Medicare's New Prospective Payment System on Facility Provision of Peritoneal Dialysis

Abstract

Background and objectives: Peritoneal dialysis is a self-administered, home-based treatment for ESKD associated with equivalent mortality, higher quality of life, and lower costs compared with hemodialysis. In 2011, Medicare implemented a comprehensive prospective payment system that makes a single payment for all dialysis, medication, and ancillary services. We examined whether the prospective payment system increased dialysis facility provision of peritoneal dialysis services and whether changes in peritoneal dialysis provision were more common among dialysis facilities that are chain affiliated, located in nonurban areas, and in regions with high dialysis market competition.

Design, setting, participants, & measurements: We conducted a longitudinal retrospective cohort study of n=6433 United States nonfederal dialysis facilities before (2006-2010) and after (2011-2013) the prospective payment system using data from the US Renal Data System, Medicare, and Area Health Resource Files. The outcomes of interest were a dichotomous indicator of peritoneal dialysis service availability and a discrete count variable of dialysis facility peritoneal dialysis program size defined as the annual number of patients on peritoneal dialysis in a facility. We used general estimating equation models to examine changes in peritoneal dialysis service offerings and peritoneal dialysis program size by a pre- versus post-prospective payment system effect and whether changes differed by chain affiliation, urban location, facility size, or market competition, adjusting for 1-year lagged facility-, patient with ESKD-, and region-level demographic characteristics.

Results: We found a modest increase in observed facility provision of peritoneal dialysis and peritoneal dialysis program size after the prospective payment system (36% and 5.7 patients in 2006 to 42% and 6.9 patients in 2013, respectively). There was a positive association of the prospective payment system with peritoneal dialysis provision (odds ratio, 1.20; 95% confidence interval, 1.13 to 1.18) and PD program size (incidence rate ratio, 1.27; 95% confidence interval, 1.22 to 1.33). Post-prospective payment system change in peritoneal dialysis provision was greater among nonurban (P<0.001), chain-affiliated (P=0.002), and larger-sized facilities (P<0.001), and there were higher rates of peritoneal dialysis program size growth in nonurban facilities (P<0.001).

Conclusions: Medicare's 2011 prospective payment system was associated with more facilities' availability of peritoneal dialysis and modest growth in facility peritoneal dialysis program size.

Podcast: This article contains a podcast at https://www.asn-online.org/media/podcast/CJASN/2018_11_19_CJASNPodcast_18_12_.mp3.

Keywords: ESKD; Health Resources; Incidence; Kidney Failure, Chronic; Medicare; Odds Ratio; Prospective Payment System; Retrospective Studies; dialysis facilities; peritoneal dialysis; policy; quality of life; renal dialysis.

Graphical abstract

Figure 1.

Growth in dialysis facilities and facility peritoneal dialysis (PD) provision in the United States by year. The facility sample is not fixed across time periods, reflecting facility closure, new entry, and changes in ownership over time.

Figure 2.

Geographic variation in changes in regional supply of dialysis facilities' provision of peritoneal dialysis: 2006 versus 2013.

Figure 3.

Dialysis facility peritoneal dialysis (PD) program size by selected year. Facility PD program size (PD census) was reported for facilities reporting any patients on PD in a year.

Figure 4.

Increased rates of dialysis facility provision of peritoneal dialysis (PD) services after implementation of 2011 Medicare Prospective Payment System (PPS) for dialysis, as shown in observed and estimated rates of PD provision and associated 95% confidence intervals by year. Estimates and 95% confidence intervals from the generalized estimating equation logistic regression model for PD provision, with all other covariates centered at mean values. A full set of regression results is available in Supplemental Table 1.