Abstract

Modalities of renal replacement therapy are categorized into in-center hemodialysis and home therapies. A subset of hemodialysis patients referred to as in-center self-care hemodialysis (ICSCHD) receive patient training as if they were going home but instead perform their dialysis in-center with minimal staff support. Preliminary data suggests ICSCHD is associated with better outcomes than traditional in-center hemodialysis. We looked at ICSCHD patients initiating maintenance dialysis from April 1, 2011 to March 30, 2014 and compared them at a 1:2 ratio to propensity-score matched controls from surrounding facilities within the same catchment area. The median follow-up was 14 months. Patients on ICSCHD had lower mortality rate (0.02 vs 0.07 per patient year; p <0.05), fewer hospitalization events (0.82 vs. 1.70 per patient year; p = 0.008) and fewer missed treatments (1.1% vs 3.8% of all treatments; p = 0.005) than matched controls. We concluded that patients on ICSCHD had lower mortality rates and fewer hospital days than well-matched controls and spent more time on dialysis and missed fewer treatments. Establishing a facility-wide culture of care promoting patient engagement in ICSCHD may have contributed to these findings.

09-16-patient-management-figure1

Introduction

Patients who elect to do home hemodialysis (HHD) are generally trained in-center with a partner before going home. The terminology of “self-care” generally refers to home therapies, but self-care hemodialysis can also be done in-center with limited assistance from health care personnel and with the patient performing the tasks of dialysis. 1, 2 While rarely reported, in-center self-care hemodialysis (ICSCHD) is performed in different areas of the globe. 3-6

When a patient starts his/her shift, they begin by setting up their machine as prescribed by their physician. Some opt to cannulate themselves. They monitor hemodynamics, record data, and administer their intravenous medications.

09-16-patient-management-figure2

Outcomes from ICSCHD have not been systematically studied, although we previously reported our experience in a limited number of patients. 1 We found that outcomes were better compared to conventional hemodialysis patients treated in Fresenius Medical Care North America (FMCNA) facilities within the same geographic area. 1 Selection bias, either through self-selection by the patients themselves and/or in combination with perception of caregivers of a type of patient who can be persuaded to practice self-care, may channel healthier patients into this therapy. However, the processes of the treatment modality itself promotes better comprehension that enhances self-motivation and reinforces positive behaviors that could, in turn, influence improvement of outcomes.

Read also: Making the case for in-center, self-care hemodialysis 

Our current study extends our previous report by comparing patients’ outcomes and treatment parameters on ICSCHD to traditional in-center HD, utilizing matched period prevalent control patients from facilities that are managed by the same dialysis provider within the same catchment area.

Methods

Study design and population

The study population included all adult (age >18) patients admitted to a self-care dialysis facility from April 1, 2011 to March 30, 2014. Patients that are transient (i.e. not a permanent patient in the dialysis facility) were excluded. The source population used for matching controls included all adult permanent patients admitted during the same period to facilities within a 60-mile radius that reported electronically to the FMCNA data warehouse.7 All patients were followed until April 2015 and were eligible for a minimum one-year follow-up period. Patients were censored upon death, withdrawal from dialysis, kidney transplant, recovery of kidney function, or permanent transfer. FMCNA system-wide data were obtained for the 12-month period from April 1, 2012–March 30, 2013 to provide context for comparison. In addition, selected 2012 US Renal System data from the 2014 Annual Data Report8 and ESRD Network 4 data from 2012 9 were also obtained for reference.

09-16-patient-management-table1

The control group was created using 1:2 matching algorithm as shown in Figure 1.10 Exact matching was performed for catheter at baseline (yes/no), incident patient status (i.e. ≤ 120 days by vintage), race (black/white) and gender (male/female). Propensity score (PS) was calculated based on laboratory values at baseline (albumin, bicarbonate, calcium, ferritin, hemoglobin, potassium, PTH-Intact, serum phosphorus and transferrin saturation), presence/absence of specific comorbid diagnosis at baseline (diabetes/DM, chronic lung disease/COPD, heart failure/CHF, cerebrovascular disease/CVD), estimated dry weight (EDW) prescribed by the fourth HD treatment, vintage, and age. Finally, we used the nearest available Mahalanobis metric matching on age, vintage and COPD within the single decimal calipers (of ≤ 0.7) defined by the logit PS. Successful matching was based on standardized differences not exceeding 10%. 11 Overall, the PS distance achieved was logit PS difference of 0.08 (i.e. 8%) between the study and control groups.

Data collection

Baseline demographic information (age, sex, race, ethnicity, diagnosis of renal disease, dialysis vintage, and height) was collected from the FMCNA Knowledge Center data warehouse. Prescription information (length of treatment, vascular access, estimated dry weight, and number of treatments per week) on the fourth treatment and weekly averages for the 120th day and 365th day were also tracked. Mean weekly Epogen dose for each patient was calculated by summing all doses administered during the baseline month, the month prior to the 120th day, and the month prior to the 365th day, then dividing by the number of hemodialysis treatments, and multiplying by three treatments/week. Comorbid diagnoses at baseline, on day 120, and day 365 were collected. Laboratory tests (albumin, bicarbonate, calcium, ferritin, hemoglobin, potassium, PTH-Intact, serum phosphorus, transferrin saturation, and spKt/V) were tracked at baseline, around day 120 and day 365. The percent of unexcused missed treatments was calculated by dividing the number of unexcused missed treatments by the total number treatments (obtained by summing all provided treatments, unexcused missed treatments, and treatments missed due to hospitalization) and multiplying by 100. Lab measurements were reported by a central laboratory (Spectra Laboratories, Rockleigh, NJ; www.spectra-labs.com). Data validation with the facility clinical manager was performed to ensure accuracy of patient admission, date of first (ever) chronic dialysis treatment, and missed treatment information.

Endpoints

The primary endpoints included death, withdrawal from dialysis, and hospitalization. The secondary endpoints included unexcused missed treatments, prescribed dialysis time, Epogen dose per treatment and 120 day/365 day laboratory values. Patients lost to follow-up contributed exposure time until kidney transplant or the last day before transfer out of the FMCNA system.

Read also: Self-care and the Tablo dialysis machine 

Statistical Methods

Descriptive analyses of patient characteristics are presented as means, medians, or percentages of the total. T-tests and Wilcoxon tests were performed to compare two groups where appropriate. We plotted Kaplan-Meier survival curves to illustrate time to death. Poisson regression models were performed to compare event rates. Sensitivity analysis were adjusted for matching variables with standardized differences exceeding 10%. All analyses were conducted using SAS v9.3 (Cary, NC; www.sas.com). The New England Institutional Review Board categorized the study as exempt from full review as a minimal risk Quality Improvement Project designed to review program outcomes.

09-16-patient-management-table2

Results

The 40 ICSCHD cohort patients were well-matched with the 80 conventional HD controls (see Table 1 and Table 2). Standardized differences exceeded 10% only for potassium (0.11), iPTH (0.12) and the proportion with CVD (0.12) in the study population. Compared to the source population, patients in the study cohort were younger, were predominantly black (due to the population mix of the ICSCHD catchment area), with slightly lower estimated dry weight, and generally similar lab values at baseline. The average age of the ICSCHD patients was 49.4 years, much younger than those in ESRD Network 4 (63.5 years) 9 and FMCNA (62.4 years) but were well-matched to controls (49.5 years).

09-16-patient-management-table3

Trends

Table 2 provides follow-up values at 120 and 365 days. ICSCHD patients had significantly higher ferritin and Hgb values at 120 days. At 365 days, ICSCHD patients had significantly higher bicarbonate, calcium, and ferritin values, and significantly lower serum phosphorous than controls. All other results exhibited no significant differences.

Trends observed during the first year for patients in both ICSCHD and control groups included:

  • Both groups saw an increase in albumin, calcium, and potassium over time.
  • The ICSHD group saw a slight increase in bicarbonate but remained unchanged in the control.
  • Parathyroid hormone fell in both groups. Calcium increased and phosphorous remained unchanged in the ICSCHD group but increased in the control group.
  • In both groups hemoglobin increased at 120 days then stabilized thereafter. Ferritin and transferrin saturation increased in both groups.

In a subset of 71 patients (24 in ICSCHD and 47 controls; data not shown) with a full year follow-up, statistically significant increases over time were found in both groups for albumin, Hgb, potassium, ferritin (with a significantly higher increase in the ICSCHD group), and transferrin saturation, similar to the observations for the entire cohort above. Significant increase in calcium was found only in ICSCHD. Significant increase in serum phosphorous was found only in controls, although the difference in trends between groups over time was not significant.

The administered Epogen doses decreased during follow-up in both groups, with lowest doses administered to ICSCHD patients at 365 days, although the percent of patients receiving Epogen were similar throughout (see Table 3). There was no difference in dialysis adequacy (spKt/V) between the two groups. At 120 days all patients in the study group were dialyzed for four or more hours compared to the control group (78.8%). At one year, 95.8% of ICSCHD patients compared to 76.6 % of control patients were dialyzed for ≥ 4 hours. Patient disposition by the end of follow-up are shown in Table 4 and discussed below.

09-16-patient-management-table4

Mortality

Over a median follow-up of 439 days there were 11 total deaths (including three withdrawals) in the control group compared to only one death for the ICSCHD cohort. The death rates were 0.02 per patient year (ppy) for ICSCHD and 0.07 ppy for controls (see Figure 2), both lower than the 0.16 ppy reported by the US Renal Data System. 8 The statistical significance of the mortality difference between cohorts was sensitive to technique, at p = 0.048 using the Wilcoxon method (planned analysis) and p = 0.053 using the log rank method (serendipitous observation).

Technique Survival

Twenty-two of the original 40 patients left ICSCHD with six switching to home therapies and four to a nocturnal in-center hemodialysis program by choice; five left due to proximity to work or home, and two opted not to continue self-care (see Table 4). Two were unable to accomplish the tasks of self-care and there were three patients who had personality clashes. One patient was unable to follow established rules.

Hospitalization

ICSCHD patients had a rate of 0.82 hospitalization events ppy compared to 1.70 events ppy in the control group (p = 0.008). ICSCHD patients had 3.9 hospital days ppy compared to 10.1 days ppy in the control group (p=0.002) with 11.0 days ppy reported by USRDS 11 and 10.8 days ppy overall in FMCNA. Time to first hospitalization was not statistically different between groups (p = 0.69).

Unexcused missed treatments

The ICSCHD cohort had 1.1% of all treatments as unexcused missed treatments compared to 3.8% of all treatments in the control group (p = 0.005).

Discussion

This study compared ICSCHD for patients with ESRD to traditional in-center hemodialysis. We previously reported favorable outcome data on a small number of ICSCHD patients 1 that had a smaller sample size and no adjustment for confounding. While not free from residual confounding, the current larger study confirmed better survival, fewer hospitalizations, and fewer missed treatments for patients treated by ICSCHD when compared to propensity score and area-based well-matched controls.

Most of the literature describing self-care refers to home modalities; there are a few older reports 3-6 and one recent study of 17 self-care patients in Canada that only reported on quality of life data in those patients.15 In the early 1980s, Bray et al. established a 6-station “ICSCHD” facility intended to train patients to go home. 2 Most of their patients had no eligible partner so were unable to go home expeditiously. These self-care hemodialysis patients were trained to perform their own dialysis with minimal staff assistance, then subsequently elected to stay in-center rather than go home. It was noted that outcomes in those patients appeared to be better than for patients who were in conventional in-center HD, 1, 2 thus cementing the concept of ICSCHD in what is now the 10-station facility described in the current study.

In our study, the annual hospitalization rates and hospital days were significantly lower in ICSCHD patients. The annual mortality rate was also lower but sensitive to the statistical test used. While the Wilcoxon test planned a priori indicated a p < 0.05 for the mortality rate difference observed (p=0.048 to be exact), a serendipitous observation from the statistical output revealed that the log rank test result was at p = 0.053. We opted to also report this test for transparency, illustrating borderline significance of the results.

Nevertheless, we were encouraged that the study was able to demonstrate better mortality despite having only 40 ICSCHD patients, which if replicated in a larger cohort, would have a higher likelihood of attaining definitive statistical significance. It should be noted that the ICSCHD mortality and hospitalization results were not only lower than matched controls, but the favorable characteristics of the cohort (e.g. younger age, etc.) were consistent with the control group’s rates that were much lower than the general ESRD population rates from USRDS 8 and FMCNA, as well as from the study facilities’ ESRD Network 4. 9 Therefore, to a certain extent, we observed persistent favorable significant incremental differences in outcomes of the ICSCHD beyond multiple clinical variables that we attempted to balance out to reduce confounding.

There are numerous potential mechanisms to account for these results, including inherent selection biases of simply being considered for treatment in the ICSCHD facility. However, the propensity-score based matching process was designed to diminish those potential biases. The ICSCHD patient population was younger than the Network and FMCNA patients in general, but the control population was well matched for age.

There is a dramatic preponderance of black race (85%). African American race has been known to portend a survival advantage relative to whites 13-15 albeit not universally so.16,17 Nevertheless, the control group was also well matched for race. Comorbidities were not a major factor in the study, in that they were well-matched between groups. The impact of unmeasured local, social, cultural, and economic factors were hopefully balanced (or at least attenuated) because the patient population from which the ICSCHD patients were drawn represented a catchment area that was identical to that for surrounding facilities in the control group.

An important characteristic demonstrated by ICSCHD patients was that 67.5% of patient did not miss treatments. Only 1.1% of all treatments had unexcused missed treatments compared with 3.8% in the control group. Studies have demonstrated that the mortality and hospitalizations are 30% higher for one missed treatment. 18,19 Establishing a facility-wide culture of self-care that is heavily supported by physicians and the facility staff may be the driver towards patient engagement, which in turn, promotes not only comprehension but more importantly, motivated compliance. Anecdotally, the ICSCHD facility environment from the Medical Director to the staff to the patients are vigilant about missed treatments with patients themselves known to “police” each other about missed and shortened treatments.

Patients on ICSCHD generally dialyzed ≥ 4 hours per session for three times per week; at 365 days, only 1 of 40 patients spent < 4 hours on dialysis. The average treatment time in the ICSCHD facility was 252 minutes. Patients in both groups were adequately dialyzed, reflected by a spKTV of 1.7-1.8. An increasing body of literature suggests that the more time spent on dialysis results in decreased hospitalization and mortality, better quality of life, and fewer cardiovascular events. 20-22 As noted above, the culture of the facility from the staff, physician, and the patients is to not tolerate shortening treatments.

Soft trends over time may also indicate compliance with clinical guidance, such as with decreasing catheter rates and lowering serum phosphorus. Although catheter rates were the same at baseline, catheter rates decreased over time and by 365 days was 16.7% in ISCHD patients compared to 23.4% in controls. Central venous catheters (CVC) have been reported to be associated with higher hospitalization rates and mortality. 23-25 Indeed, DOPPS data suggests that CVCs are the primary culprit in high mortality rates. 26,27 Similarly, while patients in both groups maintained similar estimated dry weight and good serum albumin levels, the serum phosphorus levels were lower in the ICSCHD groups after 365 days. High phosphorus levels have also been observed to be associated with worse outcomes in dialysis patients. 28-31

In essence, the longer time on dialysis, lower missed treatments, improved compliance and lower catheter rates, may play a role in the better outcomes observed in self-care patients. These findings may be a result of the “culture” created in the self-care environment of the dedicated ICSCHD facility.

A total of 23 patients left the self-care environment for various reasons with two of them unable to perform the tasks for self-care hemodialysis and two decided to no longer do self-care. Three patients could not abide by basic compliance rules within the self-care environment and one had personality conflicts with patients or staff.

Clearly self-care is not for everyone. Of the patients who left ICSCHD, one was lost to follow-up but we were able to track the other patients who left ICSCHD. Among these patients, three were transplanted and three died compared to one death among patients who stayed on ICSCHD. Notwithstanding, a substantial proportion of patients do thrive on ICSCHD.

The general philosophy of the medical team is that all patients are self-care candidates until proven otherwise. However, patients that exhibit clinically significant visual, psychological, and cognitive impairment are often excluded. Among those offered, there are patients who refuse and are unwilling to learn self-care.

Prior noncompliance is not necessarily an exclusion criterion a priori in patients who are persuaded and express willingness to try it. For this experience to be replicated and propagated, we enumerate examples of critical success factors, such as a dedicated medical director, nurses and technicians, education of staff and patients, as well as a concept of continuous education, empowerment of patients, use of a collaborative in lieu of a paternalistic approach (particularly by physicians), and a dedicated exclusive space (either entire facilities as is present in this study, an entire shift of patients, or dedicated areas within the facility).

Prior experience suggests that mixing self-care and traditional dialyzed patients and staff did not lead to diffusion of only favorable behaviors, but there was also sufficient temptation to dilute the culture of self-empowered discipline and care. Therefore, we consider the absence of controlled environment as a major barrier to establishing and maintaining a successful ICSCHD program. The culture has to be established by a strong core of staff and patients, nurtured in its infancy. Initially, an active selection process (similar to predicting favorable candidates for home dialysis) may need to be utilized, before a mature program begins open recruitment such as currently practiced in this facility.

Financial aspects of ICSCHD are beyond the scope of the current report and data are not available to investigators. However, we share logistic factors for administrators interested in exploring replication of our program. Staffing ratios are slightly better (5 patients to one patient care technician or PCT) than in our other facilities (4 patients to one PCT) although there is one RN per two PCT rather than standard one RN to 3 PCT (in the context of a 10-station facility). Fewer hospitalizations and hospital days benefit patients, use less hospital resources and cost less to insurers. We also view it as contributory to fewer missed treatments (in addition to less unexcused absences), keeping patients in their seats as scheduled and thus increasing the facility’s efficiency. Furthermore, in the current prospective payment system, eligible self-care patients are entitled to the Medicare benefit on day one of dialysis similar to home therapies. Therefore, an additional 120 days of reimbursement at a 1.51 modified rate is a financial benefit to the dialysis provider. In addition, the facility and the physician both receive training fees for patients similar to that for home therapies.

We have anecdotally noticed that roughly a third of the patients who consider self-care HD are employed at least part-time, compared to 10% in our other facilities, leading to a greater proportion of patients who keep commercial insurance. These information contain anecdotal presumptions that require further elucidation.

While this study represents the largest available and most rigorous report on outcomes associated with ICSCHD, it has several limitations. First, as an observational study, it establishes associations but cannot prove causality. Second, despite attempts to control for confounding variables, it did not eliminate (residual) confounding. Third, there was incomplete capture of information for outpatient oral medication compliance. Fourth, it was a single center experience localized to a Northeast U.S. location with patient demographics and hospitalization practices that are not necessarily generalizable to the rest of the country. Therefore, our findings are intended to educate, encourage interest, and stimulate further replication, research, and reporting on ICSCHD.

Summary

We describe here a subcategory of in-center hemodialysis in ICSCHD that appears to have better outcomes when compared to traditional in-center hemodialysis. Patients on ICSCHD spent more time on dialysis and missed fewer treatments, demonstrating lower mortality rates and fewer hospital days than well-matched controls. Patients in ICSCHD tended to have lower central-venous catheter rates and lower serum phosphorus, markers of compliance-to-care recommendations which could be contributory to beneficial outcomes. Establishing a facility-wide culture of care promoting patient engagement in ICSCHD may have led to these findings. Ideally, randomized controlled studies (RCTs) comparing ICSCHD to traditional maintenance thrice weekly in center hemodialysis are needed to confirm our findings. However, due to the logistical barriers in conducting RCTs, replication and propagation from different areas of the country will provide additional information to enrich this experience. In the absence of such information, we posit that ICSCHD represents another valuable dialytic option that is associated with better outcomes in motivated patients with ESRD.

Acknowledgements

Partial funding for this study was provided by Outset Medical Inc. We would like to thank our clinical staff and educators from self-care who are responsible for our excellent clinical outcomes.

Disclosure

E.R.J. provides consulting services to Fresenius Medical Services, Physician Choice Management, and Reliant Renal Care. He is a board member of Cytosorbent. He is a counselor to the Renal Physicians Association. At the time of the study, L.J., E.L., A.M., S.R. were full time employees of FMCNA.

References

  1. Jones ER, James LR. In-center self-care hemodialysis: An unappreciated modality in renal care. Neph News and Issues. 2011; 9:31-33,37.
  2. Bray SH. Does self-care dialysis improve quality-of-life for ESRD patients? Contemporary Dialysis and Nephrology. 2003; (May) 24-26
  3. Piccoli GB, Bermond F, Mezza E, et al. Piccoli G. Home hemodialysis à la carte: a tailor made program (1998-2003). J Nephrol.2004; 17(1):76-8
  4. McMurray MH. Seniors and self-care hemodialysis. J CANNT.1995; 5(1):13-4.
  5. Fournier G, Gaillard JL, Matha N, Man NK. Self-dialysis: a new method of treatment for end-stage renal disease. Nephrologie.1984; 5 (3):115-8.
  6. Eschbach JW, Seymour M, Potts A, Clark M, Blagg CR. A hemodialysis orientation unit. Nephron.1983; 33(2):106-10.
  7. Krishnan, M, Wilfert HM, Lacson E, Jr. In data we trust: the role and utility of dialysis provider databases in the policy process. Clin J Am Soc Nephrol. 2012 11:1891-96
  8. USRD/s: 2014 Annual Data Report Atlas of Chronic Kidney Disease and End-Stage Renal disease in the United States. Bethesda, Md. National Institutes of Health 2014
  9. Dialysis Facility Report at https/dialysisdata.org. Accessed November 4, 2015
  10. Matchit: Nonparametric preprocessing for parametric for parametric causal inference. Journal of Statistical Software. 2011: 42 (8)
  11. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Austin PC. Stat Med. 2009; 28 (25):3083-107
  12. Meers C, Singer MA, Totlelmine EB, Hopman W, McComb J, Mackensie TA. Self-delivery of hemodialysis care: A therapy in itself. Am J of Kid Dis 1996; 27(6): 844-847
  13. Wong JS, Port FK, Hulbert-Shearon TE, et al. Survival advantage in Asian American end-stage renal disease patients. Kidney Int. 1. 1999; 55:2515.
  14. Tanna MM, Vonesh EF, Korbet SM. Patient survival among incident peritoneal dialysis and hemodialysis patients in an urban setting. Am J Kidney Dis 2000; 36:1175-1182
  15. Bleyer AJ, Tell GS, Evans GW, Ettinger WH, Burkart JM. Survival of patients undergoing renal replacement therapy in one center with special emphasis on racial differences. Am J Kidney Dis 1996; 28:72-8.
  16. Mesler DE, McCarthy EP, Byrne-Logan S, Ash AS, Moskowitz MA. Does the survival advantage of nonwhite dialysis patients persist after case mix adjustment? Am J Med. 1999; 106:300-306
  17. Robinson BM, Joffe MM, Pisoni RL, Port FK, Feldman H Il. Revisiting survival differences by race and ethnicity among hemodialysis patients: the Dialysis Outcomes and Practice Patterns Study. J Am Soc Nephrol 2006; 17:2910-2918.
  18. Chan KE, Thadhani RI, Maddux FW. Adherence barriers to chronic dialysis in the United States. J Am Soc Nephrol. 2014; 25(11):2642-8.
  19. Saran R, Bragg-Gresham JL, Rayner HC, et al: Nonadherence in hemodialysis: Associations with mortality, hospitalization, and practice patterns in the DOPPS. Kidney Int 2003; 64: 254–262.
  20. Tentori F, Zhang J, Li Y, Karaboyas A, et al. Longer dialysis session length is associated with better intermediate outcomes and survival among patients on in-center three times per week hemodialysis: results from the Dialysis Outcomes and Practice Patterns Study (DOPPS). Nephrol Dial Transplant. 2012;27( 11):4180-8.
  21. Saran R, Bragg-Gresham JL, Levin NW, et al. Longer treatment time and slower ultrafiltration in hemodialysis: associations with reduced mortality in the DOPPS. Kidney Int. 2006;69(7):1222-8.
  22. Weinhandl ED, Nieman KM, Gilbertson DT, Collins AJ. Hospitalization in daily home hemodialysis and matched thrice-weekly in-center hemodialysis patients. Am J Kidney Dis 2015;65(1) 98-108.
  23. Xue H, Ix JH, Wang W, Brunelli SM, Lazarus M, Hakim R, Lacson E Jr. Hemodialysis access usage patterns in the incident dialysis year and associated catheter-related complications. Am J Kidney Dis. 2013 61(1):123-30.
  24. Lacson E Jr, Wang W, Lazarus JM, Hakim RM. Change in vascular access and hospitalization risk in long-term hemodialysis patients. Clin J Am Soc Nephrol. 2010 Nov; 5 (11):1996-2003.
  25. Lacson E Jr, Wang W, Lazarus JM, Hakim RM. Change in vascular access and mortality in maintenance hemodialysis patients. Am J Kidney Dis. 2009; 54(5):912-21.
  26. Bradbury BD, Chen F, Furniss A, et al. Conversion of vascular access type among incident hemodialysis patients: description and association with mortality. Am J Kidney Dis. 2009; 53(5):804-14.
  27. Pisoni RL, Arrington CJ, Albert JM, et al. Facility hemodialysis vascular access use and mortality in countries participating in DOPPS: an instrumental variable analysis. Am J Kidney Dis. 2009; 53(3):475-91.
  28. Lacson E Jr, Wang W, Hakim RM, Teng M, Lazarus JM. Associates of mortality and hospitalization in hemodialysis: potentially actionable laboratory variables and vascular access. Am J Kidney Dis. 2009; 53(1):79-90
  29. Umeukeje EM, Merighi JR, Browne T, et al. Self-Motivation Is Associated With Phosphorus Control in End-Stage Renal Disease. J Ren Nutr. 2015; 25(5):433-9.
  30. Block GA, Klassen PS, Lazarus JM, Ofsthun N, Lowrie EG, Chertow GM. Mineral metabolism, mortality, and morbidity in maintenance hemodialysis. J Am Soc Nephrol. 2004; 15 (8):2208-18

Lertdumrongluk P, Rhee CM, Park J, et al. Association of serum phosphorus concentration with mortality in elderly and nonelderly hemodialysis patients. J Ren Nutr. 2013; 23(6):411-21. care