Background

The impact of the surgeon and surgical center characteristics on choice of autogenous arteriovenous (AV) fistula versus artificial AV graft as permanent vascular access for hemodialysis has not been studied.

Methods

We used national data from the Department of Veterans Affairs Veterans Health Administration to measure the association of surgeon and surgical center characteristics with choice of initial permanent vascular access among patients undergoing their first vascular access placement procedure between October 1, 2000 and September 30, 2001 (fiscal year 2001). Data were analyzed using a hierarchical logistic regression model clustered for surgical center and surgeon.

Results

The study population included 1114 patients, 74 Veterans Administration Medical Centers, and 182 surgeons. Seventy-two percent of patients received an AV fistula as their initial form of permanent vascular access. After adjusting for differences in patient, center, and surgeon characteristics, odds of AV fistula placement at high volume centers (>30 procedures per year) were more than three times greater than at low volume centers [odds ratio (OR) 3.26, 95% confidence interval (95% CI) 1.37 to 7.75, P = 0.008]. In addition, a strong clustering effect was present at the level of the surgeon (OR 1.55, 95% CI 1.19 to 2.03, P = 0.001) but not at the level of the surgical center, indicating an association with surgeon practice pattern.

Conclusion

Barriers to AV fistula placement can exist at the levels of the surgeon and surgical center, respectively. Future strategies to improve AV fistula placement rates should target surgeons and surgical centers in addition to patients, nephrologists, and primary care providers.

METHODS

Procedure code search

Our goal was to identify all initial (first in each patient's lifetime) dialysis vascular access placement procedures occurring in the Veterans Health Administration during fiscal year 2001 (October 1, 2000 to September 30, 2001) by procedure code search of central Veterans Health Administration inpatient and outpatient administrative databases¹⁰. We limited our search to initial procedures because choice of subsequent procedures is often complicated by details of patient vascular anatomy and clinical status that are not easily measured. We searched inpatient and outpatient databases for the first occurrence of any access-related Current Procedural Terminology (CPT) codes¹¹ (for outpatient procedures) and International Classification of Disease, 9th revision (ICD-9) codes¹² (for inpatient procedures) occurring from fiscal years 1994 to 2001 Table 1, selecting all first procedures for fiscal year 2001 as the study population. We adopted a broad search strategy in order to accommodate procedure miscoding and capture as many initial placement procedures as possible (based on the assumption that none of these codes should be applicable until the time of initial vascular access placement).

Table 1 - Terms used for procedure code search and codes assigned to procedures included in the study sample.

Full table

Chart review

Electronic medical records, including operative notes for all patients identified in procedure code search, were requested from Veterans Administration Medical Centers across the country using the Patient Data Exchange option in the Veterans Health Administration's computerized information system, known as the Veterans Health Information System Technology Architecture (VISTA)¹³. Once obtained, we used the medical record to validate procedure type (AV fistula vs. AV graft) and to definitively exclude patients who had undergone prior initial vascular access placement either within or outside the Veterans Health Administration before fiscal year 2001, those undergoing other procedures miscoded as vascular access placement procedures, and the small number of patients for whom the procedure date recorded in the medical record did not agree with the date recorded in the administrative database and fell outside the study period. We also used the medical record to determine patient comorbidity, age, gender, race, and dialysis status at the time of vascular access placement, and to identify the attending surgeon for the procedure. Chart review was conducted independently by two coauthors (A.O. and D.N.). Discrepancies were resolved by consensus.

Ascertainment of medical centers and surgeon characteristics

Hospital procedure volume was determined by counting the number of first vascular access procedures occurring at each medical center during the study period. This variable was divided into volume increments of 10 procedures per year to include the range of center volumes encountered. Centers were grouped by their location in one of the four census regions: Northeast, South, Midwest, or West. Centers with onsite chronic hemodialysis units were identified from information published on the Department of Veterans Affairs web site by Patient Care Services (http://www.va.gov/kidneydp).

Surgeon-specific procedure volume was determined by counting the number of first vascular access procedures each surgeon performed during the study period. This variable was divided into four categories to reflect the range of surgeon procedure volumes encountered. Surgeon year of medical school graduation and self-declared surgical specialty were obtained from the American College of Surgeons' (ACS) Web site (http://web.facs.org) or, for non-ACS members, from the relevant state medical board. Years of experience at study initiation was calculated by subtracting the year of medical school graduation from the year 2000 and was categorized into those with <10 years of experience (most surgeons do not reach attending level until 6 years after medical school graduation) and 5-year increments thereafter. Surgeon specialty was coded as vascular surgery versus other. Surgeon gender was determined based on surgeon first name. The study was approved by the Institutional Review Board for the University of California, San Francisco and the San Francisco Veterans Affairs Medical Center.

Statistical analysis

We examined the association of the outcome [type of initial procedure (i.e., AV fistula versus AVgraft)] with the surgeon and surgical center characteristics after adjusting for patient characteristics. In order to incorporate surgeon and surgical center level variables into the statistical model and to accommodate the possibility that procedures performed by a single surgeon or at a single medical center may not be independent, we used hierarchical logistic regression analysis clustered at the levels of the medical center and attending surgeon, respectively. To accommodate the two levels of clustering, we used the GENMOD procedure in SAS (SAS Institute, Inc., Cary, NC, USA) to perform alternating logistic regression analysis¹⁴. This method alternately fits a logistic regression equation to the responses and a model for the pair-wise odds ratios for observations in the same center. The model for the pair-wise odds ratios is structured to accommodate clustering at the level of the surgeon in addition to the surgical center. Several center and surgeon related variables (center and surgeon procedure volume, surgeons per center, and surgeon years of training) were entered into the model as categorical variables as described above in order to accommodate nonlinear relationships while still providing a concrete and easily interpretable description of the association of these variables with the outcome.

RESULTS

Study population

Figure 1 shows the derivation of the study population. Administrative database search of vascular access–related procedure codes yielded 2020 unique patients for fiscal year 2001. Medical records were returned for 1743 (86%) of these patients. Among these, we identified 1367 true first vascular access procedures, 139 patients who had undergone prior access placement either outside the Veterans Health Administration or prior to the search period (for whom the procedure listed was therefore not an initial procedure), and 237 patients who either underwent other procedures that were miscoded as dialysis access placement or for whom the chart-recorded date of first access placement occurred later than the date recorded in the administrative database and was after September 30, 2001 (the end of the study period). The attending surgeon could be determined for 82% (1114) of the 1367 patients undergoing initial vascular access placement. The study population consisted of these 1114 patients cared for at 74 Department of Veterans Affairs' Medical Centers by 182 individual surgeons. An operative note was available for 94% of initial vascular access procedures performed on study patients.

Figure 1.

Derivation of study sample. 

Full figure and legend (44K)

Patient, center, and surgeon characteristics

Table 2 shows the characteristics of the analytic sample. Mean patient age was 64 years, 99% were male, and 43% were African American. Fifty-eight percent had diabetes, 93% had a history of hypertension, 36% had coronary artery disease, 28% had a history of congestive heart failure, 24% had peripheral vascular disease, and 18% had cerebrovascular disease.

Table 2 - Characteristics of the study sample.

Full table

Of the 1114 patients in the study sample, 72% had an AV fistula as their initial permanent vascular access procedure. This figure ranged from 17% to 100% among medical centers and from 0% to 100% among surgeons. Table 1 shows how initial vascular access procedures were coded in the administrative database for study patients. Most inpatient initial permanent access placement procedures were coded under ICD-9 codes 39.27 (arteriovenostomy for renal dialysis) or 39.93 (insertion of a vessel-to-vessel cannula) and these codes were used for both AV fistulas and AV grafts. Most outpatient initial placement procedures were coded under CPT codes 36821 [AV anastomosis (cimino type) insertion], 36825 (creation of an AV fistula, autogenous graft insertion), and 36830 (creation of an AV fistula, nonautogenous graft insertion). CPT codes distinguished better than ICD-9 codes between initial AV fistula and AV graft placement in the study sample, but discrepancies between coding and chart abstraction were still present. Furthermore, some original placement procedures were miscoded as revision procedures.

Table 2 shows that only 51% of initial vascular access procedures took place prior to initiation of dialysis. By census region, the South included the largest number of medical centers performing access procedures. Most centers (67%) performed 20 initial vascular access procedures or fewer annually, but 11% performed over 30. The number of surgeons per center ranged from one to six with most (76%) having three or fewer. Most centers (73%) had a chronic dialysis unit on site. Most surgeons were male (92%) and most were vascular surgeons (58%). The level of surgical experience ranged from 10 years (16%) to more than 25 years (29%). Most surgeons (59%) performed five initial access procedures per year or fewer and only 4% performed more than 20.

Regression analysis

Table 3 shows the association of patient-, center-, and surgeon-related variables with AV fistula placement in multivariate hierarchical logistic regression analysis clustered at the levels of the surgical center and attending surgeon, respectively. Odds of AV fistula placement were over three times as high at high-volume centers (>30 procedures per year) compared to low-volume centers (10 procedures per year). The odds of AV fistula placement increased progressively by hospital volume category (P for trend = 0.012). A strong clustering effect was apparent at the level of the surgeon [odds ratio (OR) 1.55, 95% confidence interval (95% CI) 1.19 to 2.03, P = 0.001]. This finding suggests that there were differences in access placement patterns between individual surgeons that were not explained by differences in patient characteristics or in the specific surgeon or center characteristics examined in the model. On the other hand, no statistically significant surgical center effect was detected (OR 1.14, 95% CI 0.84 to 1.55, P = 0.40). Finally, at the patient level, female gender (OR 0.14, 95% CI 0.05 to 0.44, P < 0.001), African American race (OR 0.72, 95% CI 0.52 to 0.99, P = 0.046), and placement at the time of, or after initiation of dialysis (OR 0.70, 95% CI 0.51 to 0.97, P = 0.032), were associated with decreased odds of AV fistula placement.

Table 3 - Hierarchic logistic regression model for odds of arteriovenous (AV) fistula placement as the initial form of permanent vascular access for dialysis.

Full table

DISCUSSION

We report an association between surgeon practice pattern and surgical center volume with initial choice of permanent vascular access in a cohort of all veterans undergoing this procedure in a single fiscal year. The present analysis differs from prior national and international studies of determinants of permanent vascular access^2,3,6,7,8 in several respects. First, whereas previous studies have relied on data collected for individual patients at the dialysis unit level, we collected data for individual patients at the surgical center and surgeon levels, respectively. This approach allowed us to collect detailed data on the surgical procedure performed, on the operating surgeon, and on surgical center characteristics, in addition to baseline patient characteristics, whereas prior studies have focused primarily on the association of patient characteristics with type of access. While the Dialysis Outcomes and Practice Patterns Study (DOPPS) examined the association between choice of procedure and presence or participation of a surgical trainee in the vascular access procedure⁶, this analysis did not explore surgical center–related variables or the range of surgeon characteristics examined here and was restricted to surgeries performed on study patients rather than all surgeries performed by a particular surgeon or surgical team. Second, related to this last point, our study was not restricted to incident or prevalent dialysis patients, but included all patients undergoing initial vascular access placement procedures in the Veterans Health Administration during a single fiscal year [and therefore included both pre-end-stage renal disease (ESRD) and ESRD patients]. Finally, the outcome adopted for the present analysis (type of first access placed) is not strictly equivalent to the outcomes used in prior studies (these have included type of access in place^4,7, type of access in use^3,6,8, and type of access last used⁹).

Prior studies at single hospitals have drawn attention to the importance of the operating surgeon in determining survival of permanent vascular access devices^6,15,16 but an association of surgeon practice pattern and choice of vascular access type has not been previously reported. While no specific surgeon characteristics included in the model were associated with AV fistula placement, we found that individual surgeon practice pattern was a powerful determinant of type of permanent vascular access first placed. We do not know what it is about particular surgeons that could explain the observed association with AV fistula placement, except that this association is independent of other surgeon characteristics examined in the present analysis. Structural aspects of clinical practice such as better dissemination of guidelines within the surgical community and more communication between surgeons and nephrologists may be important, but our data cannot address this relationship.

This is the first study to report an association of vascular access type with hospital volume. Patients undergoing vascular access placement at centers performing>30 initial procedures annually were more than three times as likely to receive an AV fistula as those at centers performing 10 procedures per year. While many studies have demonstrated an inverse association between hospital volume and patient mortality across a variety of surgical procedures^17,18, little is understood about the processes of care in place at high volume centers that might explain this association¹⁹. The association of surgical volume with choice of vascular access procedure may reflect the existence of superior processes of care at high-volume centers. Alternatively, this association may reflect a "practice makes perfect" phenomenon, whereby surgeons who perform a greater number of procedures develop a greater level of skill. While we did not find an association between surgeon volume of vascular access procedures and choice of procedure, the present analysis does not exclude this possibility. Because many surgeons also hold non-Veterans Affairs appointments, volume of initial vascular access procedures performed by an individual surgeon in the present analysis may not indicate overall volume for that surgeon.

The overall rate of AV fistula placement (72%) reported here for the Veterans Health Administration far exceeds the K/DOQI target openface>50% AV fistula placement⁵. While Collins et al²⁰ have reported increasing overall rates of AV fistula placement among Medicare beneficiaries in recent years, there have been no equivalent national studies of rates of initial AV fistula placement that can be compared to the present study. Comparable rates have only been reported previously for single hospitals with successful policies to increase AV fistula placement^15,16,21,22. While a relatively high rate of AV fistula placement would be expected in this predominantly male population, the rate reported here for the Veterans Health Administration is nevertheless striking because the population studied here was older than the overall United States hemodialysis population and with a higher prevalence of comorbidities that may preclude placement of AV fistulas²³. The present study raises important questions about the respective roles of health care system-wide reimbursement and administrative practices (e.g., central Department of Veterans Affairs directives) in influencing patterns of vascular access placement. With respect to the potential impact of reimbursement practices, current Medicare provider fee schedules favor AV graft over AV fistula placement. For example, 2002 provider reimbursement for CPT code 36830 (AV graft) was almost 30% higher than for CPT code 36821 (AV fistula) ($704.44 versus $540.45) (http://www.cms.hhs.gov/providers/pufdownload/rvudown.asp). In addition, overall access-related costs under Medicare during the first year after AV graft placement are higher than after AV fistula placement²⁴. Within the Veterans Health Administration, it is certainly possible that the absence of provider incentives toward AV graft placement has facilitated the high AV fistula placement rate seen in the present study.

The present analysis also demonstrates that in a national cohort of patients undergoing initial permanent vascular access placement in the Veterans Health Administration, African American patients and women were less likely to have an AV fistula placed as their initial form of permanent vascular access. These results are consistent with prior studies outside the Veterans Health Administration showing both lower AV fistula prevalence³ and placement rates in these groups^25,26,27, although it is important to note that women represented only 1% of the sample in the present study. While the lower initial AV fistula placement rates reported here for women and African American patients (even after adjustment for timing of vascular access placement) might reflect gender and racial differences in vascular anatomy precluding higher levels of AV fistula placement in these groups^1,2,3, this could also reflect race- or gender-specific preferences (perhaps at either the patient, nephrologist, or surgeon levels) with respect to vascular access choice as reported for other surgical procedures^28,29.

Late placement of permanent vascular access at the time of or after initiation of dialysis was also associated with lower odds of AV fistula placement. Furthermore, while we found high Veterans Health Administration–wide rates of AV fistula placement, rates of predialysis permanent access placement were low in this cohort, with almost 50% of initial vascular access procedures occurring at the time of or after initiation of dialysis. These findings are consistent with data from outside the Veterans Health Administration showing low rates of pre-dialysis permanent vascular access placement³⁰ and identify timing of access placement as an important target area for efforts intended to improve the overall quality of dialysis vascular access care.

Sidawy et al³¹ have drawn attention to the need for uniform terminology in the field of vascular access for hemodialysis. The present analysis identifies several problems specifically with vascular access procedure coding. First, ICD-9 procedure code terminology does not include clearly separate codes for AV fistulas and AV grafts and thus cannot be reliably used alone to determine type of permanent vascular access. Second, while the CPT system does include separate terms for AV fistulas and AV grafts, in our data set, we noted substantial procedure miscoding under this system. While misclassification may be less of a concern outside the Veterans Health Administration (where procedure coding may be more accurate due at least in part to differences in reimbursement structure), our findings suggest a need for validation studies to measure the accuracy of procedure codes whenever administrative data are used to determine type of permanent vascular access.

Use of Department of Veterans Affairs data is both a strength and limitation of the present analysis. Due to the completeness of data capture within the Department of Veterans Affairs, we were able to include both inpatient and outpatient procedures. Many databases do not lend themselves to this kind of analysis because they do not include outpatient procedures. Also, in contrast to studies limited to administrative data, this analysis was able to identify procedure miscoding because detailed medical records were available electronically and were used to exclude miscoded cases from the analysis. Finally, unlike prior studies that have been limited to data collected at the level of the dialysis unit, we were able to collect data at the level of the surgical center and individual surgeon.

However, the present study has several limitations. First, the Veterans Health Administration represents a unique health care environment with a patient population and incentive structure that are not comparable to any other system and thus our results may not be generalizeable to other practice settings. However, several of our findings (i.e., lower odds of AV fistula placement in women and African American patients) are consistent with studies in non-veteran populations. Second, we were unable to control for obesity or body mass index (BMI) because these data could not be reliably collected from the medical record for most patients. This is important because African American patients initiating dialysis have a higher mean BMI than non-African American patients³² and, thus, the association of African American race with AV fistula placement reported here could reflect confounding by BMI. Finally, due to the limitations of our data, it was not possible to include in the analysis several potentially important aspects of surgical care that may influence choice of vascular access. For example, we could not consistently and reliably determine from the medical record whether a surgical trainee was present at a given procedure or their level of participation in that procedure. We also lacked data on surgeon place of residency training and the presence of a vascular access coordinator at each site. Collection of these data elements would be important in any future prospective study.

CONCLUSION

This study demonstrates high rates of initial AV fistula placement in the Veterans Health Administration far exceeding current K\DOQI targets. After adjusting for patient, surgeon, and center characteristics, individual surgeon practice pattern and high center volume of initial vascular access procedures were strongly associated with AV fistula (versus AV graft) placement.

These findings suggest that barriers to AV fistula placement can exist at the surgeon and surgical center levels as well as at the patient level and suggest the possibility that surgeons and surgical centers should be specifically targeted in efforts to increase AV fistula placement. Our findings also underline the need for efforts to promote timely placement of vascular access.

References

1. Allon, M, Robbin, ML: Increasing arteriovenous fistulas in hemodialysis patients: Problems and solutions. Kidney Int 2002 62: 1109–1124,  | Article | PubMed | ISI |
2. Reddan, D, Klassen, P, Frankenfield, DL, et al: National profile of practice patterns for hemodialysis vascular access in the United States. J Am Soc Nephrol 2002 13: 2117–2124,  | Article | PubMed | ISI |
3. Allon, M, Ornt, DB, Schwab, SJ, et al: Factors associated with the prevalence of arteriovenous fistulas in hemodialysis patients in the HEMO study. Hemodialysis (HEMO) Study Group. Kidney Int 2000 58: 2178–2185,  | Article | PubMed | ISI | ChemPort |
4. Hirth, RA, Turenne, MN, Woods, JD, et al: Predictors of type of vascular access in hemodialysis patients. JAMA 1996 276: 1303–1308,  | Article | PubMed | ISI | ChemPort |
5. Eknoyan, G, Levin, N: NKF-K/DOQI Clinical Practice Guidelines: Update 2000. Foreword. Am J Kidney Dis 2001 37: S5–S6,  | PubMed | ISI | ChemPort |
6. Pisoni, RL, Young, EW, Dykstra, DM, et al: Vascular access use in Europe and the United States: Results from the DOPPS. Kidney Int 2002 61: 305–316,  | Article |
7. Stehman-Breen, CO, Sherrard, DJ, Gillen, D, et al: Determinants of type and timing of initial permanent hemodialysis vascular access. Kidney Int 2000 57: 639–645,  | Article | PubMed | ISI | ChemPort |
8. Young, EW, Dykstra, DM, Goodkin, DA, et al: Hemodialysis vascular access preferences and outcomes in the Dialysis Outcomes and Practice Patterns Study (DOPPS). Kidney Int 2002 61: 2266–2271,  | Article | PubMed | ISI |
9. Reddan, D, Klassen, P, Frankenfield, DL, et al: National profile of practice patterns for hemodialysis vascular access in the United States. J Am Soc Nephrol 2002 13: 2117–2124,  | Article | PubMed | ISI |
10. Murphy, PA, Cowper, DC, Seppala, G, et al: Veterans Health Administration inpatient and outpatient care data: An overview. Eff Clin Pract 2002 5:E4,
11. KIRSCHNER, C, DAVIS, S, DUFFY, C, et al: in Physicians' Current Procedural Terminology 1998, 1998 Chicago, IL, American Medical Association,
12. DEPARTMENT OF HEALTH AND HUMAN SERVICES: The International Classification of Disease, 9th revision, Clinical Modification: ICD-9-CM, 2002 Washington, DC, Government Printing Office,
13. Hynes, DM, Joseph, G, Pfeil, C: Veterans Health Information Systems and Technology Architecture (VISTA) as a research tool. VIReC Insights 2002 3: 1–8,
14. Carey, V, Zeger, SL, Diggle, P: Modelling multivariate binary data with alternating logistic regressions. Biometrika 1993 80: 517–526,  | ISI |
15. Dixon, BS, Novak, L, Fangman, J: Hemodialysis vascular access survival: Upper-arm native arteriovenous fistula. Am J Kidney Dis 2002 39: 92–101,  | PubMed | ISI |
16. He, C, Charoenkul, V, Kahn, T, et al: Impact of the surgeon on the prevalence of arteriovenous fistulas. ASAIO J 2002 48: 39–40,  | PubMed | ISI |
17. Birkmeyer, JD, Siewers, AE, Finlayson, EV, et al: Hospital volume and surgical mortality in the United States. N Engl J Med 2002 346: 1128–1137,  | Article | PubMed | ISI |
18. Dudley, RA, Johansen, KL, Brand, R, et al: Selective referral to high-volume hospitals: Estimating potentially avoidable deaths. JAMA 2000 283: 1159–1166,  | Article | PubMed | ISI | ChemPort |
19. Epstein, AM: Volume and outcome—It is time to move ahead. N Engl J Med 2002 346: 1161–1164,  | Article | PubMed | ISI |
20. Collins, AJ, Roberts, TL, St Peter, WL, et al: United States Renal Data System assessment of the impact of the National Kidney Foundation-Dialysis Outcomes Quality Initiative Guidelines. Am J Kidney Dis 2002 39: 784–795,  | PubMed | ISI |
21. Gibson, KD, Caps, MT, Kohler, TR, et al: Assessment of a policy to reduce placement of prosthetic hemodialysis access. Kidney Int 2001 59: 2335–2345,  | Article | PubMed | ISI | ChemPort |
22. Spuhler, CL, Schwarze, KD, Sands, JJ: Increasing AV fistula creation: The Akron experience. Nephrol News Issues 50, 52, 2002 16: 44–7,
23. Collins, AJ, Li, S, Ma, JZ, et al: Cardiovascular disease in end-stage renal disease patients. Am J Kidney Dis 2001 38: S26–S29,  | PubMed | ISI | ChemPort |
24. Eggers, PW, Milam, RA: Trends in vascular access procedures and expenditures in Medicare's ESRD program, in Vascular Access for Hemodialysis— VII, 2001 , edited by Mitchell L, Henry WL, Gore & Associates, Chicago, IL, Precept Press, pp 133–143
25. Allon, M, Lockhart, ME, Lilly, RZ, et al: Effect of preoperative sonographic mapping on vascular access outcomes in hemodialysis patients. Kidney Int 2001 60: 2013–2020,  | Article |
26. Rocco, MV, Bleyer, AJ, Burkart, JM: Utilization of inpatient and outpatient resources for the management of hemodialysis access complications. Am J Kidney Dis 1996 28: 250–256,  | PubMed | ISI | ChemPort |
27. Oliver, MJ, McCann, RL, Indridason, OS, et al: Comparison of transposed brachiobasilic fistulas to upper arm grafts and brachiocephalic fistulas. Kidney Int 2001 60: 1532–1539,  | Article |
28. Hawker, GA, Wright, JG, Coyte, PC, et al: Differences between men and women in the rate of use of hip and knee arthroplasty. N Engl J Med 2000 342: 1016–1022,  | Article | PubMed | ISI | ChemPort |
29. Arozullah, AM, Ferreira, MR, Bennett, RL, et al: Racial variation in the use of laparoscopic cholecystectomy in the Department of Veterans Affairs medical system. J Am Coll Surg 1999 188: 604–622,  | Article | PubMed | ISI | ChemPort |
30. Obrador, GT, Ruthazer, R, Arora, P, et al: Prevalence of and factors associated with suboptimal care before initiation of dialysis in the United States. J Am Soc Nephrol 1999 10: 1793–1800,  | PubMed | ISI | ChemPort |
31. Sidawy, AN, Gray, R, Besarab, A, et al: Recommended standards for reports dealing with arteriovenous hemodialysis accesses. J Vasc Surg 2002 35: 603–610,  | PubMed | ISI |
32. Fleischmann, E, Teal, N, Dudley, J, et al: Influence of excess weight on mortality and hospital stay in 1346 hemodialysis patients. Kidney Int 1999 55: 1560–1567,  | Article | PubMed | ISI | ChemPort |

Acknowledgments

The authors would like to thank Ellen Zufall for sharing her extensive knowledge of Department of Veterans Affairs data with us, and both Ellen Zufall and Danila Manapsal for conducting preliminary Austin database searches during grant preparation. This project was funded by the Agency for Healthcare Research and Quality (AHRQ R03 HS11471). Dr. Hynes and Dr. Stroupe and Mr. Colin are supported by VIREC VA HSR&D Service grant SDR-98-004 and VA HSR&D grant ECI-20-016. Dr. O'Hare is supported by a fellowship training grant by the National Kidney Foundation with matching support from the National Kidney Foundation of Northern California.