Archived Comments for:
Comparative mortality of hemodialysis patients at for-profit and not-for-profit dialysis facilities in the United States, 1998 to 2003: A retrospective analysis
Comparing For-Profit and Not-for-Profit Dialysis Facilities
Philip Zager, University of New Mexico Health Sciences Center
10 December 2008
We read with interest the recent article entitled “Comparative mortality of hemodialysis patients at for-profit and not-for-profit dialysis facilities in the United States, 1998 to 2003: A retrospective analysis” by Foley et al. Many of these authors have major roles in assembling the United States Renal Data System’s (USRDS) Annual Data Report (ADR). The 2007 ADR states “[standardized mortality and hospitalization ratios] remain lowest in Dialysis Clinic Inc. (DCI) units, with no other providers showing the same consistency in results.” DaVita (DV), Fresenius Medical Care of North America (FMCNA) and DCI are, by far, the largest for-profit and not-for-profit facilities, respectively, operating in the United States. According to the USRDS the standardized mortality (SMRs) and hospitalization ratios (SHRs) have been lower in DCI than in DaVita and FMCNA each year from 1999 to 2006. The published SMRs adjusted for age, gender, race, primary diagnosis and ESRD vintage during 1999, 2003 and 2006 respectively are for DCI - 0.859, 0.796, 0.898; for DV- 0.865, 0.872, 0.992 and for FMCNA- 0.896, 0.934, 1.005. Similarly, adjusted SHRs published for the same years for DCI are 0.878, 0.861, 0.891; for DV- 0.975, 0.969, 1.025 and for FMCNA- 0.967, 0.982, 1.133. Therefore, the conclusion of this article that there are no significant differences in mortality among hemodialysis patients treated in for-profit and not-for-profit facilities is surprising. These apparently conflicting results bring to light the inherent challenges associated with SMRs and SHRs and Cox regression analysis to compare clinical outcomes between different providers.
There are several factors that may have contributed to the discordant results published in the USRDS ADR and the current manuscript. First, there are important methodological differences between the calculations of SMRs, SHRs versus the use of Cox regression models, including the larger number of covariates which may be included in the latter. The unadjusted Cox models reported in this article showed significantly lower mortality among patients treated in not-for-profit facilities. Nevertheless, we agree that adjustment for case mix is warranted. However, virtually all of the co-morbid conditions were either similarly distributed or more prevalent among patients treated by not-for-profit providers. Second, the inclusion of hospital days as a covariate, while a potentially useful surrogate marker of comorbidity, may be problematic. If, not-for-profit facilities provide care in a manner that leads to fewer hospitalizations, adjusting hospitalization may be inappropriate since it may be a causal pathway through which not-for-profit providers achieve lower mortality. Third, the study population in the present study was not restricted to DV, FMCNA and DCI, but rather included 100% ESRD sample from the Medicare database to select patients who were first dialyzed between January 1, 1998, and December 31, 2003; had Medicare as primary payer throughout the exposure period; and were on hemodialysis at either a for-profit or not-for-profit dialysis facility at the end of the exposure period. Therefore, the differences in outcomes between DV, FMCNA versus DCI may have been diluted out by the inclusion of additional patients.
It is challenging to assess clinical outcomes in a rapidly evolving environment. Given the competitive economic pressures, it is highly unlikely that a prospective randomized study will ever be undertaken to compare outcomes among patients treated in for-profit versus not-for profit facilities. Therefore, it is necessary to develop novel statistical techniques to better assess outcomes from administrative databases.
For the authors: Christopher Adams, MD1
Karen S. Servilla, MD1,2
Dana Miskulin, MD3
Philip Zager, MD1,2,4
1University of New Mexico Health Sciences Center, Department of Internal Medicine, Division of Nephrology, Albuquerque, NM
2 Nephrology Section, New Mexico Veterans Affairs Health Care System, Albuquerque, NM, Albuquerque, NM
3 Tufts New England Medical Center, Boston, MA
4 Dialysis Clinic, Inc. Quality Management Division, Albuquerque, NM
Competing interests
The University of New Mexico and Tufts New England Medical Center receive salary support for Drs. Zager and Miskulin, respectively from Dialysis Clinic, Inc.
Comparing For-Profit and Not-for-Profit Dialysis Facilities
10 December 2008
We read with interest the recent article entitled “Comparative mortality of hemodialysis patients at for-profit and not-for-profit dialysis facilities in the United States, 1998 to 2003: A retrospective analysis” by Foley et al. Many of these authors have major roles in assembling the United States Renal Data System’s (USRDS) Annual Data Report (ADR). The 2007 ADR states “[standardized mortality and hospitalization ratios] remain lowest in Dialysis Clinic Inc. (DCI) units, with no other providers showing the same consistency in results.” DaVita (DV), Fresenius Medical Care of North America (FMCNA) and DCI are, by far, the largest for-profit and not-for-profit facilities, respectively, operating in the United States. According to the USRDS the standardized mortality (SMRs) and hospitalization ratios (SHRs) have been lower in DCI than in DaVita and FMCNA each year from 1999 to 2006. The published SMRs adjusted for age, gender, race, primary diagnosis and ESRD vintage during 1999, 2003 and 2006 respectively are for DCI - 0.859, 0.796, 0.898; for DV- 0.865, 0.872, 0.992 and for FMCNA- 0.896, 0.934, 1.005. Similarly, adjusted SHRs published for the same years for DCI are 0.878, 0.861, 0.891; for DV- 0.975, 0.969, 1.025 and for FMCNA- 0.967, 0.982, 1.133. Therefore, the conclusion of this article that there are no significant differences in mortality among hemodialysis patients treated in for-profit and not-for-profit facilities is surprising. These apparently conflicting results bring to light the inherent challenges associated with SMRs and SHRs and Cox regression analysis to compare clinical outcomes between different providers.
There are several factors that may have contributed to the discordant results published in the USRDS ADR and the current manuscript. First, there are important methodological differences between the calculations of SMRs, SHRs versus the use of Cox regression models, including the larger number of covariates which may be included in the latter. The unadjusted Cox models reported in this article showed significantly lower mortality among patients treated in not-for-profit facilities. Nevertheless, we agree that adjustment for case mix is warranted. However, virtually all of the co-morbid conditions were either similarly distributed or more prevalent among patients treated by not-for-profit providers. Second, the inclusion of hospital days as a covariate, while a potentially useful surrogate marker of comorbidity, may be problematic. If, not-for-profit facilities provide care in a manner that leads to fewer hospitalizations, adjusting hospitalization may be inappropriate since it may be a causal pathway through which not-for-profit providers achieve lower mortality. Third, the study population in the present study was not restricted to DV, FMCNA and DCI, but rather included 100% ESRD sample from the Medicare database to select patients who were first dialyzed between January 1, 1998, and December 31, 2003; had Medicare as primary payer throughout the exposure period; and were on hemodialysis at either a for-profit or not-for-profit dialysis facility at the end of the exposure period. Therefore, the differences in outcomes between DV, FMCNA versus DCI may have been diluted out by the inclusion of additional patients.
It is challenging to assess clinical outcomes in a rapidly evolving environment. Given the competitive economic pressures, it is highly unlikely that a prospective randomized study will ever be undertaken to compare outcomes among patients treated in for-profit versus not-for profit facilities. Therefore, it is necessary to develop novel statistical techniques to better assess outcomes from administrative databases.
For the authors: Christopher Adams, MD1
Karen S. Servilla, MD1,2
Dana Miskulin, MD3
Philip Zager, MD1,2,4
1University of New Mexico Health Sciences Center, Department of Internal Medicine, Division of Nephrology, Albuquerque, NM
2 Nephrology Section, New Mexico Veterans Affairs Health Care System, Albuquerque, NM, Albuquerque, NM
3 Tufts New England Medical Center, Boston, MA
4 Dialysis Clinic, Inc. Quality Management Division, Albuquerque, NM
Competing interests
The University of New Mexico and Tufts New England Medical Center receive salary support for Drs. Zager and Miskulin, respectively from Dialysis Clinic, Inc.