Increased Length of Stay and Costs Associatedwith Inpatient Management of Vascular

Access Failures

Abhishek Sawant,* Paul K. Mills,* and Hemant Dhingra†*Department of Internal Medicine, Community Regional Medical Center, University of California, SanFrancisco, Fresno Medical Education Program, Fresno, California, and †Division of Nephrology, University ofCalifornia, San Francisco, Fresno Medical Education Program, Fresno, California

ABSTRACT

The creation and maintenance of vascular access for he-modialysis patients is responsible for a significant amount ofmorbidity and hospital expenses which continue to escalatewith increasing population of ESRD patients. A retrospec-tive review of patient charts were performed from 2008 toMay 2011 at an academic tertiary care center who had adiagnosis of vascular access failure based on ICD 9 coding.Data regarding demographic information, length of stay(LOS), source of insurance, hospital expenses, and dischargestatus were obtained. Based on strict inclusion criteria weidentified 172 total patients. The mean age among allpatients was 60.53 ± 15.35 years and the majority ofpatients were Hispanic (n = 81). The Mean LOS was5.30 ± 4.64 days. Mean hospital costs were 41,896 ±

20,318 US$. Patients admitted for tunneled dialysis place-ment had greater length of stay (p-value = 0.011) as didpatients with hypertension (p-value = 0.030). Hospitalexpenses were significantly higher for patients admitted forarterio-venous fistula complications (55,456 ± 23,779 US$)compared with admissions for catheter or dialysis graftrelated complications (p-value = 0.004). Patients onMedicare had significantly lower length of stay (3.98 ± 3.32days) compared with patients with Medicare ⁄Medical(6.59 ± 5.69 days), p-value = 0.047. Inpatient managementof vascular access failure is associated with increased lengthof stay, and significant hospital expenses. Timely referralto vascular access centers can prevent unnecessary hospital-izations and provide cost-saving benefits.

According to the United States Renal Data System(USRDS) Dialysis Morbidity and Mortality Study2010, the estimated point prevalence of end stage renaldisease (ESRD) patients on Medicare was 453,000. Atthe same time the incidence of ESRD continues to riseand has increased to 13% per annum especially amongthe Hispanic population. Also costs for hemodialysishave increased by 9.3% to 19.4 billionUS$ from 2006 to2007. Vascular access failure (VAF) remains a majorcause for hospitalization and morbidity in patients withESRD. It contributes to approximately 15% of costs forhemodialysis patients, and can be estimated at 2.9 billiondollars. Per person costs for catheter-related vascularaccess failure was greatest at 90,000 US$ in 2008 (1).

Hospitals traditionally run radiological services at themaximum capacity, and due to low reimbursement,vascular access procedures get low priority. This subse-quently increases length of stay and inpatient hospitalcosts. Vascular access-related events were the most

common cause of hospitalization, among ESRDpatients especially during the first 3 months of dialysis,and accounted for 40% of hospital admissions amongESRD patients (2). Utilization review in hospitals needto determine feasible methods of identifying patientswith vascular access failures who do not require inpa-tient care and arrange timely referral to outpatientvascular access centers, if inpatient costs are to beminimized.

To date there has been no study to determine lengthof stay (LOS) or hospital costs ⁄expenses associated withinpatient admissions for VAF. USRDS data provideinformation about total costs of vascular access care,but do not differentiate between inpatient or outpatientexpenses. DOPPS (Dialysis Outcomes and Practice Pat-terns Study) which is a prospective, longitudinal, obser-vational study of hemodialysis patients and facilities in12 countries, in its recent annual 2010 report does notprovide data regarding inpatient vs. outpatient costs ofVAF management (3). Data from single outpatientvascular access centers have been published but nonecomparing to inpatient hospital costs andLOShavebeenconducted. To address this question we decided to per-form a retrospective observational study which focuseson two main indices (i.e. inpatient LOS and hospitalexpenses) to identify methods to reduce costs and hospi-tal stay and subsequently to increase efficiency in future.

Address correspondence to: Hemant Dhingra, MD,Primary Investigator, Division of Nephrology, University ofCalifornia, San Francisco, Fresno Medical Education Pro-gram, 568 East Herndon Avenue Suite 201, Fresno, CA93720, or e-mail: hemantdhingra@hotmail.com.

Seminars in Dialysis—2012DOI: 10.1111/j.1525-139X.2012.01083.xª 2012 Wiley Periodicals, Inc.

1

Materials and Methods

A retrospective observational study was performed atan academic tertiary care center hospital in California’sCentral Valley area. Diagnosis Related Group (DRG)coding system was used to identify patients who wereadmitted to the hospital based on certain inclusion crite-ria. Initially, we identified all patients who were diag-nosed with ESRD (DRG: 585.6) andwhowere receivingextracorporeal hemodialysis (DRG: 39.95) who wereadmitted at our tertiary care center between January 1,2008 to May 31, 2011, and selected patients who hadLOS less than 30 days. This population was cross refer-enced with all patients who had a diagnosis of vascularaccess failure based on the following three main DRGgroups: (i) Complications due to renal dialysis deviceimplant and graft (996.73), (ii) Complications due toother vascular device implant and graft (996.74), (iii)Mechanical complication of other vascular deviceimplant and graft (996.1).The resulting cross-referenced population was then

used to identify patients with exclusion criteria of eitherfluid overload (DRG: 276.6) and ⁄or presence of hyper-kalemia (DRG: 276.7) who were eliminated from thestudy. In addition, patients who had any catastrophicmedical complication during hospital stay, requiringintensive care stay or had net hospital expenses exceed-ing 100,000 US$ were also excluded from our study.The rationale behind excluding these patients stemmedfrom the underlying assumption that these patients hadother co-morbid conditions that required inpatient man-agement and would not be candidates for outpatientintervention of vascular access failures. After identifyingour study population, data were collected on demo-graphic variables, length of stay, hospital expenses,insurance subtypes, electrolyte values, clinical condi-tions, and finally any confounding variables relevant toour study population. This study was approved by theUniversity of California, San Francisco, FresnoMedicalEducation Program (UCSF Fresno MEP) institutionalreview board.

Statistical Analysis

The data collected from medical charts and labora-tory reports were tabulated using Microsoft Excelspreadsheets whereas statistical analysis was performedusing SPSS Software (PASW for Windows, Rel. 18.0.0.2009; SPSS Inc., Chicago, IL, USA) Continuous vari-ables were summarized as mean ± SD and compari-sons between continuous variables were performedusing the Student’s t-test. Categorical variables weresummarized as percentages of the group total and com-parisons between groups were analyzed using eitherFisher exact test or chi-square test where appropriate.Correlations between LOS, hospital expenses and othervariables were calculated using Spearman’s Rank Sumtest, and Pearson correlation. Logistic regression analy-sis was used to adjust for confounders during analysis todetermine if variables of interest were associated withincrease in LOS and ⁄or hospital expenses. Statistical sig-nificance was determined by a p-value<0.05.

Results

A total of 249 patients were initially identified out ofwhich 172 patients were included based on the abovementioned criteria and their baseline characteristics areoutlined in Table 1. Eighty-two (47.7%) were men and90 (52.3%) were women. Twenty-four (14%) wereCaucasian, 81 (47.1%) were Hispanic, 22 (12.8%) wereAfricanAmerican, and 15 (8.7%) were Asian. Themeanage among all patients was 60.53 ± 15.35 years. Themean LOS was 5.30 ± 4.64 days and mean hospitalcosts were 41,896 ± 20,318 US$. When compared byage groups, patients in younger age groups of20–44 years had significantly longer length of staycompared to patients in age groups of 65–74 years(p-value = 0.007) and patients in age groups of 65–74had significantly shorter length of stay compared toolder age groups of greater than 75 years (p-value =0.004). Most common reason for admission wastunnelled dialysis catheter placement (n = 58) [33.7%].Mechanical complications related to vascular accesswere present among 34 (20%) of all patients.Fifty-eight patients were admitted for catheter-related

complications with mean LOS of 6.55 ± 5.10 days,whereas 29 patients were admitted for complications ofarterio-venous fistula (AVF) with mean LOS of4.90 ± 4.16 days, and 133 patients had complicationsof dialysis graft with mean LOS of 4.91 ± 4.16 days,see Table 2. Hospital expenses were significantly higherfor patients admitted for AVF complications(55,456 ± 23,779 US$) compared to admissions forcatheter or dialysis graft-related complications(p-value = 0.004). 45 (26.2%) of patients had coronaryartery disease (CAD), and Type 2 Diabetes was identi-fied among 104 (60.5%) patients, and hypertension waspresent among 162 (94.2%) patients. Patients with

TABLE 1. Length of stay (LOS) by gender, age groups, ethnicity,

and insurance among VAF patients at academic tertiary care

center, 2008–2011

Variables N (%) LOS ± SD, days p-value

Male 82 (47.7) 5.79 ± 5.10 0.186a

Female 90 (52.3) 4.86 ± 4.15Ages 20–44 27 (15.7) 6.52 ± 4.37 0.007b

Ages 45–64 73 (42.4) 4.66 ± 3.92Ages 65–74 32 (18.6) 3.34 ± 1.79 0.004

c

Ages 75 and higher 40 (23.3) 7.2 ± 6.47Hispanic 81 (47.1) 5.37 ± 4.35Non-Hispanic 91 (52.9) 5.24 ± 4.90 0.857d

Medicare 65 (37.8) 3.98 ± 3.32 0.047e

Medical 44 (25.6) 5.36 ± 4.47 0.588Medicare ⁄Medical 51 (29.7) 6.59 ± 5.69Private Insurance 10 (5.8) 6.5 ± 5.28 0.994

Bold is to highlight statistically significant values (where p-value lessless than 0.05).

ap-value based upon Student’s t-test betweenmales and females.bp-value based upon ANOVA between age groups 20–44 and

65–74 years.cp-value based uponANOVAbetween age groups 65–74 and greater

than 75 years.dp-value based on Student’s t-test between Hispanic and non-

Hispanic patients.ep-value based upon ANOVA between Medicare and

Medicare ⁄Medical insurance groups.

2 Sawant et al.

coronary artery disease (CAD) were significantly older(67.02 ± 12.43 years) than patients without CAD(58.23 ± 15.67 years) p-value<0.001. LOS was signifi-cantly higher among hypertensive patients at 5.42 ±4.72 days, compared to LOS of 3.40 ± 2.37 daysamong non-hypertensive patients (p-value = 0.030).However, presence of CAD or diabetes had no signifi-cant influence on the LOS or hospital expenses.

Twenty-three (13.4%) patients in our study were hos-pitalized formechanical complicationsof catheterswhichwere significantly higher compared to AVF or graftrelated mechanical complications (p-value <0.001).Majorityof patients hadMedicare (n = 65), followedbyMedicare ⁄Medical (n = 51),Medical only (n = 44)andprivate insurance (n = 10) respectively. Patients withMedicare had significantly lower length of stay(3.98 ± 3.32 days) compared to patients with Medi-care ⁄Medical (6.59 ± 5.69 days),p-value =0.047.

Discussion

Recent USRDS Dialysis Morbidity and MortalityStudy showed that costs for ESRD rose by 6%, from2006 to 2007 to $24 billion, which is 5.8% of the entireMedicare budget (1). These costs are secondary toexpenses from various facets of management of ESRD,and out of them vascular access is considered as Achillesheel of dialysis care. During transition from CKD toESRD there are high expenditures due to initiation ofdialysis. Most of these high costs, not surprisingly, arerelated to hospitalizations. During the first month ofESRD, use of dialysis catheters is very high and is associ-ated with high rates of hospitalization for access failure,declotting procedures, repeated fistula placements, andinfectious complications. Also during the first 3 monthsof initiation of dialysis, hospitalizations were most com-monly due to vascular access failure, up to 36% asshown in one study (2).

The AVF should be the preferred modality for vascu-lar access among patients initiated on dialysis. However,with an increase in the mean age of ESRD patients, latereferral of patients, and increase in acute renal failureleading to ESRD, use of central venous catheters havebecome more frequent (4). Forty-six percent of patientsdo not have a permanent vascular access before initia-tion of dialysis in the United States, compared to 25%among the counterparts in Europe (5). The mostcommon causes of vascular access failure appear to be

thrombosis, stenosis, and infection. USRDS DialysisMorbidity and Mortality study data report in 2010showed that 710 (17%) patients had catheter malfunc-tion, whereas 1397 (33.6%) had AV Fistula failure, and2052 (49.3%) had graft failure. In our study 58 (26.4%)had catheter-related VAF, whereas 29 (13.2%) had AVFistula-related VAF, suggestive of higher percentage ofcatheter use on our population.

Central venous catheters have a higher rate of throm-bosis compared with prosthesis and native fistulas (6).Similar finding in our study showed higher rate ofthrombosis of venous catheters among 31 (18%)patients, compared to 19 (11%) patients withAVFistulathrombosis. The Hemodialysis (HEMO Study) grouphad shown statistically significant difference inprevalence of fistulas, which was lower among femalescompared tomales. Also prevalence of fistulas was loweramong African-American patients and among olderpatients as well (7). USRDS Dialysis Morbidity andMortality study data did not show any differencebetween gender or race in prevalence of AV Fistula,which was also seen in our study (8).

LOS or hospital expenses were not significantly differ-ent by gender or ethnicity in our study population. In2002, 45.4% patients admitted for VAF were men perUSRDS data that were similar to 47.7% found in ourstudy whereas DOPPS population showed that 55.4%weremen inDecember 2010. According toUSRDS data16% of patients admitted for VAF were Hispanic;whereas DOPPS dataset showed that only 12% ofpatients were Hispanic (9). Our study population com-prised 47%Hispanic patients, which is a reflection of thehigher prevalence of Hispanic population in California’sCentral Valley area. With incidence of ESRD amongHispanic population as high as 13%, VAF relatedinpatient admissions and expenses are bound to escalate.DOPPS dataset revealed that 61.8% patients in Decem-ber 2010 were diabetic, whereas 43.8% patients inUSRDS dataset had diabetes. Similarly 60.5% patientsadmitted inpatient for VAF in our studywere diabetic atthe time of admission.

Utilizing data from Wave II of the USRDS DialysisMorbidity and Mortality study, the incidence of anacute coronary syndrome was 2.9% per year among3374 incident dialysis patients followed for approxi-mately 2 years (10). Almost 40% of the 1846 patientsenrolled in the HEMO report were noted to have ische-mic heart disease at study initiation (11). Another studyshowed that 24% of those with chronic kidney disease

TABLE 2. LOS, hospital expenses, and medical conditions by type of vascular access failure among VAF patients at an academic tertiary

care center, 2008–2011

VariablesCatheter related

complications (n = 58)AVF related

complications (n = 29)Dialysis graft related

complications (n = 133) p-valuea

Mean length of stay (days) 6.55 ± 5.10 4.90 ± 4.16 4.91 ± 4.16 0.080Mean hospital expenses (US$) 47,092 ± 25,124 55,456 ± 23,779 40,042 ± 23,577 0.004

Coronary artery disease 12 (20.7%) 9 (31%) 38 (28.6%) 0.454Diabetes type 2 33 (56.9%) 21 (72.4%) 80 (60.2%) –Hypertension 53 (91.4%) 28 (96.5%) 126 (94.7%) –

Bold is to highlight statistically significant values (where p-value less less than 0.05).ap-value based uponANOVAbetween catheter, Arterio-Venous Fistulas (AVF), and dialysis grafts-related complications.

INCREASED LENGTH OF STAY AND COSTS 3

(CKD) and only 15% of those without CKD had car-diovascular complications (12). In our study we foundpresence of coronary artery disease among 45 (26.2%)patients. Relatively lower incidence of myocardialinfarction in blacks was observed in a study using theUSRDS database (13). However, prevalence of CADwas the same across gender and all races in our study. Asignificantly higher percentage of patients with CADhad concurrent diabetes (p-value = 0.040) as comparedwith presence of concurrent hypertension.A recent prospective trial evaluating cost of vascular

access care found that patients who were dialyzed exclu-sively with a catheter, the largest component of costswas contributed by hospitalization for access relatedcomplications (14). There were 23 (39.7%) catheter-related VAF secondary to mechanical complications,whichwere significantly higher compared toAVFistulasand grafts (p-value <0.001). Dialysis Outcome QualityInitiative (DOQI) guidelines suggest placement of anAVF as the initial choice for vascular access, however,recent study showed that vascular access costs were sig-nificantly higher for patients in whom an AVF attemptis unsuccessful but overall costs were lowest during thefirst year of dialysis (15). This could be partiallyexplained due to the diagnostic imaging costs induced byregular vascular access monitoring and surgeries associ-ated with AVF. Our study also showed that vascularaccess costs were highest due to complications of AVF(55,456 ± 23,779 US$) in spite of the fact that only13.2%of admissions were due toAVFistula failure.A study performed at a Vascular Access Center

(VAC) in Arizona, showed that 149 of 157 (95%) caseswho were referred for emergent vascular access proce-dures to the center had a functional access, showingemerging role of such resources for successful manage-ment of vascular access failures (15). Another study hadshown that outpatient VAC model offers cost savingsand reduces hospitalizations and missed dialysis treat-ments (16). They demonstrated reductions of approxi-mately >0.6 hospital days per patient year anddecreased missed treatments of >0.3 per patient yearwhich represented the effects of intense focused vascularaccess care in outpatient setting. ‘‘Fistula first’’ is abreakthrough CMS initiative with an initial goal of 66%prevalence of AVF in the United States. A dedicatedVAC, which is a part of an integrated vascular accessprogram, can help increase prevalence of AVF by effi-ciently carrying out significant number of interventionalprocedures (17).The most common issue with VAC appears to be

transportation of the patient to the VAC from dialysiscenter. Also, managed care contracting, practicepatterns, and formation of team approach are other vitalissues that need to be tackled to ensure success of a VAC(18). The largest prospective trial performed by Beat-hard and Litchfield showed that complications of endo-vascular procedures by interventional nephrologistswere lower than complications reported in surgical andradiological literature (19). Interventional nephrologistsare even performing vein obliteration and percutaneousballoon angioplasty techniques, by which they are ableto successfully salvage AVFistulas that failed tomature.

Freestanding VAC helps to minimize frequent delaysand in addition nephrologists trained in hemodialysisvascular access provide procedural care more efficiently(20). Per person per year costs for vascular access ser-vices performed by surgeons continue to fall by 39%since 1995–2007. Costs for those performed by nephrol-ogists, however, continue to rise in 2007 and were nearly10 times greater than in 2000 as published byUSRDS intheir 2010 report. With this shift in responsibilities formanagement of VAF by trained interventional nephrol-ogists it is evident that outpatient vascular access centerswill be able to efficiently and inexpensively handle VAFreducing inpatient LOS and costs.Similarly at our tertiary care center, prompt and

timely referral to a freestanding VAC for vascular accessfailures could help prevent unnecessary hospitalizationsand have huge savings of hospital costs. For 172 patientsenrolled in our study with an average hospital cost of41,000 US$, the total cost of inpatient management forthese patients amounts to more than 7.2 million dollars.With a high percentage of California’s central valleypatient population being uninsured, these expenses willadd to the fiscal strain on our tertiary care center requir-ing diversion of funds from outpatient clinics andprograms which further escalates the problem forming avicious cycle. Case management or utilization reviewcould identify these patients in the emergency room, andhelp refer them directly to the VAC. This would ensureprompt and efficient correction of the access failure andresume dialysis without any delay. This could also easethe pressure of patient load on the already overburdenedinterventional radiology department and help divert ORtime to more serious and urgent surgical issues. All saidand done, this requires clear delegation of responsibili-ties and staffing of personnel trained to identify patientswhomeet criteria for direct referral to a VAC.Our study also has several limitations. We included

costs for vascular access failure based only on hospitalbillings for inpatient stay and procedures. Physicianreimbursements for consultations and daily patient carecould not be included since it would be very difficult toobtain physician billing codes submitted to their respec-tive billing offices which is not a part of hospital records.Second, patients who were excluded from the study dueto hyperkalemia may not always require urgent dialysisand select group of patients can still be referred to outpa-tient vascular access centers. Since our study had higherpercentage of catheter related failures, which have highfrequency of mechanical complications, this could indi-rectly contribute to more frequent hospitalizations forVAF. Finally this is a single center retrospective studyand the California Central Valley population may notbe indicative of national trends.

Conclusion

Vascular access management contributes to signifi-cant expenses especially during first fewmonths of initia-tion of dialysis. This increases workload on the alreadyoverburdened hospital infrastructure and draws up costsand increases length of stay. Free standing VAC’s can

4 Sawant et al.

help ease this pressure by fixing these vascular accessrelated complications in a prompt and timely mannerpreventing unnecessary admissions. However, it requiresteam work among trained interventional nephrologists,vascular surgeons, and interventional radiologists withappropriate selection of patients who need referral foroutpatient care by trained case managers or utilizationreview personnel.

Acknowledgment

Blair Renwick, BS, Health Information Technology, Com-munity RegionalMedical Center, Fresno, CA.

Potential Conflict of Interest

None.

References

1. U SRenalData System:USRDS2010Annual Data Report: Atlas of ChronicKidney Disease and End-Stage Renal Disease in the United States. Bethesda,MD: National Institutes of Health, National Institute of Diabetes andDigestive andKidneyDiseases, 2010

2. Arora P, Kausz AT, Obrador GT, Ruthazer R, Khan S, Jenuleson CS,Meyer KB, Pereira BJ: Hospital utilization among chronic dialysis patients.JAmSocNephrol 11:740–746, 2000

3. Annual Report of theDialysis Outcomes and Practice Patterns Study:Hemod-ialysis Data 1999-2010. Ann Arbor, MI: Arbor Research Collaborative forHealth, 2010

4. SantoroA,CanovaC, Freyrie A,Mancini E: Vascular access for hemodialy-sis. JNephrol 19(3):259–264, 2006

5. PisoniRL, YoungEW,Dykstra DM,GreenwoodRN,Hecking E,GillespieB, Wolfe RA, Goodkin DA, Held PJ: Vascular access use in Europe and intheUnited States: results from theDOPPS.Kidney Int 61:305–316, 2002

6. DepnerTA:Catheter performance. SeminDial 14:425–431, 2001

7. AllonM, Ornt DB, Schwab SJ, RasmussenC, Delmez JA,Greene T, KusekJW,Martin AA,Minda S: Factors associated with the prevalence of arterio-venous fistulas in hemodialysis patients in the HEMO study. Hemodialysis(HEMO)StudyGroup.Kidney Int 58:2178–2185, 2000

8. Hirth RA,TurenneMN,Woods JD, Young EW,Port FK, PaulyMV,HeldPJ: Predictors of type of vascular access in hemodialysis patients. JAMA276:1303–1308, 1996

9. Robinson B, Fuller D, Zinsser D, Albert J, Gillespie B, Tentori F, TurenneM, Port F, Pisoni R: The Dialysis Outcomes and Practice Patterns Study(DOPPS) Practice Monitor: Rationale and Methods for an Initiative toMonitor the New US Bundled Dialysis Payment System. AJKD 57(6):822–831, 2011

10. Trespalacios FC, Taylor AJ, Agodoa LY, Abbott KC: Incident acute coro-nary syndromes in chronic dialysis patients in the United States. Kidney Int62(5):1799–1805, 2002

11. Cheung AK, Sarnak MJ, Yan G, Berkoben M, Heyka R, Kaufman A,Lewis J, Rocco M, Toto R,Windus D, Ornt D, Levey AS: Cardiac diseasesin maintenance hemodialysis patients: results of the HEMO Study. HEMOStudyGroup.Kidney Int 65(6):2380–2389, 2004

12. Collins AJ, Li S, Gilbertson DT, Liu J, Chen S-C, Herzog CA: Chronickidney disease and cardiovascular disease in the medicare population.Kidney Int 64(Suppl. 87):S24–S31, 2003

13. Young BA, Rudser K, Kestenbaum B, Seliger SL, Andress D, BoykoEJ: Racial and ethnic differences in incident myocardial infarction inend-stage renal disease patients: The USRDS. Kidney Int 69(9):1691–1698,2006

14. Manns B, Tonelli M, Yilmaz S, Lee H, Laupland K, Klarenbach S,Radkevich V,Murphy B: Establishment andmaintenance of vascular accessin incident hemodialysis patients: a prospective cost analysis. J Am SocNephrol 16(1):201–209, 2005

15. KianK, TakesianK,Wyatt C, Vassalotti J,Assalotti J,MishlerR, SchonD:Efficiency and outcomes of emergent vascular access procedures performedat a dedicated outpatient vascular access center. Semin Dial 20(4):346–350,2007

16. MishlerR, Sands JJ, OfsthunNJ,TengM,SchonD,Lazarus JM:Dedicatedoutpatient vascular access center decreases hospitalization and missed out-patient dialysis treatments.Kidney Int 69(2):393–398, 2006

17. Jackson J, Litchfield TF: How a dedicated vascular access center can pro-mote increased use of fistulas.Nephrol Nurs J 33(2):189–196, 2006

18. Perry Arnold W: Improvement in hemodialysis vascular access outcomes ina dedicated access center. SeminDial 13(6):359–363, 2000

19. Beathard GA, Litchfield T: Effectiveness and safety of dialysis vascularaccess procedures performed by interventional nephrologists. Kidney Int66:1622–1632, 2004

20. Asif A,Merrill D, Briones P, Roth D, Beathard GA:Hemodialysis vascularaccess: percutaneous interventions by nephrologists. Semin Dial 17(6):528–534, 2004

INCREASED LENGTH OF STAY AND COSTS 5