Abstract
The objective of this study is to examine the association between surgeon characteristics, procedural volume and short-term outcomes of hemodialysis vascular access. A retrospective cohort study was performed using Medicare Part A and B data from 2007 through 2014 merged with American Medical Association Physician Masterfile surgeon data. 29,034 procedures met the inclusion criteria: 22,541 (78%) AVF and 6,493 (22%) AVG. 13,110 (45.2%) were performed by vascular surgeons, 9,398 (32.3%) by general surgeons, 2,313 (8.0%) by thoracic surgeons, 1,517 (5.2%) by other specialties and 2,696 (9.3%) were unknown. Every 10-year increase in years in practice was associated with a 6.9% decrease in the odds of creating AVF vs AVG (P=0.02). Surgeon characteristics were not associated with likelihood of vascular access failure. Every 10-procedure increase in cumulative procedure volume was associated with a 5% decrease in the odds of vascular access failure (P=0.007). There was no association of provider characteristics or procedure volume with survival free of repeat AVF/AVG or TC placement at 12 months. A significant portion of the variability in likelihood of creating AVF vs AVG is attributable to the provider-level variation. Increase in procedure volume is associated with decreased odds of vascular access failure.
Introduction
Prior research on dialysis access outcomes has been primarily limited to traditional patient demographic factors (sex, race, presence/absence of co-morbidities such as diabetes, peripheral and coronary artery disease).1–3 Beyond such patient variables, potentially modifiable environmental factors that may influence surgical outcomes include surgeons’ specialty and procedure volume. For example, patients whose operations for rectal cancer are performed by colorectal surgeons have better long-term survival vs. those whose operations were by non-colorectal surgeons,4 and in vascular surgery, high surgeon procedure volume and specialization are similarly associated with lower mortality after intact abdominal aortic aneurysm repair.5,6
Some evidence suggests that increased surgical training in fistula creation improves maturation rates.7 In the US, surgical specialties that include training in dialysis access operations are primarily vascular and transplant surgery. However, availability of these two specialties varies widely regionally, and other surgeons, such as cardiac and general surgeons, may perform dialysis access surgery. The objective of this study is to examine the association between surgeon characteristics, procedural volume and short-term outcomes of hemodialysis vascular access in the United States.
Methods
Medicare Outpatient, Inpatient and Carrier files from 2007 through 2014 were queried for upper extremity vascular access procedures identified by the current procedural terminology (CPT) codes for native arteriovenous fistula and prosthetic arteriovenous graft (Table 1). Index fistula/graft creation was defined as the first occurrence of the CPT code per beneficiary during the study period. Repeat fistula/graft creation was defined as any new fistula/graft creation, identified by CPT code, performed in the twelve months subsequent to index fistula/graft, regardless of whether the index procedure was a fistula or a graft. Tunneled hemodialysis catheter (THC) placement procedures were identified by CPT code.
Table 1.
Current Procedural Terminology (CPT) codes and procedure descriptions
| CPT code | Procedure description |
|---|---|
| Fistula | |
| 36818 | arteriovenous anastomosis, open; by upper arm cephalic vein transposition |
| 36819 | arteriovenous anastomosis, open; by upper arm basilic vein transposition |
| 36820 | arteriovenous anastomosis, open; by forearm vein transposition |
| 36821 | arteriovenous anastomosis, opendirect, any site |
| 36825 | creation of arteriovenous fistula by other than direct arteriovenous anastomosis; autogenous graft |
| Graft | |
| 36830 | creation of arteriovenous fistula by other than direct arteriovenous anastomosis; non autogenous graft |
| Tunneled Catheter | |
| 36565 | Insertion of tunneled centrally inserted central venous access device, requiring 2 catheters via 2 separate venous access sites; without subcutaneous port or pump (e.g., Tesio type catheter) |
| 36558 | Insertion of tunneled centrally inserted central venous catheter, without subcutaneous port or pump; age 5 years or older |
Inclusion criteria were age ≥66 and dialysis dependent at time of index fistula/graft creation, qualification for Medicare exclusively by virtue of age, complete demographic and co-morbidity data, continuous enrollment in fee-for-service Medicare for twelve months before and after index fistula/graft creation. By excluding patients who qualified for Medicare by virtue of ESRD and limiting the patient population to those who were enrolled in Medicare for twelve months before index fistula/graft creation, we limit the cohort to patients who were undergoing first time dialysis vascular access creation to the best of our ability. Patients who died within twelve months after index fistula/graft creation were included. Co-morbidities were defined by having ever met the claims criteria of the Chronic Condition Data Warehouse as of the date of index fistula/graft creation. Chronic dialysis dependence was defined as at least one occurrence of dialysis services during the 12 months preceding index fistula/graft month or within 30 days after creation of index fistula/graft.8
Medicare data was merged with American Medical Association Physician Masterfile surgeon data using the National Provider Identifier (NPI) number. Surgeon practice type (office based, hospital staff, academic), specialty, gender and years in practice were obtained from the Physician Masterfile. Providers with vascular surgery listed as a specialty, even if the provider had more than one specialty, were categorized as a vascular surgeon. Providers with general surgery and a surgical specialty that did not include vascular surgery were categorized as the specialty.
Statistical analysis was conducted using Stata, version 13.1 (StataCorp, College Station, Texas). Likelihood of fistula/graft was analyzed using logistic regression, controlling for sociodemographic factors, co-morbidities and AVF/AVG. Logistic regressions of early failure were performed based on index fistula versus graft, controlling for the factors described in the various models. Early failure was defined as creation of another fistula/graft and/or THC placement in the 12 months following index fistula/graft. 12-month failure-free survival was defined by the patient surviving for 12 months after index fistula/graft creation without occurrence of another fistula/graft creation or THC placement. Survival analysis was performed. All models were adjusted for all available patient demographic, co-morbidity and zip code sociodemographic factors.
This research was deemed exempt from review by the Institutional Review Board as the study was retrospective, the Medicare data utilized was de-identified and there was no contact with human subjects for this research.
Results
29,034 procedures met the inclusion criteria: 22,541 (78%) AVF and 6,493 (22%) AVG. 13,110 (45.2%) were performed by vascular surgeons, 9,398 (32.3%) by general surgeons, 2,313 (8%) by thoracic surgeons, 1517 (5.2%) by other specialties and 2,696 (9.3%) were unknown. The mean age of patients who underwent graft was slightly older than patients who underwent fistula. (Table 2) A smaller proportion of females underwent fistula. (Table 2). Patients undergoing graft had higher incidence of diabetes, stroke, coronary artery disease and chronic obstructive pulmonary disease compared to patients undergoing fistula for all specialties. (Table 2) Surgeon characteristics also varied by specialty (Table 3).
Table 2.
Demographics and Co-morbidities by access type and specialty
| Vascular Surgery | General Surgery | Thoracic Surgery | Other | |||||
|---|---|---|---|---|---|---|---|---|
| Fistula (%) | Graft (%) | Fistula (%) | Graft (%) | Fistula (%) | Graft (%) | Fistula (%) | Graft (%) | |
| Age (mean) | 77 | 78 | 77 | 78 | 77 | 77 | 77 | 79 |
| Female | 5586 (44) | 2258 (59) | 2332 (46) | 845 (61) | 632 (46) | 247 (61) | 292 (43) | 140 (62) |
| Diabetes | 9943 (74) | 2957 (78) | 3716 (73) | 1063 (77) | 1085 (79) | 329 (81) | 509 (74) | 186 (82) |
| Stroke | 3738 (25) | 1221 (32) | 1196 (24) | 432 (31) | 342 (25) | 124 (31) | 178 (26) | 80 (35) |
| Ischemic Heart Disease | 10774 (81) | 3170 (83) | 3887 (76) | 1106 (84) | 1104 (80) | 349 (86) | 530 (7) | 191 (84) |
| Chronic Obstructive Pulmonary Disease | 5955 (45) | 1833 (48) | 2180 (43) | 639 (46) | 638 (46) | 181 (45) | 293 (43) | 111 (49) |
Table 3.
Provider characteristics by specialty
| Vascular Surgery | General Surgery | Thoracic Surgery | Other | |
|---|---|---|---|---|
| Mean Years in Practice (SD) | 22.3 (10.1) | 26.6 (9.2) | 26.9 (10.7) | 33.8 (9.0) |
| Practice type Office Based (%) Hospital Staff (%) |
11236 (86) 1239 (9) |
7429 (79) 1077 (11) |
1865 (81) 237 (10) |
* |
| Female (%) | 585 (4) | 309 (3) | 21 (1) | * |
| Mean yearly procedure volume (SD) | 12.3 (11.8) | 13.3 (14.2) | 9.9 (7.4) | 12.1 (8.1) |
data not available
Surgeon practice type, surgeon sex and procedure volume were not associated with odds of creating AVF vs AVG. General surgeons were 17% more likely than vascular surgeons to create AVF vs AVG (OR1.17, 95%CI 1.02–1.33, P=0.02). Every 10-year increase in years in practice was associated with a 6.9% decrease in the odds of creating AVF vs AVG (OR 0.93, 95%CI 0.88–0.99, P=0.02). When a random effect was included in the model for individual provider, 24.7% of the unexplained variability in the odds of creating AVF vs AVG was attributable to the individual provider. Surgeon characteristics were not associated with likelihood of vascular access failure. Every 10-procedure increase in cumulative procedure volume was associated with a 5% decrease in the odds of vascular access failure (OR 0.95, 95%CI 0.92–0.99, P=0.007). There was no association of practice type, surgeon sex or years in practice with survival free of repeat AVF/AVG or TC placement at 12 months. Thoracic surgery specialty had a 7% increase risk of loss of failure free survival that approached statistical significance (HR 1.07, 95%CI0.99–1.14, P=0.06). AVF vs AVG was the most significant factor in vascular access failure (OR 1.58, 95%CI 1.47–1.69, P<0.001) and loss of failure-free survival (HR 1.15, 95% CI 1.1–1.2, P<0.001).
Discussion
A growing body of literature exists regarding the association between patient factors including age, sex, vein diameter, co-morbidities and outcomes of hemodialysis vascular access operations.9–12 Despite an abundance of evidence in other specialties and even other vascular surgery operations that provider characteristics are associated with surgical outcomes, provider characteristics have not been well studied in association with outcomes of hemodialysis vascular access in the US.4–6 A previous study performed in 1995 demonstrated wide variation in outcomes of vascular access operations among individual surgeons, however, it did not investigate the specific surgeon factors that were associated with outcomes.13
Our analysis demonstrated a reduced risk of vascular access failure in association with increased surgeon vascular access procedure volume. This is consistent with findings in a number of other operations. In a study of abdominal aortic aneurysm repair, high volume surgeons had a 40% reduction in in-hospital mortality compared to low volume surgeons.5 Similarly, patients undergoing percutaneous coronary intervention by high-volume physicians had lower risk of requiring coronary artery bypass graft surgery during the index hospitalization and lower risk of 30-day mortality.14
We found that surgeon specialty was not associated with likelihood of access failure or failure-free survival. This may be due to the fact that we are unable to determine details regarding specific surgical training in vascular access operations. A previous study has demonstrated that a greater number of fistulas performed during training is associated with increased likelihood of fistula success.7 Specific operations performed in surgical training in the US can vary widely among various training programs. Our data demonstrates that in the Medicare population, vascular and general surgeons perform the great majority of vascular access operations. However, vascular access operations are not a defined category in vascular or general surgery training. As such, there is no minimum requirement of vascular access cases for vascular or general surgery trainees to perform. Interestingly, individual provider accounted for a significant amount of variation in access type. This would suggest that the same set of patient characteristics are interpreted differently by individual providers in determining the ideal access type for that patient.
There are a number of other limitations to this study. The most significant limitation is that it is an observational study. Although all readily available confounders measured in claims data were controlled for, it is possible that lower rates of access failure can be attributed to factors other than surgeon procedure volume that were unmeasured. The absolute measure of procedure volume is inaccurate in that not all cases performed by a provider are represented in the Medicare claims database. There is no single database that contains all of the dialysis access procedures performed by provider. Nevertheless, all dialysis dependent patients are eligible for Medicare coverage starting on the first day of their fourth month of dialysis with Medicare fee-for-service representing coverage of 75% of the end-stage renal disease patients in the United States. As such, the relative procedure volume using Medicare claims data is largely representative.
Using claims data, we are unable to identify whether the access was used. As such, the definition of failure is a surrogate measure. By including only dialysis-dependent patients in the cohort, we maximize the likelihood that we capture failure using creation of a new access or placement of a THC since dialysis-dependent patients must have vascular access. Nevertheless, using this method, patients who initiate hemodialysis with a THC, who then have a failed access, and do not have another access created or another THC placed would not be captured as failures. However, the population of patients who dialyze permanently through a THC is <10% and should represent a small under-estimation of access failure in this study.
Conclusion
A significant portion of the variability in likelihood of creating fistula vs graft is attributable to the provider-level variation. Increase in procedure volume is associated with decreased odds of vascular access failure. Future studies on vascular access outcomes should include extent of surgeon training and experience with vascular access.
References
- 1.Chan MR, Sanchez RJ, Young HN, Yevzlin AS. Vascular access outcomes in the elderly hemodialysis population: A USRDS study. Semin Dial. 2007;20(6):606–610. [DOI] [PubMed] [Google Scholar]
- 2.Lazarides MK, Georgiadis GS, Antoniou GA, Staramos DN. A meta-analysis of dialysis access outcome in elderly patients. J Vasc Surg. 2007;45(2):420–426. [DOI] [PubMed] [Google Scholar]
- 3.Lok CE, Oliver MJ, Su J, Bhola C, Hannigan N, Jassal SV. Arteriovenous fistula outcomes in the era of the elderly dialysis population. Kidney Int. 2005;67(6):2462–2469. [DOI] [PubMed] [Google Scholar]
- 4.Etzioni DA, Young-Fadok TM, Cima RR, et al. Patient survival after surgical treatment of rectal cancer: impact of surgeon and hospital characteristics. Cancer. 2014;120(16):2472–2481. [DOI] [PubMed] [Google Scholar]
- 5.Dimick JB, Cowan JA Jr., Stanley JC, Henke PK, Pronovost PJ, Upchurch GR Jr. Surgeon specialty and provider volumes are related to outcome of intact abdominal aortic aneurysm repair in the United States. J Vasc Surg. 2003;38(4):739–744. [DOI] [PubMed] [Google Scholar]
- 6.Tu JV, Austin PC, Johnston KW. The influence of surgical specialty training on the outcomes of elective abdominal aortic aneurysm surgery. J Vasc Surg. 2001;33(3):447–452. [DOI] [PubMed] [Google Scholar]
- 7.Saran R, Elder SJ, Goodkin DA, et al. Enhanced training in vascular access creation predicts arteriovenous fistula placement and patency in hemodialysis patients: results from the Dialysis Outcomes and Practice Patterns Study. Ann Surg. 2008;247(5):885–891. [DOI] [PubMed] [Google Scholar]
- 8.Clement FM, James MT, Chin R, et al. Validation of a case definition to define chronic dialysis using outpatient administrative data. BMC medical research methodology. 2011;11:25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Woodside KJ, Bell S, Mukhopadhyay P, et al. Arteriovenous Fistula Maturation in Prevalent Hemodialysis Patients in the United States: A National Study. Am J Kidney Dis. 2018;71(6):793–801. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Wilmink T, Corte-Real Houlihan M. Diameter Criteria Have Limited Value for Prediction of Functional Dialysis Use of Arteriovenous Fistulas. Eur J Vasc Endovasc Surg. 2018. [DOI] [PubMed] [Google Scholar]
- 11.Voorzaat BM, van der Bogt KEA, Janmaat CJ, van Schaik J, Dekker FW, Rotmans JI. Arteriovenous Fistula Maturation Failure in a Large Cohort of Hemodialysis Patients in the Netherlands. World journal of surgery. 2018;42(6):1895–1903. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Shah S, Leonard AC, Meganathan K, Christianson AL, Thakar CV. Gender and Racial Disparities in Initial Hemodialysis Access and Outcomes in Incident End-Stage Renal Disease Patients. Am J Nephrol. 2018;48(1):4–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Prischl FC, Kirchgatterer A, Brandstatter E, et al. Parameters of prognostic relevance to the patency of vascular access in hemodialysis patients. J Am Soc Nephrol. 1995;6(6):1613–1618. [DOI] [PubMed] [Google Scholar]
- 14.McGrath PD, Wennberg DE, Dickens JD Jr., et al. Relation between operator and hospital volume and outcomes following percutaneous coronary interventions in the era of the coronary stent. Jama. 2000;284(24):3139–3144. [DOI] [PubMed] [Google Scholar]