Bundled Payment Reform Effects Rates of Vitamin D Therapy Initiation

Beginning in 1972, all patients in the United States with end-stage renal disease (ESRD) have been eligible for Medicare. The number of patients with ESRD requiring maintenance hemodialysis has increased from 16,000 to >400,000 in the years since. Individuals in that patient population represent <1% of all Medicare beneficiaries, but costs for those on maintenance hemodialysis account for 7.4% of all Medicare expenditures.

In the past, Medicare reimbursed the costs of injectable drugs in patients with ESRD separately from hemodialysis services. Between 2007 and 2010, vitamin D therapy for patients on hemodialysis was based on the intravenous (IV) vitamin D analogue paricalcitol, which was the second largest driver of cost, accounting for 4.3% of the reimbursed cost per hemodialysis session (the largest cost driver was erythropoiesis-stimulating agents, 23% of the reimbursed cost).

In January 2011, the US Congress implemented a bundled prospective payment system (PPS) for ESRD. Under the system, 11 injectable drugs, including erythropoiesis-stimulating agents and vitamin D analogues, are included into a capitated payment per in-center hemodialysis session. Studies of the effect of the PPS on overall vitamin D therapy have resulted in uncertainty and incident vitamin D therapy since implementation of the PPS has not been examined.

Julia Spoendlin, PhD, MPH, and colleagues conducted interrupted time-series analyses to examine the association between the PPS and overall use patterns and new use patterns of vitamin D therapy in patients with ESRD. To account for temporal changes in prescribing practices for patients with ESRD, phosphate-binder use patterns were used as a negative control outcome, because they are not yet included in the bundle. Results of the analyses were reported in the American Journal of Kidney Diseases [2018;72(2):178-187].

The researchers established two mutually exclusive study populations: (1) prevalent hemodialysis patients and (2) incident hemodialysis patients. In population 1, the researchers examined overall use of vitamin D before and after implementation of PPS. In population 2, they quantified the treatment initiation rate with different vitamin D products, cinacalcet, and phosphate binders (negative control) within 90 days of hemodialysis therapy initiation, as well as the average starting dose of IV vitamin D analogues over time. Data from the US Renal Data System (USRDS) from 2008 through 2013 were used in the analyses.

Prevalent Study Population

There were 359,600 unique patients undergoing in-center hemodialysis in the USRDS database between 2008 and 2013. Of those, an annually increasing number of patients contributed data to each yearly cohort (143,845 to the 2008 cohort and 181,666 to the 2013 cohort). In 2008, 88.9% of patients received any vitamin D therapy; in 2013, 84.6% received vitamin D therapy. Most patients were treated with IV vitamin D analogues, and there was a transition from paricalcitol (69.4% in 2008 to 45.0% in 2013) to doxercalciferol (26.2% in 2008 to 54.7% in 2013). Use of phosphate binders remained constant (77.3% in 2008 to 76.4% in 2013); sevelamer carbonate largely replaced sevelamer hydrochloride over time.

In quartile 1 in 2008, the average number of dispensations of IV vitamin D was 25.86 per patient. The trend remained largely constant prior to implementation of PPS; there was no change in trend seen following implementation of PPS. On average, patients received 186.5 µg of paricalcitol equivalents per hemodialysis session in the first quarter of 2008; prior to PPS implementation, the trend remained constant. The average dose of IV vitamin D analogues changed immediately following PPS implementation: in the second quarter of 2011, the approximate average dose was 4.4 µg per hemodialysis session. The average dose remained consistently lower in the post-policy period.

Incident Study Population

Data analyses were limited to patients with ≥3 months of continuous Medicare enrollment before initiation of hemodialysis therapy. Most of the patients were ≥65 years of age. Time series analyses of rates of treatment initiation were conducted within three cohorts: (1) nonprevalent vitamin D users (n=99,970); (2) nonprevalent cinacalcet users (n=119,388); and (3) nonprevalent phosphate-binder users (n=106,592).

Of the patients in cohort 1 who started hemodialysis therapy in quarter 1, 2008, 64.0% initiated vitamin D therapy within the first 90 days after initiation of hemodialysis, resulting in a rate of vitamin D therapy initiation of 44.88 per 100 person-months within 90 days after onset of hemodialysis therapy. That rate decreased significantly prior to implementation of PPS. Following implementation of PPS, there was a significant immediate decrease between quarter 4 in 2010 and quarter 1 in 2011; the decline continued at a similar rate post-PPS implementation compared with the pre-PPS slope.

In cohort 2, 4.1% of patients who started hemodialysis therapy in quarter 1 2008 initiated cinacalcet therapy within the first 90 days, resulting in an initiation rate of 2.01 per 100 person-months. There was a small increase of 0.43 initiations per 100 person-months immediately following implementation of PPS (between quarter 4, 2010 and quarter 1, 2011). The rate continued to increase slightly over time compared with the pre-PPS slope. Overall use of cinacalcet during early hemodialysis therapy remained low following PPS implementation.

A total of 41.6% of patients in cohort 3 who began hemodialysis therapy in quarter 1, 2008 initiated phosphate-binder therapy within the first 90 days, resulting in a phosphate-binder initiation rate of 24.13 per 100 person-months; the rates were not affected by the introduction of PPS.

There were some limitations to the study cited by the authors, including restricting the analyses of vitamin D therapy initiation to patients with available pre-ESRD Medicare data (mainly to individuals ≥65 years of age), assessing initiation rates within 90 days of hemodialysis therapy initiation only, an over representation of black patients in the yearly cohorts, and lack of information on the proportion of small dialysis organizations that opted for a 4-year phase-in of the PPS.

“In summary, implementation of the PPS in January 2011 was associated with an immediate 7% decline in the average dose and starting dose of IV vitamin D analogues. PPS implementation was further associated with an immediate 10% decrease and a continuously decreasing rate of initiation of vitamin D therapy in patients with ESRD starting hemodialysis therapy. Use of oral vitamin D analogues did not increase between 2011 and 2013,” the researchers said.

Takeaway Points

  1. In 2011, the Centers for Medicare & Medicaid implemented the prospective payment system (PPS) to curb overuse of separately billable injectable drugs. Intravenous vitamin D analogues were the second biggest drivers of drug costs in patients receiving hemodialysis therapy.
  2. Researchers conducted interrupted time-series analyses to test the hypothesis that implementation of PPS would results in a reduction in the use of IV vitamin D analogues in favor of their oral equivalents or calcitriol, or in favor of cinacalcet.
  3. Following implementation of PPS, there was a 7% reduction in the average dose and starting dose of IV vitamin D analogues and a 10% reduction in the rate of initiation of vitamin D therapy. There was no increase in the use of oral vitamin D analogues.