Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Vasc Surg. Author manuscript; available in PMC 2009 October 1.
Published in final edited form as:
PMCID: PMC2582051

Insurance Status Predicts Access to Care and Outcomes of Vascular Disease



To determine if insurance status predicts severity of vascular disease at the time of treatment or outcomes following intervention.


Hospital discharge databases from Florida and New York from 2000–2005 were analyzed for lower extremity revascularization (LER, n=73,532), carotid revascularization (CR, n=116,578), or abdominal aortic aneurysm repair (n=35,593), using ICD-9 codes for diagnosis and procedure. The indications for intervention as well as the post-operative outcomes were examined assigning insurance status as the independent variable. Patients covered under a variety of commercial insurers, as well as Medicare, were compared to those who either had no insurance or were covered by Medicaid.


Patients without insurance or with Medicaid were at significantly greater risk of presenting with a ruptured abdominal aortic aneurysm (AAA) compared to insured (non-Medicaid) patients; while insurance status did not seem to impact post-operative mortality rates for elective and ruptured AAA repair. The uninsured or Medicaid recipients presented with symptomatic carotid disease nearly twice as often as the insured, but stroke rates after CR did not differ significantly based on insurance status. Patients with Medicaid or without insurance were more likely to present with limb threatening ischemia than claudication. In contrast to AAA repair and CR, the outcomes of LER were worse in the uninsured and Medicaid beneficiaries, who had higher rates of post-revascularization amputation compared to the insured (non-Medicaid) group.


Insurance status predicts disease severity at the time of treatment, but once treated, the outcomes are similar among insurance categories, with the exception of lower extremity revascularization. This data suggests inferior access to preventative vascular care in the Medicaid and the uninsured populations.


It is well-known that health insurance is a critical determinant of adequate access to preventative and therapeutic care. Research conducted in various fields of medicine and surgery has revealed disparities in the utilization of medical care, treatment of disease and outcome of treatment as a function of insurance status.17 Medicaid patients and the uninsured are much less likely to undergo CABG surgery or PTCA for coronary atherosclerosis and are at increased risk of in-hospital mortality after an MI.4, 7 It has also been demonstrated that Medicaid recipients and the uninsured are found to be in the more advanced stages of cancer at diagnosis and are at increased risk of perioperative mortality following colorectal surgery for carcinoma when compared to those with other insurance coverage.1, 3, 5, 6 Studies have also shown that uninsured individuals are less likely to have a regular source of care, such as a primary care physician, and instead use the emergency room to fill this role. The uninsured are less likely to use preventative services, less likely to fill prescriptions and ultimately have worse health outcomes.710 The purpose of this study was to further explore these disparities in the context of vascular disease.

There have been few studies that comprehensively examine the role of health insurance status in outcomes of patients with peripheral arterial disease. The factors involved in the development and progression of arterial pathologies are numerous and complex. However, it is probable that risk factor modification and early detection could reduce the severity of disease at presentation and improve the outcome of treatment. The ability to provide the proper services to patients such as education, screening, frequent monitoring and early intervention may depend on adequate access to care. The objectives of this analysis were to determine if insurance status predicts disease severity at the time of treatment for three common vascular diseases (abdominal aortic aneurysm (AAA), carotid stenosis and lower extremity occlusive disease), and to determine whether it predicts outcome following treatment.


Data sources

Databases for inpatient hospitalizations from New York State health department statewide planning and research cooperative system (SPARCS) and Florida state agency for healthcare administration, years 2000–2005, were queried for patient information including; primary and secondary diagnoses, and primary and secondary procedures by International Classification of Disease, 9th Revision, Clinical Modification (ICD-9-CM) codes. These are publicly available discharge databases which include all discharges in these two states over the six years specified. New York and Florida are populous states which adequately represent the national population distribution within insurance categories. These databases contain coded clinical and demographic information about each hospital discharge, including the ethnicity of the patient.

Patient population

Patients were designated into insurance groups by principle payer codes. Three groups were analyzed: insured (non-Medicaid), Medicaid and uninsured. The insured (non-Medicaid) group included patients with the following principle payer: Medicare, Medicare HMO, Medicaid HMO, commercial indemnity plan, commercial HMO, commercial PPO, Worker's compensation, C.H.A.M.P.U.S. (The Civilian Health and Medical Program of the Uniformed Services) Veterans Affairs, Blue Cross, other state/local government insurance and other federal programs. Medicaid HMO reimbursement is typically higher than Medicaid fee-for-service and more comparable to payers in the insured group. Moreover, HMOs are typically organized based on preventative care model which further distinguish the from non-HMO Medicaid. (To simplify presentation, we will refer to this group as insured.) Patients were designated as Medicaid beneficiaries if their principle payer was Medicaid or uninsured if their principle payer code was self pay. Patients with payer code designated no charge or charity were excluded from this analysis.

Treatment groups were identified by matching all relevant ICD-9 procedure codes with the related primary and secondary diagnosis codes. A complete list of procedure and diagnosis codes is available online (Table I, online only).The indication for, and the outcomes of three procedures were analyzed including abdominal aortic aneurysm repair, carotid revascularization (CR) and lower extremity revascularization (LER). Disease severity at time of treatment for patients undergoing AAA repair was categorized as ruptured (emergent) (441.3) or elective (441.4). We examined post-operative mortality of all AAA repairs and also stratified by disease severity. Patients who underwent carotid procedures were divided into two groups: symptomatic and asymptomatic. The symptomatic patients presented preoperatively with any one of the following diagnoses (in the primary or any secondary coding positions): occlusion and stenosis of carotid artery with cerebral infarction (433.11), occlusion and stenosis of multiple and bilateral arteries with cerebral infarction (433.31), unspecified transient cerebral ischemia (435.9), retinal vascular occlusion (362.3) and retinal ischemia (362.84). Asymptomatic patients presented with one of the following two diagnoses (primary or any secondary positions): occlusion and stenosis of carotid or multiple and bilateral arteries without mention of cerebral infarction respectively (433.10 or 433.30) and none of the symptomatic codes. We analyzed post-operative stroke (997.02) for all patients undergoing carotid revascularizations and also stratified groups by symptomatology. Patients who underwent LER with diagnoses of rest pain (440.22), ulceration (440.23 or 707.1) or gangrene (440.24 or 785.4) were included in the limb threatening ischemia group and were compared to patients with a diagnosis of claudication (440.21) alone. LER followed by major amputation during the same hospitalization were considered failed LER. The absence of major amputation (84.13- 84.17) following LER was considered a successful procedure.

Table I
Patient demographics and comorbidities and their association with insurance status.

We assessed the relationship between the following comorbidities and outcome: diabetes, hypertension, coronary artery disease, renal failure, emphysema and disorders of lipid metabolism. A list of ICD9 diagnosis codes representing comorbidities is available from our earlier publications.11, 12

Statistical analysis

Statistical analysis was performed using the SAS system software (SAS Institute Inc., Cary, NC). Univariate analyses of proportions was performed using the chi-square test or the Fisher exact test, where appropriate. Means were compared with the Student t-test. Statistical significance was expressed as both p-values and 95% confidence intervals (CI). Confidence intervals for proportions were calculated using normal approximation to the binomial distribution. Predictors of clinical manifestation of disease and major outcomes after surgical interventions were identified with multivariable logistic regression analysis. Consideration was given to the unequal age distribution across insurance groups and was adjusted for in our regression models. Patients' demographics and baseline comorbid conditions with significant univariate association to outcomes were included in the models. Results of the multivariate logistic regression are presented as odds ratios (OR) with the appropriate 95% CI. The insured group was chosen as the control.


We identified significant differences in demographics between the insurance groups undergoing the selected vascular interventions (Table 2). Patients without insurance were significantly younger compared to the insured group, with the largest difference found among LER patients (approximately 11 years younger). The percentage of blacks and Hispanics was highest in the Medicaid group (P<0.05), while the proportion of white non-Hispanics was significantly greater in the insured population (P<.0001).

Table II
Abdominal aortic aneurysm repair

We also identified significant differences in comorbidities between the insurance groups (Table 2). For all three arterial disorders, the prevalence of cardiovascular comorbidities, including hypertension and hyperlipidemia, was uniformly less in the uninsured versus the insured or the Medicaid populations. Our data also show a substantially higher proportion of diabetics among Medicaid patients. This was the case for all three arterial disorders, but most prominent in CR and LER.

Adjusting for patients' demographics and comorbidities we examined whether insurance status was an independent predictor of disease severity at time of treatment and of outcome of intervention. Table 3, Table 4 and 5 illustrate the outcomes of this analysis for abdominal aortic aneurysm repair, carotid revascularization and lower extremity revascularization, respectively.

Table III
Carotid revascularization
Table IV
Lower extremity revascularization

For aneurysmal disease, Medicaid recipients were found to be 1.68 times more likely to present with a AAA that had already ruptured, while uninsured patients were 4.17 times more likely to have ruptured prior to surgery. For Medicaid patients, the overall risk of post-operative death following AAA repair (both elective and emergent) was not significantly increased compared to insured patients; however, uninsured patients were 3.1 times more likely to die following AAA repair compared to insured patients (Table 3). When stratified by procedure type (elective or emergent, (Table 2)), the odds of post-operative death after elective AAA were not higher for the uninsured: OR=1.016, 95% CI (0.44–2.33), or the Medicaid population: OR=0.92, 95% CI (0.47–1.84) when compared to the insured.

Similarly, Medicaid patients and those without insurance had a higher probability of symptoms, such as stroke, transient ischemic attack or retinal artery occlusion, prior to carotid revascularization (OR=1.61, 95% CI (1.40–1.84), OR=1.93, 95% CI (1.63–2.30) respectively) without significantly increasing the risk of post-operative stroke (Table 2 and Table 3).

When analyzing lower extremity peripheral vascular disease, we found Medicaid recipients to be at the highest risk (OR=2.34, 95% CI (2.14–2.56)) of presenting with limb threatening ischemia, including rest pain, ulceration or gangrene, even higher than the uninsured (TABLE 5). Those with Medicaid insurance were also at the highest risk of having amputation following lower extremity revascularization (OR=1.52, 95% CI (1.34–1.74)). While uninsured status was also associated with higher likelihood of presenting with limb threatening ischemia (OR=2.23, 95% CI (1.88–2.65)), this group did not appear to have a higher risk of amputation following bypass (Table 2).

With regard to demographic characteristics and comorbidities we found across all three procedures, that renal failure and black ethnicity significantly predicted presentation with more severe disease, as well as worse outcome with intervention. Although Hispanics with Medicaid fared better than Hispanics without insurance; Hispanic ethnicity significantly increased the likelihood of presentation with advanced carotid and lower extremity disease, increased the risk of death following AAA repair and amputation following LER. Diabetes reduced the likelihood that an aneurysm was found to be ruptured at time of treatment (OR=0.81, 95% CI (0.72–0.91)) and reduced the risk of mortality following AAA repair (OR=0.85, 95% CI (0.75–0.98)). However, this disease significantly increased the likelihood that a limb's viability was threatened at time of treatment and increased the risk of amputation following LER. Both of these finding findings have been made by others. 1318


We chose to compared three groups based on insurance status, and analyzed three common arterial pathologies, (AAA, carotid and lower extremity occlusive disease), in order to determine if insurance status predicts disease severity at time of treatment and outcome following treatment. While our results with respect to insurance status do not definitively prove that Medicaid patients and the uninsured have worse access to vascular care, they support this hypothesis and add to insights derived from earlier studies.1921 While the findings are clear, the explanation for these findings remains to be fully determined. Indeed, the disparate results may be due to different factors for different pathological entities.

Among AAA patients, we found that Medicaid beneficiaries and the uninsured were at significantly higher risk of emergent repair than those with other coverage. In fact, the risk of presenting with rupture is two to three times as likely in the uninsured. One explanation for this finding is that inferior access to screening in these populations decreases early detection of AAA, and hence fewer elective repairs. Disease severity may be the best predictor of outcomes following AAA repair, as the mortality rate for emergent repair consistently approaches 50% compared to 5% or less for elective repair.22, 23 Therefore, it is not surprising that the uninsured, who are more likely to be treated for an already ruptured AAA, have an overall post-operative mortality rate more than double seen in patients with insurance. This finding may well be related to the high number of ruptured AAAs in the uninsured cohort and the anticipated poor outcome of the emergent procedure. The excess mortality and healthcare costs associated with the higher rate of emergent AAA repairs has not been quantified, however one might anticipate it to have a considerable public health impact. Of note, the uninsured, Medicaid and insured patients appear to have equivalent outcomes following treatment of AAA when the diagnosis is electively made. Boxer et al have previously reported a significant relationship between insurance status, aneurysm presentation and operative mortality using The Nationwide Inpatient Sample (NIS) database.24 Medicare beneficiaries were excluded from that analysis and consequentially, only 5,363 patients, all less than 65 years of age were included in their study. Similar to the findings reported here, the authors described an increased risk of rupture (OR=2.3, 95% CI (1.5–3.5)) in patients who were uninsured.24 However, they did not observe Medicaid insurance to be associated with an increased risk of rupture. While our results show that patients with Medicaid or no insurance are more likely to rupture and die from AAA it is also clear that such patients have the same potential as insured patients for favorable outcomes if they are treated in a timely manner.

In carotid occlusive disease, we found that Medicaid recipients and the uninsured have more strokes, transient cerebral ischemia and occular complications prior to receiving carotid revascularization. This finding of advanced symptomatology at time of treatment, among Medicaid and uninsured patients, may be due to poor access to preventative education regarding modifiable health factors, as well as lower rates of screening for asymptomatic carotid disease. Shen and Washington recently reported that carotid endarterectomy, in the absence of acute stroke symptoms, is more frequently performed in those with insurance coverage other than Medicaid, and concluded that access to preventative care is greater in those with private insurance.21 Proper access to screening and prophylactic intervention in the underinsured may significantly reduce strokes by allowing for treatment of asymptomatic patients with high-risk lesions. Although the risk of late or symptomatic presentation was increased in Medicaid beneficiaries and the uninsured, post-revascularization stroke risks were not significantly increased when compared to the insured cohort. Once again, when patients are detected prior to symptoms, they tend to have similar outcomes regardless of their insurance status.

Our analysis of LER revealed the Medicaid population to be at the highest risk for presenting with limb threatening ischemia prior to intervention and the most likely to require subsequent amputation. Conversely, those with insurance were most often treated for claudication as the indication for LER. The uninsured also present with a higher incidence of limb threatening ischemia, but there was no statistical difference in their risk of amputation following bypass compared to the insured. A recent study by Eslami et al using the NIS dataset, also found an association between payer, presentation and outcome of lower extremity peripheral vascular disease, but the uninsured were excluded in their analysis. This group also found that Medicaid beneficiaries presented more commonly with ischemic gangrene and were at 1.91 times the risk for primary amputation when compared to those on Medicare or privately insured.25 Although, in each of these analyses a significantly higher prevalence of diabetes was demonstrated within the Medicaid population, insurance status remained a predictor of limb threatening ischemia and amputation after controlling for coexisting comorbidities. The obvious explanation for the increased rate of amputation following revascularization in Medicaid patients is that these patients present with more advanced disease requiring more complex difficult revascularization and may have tissue loss too severe to permit limb salvage. The younger age associated with patients in the Medicaid group makes sense given the distribution of coverage based on age, however it remains a factor which may predict poorer outcomes. Patients presenting at younger age with limb threatening ischemia may have more severe disease and poorly controlled risk factors, thus making them less likely to have successful limb salvage.

Although discrepancies in the distribution of demographic variables and comorbidities existed between the insurance groups, as well as between the three arterial pathologies, certain trends were quite apparent: blacks and Hispanics were more likely to be uninsured or covered by Medicaid than white non-Hispanics. Moreover, black or Hispanic ethnicity was often associated with an increased risk of late presentation of disease or poor outcome. Interestingly, hypertension, hyperlipidemia or coronary disease were almost uniformly protective of poor outcomes and of advanced disease presentation. These diagnoses were most prevalent in patients with insurance coverage. The reason uninsured patients would have fewer cardiovascular risk factors is unclear, albeit one possibility is that these comorbidities existed but were undiagnosed in these patients. The increased prevalence of these three illnesses in the insured population may reflect two facts. First, in order to be diagnosed and coded for one of these comorbidities a patient must already have access to care, and second, effective treatment of these comorbidities ultimately protects patients; i.e. medications used to treat these conditions, such as beta-blockers and statins, are protective against progression of vascular disease and improve outcomes.26, 27 Therefore, the higher prevalence of these seemingly beneficial diagnoses in the insured is likely a surrogate for the greater access of these patients to medical care. These findings underscore the importance of insurance status for the delivery of preventative vascular care.

There are several limitations inherent in analyzing an administrative dataset. Although these datasets are maintained by the New York and Florida Departments of Health and capture all hospitalizations in these states, diagnosis codes are broad and vague and provide limited detail. Illness severity may be inconsistently recorded resulting in coding inaccuracies and oversights. For example, in our analysis of carotid occlusive disease, assigning symptomatic status or post-operative stroke prior to 2005 when the specific code was introduced was particularly difficult, yet it is probable that if errors in coding did occur, the occurred at similar frequencies across insurance groups. Another possible explanation for the implausible protective affect of certain comorbidities in our logistic regression analysis may be that comorbidities are coded less commonly in patients who die than in patients discharged alive.28 In addition, discharge databases lack physiologic information and do not provide insight regarding patient preferences for treatment options or compliance. Moreover, patient identifiers are typically absent, precluding longitudinal analyses. Given the large numbers of patients in these databases, it may be possible to detect significant differences simply because of large sample size. The use of multivariate analysis should eliminate the impact of detecting positive findings based solely on sample size. All of these factors limit the precision of analysis when using large datasets. Despite these limitations, these findings which remain strongly positive after multivariate analysis, can provide powerful evidence of trends and allow hypotheses to be developed and tested with more rigid scientific methods. In conclusion, our data support and expand on previously reported relationships between insurance status and disease severity at the time of treatment for aneurysmal, lower extremity, and carotid artery disease. There also appears to be a relationship between insurance status and amputation as the outcome of LER. Vascular patients with Medicaid or without insurance have increased mortality, pre-operative stroke and limb loss following LER. The advanced symptomatology and unfavorable outcomes in these patients, which are likely related to delayed diagnosis, are liable to negatively affect their quality of life and to result in substantially higher cost of care.

The public health implications surrounding these disparities in vascular disease are significant and preventable. Perhaps this paper may encourage efforts towards public education and improved awareness of warning signs related to aneurysmal, lower extremity and carotid diseases. Importantly this study illustrates that Medicaid insurance may not be adequate coverage for vascular patients and enrollment may not overcome the barriers to vascular care associated with being uninsured. This analysis raises important questions regarding the inequalities in levels of care given to those with and without insurance coverage.

Table thumbnail
line only. ICD9 procedure and diagnosis


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


1. Bradley CJ, Given CW, Roberts C. Disparities in cancer diagnosis and survival. Cancer. 2001 Jan 1;91(1):178–188. [PubMed]
2. Hadley J, Steinberg EP, Feder J. Comparison of uninsured and privately insured hospital patients. Condition on admission, resource use, and outcome. Jama. 1991 Jan 16;265(3):374–379. [PubMed]
3. Kelz RR, Gimotty PA, Polsky D, Norman S, Fraker D, DeMichele A. Morbidity and mortality of colorectal carcinoma surgery differs by insurance status. Cancer. 2004 Nov 15;101(10):2187–2194. [PubMed]
4. Kuykendall DH, Johnson ML, Geraci JM. Expected source of payment and use of hospital services for coronary atherosclerosis. Med Care. 1995 Jul;33(7):715–728. [PubMed]
5. Roetzheim RG, Pal N, Gonzalez EC, Ferrante JM, Van Durme DJ, Krischer JP. Effects of health insurance and race on colorectal cancer treatments and outcomes. Am J Public Health. 2000 Nov;90(11):1746–1754. [PubMed]
6. Roetzheim RG, Pal N, Tennant C, et al. Effects of health insurance and race on early detection of cancer. J Natl Cancer Inst. 1999 Aug 18;91(16):1409–1415. [PubMed]
7. Sada MJ, French WJ, Carlisle DM, Chandra NC, Gore JM, Rogers WJ. Influence of payor on use of invasive cardiac procedures and patient outcome after myocardial infarction in the United States. Participants in the National Registry of Myocardial Infarction. J Am Coll Cardiol. 1998 Jun;31(7):1474–1480. [PubMed]
8. Billings J, Teicholz N. Uninsured patients in District of Columbia hospitals. Health Aff (Millwood) 1990 Winter;9(4):158–165. [PubMed]
9. Hadley J. 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 Jun;60(2 Suppl):3S–75S. discussion 76S–112S. [PubMed]
10. 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 Mar 14;297(10):1073–1084. [PubMed]
11. Greco G, Egorova N, Anderson PL, et al. Outcomes of endovascular treatment of ruptured abdominal aortic aneurysms. J Vasc Surg. 2006 Mar;43(3):453–459. [PubMed]
12. Nowygrod R, Egorova N, Greco G, et al. Trends, complications, and mortality in peripheral vascular surgery. J Vasc Surg. 2006 Feb;43(2):205–216. [PubMed]
13. Arora S, Pomposelli F, LoGerfo FW, Veves A. Cutaneous microcirculation in the neuropathic diabetic foot improves significantly but not completely after successful lower extremity revascularization. J Vasc Surg. 2002 Mar;35(3):501–505. [PubMed]
14. Hertzer NR, Bena JF, Karafa MT. A personal experience with the influence of diabetes and other factors on the outcome of infrainguinal bypass grafts for occlusive disease. J Vasc Surg. 2007 Aug;46(2):271–279. [PubMed]
15. Kanagasabay R, Gajraj H, Pointon L, Scott RA. Co-morbidity in patients with abdominal aortic aneurysm. J Med Screen. 1996;3(4):208–210. [PubMed]
16. LaMorte WW, Scott TE, Menzoian JO. Racial differences in the incidence of femoral bypass and abdominal aortic aneurysmectomy in Massachusetts: relationship to cardiovascular risk factors. J Vasc Surg. 1995 Mar;21(3):422–431. [PubMed]
17. Lederle FA, Johnson GR, Wilson SE, et al. Prevalence and associations of abdominal aortic aneurysm detected through screening. Aneurysm Detection and Management (ADAM) Veterans Affairs Cooperative Study Group. Ann Intern Med. 1997 Mar 15;126(6):441–449. [PubMed]
18. Wong DR, Willett WC, Rimm EB. Smoking, hypertension, alcohol consumption, and risk of abdominal aortic aneurysm in men. Am J Epidemiol. 2007 Apr 1;165(7):838–845. [PubMed]
19. Mueller KJ, Patil K, Boilesen E. The role of uninsurance and race in healthcare utilization by rural minorities. Health Serv Res. 1998 Aug;33(3 Pt 1):597–610. [PMC free article] [PubMed]
20. Schoen C, DesRoches C. Uninsured and unstably insured: the importance of continuous insurance coverage. Health Serv Res. 2000 Apr;35(1 Pt 2):187–206. [PMC free article] [PubMed]
21. Shen JJ, Washington EL. Disparities in outcomes among patients with stroke associated with insurance status. Stroke. 2007 Mar;38(3):1010–1016. [PubMed]
22. Bown MJ, Sutton AJ, Bell PR, Sayers RD. A meta-analysis of 50 years of ruptured abdominal aortic aneurysm repair. Br J Surg. 2002 Jun;89(6):714–730. [PubMed]
23. Dueck AD, Kucey DS, Johnston KW, Alter D, Laupacis A. Survival after ruptured abdominal aortic aneurysm: effect of patient, surgeon, and hospital factors. J Vasc Surg. 2004 Jun;39(6):1253–1260. [PubMed]
24. Boxer LK, Dimick JB, Wainess RM, et al. Payer status is related to differences in access and outcomes of abdominal aortic aneurysm repair in the United States. Surgery. 2003 Aug;134(2):142–145. [PubMed]
25. Eslami MH, Zayaruzny M, Fitzgerald GA. The adverse effects of race, insurance status, and low income on the rate of amputation in patients presenting with lower extremity ischemia. J Vasc Surg. 2007 Jan;45(1):55–59. [PubMed]
26. Aziz IN, Lee JT, Kopchok GE, Donayre CE, White RA, de Virgilio C. Cardiac risk stratification in patients undergoing endoluminal graft repair of abdominal aortic aneurysm: a single-institution experience with 365 patients. J Vasc Surg. 2003 Jul;38(1):56–60. [PubMed]
27. Kertai MD, Boersma E, Klein J, et al. Optimizing the prediction of perioperative mortality in vascular surgery by using a customized probability model. Arch Intern Med. 2005 Apr 25;165(8):898–904. [PubMed]
28. Finlayson EV, Birkmeyer JD, Stukel TA, Siewers AE, Lucas FL, Wennberg DE. Adjusting surgical mortality rates for patient comorbidities: more harm than good? Surgery. 2002 Nov;132(5):787–794. [PubMed]