed the Society for Vascular Surgery Vascular Quality Initiative Carotid Endarterectomy dataset from January 2012 to January 2021. Univariable and multivariable methods were used to analyze the differences across the insurance types for the primary outcome variable: CEA urgency. The analyses were limited to patients aged <65 years to minimize age confounding across insurers. We also examined differences in preoperative medical optimization and symptomatic disease and postoperative outcomes. A secondary analysis was performed to examine the effect of CEA urgency on the postoperative outcomes.
Results:
A total of 27,331 patients had undergone first-time CEA. Of these patients, 4600 (17%) had Medicare, 3440 (13%) had Medicaid, 17,917 (65%) had commercial insurance, and 1374 (5%) were uninsured. The Medicaid and uninsured patients had higher rates of urgent operation compared with Medicare (20.0% and 34.7% vs 14.4%; P < .001), with no differences in the commercial group vs the Medicare group. Additionally, Medicaid and uninsured patients had lower rates of aspirin, statin, and/or antiplatelet use (93.6% and 93.5% vs 95.8%; P < .001) and higher rates of symptomatic disease (42.1% and 57.6% vs 36.2%; P < .001) compared with Medicare patients. The rate of perioperative stroke/death was higher for the Medicaid and uninsured patients than for the Medicare patients (1.63% and 1.89% vs 1.02%; P = .017 and P = .01, respectively), with no differences in the commercial group. Multivariable analysis demonstrated that compared with Medicare, Medicaid and uninsured status were associated with increased odds of an urgent operation (odds ratio [OR], 1.3; 95% confidence interval [CI], 1.1-1.5; and OR, 2.3; 95% CI, 2.0-2.7, respectively), symptomatic disease (OR, 1.2; 95% CI, 1.1-1.4; and OR, 2.2; 95% CI, 1.9-2.5, respectively), and perioperative stroke/death (OR, 1.6; 95% CI, 1.1-2.4; and OR, 1.8; 95% CI, 1.1-3.0, respectively) and a decreased odds of aspirin, statin, and/or antiplatelet use (OR, 0.71; 95% CI, 0.6-0.9; and OR, 0.76; 95% CI, 0.6-0.99, respectively). Additionally, the rates of perioperative stroke/death were higher for patients who had required urgent surgery compared with elective surgery (2.8% vs 1.0%; P < .001). Multivariable analysis demonstrated increased odds of perioperative stroke/death for patients who had required urgent surgery (OR, 2.4; 95% CI, 1.9-3.1).
Conclusions:
Medicaid and uninsured patients were more likely to require urgent CEA, in part because of poor preoperative medical optimization. Additionally, urgent operation was independently associated with worse postoperative outcomes. These results highlight the need for improved preoperative follow-up for underinsured populations.
Keywords: Carotid artery disease, Carotid endarterectomy, Insurance status, Medicaid
Health insurance is a key social determinant of health and is critical for access to both preventative care and intervention. Uninsured patients are less likely to seek preventative care and undergo surgical interventions and will experience worse overall health outcomes.1,2 Insured patients are more likely to receive preventative care and be medically optimized for cardiovascular comorbidities.3,4 Specifically, for carotid artery disease, uninsured and minority patients are less likely to receive carotid revascularization for asymptomatic disease.5 A recent study using the Nationwide Inpatient Sample showed that patients in centers with a >40% racial minority population had higher rates of Medicaid status and were less likely to undergo revascularization even after ischemic stroke due to carotid disease.6 In addition, both Medicaid and uninsured patients have been less likely to receive care at high-volume centers and are more likely to experience delays in referrals for surgery.7,8
Previous work has also elucidated an association between insurance status and the acuity of presentation for those with vascular disease. Underinsured patients have been demonstrated to be more likely to present with a requirement for nonelective thoracic aortic repair, which carries adverse postoperative outcomes.9,10 A study from 2008 using hospital discharge databases in Florida and New York showed that uninsured and Medicaid patients had a higher risk of presenting with ruptured abdominal aortic aneurysms and symptomatic carotid artery disease.11 However, few studies have used nationwide data in more recent years to examine the effect of insurance on both the acuity of presentation and the postoperative outcomes after carotid revascularization. In the present study, we used a robust database to examine the interrelations among insurance status, preoperative medical optimization, operative acuity, and postoperative outcomes. We hypothesized that Medicaid and uninsured patients would be less likely to be receiving optimal medical therapy for carotid disease and would experience a greater risk of requiring nonelective carotid endarterectomy (CEA) and worse postoperative outcomes.
METHODS
Database and study population.
The Vascular Quality Initiative represents a multidisciplinary collaboration of >900 participating centers across the United States and Canada. We included all the patients who had undergone CEA from January 2012 to January 2021. Patients with prior CEA or carotid artery stenting, with military or non-U.S. insurance, and those who had undergone concomitant coronary artery bypass grafting (CABG) were excluded. Finally, because of the inherent age differences between Medicare and Medicaid patients, we limited our analysis to patients aged <65 years.
Patient factors, surgical characteristics, and hospital volume.
The patient characteristics were analyzed as either continuous or categorical variables and included age, sex (male vs female), race, ethnicity (Hispanic vs non-Hispanic), World Health Organization body mass index classification, ambulatory status (ie, ambulatory, ambulatory with assistance, nonambulatory), and living status (home vs nursing home). Race was categorized as White, Asian, Black, other (Native American, Hawaiian Native, Pacific Islander, Alaskan Native), or unknown. Patient comorbidities included hypertension, diabetes, congestive heart failure, chronic obstructive pulmonary disease, prior percutaneous coronary intervention or CABG in the previous 5 years, and end-stage renal disease (ESRD; defined as a preoperative dialysis requirement).
We also recorded preoperative aspirin, statin, and/or platelet inhibitor use (ie, clopidogrel, prasugrel, ticlopidine, ticagrelor), the presence of contralateral carotid occlusion, and the presence of symptomatic disease (defined as ipsilateral transient ischemic attack or stroke). Laboratory tests included preoperative hemoglobin and creatinine. Surgical characteristics included CEA type (conventional vs eversion), shunt use, patch use, and operative time. Hospital volume was assessed by stratification into quintiles according to the average annual volume of procedures.
Outcomes.
The main exposure of interest in the present study was insurance status at the time of procedure, which was categorized into four groups: Medicare, Medicaid, commercial, and uninsured. Our primary outcome variable was CEA urgency, which was defined as elective vs nonelective. Nonelective procedures included those classified as both urgent (within 24 hours of admission) and emergent (within 6 hours of admission). The secondary outcome variables included preoperative aspirin, statin, or platelet inhibitor use, symptomatic disease (preoperative ipsilateral transient ischemic attack or stroke in the previous 6 months), preoperative Rankin score (none or slight disability vs moderate or severe disability), perioperative stroke/death (any ipsilateral in-hospital stroke or any death <30 days), postoperative myocardial infarction (MI), postoperative length of stay (LOS) >2 days, and any return to the operating room during the admission. Stroke was defined as any new motor or sensory loss or speech abnormality or new documented neurologic symptoms lasting ≥24 hours during the admission. The postoperative LOS was the number of days between the index operation and discharge. A secondary analysis was performed to investigate the effect of CEA urgency on perioperative stroke/death, postoperative MI, LOS, and reoperation.
Statistical analysis.
Continuous variables are reported as the median and interquartile range and categorical variables as counts and percentages. To assess the differences across insurance categories in the univariate analyses of continuous and categorical variables, we used the Mann-Whitney U test and χ2 test, respectively.
The following outcome variables, CEA urgency, symptomatic disease, Rankin score, and preoperative aspirin, statin, and/or platelet inhibitor use, were assessed with the following covariates selected a priori for input into a stepwise forward selection model: age, sex, race, ethnicity, body mass index class, ambulatory status, living status, smoking status, hypertension, diabetes, prior percutaneous coronary intervention/CABG in the previous 5 years, congestive heart failure, chronic obstructive pulmonary disease, and ESRD. Additionally, the covariates for preoperative aspirin, statin, and/or platelet inhibitor use and symptomatic disease were included when these were not the outcome variables of interest. When assessing the postoperative outcome variables, additional covariates added to the stepwise forward selection model included hospital volume, CEA type (conventional vs eversion), shunt and patch use, operative time, preoperative hemoglobin and creatinine, and contralateral carotid occlusion. The criteria for inclusion in the final multivariable logistic regression model was based on the stepwise forward selection method with a cutoff value of P = .1. When assessing the postoperative outcome variables, a mixed effects multivariable logistic regression was used, clustering around the surgeon and hospital to account for random effects. The results of the multivariate logistic regressions are presented as odds ratios (ORs), with 95% confidence intervals (CIs). Statistical significance was defined as P < .05. We did not adjust for multiple comparisons in accordance with previously described methods.12,13 Statistical analyses were performed using Stata, version 17 (StataCorp LLC, College Station, TX). The NewYork-Presbyterian Columbia University Irving Medical Center institutional review board exempted the present study from the requirement for review and approval.
RESULTS
We identified 27,331 patients who had undergone first-time CEA and met our inclusion and exclusion criteria (Table I). Of the 27,331 patients, 4600 (17%) had Medicare insurance, 17,917 (66%) had commercial insurance, 3440 (12%) had Medicaid, and 1374 (5%) were uninsured. The median age across all insurance groups was 60 years and was lower (58 years) in the Medicaid and uninsured groups. Of the total population, 88% of the patients were White, with the non-White proportion higher in the Medicaid (20%) and uninsured (16%) groups compared with the Medicare and commercially insured groups (13% and 10%, respectively). The rate of comorbidities (ie, hypertension, diabetes, congestive heart failure, chronic obstructive pulmonary disease, and ESRD) was highest in the Medicare population. Among the insurer types, the differences across all patient characteristics, apart from intraoperative patch use, were statistically significant (Table I).
Table I.
Patient demographics stratified by insurance status
| Insurer |
||||||
|---|---|---|---|---|---|---|
| Variable | All patients (N = 27,331) |
Medicare (n = 4600) |
Commercial (n = 17,917) |
Medicaid (n = 3440) |
Uninsured (n = 1374) |
P value |
| Age, years | 60 (56-62) | 60 (57-63) | 60 (56-62) | 58 (54-61) | 58 (53-61) | <.001 |
| Sex | <.001 | |||||
| Male | 16,216 (59.3) | 2579 (56.1) | 10,986 (61.3) | 1834 (53.3) | 817 (59.5) | |
| Female | 11,115 (40.7) | 2021 (43.9) | 6931 (38.7) | 1606 (46.7) | 557 (40.5) | |
| Race | <.001 | |||||
| White | 24,061 (88.1) | 3985 (86.7) | 16,160 (90.2) | 2762 (80.4) | 1154 (84.2) | |
| Asian | 272 (1.0) | 21 (0.5) | 182 (1.0) | 56 (1.6) | 13 (0.9) | |
| Black | 1867 (6.8) | 421 (9.2) | 937 (5.2) | 379 (11.0) | 130 (9.5) | |
| Other | 152 (0.6) | 37 (0.8) | 78 (0.4) | 30 (0.9) | 7 (0.5) | |
| Unknown | 958 (3.5) | 133 (2.9) | 549 (3.1) | 210 (6.1) | 66 (4.8) | |
| Ethnicity | <.001 | |||||
| Non-Hispanic | 26,082 (95.8) | 4380 (95.5) | 17,219 (96.5) | 3209 (93.5) | 1274 (93.5) | |
| Hispanic | 1139 (4.2) | 208 (4.5) | 620 (3.5) | 222 (6.5) | 89 (6.5) | |
| BMI class | <.001 | |||||
| Underweight (<18.5 kg/m2) | 590 (2.2) | 129 (2.8) | 314 (1.8) | 103 (3.0) | 44 (3.2) | |
| Normal (18.5-24.9 kg/m2) | 6188 (22.6) | 1020 (22.2) | 3800 (21.2) | 984 (28.6) | 384 (27.9) | |
| Overweight (25.0-29.9 kg/m2) | 9520 (34.8) | 1514 (32.9) | 6428 (35.9) | 1095 (31.8) | 483 (35.2) | |
| Obese (≥30 kg/m2) | 11,033 (40.4) | 1937 (42.1) | 7375 (41.2) | 1258 (36.6) | 463 (33.7) | |
| Living status | <.001 | |||||
| Home | 27,041 (99.2) | 4503 (98.2) | 17,801 (99.5) | 3371 (98.3) | 1366 (99.6) | |
| Nursing home | 230 (0.8) | 83 (1.8) | 81 (0.5) | 60 (1.7) | 6 (0.4) | |
| Contralateral carotid occlusion | 1651 (6.0) | 327 (7.1) | 954 (5.3) | 257 (7.5) | 113 (8.2) | <.001 |
| Smoking status | <.001 | |||||
| Never/prior | 14,275 (52.3) | 2134 (46.5) | 10,337 (57.8) | 1289 (37.5) | 515 (37.6) | |
| Current smoker | 13,018 (47.7) | 2456 (53.5) | 7560 (42.2) | 2146 (62.5) | 856 (62.4) | |
| Hypertension | 23,236 (85.1) | 4069 (88.6) | 15,042 (84.0) | 2965 (86.3) | 1160 (84.7) | <.001 |
| Diabetes | 9869 (37.2) | 2110 (46.0) | 6030 (33.7) | 1282 (37.3) | 447 (32.6) | <.001 |
| Congestive heart failure | 2365 (8.7) | 682 (14.8) | 1219 (6.8) | 383 (11.1) | 81 (5.9) | <.001 |
| COPD | 6225 (22.8) | 1731 (37.7) | 3238 (18.1) | 1024 (29.8) | 232 (16.9) | <.001 |
| ESRD | 389 (1.4) | 239 (5.2) | 116 (0.6) | 30 (0.9) | 4 (0.3) | <.001 |
| Prior PCI/CABG | 4601 (16.8) | 931 (20.2) | 2866 (16.0) | 637 (18.5) | 167 (12.2) | <.001 |
| Ambulatory status | <.001 | |||||
| Ambulatory | 25,037 (92.0) | 3862 (84.4) | 16,857 (94.4) | 3039 (88.7) | 1279 (93.4) | |
| With assistance | 1799 (6.6) | 577 (12.6) | 833 (4.7) | 313 (9.1) | 76 (5.6) | |
| Nonambulatory | 389 (1.4) | 136 (3.0) | 165 (0.9) | 74 (2.2) | 14 (1.0) | |
| CEA type | .001 | |||||
| Conventional | 24,470 (89.9) | 4075 (89.1) | 16,098 (90.2) | 3039 (88.7) | 1258 (91.8) | |
| Eversion | 2757 (10.1) | 501 (10.9) | 1754 (9.8) | 389 (11.3) | 113 (8.2) | |
| Shunt use | 14,104 (51.7) | 2497 (54.5) | 9067 (50.7) | 1782 (51.9) | 758 (55.4) | <.001 |
| Patch use | 24,425 (89.6) | 4116 (89.9) | 16,024 (89.6) | 3064 (89.2) | 1221 (89.2) | .75 |
| Center volume | <.001 | |||||
| Lowest quintile | 5480 (20.1) | 941 (20.5) | 3480 (19.4) | 788 (22.9) | 271 (19.7) | |
| Highest quintile | 5087 (18.6) | 848 (18.4) | 3432 (19.2) | 594 (17.3) | 213 (15.5) | |
| Operative time, minutes | 112 (88-141) | 112 (87-141) | 111 (87-140) | 116 (90-146) | 112 (87-142) | <.001 |
| Preoperative hemoglobin, g/dL | 13.6 ± 1.9 | 13.2 ± 2.0 | 13.7 ± 1.8 | 13.3 ± 2.0 | 13.6 ± 1.9 | <0.001 |
| Preoperative creatinine, mg/dL | 1.0 ± 0.5 | 1.0 ± 0.6 | 1.0 ± 0.4 | 0.9 ± 0.5 | 0.9 ± 0.5 | <0.001 |
BMI, body mass index; CABG, coronary artery bypass graft; CEA, carotid endarterectomy; COPD, chronic obstructive pulmonary disease; ESRD, end-stage renal disease; PCI, percutaneous coronary intervention.
Data presented as median (interquartile range), number (%), or mean ± standard deviation.
On univariate analysis, Medicare patients had the lowest rate of nonelective CEA (14.4%), followed by the commercially insured group, and the difference was not statistically significant (14.6%; P = .647; Fig 1). Compared with the Medicare group, the rate of nonelective CEA was higher for the Medicaid and uninsured populations (20.0% and 34.7%, respectively, vs 14.4% for Medicare; P < .001). The rate of preoperative aspirin, statin, and/or platelet inhibitor use was highest in the Medicare group (95.8%), followed by the commercially insured group, with no statistically significant difference (95.2%; P = .130). Compared with the Medicare group, this rate was lower for the Medicaid and uninsured groups (95.8% vs 93.6% and 93.5%, respectively; P < .001). The rate of symptomatic disease was also lowest for the Medicare patients (36.2%). This was slightly higher in the commercially insured group (38.2% vs 36.2%; P = .012). However, the differences were greater for the Medicaid and uninsured population (42.1% and 57.6% vs 36.2%; P < .001). Additionally, the Medicaid and uninsured patients had a higher rate of moderate to severe disability using the Rankin score, although this difference was statistically significant for the Medicaid patients only (10.6% and 9.50% vs 8.73%; P = .008 and P = .404, respectively). The commercially insured patients had a lower rate of moderate to severe disability on the Rankin scale compared with the Medicare group (4.61% vs 8.73%; P < .001).
Fig 1.
Unadjusted rates of outcomes stratified by insurer. Medicaid and uninsured patients experienced higher unadjusted rates of nonelective carotid endarterectomy (CEA), symptomatic disease, and moderate to severe preoperative Rankin scores, with lower rates of preoperative aspirin, statin, and/or antiplatelet use compared with the Medicare patients. Medicaid and uninsured patients also experienced higher rates of adverse postoperative outcomes, including perioperative stroke/death, prolonged length of stay (LOS), and a return to the operating room. No significant differences were found in the incidence of postoperative myocardial infarction (MI).
On univariate analysis of the postoperative outcomes, the Medicare patients also had the lowest rate of perioperative stroke/death (1.02%; Fig 1). There was no statistically significant difference in the commerically insured group (1.25% vs 1.02%; P = .205). However, the rate of perioperative stroke/death was higher for the Medicaid and uninsured groups than for the Medicare group (1.63% and 1.89% vs 1.02%; P = .017 and P = .010, respectively). No statistically significant differences were found in the rate of postoperative MI across the insurance groups. The Medicaid and uninsured patients had a higher rate of a LOS >2 days compared with the Medicare group (20.3% and 22.9% vs 18.1%; P = .013 and P < .001, respectively). The rate of a LOS >2 days was lower for the commercially insured group (11.6% vs 18.1%; P < .001). The rate of a return to the operating room was also higher for the Medicaid population (2.65% vs 1.96%; P = .04). The reoperation rate was higher for the uninsured group, although the difference was not statistically significant (2.41% vs 1.96%; P = .307).
On multivariable analysis, after adjusting for confounders, an association remained between the insurer type and CEA urgency (Table II). Medicaid and uninsured status were associated with increased odds of nonelective CEA compared with Medicare as the referent (OR, 1.3; 95% CI, 1.1-1.5; P < .001; and OR, 2.3; 95% CI, 2.0-2.7; P < .001, respectively). The results of the full multivariable logistic regression analyzing CEA urgency are presented in the Supplementary Table (online only). Preoperative aspirin, statin, and/or antiplatelet use was independently associated with a lower risk of requiring nonelective CEA (OR, 0.4; 95% CI, 0.4-0.5; P < .001), and symptomatic disease was independently associated with an increased risk of requiring nonelective CEA (OR, 5.5; 95% CI, 5.1-6.0; P < .001; Supplementary Table, online only). We observed decreased odds of preoperative aspirin, statin, and/or platelet inhibitor use in the Medicaid and uninsured groups (OR, 0.71; 95% CI, 0.6-0.9; P = .001; and OR, 0.76; 95% CI, 0.6-0.99; P = .041). The Medicaid patients had 20% greater odds of having symptomatic disease and 40% greater odds of a moderate to severe Rankin score, and the uninsured patients had more than twofold greater odds of having symptomatic disease and 50% greater odds of a moderate to severe Rankin score (Table II). Commercially insured patients were less likely to have a worse Rankin score compared with the Medicare patients (OR, 0.73; 95% CI, 0.6-0.8; P < .001). The commercial insurance group did not differ from the Medicare group in the likelihood of nonelective CEA or preoperative aspirin, statin, and/or antiplatelet use. However, the former group was associated with a mildly increased risk of symptomatic disease (OR, 1.1; 95% CI, 1.0-1.2; P = .034).
Table II.
Multivariable analysis of outcomes stratified by insurance statusa
| Commercial |
Medicaid |
Uninsured |
||||
|---|---|---|---|---|---|---|
| Variable | OR (95% CI) | P value | OR (95% CI) | P value | OR (95% CI) | P value |
| Nonelective CEA | 1.0 (0.9-1.1) | .909 | 1.3 (1.1-1.5) | <.001 | 2.3 (2.0-2.7) | <.001 |
| Preoperative aspirin, statin, antiplatelet use | 0.90 (0.8-1.1) | .243 | 0.71 (0.6-0.9) | .001 | 0.76 (0.6-0.99) | .041 |
| Symptomatic disease | 1.1 (1.0-1.2) | .034 | 1.2 (1.1-1.3) | <.001 | 2.2 (1.9-2.5) | <.001 |
| Moderate to severe Rankin score | 0.73 (0.6-0.8) | <.001 | 1.4 (1.2-1.6) | <.001 | 1.5 (1.2-1.9) | .001 |
| Perioperative stroke/death | 1.4 (1.0-2.0) | .050 | 1.6 (1.1-2.4) | .023 | 1.8 (1.1-3.0) | .018 |
| Postoperative MI | 1.1 (0.6-1.7) | .878 | 1.3 (0.7-2.4) | .426 | 0.8 (0.3-2.2) | .613 |
| LOS >2 days | 0.7 (0.7-0.8) | <.001 | 1.2 (1.0-1.3) | .008 | 1.4 (1.2-1.6) | <.001 |
| Return to operating room | 1.1 (0.9-1.4) | .349 | 1.4 (1.0-1.9) | .031 | 1.3 (0.9-2.0) | .255 |
CEA, Carotid endarterectomy; CI, confidence interval; LOS, length of stay; MI, myocardial infarction; OR, odds ratio.
Boldface P values represent statistical significance.
Models were adjusted for age, sex, race, ethnicity, body mass index, ambulatory status, living status, preoperative aspirin, statin, and/or antiplatelet use,b symptomatic disease,b smoking history, hypertension, diabetes, prior percutaneous coronary intervention/coronary artery bypass grafting, congestive heart failure, chronic obstructive pulmonary disease, end-stage renal disease, CEA type, hospital volume, shunt use, patch use, operative time, preoperative hemoglobin and creatinine, and contralateral carotid occlusion.
Variables were included in the model when not the outcome of interest.
Regarding the postoperative outcomes on adjusted analysis, the Medicaid and uninsured patients had greater odds of perioperative stroke/death (OR, 1.6; 95% CI, 1.1-2.4; P = .023; and OR, 1.8; 95% CI, 1.1-3.0; P = .018, respectively). Uninsured status was associated with increased odds of postoperative LOS >2 days (OR, 1.4; 95% CI, 1.2-1.6; P < .001) and Medicaid was associated with greater odds of both a LOS >2 days (OR, 1.2; 95% CI, 1.0-1.3; P = .008) and a return to the operating room (OR, 1.4; 95% CI, 1.0-1.9; P = .031). No differences were found in the odds of postoperative MI across the insurance groups. Compared with Medicare, commercial insurance was not associated with any increased risk of adverse postoperative outcomes.
We performed a secondary analysis to determine the association between CEA urgency and postoperative outcomes (Fig 2). The rate of perioperative stroke/death was more than twofold greater for patients who had required nonelective CEA (2.79% vs 0.99%; P < .001). Nonelective CEA was also associated with a threefold greater rate of a prolonged LOS >2 days (31.6% vs 10.9%; P < .001) and twofold greater rate of a return to the operating room (3.62% vs 1.71%; P < .001; Fig 2). The rate of postoperative MI was unchanged in the nonelective group. These effects persisted despite controlling for confounders on multivariable analysis. Nonelective CEA was associated with a twofold greater risk of perioperative stroke/death and a return to the operating room and a threefold greater risk of a prolonged LOS >2 days (Table III).
Fig 2.
Unadjusted rates of outcomes stratified by carotid endarterectomy (CEA) urgency. Patients undergoing nonelective CEA experienced greater rates of perioperative stroke/death, prolonged length of stay (LOS), and a return to operating room. No significant differences were found in the incidence of postoperative myocardial infarction (MI).
Table III.
Multivariable analysis of outcomes stratified by carotid endarterectomy (CEA) urgencya
| Outcome | Nonelective CEA |
|
|---|---|---|
| OR (95% CI) | P value | |
| Perioperative stroke/death | 2.4 (1.9-3.0) | <.001 |
| Postoperative MI | 1.1 (0.7-1.8) | .713 |
| LOS >2 days | 3.5 (3.2-3.8) | <.001 |
| Return to operating room | 2.1 (1.7-2.6) | <.001 |
CI, Confidence interval; LOS, length of stay; MI, myocardial infarction; OR, odds ratio.
Boldface P values represent statistical significance.
Models were adjusted for age, sex, race, ethnicity, insurer, body mass index, ambulatory status, living status, preoperative aspirin, statin, and/or antiplatelet use, symptomatic disease, smoking history, hypertension, diabetes, prior percutaneous coronary intervention/coronary artery bypass grafting, congestive heart failure, chronic obstructive pulmonarydisease, end-stage renal disease, CEA type, hospital volume, shunt use, patch use, operative time, preoperative hemoglobin and creatinine, and contralateral carotid occlusion.
DISCUSSION
In the present study, we examined the association between payer status and the acuity of carotid artery disease at presentation and the subsequent effects on postoperative outcomes. Our analysis was limited to patients aged <65 years, primarily owing to the intrinsic age differences across the payer groups. Medicare patients will be inherently older, and we recognized age as a significant effect modifier, particularly for postoperative outcomes. Thus, we limited our analysis to minimize the confounding effects of age. We acknowledge that our findings should not be extrapolated to patients aged >65, because that older age stratification will also capture those Medicaid and commercially insured patients who have transitioned to Medicare insurance, causing sampling of our variables of interest to regress to the mean. Instead, our study used this younger cohort to examine the payer-driven disparities and the age-related effects were likely to be minimal. Given that the Vascular Quality Initiative provides data collected only at the operative intervention, we decided, a priori, to use CEA urgency, preoperative aspirin, statin, and/or antiplatelet use, symptomatic disease, and the Rankin score as proxies for preoperative follow-up quality. We found that Medicare patients had the lowest rates of requiring nonelective CEA, with no significant difference compared with the commercial insurance group. In contrast, the Medicaid and uninsured patients had significantly higher rates of requiring urgent or emergent CEA. These patients were less likely to have received optimization with appropriate medical therapy and were more likely to have had symptomatic disease and a greater degree of disability using the Rankin scale. The Medicare patients also had the lowest rate of perioperative stroke/death, with no significant difference compared with the commercially insured group. However, Medicaid and uninsured status carried a greater risk of perioperative stroke/death, in addition to a prolonged hospital LOS and reoperation. We also performed a secondary analysis to examine the clinical importance of CEA urgency. Even after controlling for insurer type and all other confounders, CEA urgency remained independently associated with a more than twofold greater risk of perioperative stroke/death and reoperation and a more than threefold greater risk of a prolonged LOS.
Previous studies have demonstrated that significant socioeconomic disparities exist regarding interventions for carotid artery disease. White race is an independent predictor of revascularization for carotid disease and minority patients are less likely to undergo carotid interventions even after sustaining an ischemic stroke event and have experienced worse outcomes likely owing to inequalities in socioeconomic status.6,14,15 Our finding that Medicaid and uninsured patients were more likely to have symptomatic disease and require urgent intervention suggests that payer type could be a proxy of socioeconomic status, because minority patients in the lower socioeconomic strata are more likely to be covered by Medicaid or uninsured altogether. In accordance with this, our baseline characteristics have demonstrated a greater proportion of non-White patients in the Medicaid and uninsured groups (20% and 16%, respectively) compared with the Medicare and commercially insured groups (13% and 10%, respectively).
This does not answer the question of the underlying reason why these patients have tended to present with higher acuity and more severe disease. It has been suggested that uninsured patients will be less likely to have appropriate outpatient follow-up and, thus, will be more likely to seek care in the emergency department, which is inadequate for appropriate screening or maintenance care.16 Further exacerbating this issue, both uninsured and Medicaid patients have been shown to receive fewer services in the emergency department and lower imaging relative value units, even when adjusted for the level of acuity.17 This suggests a bias toward lower valued interventions and imaging modalities for patients lacking insurance coverage. Even in an outpatient setting, Medicare and commercial insurance have been demonstrated to be protective against loss to follow-up for duplex ultrasound imaging for carotid stenosis patients.18 Our study has demonstrated a disparity between Medicare and commercially insured patients vs those with Medicaid or uninsured status regarding optimal medical therapy. It is likely that underinsured patients are less likely to seek appropriate preventative care before presentation, as evidenced by the greater rate of symptomatic disease and more severe preoperative Rankin scores in our study.
Other bodies of work have also examined the effect of insurance status on postoperative outcomes. Uninsured and underinsured patients have been shown to have worse perioperative outcomes after CEA, including mortality, stroke, and postoperative LOS.19-21 To date, the most comprehensive study examining the interplay among race, socioeconomic status, and payer type and their effects on postoperative outcomes after carotid procedures found disparities when stratified by race but not by insurance status.22 Two methodologic differences might explain the variations in our findings. First, the prior study had used data from the state inpatient databases of five states: California, Florida, Kentucky, Maryland, and New York. All but one of these states have incorporated Medicaid expansions, which might have ameliorated the disparities among the payer types. Our study used a robust nationwide clinical database that includes patients from 46 states, many of which have not implemented Medicaid expansions. Second, the prior study had included Medicare patients of all ages, many of whom were older and at a greater risk of adverse postoperative outcomes. Our study, instead, sought to capture disparities among insurance types that affect a younger cohort of patients.
Finally, our study also demonstrated that not only is Medicaid and uninsured status associated with an increased likelihood of a nonelective CEA, but also that CEA urgency is independently associated with worse postoperative outcomes. This is consistent with the general consensus in the literature that CEA should be performed within 2 weeks of neurologic symptoms to minimize the risk of recurrent stroke.23,24 However, more urgent carotid revascularization within 48 hours of symptom onset has been associated with an increased risk of perioperative stroke and death.25,26 In our study, nonelective CEA was defined as operative intervention within 48 hours of admission and, in some cases, even within 6 hours. Two explanations are likely for Medicaid and uninsured patients having higher rates of nonelective interventions. First, it is possible that these patients had presented with more advanced disease and were more likely to present with neurologically unstable or evolving symptoms. Second, these underinsured patients might have been seeking care via admission from the emergency department rather than through a conventional outpatient course. Regardless of the explanation, Medicaid and uninsured status are likely associated with more severe carotid artery disease, indirectly influencing the postoperative outcomes, or are potentially related to poorer outpatient follow-up for medical optimization prior to operative intervention. This was further evidenced in our study by the more severe Rankin scores for the Medicaid and uninsured cohorts.
The present study had several limitations. First, inherent differences were present in the patient characteristics across the insurer groups. We designated the Medicare group as the statistical comparator because these patients had demonstrated the lowest rates of key outcomes, including nonelective CEA and perioperative stroke/death. However, we recognize that for these patients to quality for Medicare, they must also have additional comorbidities. Therefore, in all multivariable analyses, we controlled for comorbidities including ESRD and disabilities. Also, despite the greater prevalence of underlying health issues in this population, we still observed worse outcomes for the Medicaid and uninsured groups. We also acknowledge that although the greater prevalence of comorbidities in the Medicare patients was likely valid owing to their qualification despite their younger age, this could have been confounded by a greater rate of diagnosis. Second, because of the large administrative nature of the database, inaccuracies could have been present in the diagnosis codes and timing of symptoms that influenced whether patients were classified as asymptomatic or symptomatic, which could have affected our results. Third, great variation exists among the states in the quality of Medicaid coverage owing to differences in how providers are reimbursed and their willingness to provide care to Medicaid patients. Although it was outside the scope of our study to examine these statewide differences, this will be an important next step to further elucidate the deficiencies in coverage across geographic regions.
CONCLUSIONS
We found that compared with Medicare, Medicaid and uninsured status were associated with a greater likelihood of urgent CEA, symptomatic disease, and worse Rankin score and a decreased likelihood of preoperative aspirin, statin, and/or antiplatelet use, which, taken together, are representative of worse preoperative follow-up. Nonelective CEA also resulted in a significantly greater risk of postoperative stroke/death, prolonged LOS, and reoperation. These findings suggest that payer status might be a mediator of the socioeconomic disparities seen among patients with carotid disease, especially in a younger cohort of patients. For the most disadvantaged patients, Medicaid alone might not be sufficient to compensate for other social barriers to access appropriate preventative and interventional care. In the present study, we have elucidated the interrelationship of payer type with both the quality of preoperative care and the postoperative outcomes. From these results, we have extrapolated a few actionable changes: (1) to increase the emphasis on preventative care in underinsured patients; (2) to ensure appropriate medical therapy for underinsured patients with carotid disease; and (3) to inform future investigation within Medicaid-expansion states to determine whether these disparities have resulted solely from the quality of insurance coverage or are also secondary to socioeconomic disparities.
Supplementary Material
ARTICLE HIGHLIGHTS.
Type of Research: A retrospective analysis of prospectively collected Vascular Quality Initiative data
Key Findings: Among patients undergoing carotid endarterectomy (CEA), when compared to Medicare, Medicaid and uninsured status increased the adjusted odds ratio (OR) of requiring a nonelective intervention (OR, 1.3; 95% confidence interval [CI], 1.1-1.5; and OR, 2.3; 95% CI, 2.0-2.7, respectively) and perioperative stroke and death (OR, 1.6; 95% CI, 1.1-2.4; and OR, 1.8; 95% CI, 1.1-3.0; respectively). Additionally, urgent CEA was independently associated with greater adjusted OR of perioperative stroke/death (OR, 2.4; 95% CI, 1.9-3.1).
Take Home Message: Medicaid and uninsured patients are more likely to require urgent CEA, which was also independently associated with worse postoperative outcomes.
Acknowledgments
P.C. is supported by the Columbia University Research Training in Cardiovascular Disease (National Institutes of Health T32 grant 5T32HL007854-27).
Footnotes
Author conflict of interest: none.
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.
Presented as a poster at the 2022 Eastern Vascular Society Thirty-sixth Annual Meeting, Philadelphia, PA, September 29 to October 1, 2022.
Additional material for this article may be found online at www.jvascsurg.org.
REFERENCES
- 1.Hadley J, Ayanian JZ, Butler S, Davis K, Kronick R. Sicker and poorer—the consequences of being uninsured: a review of the research on the relationship between health insurance, medical care use, health, work, and income. Med Care Res Rev 2003;60(Suppl): 3S–75S; discussion: 76S-112S. [DOI] [PubMed] [Google Scholar]
- 2.Hadley J Insurance coverage, medical care use, and short-term health changes following an unintentional injury or the onset of a chronic condition. JAMA 2007;297:1073–84. [DOI] [PubMed] [Google Scholar]
- 3.Gandelman G, Aronow WS, Varma R. Prevalence of adequate blood pressure control in self-pay or Medicare patients versus Medicaid or private insurance patients with systemic hypertension followed in a university cardiology or general medicine clinic. Am J Cardiol 2004;94:815–6. [DOI] [PubMed] [Google Scholar]
- 4.Yeh D, Jones M, Schulman C, Karmacharya J, Velazquez OC. Uninsured South Florida vascularsurgery patients are less likely to receive optimal medical management than their insured counterparts. J Vasc Surg 2010;51(Suppl):4S–8S. [DOI] [PubMed] [Google Scholar]
- 5.Brinjikji W, El-Sayed AM, Kallmes DF, Lanzino G, Cloft HJ. Racial and insurance based disparities in the treatment of carotid artery stenosis: a study of the Nationwide Inpatient Sample. J Neurointerv Surg 2015;7:695–702. [DOI] [PubMed] [Google Scholar]
- 6.Faigle R, Cooper LA, Gottesman RF. Lower carotid revascularization rates after stroke in racial/ethnic minority-serving US hospitals. Neurology 2019;92:e2653–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Adamson AS, Zhou L, Baggett CD, Thomas NE, Meyer AM. Association of delays in surgery for melanoma with insurance type. JAMA Dermatol 2017;153:1106–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Mabeza RM, Chervu N, Sanaiha Y, Hadaya J, Tran Z, de Virgilio C, et al. Demystifying the outcome disparities in carotid revascularization: utilization of experienced centers. Surgery 2022;172:766–71. [DOI] [PubMed] [Google Scholar]
- 9.Andersen ND, Brennan JM, Zhao Y, Williams JB, Williams ML, Smith PK, et al. Insurance status is associated with acuity of presentation and outcomes for thoracic aortic operations. Circ Cardiovasc Qual Outcomes 2014;7:398–406. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kalesan B, Cheng TW, Farber A, Zuo Y, Kalish JA, Jones DW, et al. Readmissions after thoracic endovascular aortic repair. J Vasc Surg 2018;68:372–82.e3. [DOI] [PubMed] [Google Scholar]
- 11.Giacovelli JK, Egorova N, Nowygrod R, Gelijns A, Kent KC, Morrissey NJ. Insurance status predicts access to care and outcomes of vascular disease. J Vasc Surg 2008;48:905–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Rothman KJ. No adjustments are needed for multiple comparisons. Epidemiology 1990;1:43–6. [PubMed] [Google Scholar]
- 13.Althouse AD. Adjust for multiple comparisons? It’s not that simple. Ann Thorac Surg 2016;101:1644–5. [DOI] [PubMed] [Google Scholar]
- 14.Cui CL, Yei KS, Janssen CB, Ramachandran M, Siracuse JJ, Malas MB. Differences in perioperative outcomes after carotid revascularization between white and non-white patients. Ann Vasc Surg 2022;79:31–40. [DOI] [PubMed] [Google Scholar]
- 15.Amaranto DJ, Abbas F, Krantz S, Pearce WH, Wang E, Kibbe MR. An evaluation of gender and racial disparity in the decision to treat surgically arterial disease. J Vasc Surg 2009;50:1340–7. [DOI] [PubMed] [Google Scholar]
- 16.Cunningham PJ. What accounts for differences in the use of hospital emergency departments across U.S. communities? Health Aff (Millwood) 2006;25:w324–36. [DOI] [PubMed] [Google Scholar]
- 17.Moser JW, Applegate KE. Imaging and insurance: do the uninsured get less imaging in emergency departments? J Am Coll Radiol 2012;9:50–7. [DOI] [PubMed] [Google Scholar]
- 18.Boelitz K, Jirka C, Eberhardt RT, Kalish JA, Siracuse JJ, Farber A, et al. Inadequate adherence to imaging surveillance and medical management in patients with duplex ultrasound-detected carotid artery stenosis. Ann Vasc Surg 2021;74:63–72. [DOI] [PubMed] [Google Scholar]
- 19.Schneider EB, Black JH, Hambridge HL, Lum YW, Freischlag JA, Perler BA, et al. The impact of race and ethnicity on the outcome of carotid interventions in the United States. J Surg Res 2012;177:172–7. [DOI] [PubMed] [Google Scholar]
- 20.Vogel TR, Kruse RL, Kim RJ, Dombrovskiy VY. Racial and socioeconomic disparities after carotid procedures. Vasc Endovascular Surg 2018;52:330–4. [DOI] [PubMed] [Google Scholar]
- 21.Fry DE, Nedza SM, Pine M, Reband AM, Huang CJ, Pine G. Risk-adjusted hospital outcomes in elective carotid artery surgery in patients with Medicare. Surgery 2018;163:606–11. [DOI] [PubMed] [Google Scholar]
- 22.Panchap L, Safavynia SA, Tangel V, White RS. Socioeconomic disparities in carotid revascularization procedures. J Cardiothorac Vasc Anesth 2020;34:1836–45. [DOI] [PubMed] [Google Scholar]
- 23.Coelho A, Peixoto J, Mansilha A, Naylor AR, de Borst GJ. Editor’s choice – timing of carotid intervention in symptomatic carotid artery stenosis: a systematic review and meta-analysis. Eur J Vasc Endovasc Surg 2022;63:3–23. [DOI] [PubMed] [Google Scholar]
- 24.Ricotta JJ, Aburahma A, Ascher E, Eskandari M, Faries P, Lal BK. Updated Society for Vascular Surgery guidelines for management of extracranial carotid disease: executive summary. J Vasc Surg 2011;54:832–6. [DOI] [PubMed] [Google Scholar]
- 25.Huang Y, Gloviczki P, Duncan AA, Kalra M, Oderich GS, DeMartino RR, et al. Outcomes after early and delayed carotid endarterectomy in patients with symptomatic carotid artery stenosis. J Vasc Surg 2018;67:1110–9.e1. [DOI] [PubMed] [Google Scholar]
- 26.Blay E, Balogun Y, Nooromid MJ, Eskandari MK. Early carotid endarterectomy after acute stroke yields excellent outcomes: an analysis of the procedure-targeted ACS-NSQIP. Ann Vasc Surg 2019;57:194–200. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.