Correspondence
Ownership of Dialysis Facilities and Patients' Survival
N Engl J Med 2000; 342:1053-1056April 6, 2000
- Article
To the Editor:
Garg et al. (Nov. 25 issue)1 reported on the effect of the ownership of dialysis facilities on patients' survival and referral for transplantation. They described large differences in outcomes among patients undergoing dialysis who had been grouped according to the type of dialysis facility in a sample of facilities. Although they are not cited by the authors, the U.S. Renal Data System (USRDS) has three times published standardized mortality ratios for all facilities, which were grouped according to the three types considered by the authors: free-standing for-profit, free-standing not-for-profit, and hospital facilities.2,3 The standardized mortality ratio, which adjusts for differences in age, sex, race, and diabetes as a cause of end-stage renal disease,4,5 was 5 to 7 percent higher in for-profit facilities than in not-for-profit facilities for the period from 1991 through 1995 (a period that is concurrent with that used in the study by Garg et al.). The standardized mortality ratio is used to evaluate mortality rates for both new and continuing patients at a facility during a specific period.
Because the difference in the mortality rates reported by the USRDS (5 to 7 percent) is substantially smaller than that reported by Garg et al. (20 percent), we analyzed recent national data from USRDS standard-analysis files (which were also used by Garg et al.). We included in the analysis all patients who began receiving dialysis in 1995 or 1996 at all facilities and who were followed through 1997. We used the Cox model to assess the relative risk of death at 60 days after the start of end-stage renal disease; we censored data on patients when they underwent transplantation, and we adjusted for age, race, sex, cause of end-stage renal disease, geographic region, and the presence or absence of 15 coexisting conditions (in 93,383 patients at 3046 dialysis facilities). The relative risk was 1.06 (95 percent confidence interval, 1.01 to 1.12, when the clustering of patients in facilities was accounted for) for patients treated in for-profit facilities as compared with not-for-profit facilities.
Using these same national data, we also analyzed the rates of transplantation among patients in all free-standing dialysis facilities according to profit status. We found the adjusted rate of placement on a waiting list for a renal transplant to be 15 percent lower in for-profit facilities than in not-for-profit facilities — not 26 percent lower, as reported by Garg et al. More important to patients is the actual rate of transplantation, which was 8 percent lower in for-profit units than in not-for-profit units; this difference was not significant.
These national data confirm earlier findings by the USRDS of a significantly higher risk of death (6 percent higher) in for-profit dialysis facilities than in not-for-profit facilities. These more reliable estimates suggest that the reports by Garg et al. of a 20 percent higher mortality rate and a 26 percent lower rate of placement on a waiting list for a transplant are substantially overestimated.
5 ReferencesFriedrich K. Port, M.D.
Robert A. Wolfe, Ph.D.
University of Michigan, Ann Arbor, MI 48103Philip J. Held, Ph.D.
University Renal Research and Education Association, Ann Arbor, MI 481031
Garg PP ,Frick KD ,Diener-West M ,Powe NR . Effect of the ownership of dialysis facilities on patients' survival and referral for transplantation. N Engl J Med 1999;341:1653-1660
Full Text | Web of Science | Medline2
USRDS 1995 annual data report. Bethesda, Md.: National Institutes of Health, 1995.
3
USRDS 1997 annual data report. Bethesda, Md.: National Institutes of Health, 1997.
4
Wolfe RA ,Gaylin DS ,Port FK ,Held PJ ,Wood CL . Using USRDS generated mortality tables to compare local ESRD mortality rates to national rates. Kidney Int 1992;42:991-996
CrossRef | Web of Science | Medline5
Wolfe RA . The standardized mortality ratio revisited: improvements, innovations, and limitations. Am J Kidney Dis 1994;24:290-297
Web of Science | Medline
To the Editor:
After critical review, we find the results of the study by Garg et al. erroneous and the conclusions biased and unsubstantiated. Their main conclusions, a higher mortality rate and a lower rate of placement on a waiting list for a transplant — and therefore poorer quality of care — among patients in for-profit dialysis units, are not current and are not supported with proper interpretation of the data. A much larger study of the same source of data — the 1999 USRDS annual data report1 — analyzed mortality for the entire population of patients with end-stage renal disease as a standardized mortality ratio that adjusted for age, sex, diagnosis, and race and found the differences in the relative risk of mortality among patients in not-for-profit units (0.94), for-profit units (1.01), and hospital units (1.04) significantly smaller than those found by Garg et al. The USRDS also evaluated standardized hospitalization ratios in the various categories of ownership. The ratio was higher among patients in for-profit units, and the USRDS suggested that this finding could indicate that for-profit units treated sicker patients.
The sample used in the study by Garg et al. was very small (less than 3 percent of the available patients in not-for-profit facilities), was derived from multiple data subsets never intended for this type of analysis, and therefore involved a larger-than-acceptable chance for sampling error, selection bias, or both. In addition, we believe that Garg et al. failed to adjust appropriately for the coexisting factors (age, diabetes, and nursing home placement) in patients cared for in for-profit units and did not consider other relevant coexisting conditions. They did not mention other measures of quality, which were readily available in the USRDS data base. Did they find any differences?
The crude mortality rates for both the not-for-profit and for-profit facilities were lower than the national average (according to USRDS data). This suggests that the data set used by Garg et al. was not representative. The authors also failed to consider the rapid improvement in the case-mix–adjusted mortality rate in the recent past (from 32 percent in 1986 to 21.5 percent in 1996), during which time for-profit ownership and consolidation increased dramatically.1
A decision regarding transplantation is most often made before the patient's arrival at a free-standing facility. It is overwhelmingly a matter between the nephrologist and the patient and is a complex interplay of religious beliefs, economic situations, education, race, geographic location, and family dynamics. Ayanian et al.2 examined this issue and found no difference between for-profit and not-for-profit facilities in the rate of referral for transplantation. Wolfe et al.3 noted that patients placed on waiting lists for a transplant are very selected and have a death rate significantly lower than that of patients left behind in dialysis units. Referral for a transplant evaluation is a much more ap-propriate measure than placement on a waiting list for a transplant, because the latter reflects yet another selection process.
3 ReferencesSteven J. Bander, M.D.
Gambro Healthcare, St. Louis, MO 63112J. Michael Lazarus, M.D.
Fresenius Medical Care North America, Lexington, MA 02420-9192Stan M. Lindenfeld, M.D.
Total Renal Care, Torrance, CA 905031
USRDS 1999 annual data report. Bethesda, Md.: National Institutes of Health, 1999:76.
2
Ayanian JZ ,Cleary PD ,Weissmen JS ,Epstein AM . The effect of patients' preferences on racial differences in access to renal transplantation. N Engl J Med 1999;341:1661-1669
Full Text | Web of Science | Medline3
Wolfe RA ,Ashby VB ,Milford EL , et al. Comparison of mortality in all patients on dialysis, patients on dialysis awaiting transplantation, and recipients of a first cadaveric transplant. N Engl J Med 1999;341:1725-1730
Full Text | Web of Science | Medline
To the Editor:
The findings of Garg et al. are based on data that are 7 to 10 years old. Before the findings reported herein are extrapolated to the current state of dialysis care, it is critical that they be validated in a contemporary data set that reflects the recent and substantial changes in health care delivery and the improvement in quality in the End Stage Renal Disease Program.1 Since 1993, considerable improvement in patients' care has been achieved by identifying relevant intermediate outcomes and by facilitating the delivery, monitoring, and feedback of these to dialysis providers. Examples of core indicators of care are the dose of hemodialysis, the extent of anemia correction, and nutritional status; each is an independent predictor of mortality among patients undergoing dialysis and reflects the processes of care.2 The Health Care Financing Administration (HCFA), End Stage Renal Disease networks, and professional societies developed strategies, based on studies from the early 1990s, to facilitate improvements in care using the methods of continuous quality improvement.1
Consequently, from 1993 to 1999, the proportion of patients receiving an adequate dose of hemodialysis increased from 43 percent to 81 percent and the mean hematocrit increased from 30.5 percent to 34.3 percent.3 From 1990 to 1997, concurrent with these improvements in care, the annual adjusted mortality rate decreased from 24.9 percent to 22.8 percent, as reported by the USRDS.4 Notably, this decrease in mortality occurred despite the concurrent trend toward the consolidation of free-standing dialysis units into for-profit corporate providers. Now approximately 65 percent of dialysis units are for-profit. The observed improvement in patients' survival during this period seriously calls into question the external validity of the observations of Garg et al.
Putative associations between the ownership of dialy-sis facilities and patients' outcomes need to be evaluated with contemporary data sets. In contrast to the results in periods of observation that preceded the development of relevant clinical-practice guidelines and their translation into measures of performance, the results may now be substantially different. After all, from a mercantile perspective, a live and healthy patient undergoing dialysis is a paying customer.
4 ReferencesAllen R. Nissenson, M.D.
William F. Owen, Jr., M.D.
Renal Physicians Association, Rockville, MD 208521
Nissenson AR ,Rettig RA . Medicare's End Stage Renal Disease Program: current status and future prospects. Health Aff (Millwood) 1999;18:161-179
CrossRef | Web of Science | Medline2
Madore F ,Lowrie EG ,Brugnara C , et al. Anemia in hemodialysis patients: variables affecting this outcome predictor. J Am Soc Nephrol 1997;8:1921-1929
Web of Science | Medline3
Health Care Financing Administration. 1999 Annual report, End Stage Renal Disease Core Indicators Project, Department of Health and Hu-man Services. Baltimore: Health Care Financing Administration (inpress).
4
USRDS 1999 Annual data report. Bethesda, Md.: National Institutes of Health, April 1999.
To the Editor:
Don't all facilities, including university-affiliated centers, ultimately use the revenue generated by dialysis as profit? In a university-based service, reimbursements for dialysis, although not used for stockholders, are nonetheless used for the benefit of the institution. Garg et al. presented no data to support their implied contention that patients in not-for-profit dialysis facilities received better dialysis than patients in for-profit centers. The only potential measure of the adequacy of dialysis used was albumin concentration, which can be affected by a multitude of factors. Why were standard measures of the adequacy of dialysis, such as KT/V or urea reduction ratio, which can be easily obtained from USRDS data bases, not used in comparing the two groups? If for-profit dialysis units had higher mortality rates, they should have had quantifiable differences in the time spent undergoing dialysis or the dose of dialysis delivered.
With regard to the issue of referral for transplantation, the authors failed to emphasize the effect of logistic factors such as the distance from a major transplantation center. In south Texas, the referral rate and the length of time before a patient is placed on the United Network for Organ Sharing waiting list for a cadaveric transplant depend less on whether the patient's primary referral facility is classified by HCFA as for profit or not for profit and more on its geographic distance from the transplantation center. This may help explain the finding by Garg et al. that for-profit units closer to university-based centers have better rates of referral than free-standing for-profit units farther away.
Wajeh Y. Qunibi, M.D.
Daniel J. Riley, M.D.
Hanna E. Abboud, M.D.
University of Texas Health Science Center at San Antonio, San Antonio, TX 78229To the Editor:
Garg et al. did not consider or discuss the possible effects of HCFA's classification of free-standing dialysis units. Both for-profit and not-for-profit free-standing units may be affiliated with large medical centers or academic teaching programs; we believe, however, that at the time of this study, such medical centers or academic programs were much more often affiliated with not-for-profit units. For the most part, most not-for-profit units in 1990 and 1993 were affiliated with large medical centers and academic programs.
As shown by the substudy by Garg et al., many for-profit units were in counties with no competing hospital-based or not-for-profit units. We suspect that the finding of only for-profit units in some counties represents the placement of for-profit units in sparsely populated and poorly served counties where there are no major medical centers or academic programs. These facilities are often smaller than those near major medical centers, and their small size may be associated with excess mortality because a disproportionate number of high-risk patients may be enrolled.1
In the substudy, in counties where there was competition — that is, both for-profit and not-for-profit units — there was no difference in either the mortality rate or the rate of placement on a waiting list for a transplant. The authors speculate that the not-for-profit competition forced the for-profit units to improve; however, a much more likely and reasonable explanation is that these counties contained large medical centers and academic programs in which both for-profit and not-for-profit units exist. We would argue that units closely affiliated or associated with such centers, which have a large staff, readily available subspecialists, and sophisticated emergency rooms, are likely to have better outcomes. The issue is not ownership but the relation with large medical centers or academic programs, which are also the sites of transplantation programs. One must question whether the findings by Garg et al. were a result of ownership or of variation in the regional patterns of practice of the medical community at large. The authors did not consider other such correlates of free-standing units or the implications of their being placed in communities to serve the needs of patients better.
1 ReferencesJoseph John Ruma, B.A.
Gail S. Wick, R.N., C.N.N.
Fresenius Medical Care North America, Lexington, MA 02420-91921
Flanigan M . Excess mortality in small dialysis centers: the result of dialyzing high-risk patients. ASAIO J 1995;41:177-181
Medline
Author/Editor Response
The authors reply:
To the Editor: The results of Port et al., which were based on more recent data, corroborate our findings of higher mortality rates and lower rates of placement on the waiting list for a transplant in for-profit dialysis centers. We assume that they obtained data on coexisting diseases from the HCFA Medical Evidence Report (Form 2728) and not from medical records, as was done in our study. The misclassification of coexisting conditions on Form 2728 occurs and may bias associations toward null findings.1 In part, this could explain the smaller effects found by Port et al. The failure to consider patients who change providers over time might also blunt the measures of association.
Bander et al. question the appropriateness of the study cohort; however, as illustrated by Figure 1 in our article, the USRDS was very careful in its sampling strategy to ensure nationally representative groups of patients and facilities. Unmeasured differences in case mix may exist between for-profit and not-for-profit centers. However, the standardized mortality ratios cited by both Bander et al. and Port et al. account for few of the coexisting conditions that we included. Our hypotheses examined only mortality and placement on the waiting list for a transplant because these are unambiguous, important outcomes and are more reliable than dose of dialysis,2 which is currently subject to differences among facilities in terms of the methods of measurement (e.g., the timing of postdialysis measurements of blood urea nitrogen). Placement on the waiting list is relevant because it is an essential step toward the receipt of a cadaveric transplant. In addition, by using waiting-list placement rather than an antecedent end point such as referral for evaluation for transplantation, we conservatively biased our findings toward the null.
The efforts cited by Nissenson and Owen to improve the care and outcomes of patients with end-stage renal disease (through the Dialysis Outcomes Quality Initiative3 and the Core Indicators Project4) should be applauded. However, even if national care is improving, it would be desirable to erase the differences that persist in the care provided by for-profit and not-for-profit facilities.
Qunibi et al. and Ruma and Wick suggest that the proximity of some facilities to transplantation centers or academic medical centers explains our findings. We examined distance to the nearest transplantation center and controlled for it when it was a confounding factor. Affiliations with academic health centers also may influence the quality of care and should be assessed in future research. We hope that our findings regarding facility ownership, as well as those by Port et al., will stimulate such efforts so that we can understand and reduce differences in outcomes for patients with chronic renal disease.
4 ReferencesPushkal P. Garg, M.D.
Harvard Medical School, Boston, MA 02115-5899Kevin D. Frick, Ph.D.
Neil R. Powe, M.D., M.P.H., M.B.A.
Johns Hopkins University, Baltimore, MD 21205-19011
Longnecker JC ,Coresh J ,Klag MJ , et al. Validation of comorbid conditions on the end-stage renal disease Medical Evidence Report: the CHOICE study. J Am Soc Nephrol 2000;11:520-529
Web of Science | Medline2
Owen WF Jr ,Chertow GM ,Lazarus JM ,Lowrie EG . Dose of hemodialysis and survival: differences by race and sex. JAMA 1998;280:1764-1768
CrossRef | Web of Science | Medline3
National Kidney Foundation-Dialysis Outcomes Quality Initiative clinical practice guidelines. New York: National Kidney Foundation, 1997.
4
McClellan WM ,Soucie JM ,Krisher J ,Caurana R ,Haley W ,Farmer C . Improving the care of patients treated with hemodialysis: a report from the Health Care Financing Administration's ESRD Core Indicators Project. Am J Kidney Dis 1998;31:584-592
CrossRef | Web of Science | Medline
- Citing Articles (5)
Citing Articles
1
Donald K. K. Lee, Glenn M. Chertow, Stefanos A. Zenios. (2010) Reexploring Differences among For-Profit and Nonprofit Dialysis Providers. Health Services Research 45:3, 633-646
CrossRef2
John M. Brooks, Christopher P. Irwin, Lawrence G. Hunsicker, Michael J. Flanigan, Elizabeth A. Chrischilles, Jane F. Pendergast. (2006) Effect of Dialysis Center Profit-Status on Patient Survival: A Comparison of Risk-Adjustment and Instrumental Variable Approaches. Health Services Research 41:6, 2267-2289
CrossRef3
L A Szczech, P S Klassen, B Chua, S S Hedayati, M Flanigan, W M McClellan, D N Reddan, R A Rettig, D L Frankenfield, W F Owen. (2006) Associations between CMS's Clinical Performance Measures project benchmarks, profit structure, and mortality in dialysis units. Kidney International 69:11, 2094-2100
CrossRef4
Hacer Ozgen, Yasar A. Ozcan. (2002) A National Study of Efficiency for Dialysis Centers: An Examination of Market Competition and Facility Characteristics for Production of Multiple Dialysis Outputs. Health Services Research 37:3, 711-732
CrossRef5
Pushkal P. Garg, Neil R. Powe. (2001) Ethical Issues in Dialysis, Series Editor: Aaron Spital,: Profit-Making in the Treatment of Chronic Kidney Disease: Truth and Consequences. Seminars in Dialysis 14:3, 153-156
CrossRef
