Published ahead of print on September 6, 2006
Clin J Am Soc Nephrol 1: 1234-1240, 2006
© 2006 American Society of Nephrology
doi: 10.2215/CJN.01210406
Hospital Resource Utilization That Occurs with, Rather than Because of, Kidney Failure in Patients with End-Stage Renal Disease
Edward A. Ross*,
Rita E. Alza
, and
Neerav N. Jadeja
* Division of Nephrology, Hypertension and Transplantation, University of Florida; and Shands at the University of Florida, Gainesville, Florida
Address correspondence to: Dr. Edward A. Ross, Division of Nephrology, Hypertension & Transplantation, University of Florida, Box 100224, Gainesville, FL 32610-0224. Phone: 352-392-4007; Fax: 352-392-3581; E-mail: rossea{at}medicine.ufl.edu
 |
Abstract
|
|---|
More than $18 billion annually is attributed to care of patients with ESRD, with the perception of high renal costs for a relatively small population. It was proposed that accounting methods exaggerate resource utilization that often occurs with rather than because of kidney failure. The dialysis patients in this study had nearly all of their care at university facilities with one financial database. For 1 yr, 112 chronic hemodialysis patients were studied using demographic, insurance, and hospital facility (diagnoses, length of stay, charges, costs, and net income) variables. Substantial inpatient costs and hospitalizations were for nonrenal primary diagnoses, including malignancies, substance abuse, trauma, HIV, and psychiatric diseases: 37% of admissions, 36% of inpatient days, and 32% of charges. Dialysis patients were healthier than indicated by averaged length of stay and cost data, because results were very skewed: Mean 17.3 inpatient days but median only 2.4 d; 43% of patients had 0 to 1 inpatient days (1.3% of charges), 23% had 2 to 7 d (charges 7.6%), 18% had 8 to 30 d (charges 26%), and 16% had >30 d (charges 66%). Lengthy hospitalizations had disproportionately high operating room and respiratory care costs. The large group of relatively healthy outpatients did not avoid hospitalization by high use of facility resources. The true costs for medical care that results from ESRD are not as high as publicized, as a result of misclassification of inpatient expenses from nonrenal comorbidities. When not confounded by analyses that use data means, it is clear that substantial numbers of hemodialysis patients have very brief hospitalizations with low resource utilization.
|
Introduction
|
|---|
The cost that is attributed to the care of patients with ESRD has soared during the past decade in the United States and now is thought to exceed $18 billion annually (1). For the treatment of what is viewed as a relatively small segment of the population (<400,000 individuals), the high dollar figure has a profound impact on health care fiscal policy, especially as it relates to the distribution of federal funding. We believe that this amount may be very misleading, however, because it would include a wide variety of health interventions all grouped into the Renal Medicare category of costs. Therefore, we thought that many illnesses and their resultant costs were occurring in patients with renal failure rather than because of kidney disease. The problem of expense classification would be particularly important for costly hospitalizations and would not be evident in the many previous financial analyses that focused on outpatient dialysis care. Those reports used chronic dialysis data registries (1,2) and potentially could have led to additional financial misperceptions that pertain to this heterogeneous patient population. Specifically, rates of hospitalization that were extracted from those data typically are presented as means. We thought that the bulk of inpatient resources were being used by a small group of exceptionally ill patients and often were due to their comorbid diseases. Using data means when the results are not normally distributed could attribute inappropriately excess illness to the broader dialysis population. The impact of advances in renal replacement and associated therapies are apt to be underappreciated when outcome analyses are based on crude fiscal parameters that are confounded by smaller subsets of patients with severe nonrenal comorbidities.
The challenge to a global analysis of use of hospital-based services by patients with ESRD is obtaining a robust database that captures the full spectrum of health care costs, and there have been few investigations in this regard (3–8). We had a unique opportunity in that nearly all of the patients with ESRD in our academic practice with a single dialysis unit had their health care at two facilities in our medical enterprise, with comprehensive financial records residing within one computer system. Unlike some previous studies, we were able to investigate the use of all facility-based resources (inpatient, outpatient, and emergency department [ED]) by 112 chronic hemodialysis patients during a 1-yr interval, with a particular focus on the renal versus nonrenal causes for hospitalization. Analyses were performed to address our hypotheses that (1) substantial inpatient Renal Medicare costs are not due primarily to kidney failure and (2) a small fraction of patients used a disproportionately high portion of resources, often unrelated to their kidney disease, so that the monitoring of simple data means distorts the perception of illnesses in this population.
|
Materials and Methods
|
|---|
Data Collection
The calendar year 2002 was chosen for analysis because a complete data set was available for all of our outpatients who were on hemodialysis, as well as for the two local hospitals in our system (Shands at the University of Florida and Shands at Alachua General Hospital, Gainesville, FL). Hospital care and financial information for both facilities resided within a common software system. The project was approved by the University of Florida Investigational Review Board and complied with Health Insurance Portability and Accountability Act regulations. All chronic patients who were undergoing in-center or home hemodialysis were included, with the criteria that they had to have been in our program for 90 consecutive days as an outpatient before any hospitalization. It was believed that this permitted a more accurate analysis of health care events that could be attributed to our institution’s care. Therefore, patient data for January 2002 were included only when the patient had been dialyzing with us as an outpatient since at least October 2001. This criterion intentionally excluded the many transient dialysis patients in Florida who might have been under our care for days or hours before developing acute illnesses. Of the 112 patients who were assessable, 107 had all of their facility-based care at one or both of our hospitals. Five instead went to our university-associated Veterans Affairs (VA) hospital. For those patients, we were able to track their VA inpatient days but did not have access to any other details from that facility. Therefore, depending on the variables being analyzed, all patients (n = 112) or just those with the full resource utilization data set (n = 107) were included.
Patient Demographics
Review of the medical charts and facility records provided data for patient age, gender, race, cause of ESRD (as indicated on the Health Care Financing Administration #2728 form), date of first dialysis treatment, date ranges of dialyses in any of our facilities, whether diabetes was present, insurance coverage, and geographic region of residence. The primary diagnoses in the diagnosis-related group (DRG) listing for the facility’s charges were obtained and divided into nine broad categories of illnesses: Noninfectious access related, infectious access related, other infections, cardiac (with or without cardiac surgery), other vascular disorders, nonvascular surgeries (i.e., orthopedics and trauma), other internal medicine disorders (i.e., gastrointestinal, hematology, oncology, endocrinology, and rheumatology), neurology and psychiatry, and all other illnesses. Diagnosis and comorbidity data were obtained by hospital coders who were not involved with the study. The dates and causes of all deaths also were obtained from the medical records.
Hospital-Based Services Analyzed
The unique patient account numbers of every chronic outpatient were used to query all hospital financial records, including all outpatient tests; ED visits; and pharmacy, laboratory, and radiology services. Outpatient care encompassed all hospital-based procedures (endoscopy, bronchoscopy, cardiac catheterizations, and physical therapy). Inpatient service data included all associated care and the principle diagnosis (DRG), and we also captured not just the length of stay (LOS) but also the number of days spent in any of the many intensive care units. Admission diagnoses were classified as to whether the hospitalization was due primarily to a nonrenal reason versus one that could be related to kidney failure. For example, all admissions for medical or surgical vascular (cardiac, peripheral, or cerebral) problems; catheter-associated complications and infections; fever of unknown origin; urologic or renal malignancies; anemia-associated complications; all fluid, electrolyte, or acid-base disorders; or nontraumatic orthopedic (e.g., hip fracture) or joint diseases were grouped with the renal causes because the pathophysiology of those disorders could have been exacerbated by ESRD.
For each service, we identified its billed charges, the total costs, and the gross collections and calculated the net income (i.e., profit/loss). Of note, the value that was used for the "cost" represents the sum of the direct and indirect costs. The indirect costs, also referred to as overhead costs, cover many expenditures that are associated with (and necessary for) academic medical centers and that are incurred at a greater scale than nonteaching entities. Care and expenses within the outpatient dialysis unit were not included because they have been the subject of other investigations and because hospitals typically have no control over those areas when they are under the ownership or administration of proprietary dialysis chains. Physician professional fees and reimbursement could not be included in this analysis of hospital-based resource utilization for three reasons: (1) A complete data set of charges and collections was not available from both institutions; (2) unlike the case of the facilities in which we had both cost and expense data, the physician costs (salaries) were not accessible to us; and (3) there was a concern that surgical global payments that covered 90 d after operations might cross the data inclusion window dates and be difficult to track.
Statistical Analyses
Statistics were performed using SAS software (version 9.2; SAS Institute, Cary, NC). Descriptive statistics were expressed as means ± SD. Data for inpatients also were analyzed for specified ranges of admissions duration: 0 to 1 d, 2 to 7 d, 8 to 30 d, and >30 d. The rationale for choosing a "0 to 1 d" group is that this defines a set of admissions that arguably could be considered "short stays" or outpatient observation days by some insurance carriers. This group also captured hospitalization for only a few hours, as the database labeled them as a single admission day. Each hospital admission was analyzed separately, and calculations were made for the yearly sums as well. Total annual values were used for outpatient (procedural, laboratory) and ED services. For patients who were not in our program for the full year, the actual number of days was used to extrapolate arithmetically to 365 d so as to be able to compare all patients meaningfully and allow us to express the hospital days per patient year. However, because normalization potentially could introduce errors, we performed similar analyses for the subset of 58 patients who were in our program for the full year. No extrapolation was needed when the financial data were analyzed with respect to individual admissions. To explore factors that might be associated with morbidity, we performed ANOVA with all terms in the model. All NS terms were dropped from the model and the analysis was performed again. Differences between groups and linear trends were tested using contrasts. P < 0.05 was considered statistically significant.
|
Results
|
|---|
Patient Characteristics
For the total data set of 112 patients (Table 1), there were 57 (50.9%) men with a median age of 53 yr. They had an ESRD duration of 6.1 ± 6.1 yr (range 0.3 to 29.5 yr), with an average of 319 d under our care in 2002. The cause of renal failure was hypertension (40%), diabetes (24%), glomerulonephritis (15%), polycystic kidney disease (4%), and other or unknown causes (17%); 37% carried the diagnosis of diabetes. The racial distribution was 63% black, 33% white, and 4% other heritage. Primary insurance coverage was by federal programs (e.g., Medicare, VA) in 77%, Medicaid in 9%, and commercial (including managed care) companies in 11%. Approximately 4% were totally uncovered. A total of 97% resided in our regional district, and one patient was from each of three other districts. Five patients sought their hospital facility care at the VA Medical Center, and all were male.
Hospital Admissions
According to the regression analysis, the reported cause of renal failure and the duration of ESRD were not statistically significant with any of the LOS or financial variables. The most common reasons for hospital admissions in each LOS group are detailed in Table 2: There were the expected vascular access and cardiovascular causes, as well as a variety of distinct medical and surgical illnesses that were not attributable to the renal failure (e.g., malignancies, chronic lung disease, orthopedic and gastrointestinal surgeries, longstanding psychiatric illness). There was a trend for patients in the longest LOS groups to have diabetes, but this was not statistically significant (P = 0.20). There was no significant relationship between the primary insurer or regional district and these LOS or financial measures.
There were a total of 304 hospital facility visits (including ED and observation stays), 189 of which were inpatient admissions defined as being at least 1 d duration. The reasons for these hospitalizations were vascular access related (16%; noninfectious 9%, infectious 7%), infections (not access related; 9%), cardiovascular (17%; cardiac 11%, noncardiac 6%), noncardiac medical disorders (20%), surgical (noncardiac; 17%), neurology and psychiatry (10%), and other illnesses (11%). Table 3 demonstrates resource utilization using our conservative criteria to differentiate inpatient stays for renal-associated diagnoses from those that were unrelated to kidney failure. The nonrenal causes for admission included HIV complications, diverticulitis, cholecystitis, inflammatory bowel disease, sickle cell disease, colon cancer, multiple myeloma, substance abuse, psychiatric disease predating ESRD, trauma, and continuous obstructive pulmonary disease exacerbations including pneumonia. These accounted for 37% of the hospitalizations (23% of all of the facility visits), 36% of the inpatient days, 32% of the total charges, and 27% of the collections. They incurred a disproportionately low 13% of operating room costs (likely as a result of exclusion of all vascular operations) and, overall, led to a net loss of income compared with the fiscally favorable admissions for renal-associated diagnoses.
LOS
The LOS results (for all 112 patients with a total of 1112 d) were markedly skewed. After each patient was normalized to a full year of care, the mean value was 17.3 inpatient days (±37.6 d; range 1 to 180 d), and the median was much lower at only 2.4 d. Using the normalized data (Table 4), remarkably, 42.9% of the patients were in the range of 0 to 1 d; 22.3% of patients were in the range of 2 to 7 d, 17.9% were in the range of 8 to 30 d, and 17.0% had >30 d. Similar findings were evident in the separate subanalysis of the 58 patients for whom we had 365 d of data and thus demonstrated that there was no bias from extrapolations to a full year of care in the analysis of the entire patient group. For example, in the subset analysis, 15.5% of the patients had an LOS of >30 d (and they generated 63.1% of the charges).
The LOS was not associated with any identifiable factor other than the primary DRG (P = 0.022). The diagnosis groups with the longest LOS were noncardiac nonaccess vascular surgery (mean 46 d), noncardiac medical problems (21 d), other surgeries (15 d, including trauma and diverse procedures, e.g., gastrointestinal operations), access infections (12 d), and other infections (11 d). There were a total of 12 deaths, and the causes were known in 11. At least four deaths were not attributed primarily to kidney failure and were due to malignancies, substance abuse, and severe cardiomyopathy predating the ESRD.
Inpatient Financial Parameters
For the subset of 107 patients (95.5% of the total group) for whom we had full financial data (Table 3), the 43% of patients who had the low 0 to 1 inpatient days incurred only 1.3% of the hospital charges, which resulted in a small financial gain (mean $606/patient). The 23% of patients in the relatively short 2 to 7 inpatient day group had an average of two admissions and produced only 7.6% of the hospital charges, which generated a net income (mean $8538/patient). Taken together, these individuals with highly favorable hospital durations of 
7 d represented 66% of our patient population and incurred only 8.9% of the charges. The 18% of patients who were in the longer 8 to 30 inpatient day group had a mean of three admissions and were responsible for 25.6% of the hospital charges, resulting in net income (mean $6373/patient).
The smaller (16%), very ill group of patients with the long >30 inpatient days had an average of six admissions, incurred a large 65.5% of the hospital charges, and generated 53.4% of the gross collections. However, because they had such resource-intense care (responsible for a disproportionate 65.8% of all of the hospital costs), there was net loss of money (mean $11,611/patient). There was not only greater respiratory (including tracheotomy) care but also laboratory and pharmacy resource utilization during these prolonged hospitalizations but proportionally less costs from hemodialysis treatments.
Outpatient Hospital-Based and ED Financial Parameters
To characterize the remarkable 43% of patients who were in the brief 0 to 1 inpatient day group, their outpatient and ED utilization was analyzed (Table 3). They used only 32% of the hospital-based outpatient services and 11.9% of the ED services. Thus, attaining the goal of near-avoidance of inpatient services was not achieved by virtue of excess or disproportionate use of outpatient or emergency resources. Taken together with the small profit from inpatient services, this healthy group was responsible for a small net loss to the medical center of $345 per patient year.
A similar analysis was performed for the 23% of patients with the short 2 to 7 inpatient days. They were responsible for 28% of the outpatient services and 28.3% of the ED costs. Despite substantial (although nearly budget-neutral) outpatient and emergency utilization of resources, because of the short profitable hospitalizations, the net effect was a positive income of $8554 per patient year.
The sicker 18% of patients with 8 to 30 inpatient days had disproportionately high utilization in the outpatient (23.1% of all costs) and emergency (34.7%) settings. Despite losses in these two areas, because of the profitable inpatient services, there still was a net income of a mean of $7686 per patient year.
The 16% of patients who were very ill with prolonged hospitalizations (>30 d) were high resource users in all venues. They incurred 17.2% of the outpatient charges and 25.1% of the ED costs. Taking into account the financial shortfall from their hospitalizations, this increased the net loss to $13,225 per patient year for the hospital system.
|
Discussion
|
|---|
With limited national resources available for medical care, we believe that it is important to have an accurate perception of the general health of patients with ESRD and how renal failure per se contributes to the overall expenses for treating these individuals. At a time when the nephrology community wants to maintain, if not increase, its share of a limited pool of funding, it would be helpful to appreciate that some of the $18 billion annually to care for the relatively small ESRD population is really due to nonrenal comorbidities. We provided evidence in this regard for costly hospitalizations and also demonstrated that common data-reporting methods are apt to lead to an underappreciation of favorable limited resource utilization by large portions of relatively healthy dialysis patients. Whether this cost analysis approach can be used to adjust inpatient or outpatient reimbursement remains to be studied. Hopefully, our single-site study will prompt further investigation of larger national databases, as the well-established scrutiny of outpatient expenses and quality of care (9) has not yet rigorously been extended to inpatients.
Our results stress the frequency and expenses (37% of the hospitalizations and 32% of the total charges) from the many admission (and death) diagnoses that are not primarily related to kidney failure, such as malignancies; trauma; substance abuse; diabetes; and gastrointestinal, chronic lung, and psychiatric diseases. The high hospital costs of operating rooms and ventilator and respiratory care during prolonged hospitalizations for those illnesses increase out of proportion to those for renal replacement therapy. Even though many of the reported expenses really are due to nonrenal diseases or comorbidities, they are misclassified when the patients’ entire fiscal account is labeled as being in the ESRD category ("Renal Medicare"). Note that we were careful to use very conservative criteria when attributing admissions to nonrenal primary diagnoses; however, we acknowledge the possibility that ESRD contributed to the severity of some of the illnesses that prompted the hospitalization. This emphasizes the need for prospective studies of these issues, ideally with control groups of patients without ESRD and with similar admission disorders.
Diagnosis and DRG data need to be viewed with caution because of the difficulty of coding all of the comorbid conditions and procedures for patients who have multiorgan diseases. We believe that our facility’s rigorous chart extraction and software tools overcame some of the coding difficulties that have been present in the past. This helps explain why DRG were reported previously to reflect inadequately the true costs of caring for patients with ESRD (10) and why only some series showed an impact of diabetes (11) or cardiac problems (12) on admissions.
In addition to misclassified costs of inpatient illnesses, some methods for reporting the number of hospitalization days per year can distort further the perception of the patients’ overall health. Indeed, an outpatient facility’s LOS track record is one of the main published variables that are used to compare dialysis units (1). That value is an arithmetic mean and has great influence on evaluating a unit’s quality, guiding patients’ choice of a facility, and securing contracts with insurance carriers. We confirmed our hypothesis that simple calculations of LOS are very misleading because the outcomes did not have a normal distribution and therefore could be problematic when used to compare outpatient dialysis units. Although the mean number of inpatient days was 17.3, the median was a remarkable 2.4 d. Another consequence of using mean values is that they could make it difficult to discern improvements in medical outcomes as a result of advances in renal replacement therapy. Notably, 43% of our patients were hospitalized only 0 to 1 d, and they incurred only 1% of the in-hospital and 12% of the emergency charges. This compares very favorably with data provided by the Florida ESRD Network, which is based on University of Michigan Kidney Epidemiology and Cost Center analyses. Those data showed that approximately 16% of state, regional, and national admissions were single-day stays. Our patients who were admitted for >1 wk/yr accounted for 34% of the population but generated 92% of the hospital charges. The 16% with >30 d of hospitalization had 66% of those charges. In addition, the ongoing transition of much health care from the inpatient to the outpatient or short-stay setting (i.e., "day surgery") will complicate further these in-hospital data analyses. Indeed, the US Renal Data System analyses do not look specifically at single-day admissions. Local ESRD networks may provide this useful information, but the data might be weak as a result of a lack of standardization of the nomenclature for these brief patient encounters.
Although not the primary focus of this investigation, more extensive analyses of the charges, costs, collections, and net income for these and other patient groups could be very revealing. Indeed, it was disturbing but not surprising that the combination of the payer mix and reimbursement contracts led to a situation in which our hospital was not compensated adequately for the care of the relatively small percentage of outliers who have a disproportionately enormous consumption of resources. Taken in the totality of costs and collections, our medical center essentially was budget-neutral for the facility-based (inpatient, outpatient, and ED) care of our entire outpatient hemodialysis population. Future federal or insurance carrier strategies that would reduce the alleged "profits" from the subset of our patients with low or moderate illness acuity (without a compensatory increase in payments for the sickest group) could have the unintended result of a financial loss in caring for the entire population spectrum. Although dependent on their payer mix, many academic centers with high direct and indirect (overhead) costs as a result of the necessities of the scholastic mission can at least take comfort that the current reimbursement scheme can lead to a break-even situation.
The retrospective nature of our study may have influenced these analyses, as well as our ability to ascertain modifiable factors that were associated with high resource utilization. Some areas for improvement are in the realm of renal-related diseases (e.g., complications of vascular access grafts as opposed to native arteriovenous fistulas [13]), whereas others pertain to independent medical or surgical comorbidities. It is our hope that identifying the disproportionately small group of patients who are at risk for long, costly hospitalizations could lead to preemptive strategies in outpatient clinics or even the ED. The latter setting has received little investigation (14), and our experience concurs with the observation that those costs are high in sick individuals who need admission. We did not detect a pattern of ED utilization suggestive of its having led to the prevention of hospitalizations, although the study was not designed to address that specific question. Our findings also need to be interpreted in light of our patient population’s being somewhat atypical (e.g., younger age) than that described by the US Renal Data System. Additional weaknesses were described in detail herein and included the exclusion of events that occurred within the first 90 d of our care and physician cost data.
|
Conclusion
|
|---|
The true costs for the medical care of patients with ESRD are not as high as publicized, as a result of misclassification of expenses from nonrenal comorbidities and diseases. When not confounded by analyses that use data means, it is clear that substantial numbers of hemodialysis patients have very brief hospitalizations with low resource utilization.
|
Acknowledgments
|
|---|
Results were presented in part at the 38th Annual Meeting of the American Society of Nephrology; November 8 through 13, 2005; Philadelphia, PA.
|
Footnotes
|
|---|
Published online ahead of print. Publication date available at www.cjasn.org.
|
References
|
|---|
- US Renal Data System:
USRDS 2005 Annual Data Report: Atlas of End-Stage Renal Disease in the United States, Bethesda, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases,2005
- Wolfe RA, Held PJ, Hulbert-Shearon TE, Agodoa LY, Port FK: A critical examination of trends in outcomes over the last decade.
Am Kidney Dis32[Suppl 4]
:S9
–S15,1998
- Bruns FJ, Seddon P, Saul M, Zeidel ML: The cost of caring for end-stage kidney disease patients: An analysis based on hospital financial transaction records.
J Am Soc Nephrol9
:884
–890,1998[Abstract]
- Mix TC, St. Peter WL, Ebben J, Xue J, Pereira BJ, Kausz AT, Collins A: Hospitalization during advancing chronic kidney disease.
Am J Kidney Dis42
:972
–981,2003[CrossRef][Medline]
- Arora P, Kausz AT, Obrador GT, Ruthazer R, Khan S, Jenuleson CS, Meyer KB, Pereira BJ: Hospital utilization among chronic dialysis patients.
J Am Soc Nephrol11
:740
–746,2000[Abstract/Free Full Text]
- Khan SS, Kazmi WH, Abichandani R, Tighiouart H, Pereira BJ, Kausz AT: Health care utilization among patients with chronic kidney disease.
Kidney Int62
:229
–236,2002[CrossRef][Medline]
- Ploth DW, Shepp PH, Counts C, Hutchison F: Prospective analysis of global costs for maintenance of patients with ESRD.
Am J Kidney Dis42
:12
–21,2003[CrossRef][Medline]
- Kshirsagar AV, Hogan SL, Mandelkehr L, Falk RJ: Length of stay and costs for hospitalized hemodialysis patients: Nephrologists versus internists.
J Am Soc Nephrol11
:1526
–1533,2000[Abstract/Free Full Text]
- National Kidney Foundation: Kidney Disease Outcome Quality Initiative, Clinical Practice Guidelines. Available: http://www.kidney.org/professionals/kdoqi/guidelines.cfm. Accessed June 12,2006
- Chazan JA, London MR, Pono L: The impact of diagnosis-related groups on the cost of hospitalization for end-stage renal disease patients at Rhode Island Hospital from 1987 to 1990.
Am J Kidney Dis19
:523
–525,1992[Medline]
- Becker BN, Coomer RW, Fotiadis C, Evanson J, Shyr Y, Hakim RM: Risk factors for hospitalization in well-dialyzed chronic hemodialysis patients.
Am J Nephrol19
:565
–570,1999[CrossRef][Medline]
- Carlson DM, Duncan DA, Naessens JM, Johnson WJ: Hospitalization in dialysis patients.
Mayo Clin Proc59
:769
–775,1984[Medline]
- Rocco MV, Bleyer AJ, Burkart JM: Utilization of inpatient and outpatient resources for the management of hemodialysis access complications.
Am J Kidney Dis28
:250
–256,1996[Medline]
- Loran MJ, McErlean M, Eisele G, Raccio-Robak N, Verdile VP: The emergency department care of hemodialysis patients.
Clin Nephrol57
:439
–443,2002[Medline]
Top
Abstract
Introduction
