Eleven: Costs of ESRD
Total Medicare spending in 2006 was nearly $355 billion, while ESRD costs rose to $23 billion — 6.4 percent of the entire Medicare budget, a level unchanging over the past four years. This stability of costs is a result of comparable growth in both the Medicare and ESRD programs, which has kept proportional costs constant.
Recent attention given to therapies with erythropoiesis stimulated agents (ESAs) has raised awareness of the costs of this service to the healthcare system. On the next page we illustrate total Medicare expenditures for patients with complete Medicare coverage and those covered by an employer group health plan (EGHP), along with the percent of these costs attributed to ESA therapy. ESAs account for nearly 10 percent of Medicare expenditures, but just 3 percent of EGHP costs. The latter group does contain a high proportion of transplant patients, who receive little ESA therapy compared to patients on dialysis.
Data on the transition from CKD to ESRD show very high expenditures in the month of dialysis initiation, reaching nearly $15,000. This year we also examine the same period in patients with EGHP coverage, showing the premium paid to providers by commercial health plans compared to Medicare. Most of this amount, not surprisingly, is related to hospitalization. As shown in Chapter Three, catheter use is very high in the first month, and is associated with high rates of hospitalization for access failure, declotting procedures, repeated fistula placements, and infectious complications. Figure 11.11 compares inpatient vascular access costs during the transition to those for cardiovascular hospitalizations.
Comparisons between Medicare per person per month (PPPM) expenditures and those for EGHP patients show that hospital and outpatient costs for dialysis services have grown 24 and 39 percent, respectively, between 2000 and 2006. Although EGHP patients are younger, their costs for inpatient and outpatient services are higher — 33 percent higher for transplant patients, and two to three times higher for those on dialysis or with a functioning graft. Physician/supplier payments have declined, and, in contrast to outpatient and institutional payments, are lower in the EGHP population. These differences may represent cost shifting between Medicare and EGHP payors, but could also illustrate differences in the ability of smaller payors to negotiate pricing compared to Medicare.
Services delivered to the dialysis population on a provider level continue to be a focus of the ADR. As reported in prior ADRs, per person per month expenditures for ESAs, IV iron, IV vitamin D, and laboratory services continue to vary across providers. The mortality and morbidity outcomes illustrated in Chapter Ten, however, show little difference by provider, with some providers having lower mortality and hospitalization ratios but also lower expenditures for injectables. Other factors may be better predictors of improved outcomes, and will be assessed in the future.
Costs for vascular access have been an area of investigation in the ADR for several years. Per person per year (PPPY) expenditures have declined, paralleling the decrease in placement rates (noted in Figure hp.13). The sharp decline in payments noted in 2004 reflects a change in CPT codes for billing of catheters placements. Billings for vascular access services by radiologists now exceed those for surgeons, a new finding, while costs for nephrologist services have increase eight-fold since 1998, the first full year after the publication of the K/DOQI guidelines. On a larger scale, PPPY expenditures for patients with a dialysis catheter are 30 percent higher than in individuals with a fistula. These differences likely represent important severity of disease factors which may lead to the use of dialysis catheters, but the rising use of fistulas also suggests that many patients could use this type of access, which could lead to lower complications and costs.
At the end of the chapter we revisit the new Medicare Advantage payment system, which more completely addresses the chronic disease burden of the population, improving the predictive power and matching of actuarial versus predicted costs. Table 11.b presents PPPM payments from the actuarial as-treated model, the Average Annual Per Capita Cost (AAPCC) basic payment model — used by Medicare in the past for managed care populations, and the current Medicare Advantage hierarchical payment model (HCC). Generally, the HCC model is closer to the actuarial payments than the AAPCC historical system, a marked improvement. The increasing burden of congestive heart failure and diabetes in patients with ESRD, however, means the HCC model still underperforms, which may place providers at greater financial risk.
One challenge to the payment models is understanding that the factors driving costs in the dialysis population may be different than those for general Medicare patients. Dialysis services under the composite rate are not driven by disease, as all patients need a dialyzer, blood lines, heparin, and staff. Other expenditures, including IV iron, epoetin alpha, and vitamin D, may be more sensitive to infectious complications and other diseases not included in the chronic disease adjustment method. New approaches to the composite rate bundle of included dialysis services have been proposed by Congress and CMS, attempting to address the overutilization incentives that exist in the fee-for-service payment system.
The economics of the ESRD population are clearly complex, with considerable variation in provider billings to Medicare. Yet data show little correlation to patient outcomes. Such findings continue to raise concerns about the dialysis payment system, provider incentives, and the cost-effective use of Medicare trust funds. More comprehensive studies are needed to assess the relationship of patient morbidity and mortality to the level of care indicated by expenditures. The USRDS will continue to examine provider-specific practices, along with associated costs and outcomes, in future ADRs.
figure 11.1 Total costs for ESRD, & percent spent on ESAs in Medicare & Medstat (EGHP) patients
period prevalent ESRD patients; Medicare all ages, Medstat age <65. ESA costs identified from revenue & HCPCS codes in Medicare & Medstat claims.
figures 11.2, 11.3, 11.4, 11.5, & 11.6 ESRD’s proportion of Medicare expeditures has remained at 6.3–6.5 percent since 2000, reaching $22.7 billion in 2006. The estimated number of prevalent Medicare ESRD patients rose 3.5 percent between 2005 and 2006, reaching 422,000, while the non-Medicare ESRD population grew nearly 8 percent, to 84,000. Expenditures for non-Medicare patients remained stable in this period, at slightly less than $7.2 billion, while Medicare costs grew at a slightly slower pace than in 2005, reaching $21 billion, $3.7 billion, and $1.7 billion, respectively, for paid claims, patient obligations, and Medicare HMO costs. Overall, the one-year change in total Medicare spending on ESRD was 3.3 percent, down from 7.8 in 2005, and PPPY costs grew just 1.6 percent, compared to 4.3 the year before. By type of service, 38.5 percent of Medicare’s ESRD dollars are now being spent on outpatient services, compared to 34.6 percent on services in the inpatient setting.
figures 11.7, & 11.8 Total Medicare expenditures for hemodialysis totaled nearly $17 billion in 2006, while costs for peritoneal dialysis approached $1 billion and those for transplant reached 1.8 billion, just under 10 percent of total dollars spent. Expenditures per patient year show parallel trends, with hemodialysis costs at $71,889 in 2006, compared to $53,327 and $24,951, respectively, for peritoneal dialysis and transplant.
figures 11.9, 11.10, & 11.11 Per member per month costs in 2005 increased during the transition to ESRD in both the Medicare and Medstat (EGHP) populations, a rise which appears to be associated primarily with inpatient services. Overall transition costs for Medicare patients, for example, r0se from $6,701 in the month prior to ESRD treatment to $14,461 following initiation, and $9,588 of these costs were associated with the inpatient setting. Increased costs were also evident for patients covered by EGHPs. Transition costs for this group were noticeably higher than those incurred by Medicare patients, increasing by nearly a factor of six from $4,834 to $28,718, and with $20,792 resulting from inpatient services. Vascular access hospitalization costs in the transition increased from $418 to $1,509 in Medicare patients, and from $255 to $1,359 in the Medstat population. And while cardiovascular hospitalization costs during the transition nearly doubled for the Medicare population (from $2,096 to $3,478), costs for employed patients rose much more dramatically, from $866 in the month prior to initiation to $6,885 at one month post-initiation—nearly an eight-fold increase.
figure 11.12 organ acquisition costs computed from CMS hospital reports (CMS form 2552-96). Transplant costs estimated from the 5 percent Medicare sample; include inpatient costs only.
figures 11.13, & 11.14 Per person per month ESRD costs vary considerably depending on insurance coverage. Monthly inpatient/outpatient expenditures for patients on hemodialysis who are covered by an employer group health plan (EGHP), for example, are nearly twice those for patients with Medicare as their primary payor, at $8,340 versus $4,300. Likewise, costs for patients with transplant events and with functioning grafts are higher for EGHP patients, at $9,894 versus $7,410 and $2,404 versus $762, respectively.
table 11.a Per person per month (PPPM) costs for combined inpatient and outpatient services in the ESRD population reached $4,027 in 2006, while physician/supplier costs were $1,071. Expenditures for both types of service vary widely by a patient’s diabetic status. Costs for inpatient/outpatient services, for example, were $4,578 for diabetics, 25 percent greater than the $3,660 incurred by non-diabetic patients. And physician/supplier costs in the diabetic population were 36 percent greater, at $1,273 compared to $936.
figures 11.15, 11.16, & 11.17 After rising steadily through the 1990s and into the early 2000s, total Medicare spending on erythropoeitin stimulating agents (ESAs) appears to have plateaued, as demonstrated by a cost increase of only 0.5 percent between 2005 and 2006. Costs for IV vitamin D, on the other hand, which had appeared to be decreasing (11.7 percent from 2004–2005), rose by 8.9 percent in 2006, and it is too early to predict if this upward movement will continue into future years. After falling more than 17 percent between 2004 and 2005, spending for IV iron grew 5 percent in 2006, perhaps indicating a shift in providers’ clinical practices as they relate to anemia management. Zemplar appears to be the drug of choice for treating bone and mineral disorders in dialysis patients. Per person per year costs reached nearly $2,000 in 2006; cost for Hectorol and Calcijex were far less, at $1,073 and $517, respectively. IV iron costs appear to be fairly evenly distributed between Ferrlicit and Venofer, at $772 and $685 per person per year, respectively, while costs for INFeD are considerably lower, at $439. PPPM costs for ESAs and IV iron show a sharp geographic pattern, with costs highest along the Gulf Coast and the Eastern Seaboard, and lowest in the western half of the country. Costs for IV vitamin D, in contrast, vary far more widely across the U.S., though they too are high in the southeastern states.
figures 11.18, & 11.19 Per person per month (PPPM) dialyis costs continued their upward trend in 2006, totaling $1,418, an increase of 8.6 percent from the previous year. Laboratory services also showed a slight increase of 3.3 percent, totaling $134, while clinical service costs for ESAs, IV iron, IV vitamin D, and other injectables were down by 3.6, 11.0, 8.6, and 7.2 percent, respectively. By unit affiliation, overall PPPM costs for dialysis services were $1,397 in 2006, and were highest for units owned by DaVita, at $1,439, and lowest in those owned by DCI, at $1,326. Costs for ESAs ranged from $469 to $623, with overall costs totalling $530. PPPM costs for IV iron were $53 overall, and ranged from a low of $40 in hospital-based units to $64 in units owned by DaVita. Costs for vitamin D hormone vary widely by ownership. In units owned by DCI, for example, PPPM costs totaled $46 in 2006, while in DaVita/Gambro* and DaVita units costs were $133 and $155, respectively.
figure 11.20 Per person per month preventive care costs do not vary significantly by type of ownership, generally showing only small differences across unit types. Hepatitis B vaccination costs in 2006, for example, ranged from $1.98 in units owned by DCI to $2.81 and $2.82 in units owned by DaVita and in those affiliated with small dialysis organizations (SDOs). Costs for lipid testing, however, varied by more than $2 depending on ownership. Monthly costs totaled $0.89 in units owned by DCI, compared to costs of $3.03 in units that are independently owned. Diabetic eye examinations are the most costly preventive care service, in 2006 ranging from $6.71 PPPM in units owned by DCI to $9.75 in those affiliated with an SDO; overall costs totaled $8.17. Costs for comprehensive diabetic testing (see caption for description) generally averaged less than one dollar per person per month across all unit types, with the exception of costs of $1.14 in units owned by DCI. Influenza vaccination costs are insignificant at well under $1 per month,in comparison to pneumococcal vaccination costs, which range from $4.27 in units owned by an SDO to $6.69 in those owned by DaVita/Gambro.
figures 11.21 Physician/supplier costs for vascular access reflect recent efforts (K/DOQI, CMS’s “Fistula First” initiative) to establish arteriovenous (AV) fistulas as the preferred vascular access choice for hemodialysis. Costs for catheters, for example, have fallen nearly 30 percent since 2003, while AV fistula costs have increased 41 percent. AV graft costs have fallen 35 percent over the same period.
figures 11.22, & 11.23 Overall per person per year (PPPY) expenditures totaled $65,937 in 2006. Patients with catheters and grafts have the highest costs, at $77,093 and $71,616, respectively, while $59,347 PPPY is spent on those with AV fistulas and $53,470 on those with peritoneal dialysis catheters. Hemodialysis access costs are lowest for patients with AV fistulas, at $3,284, and comparable for those with catheters and grafts, at $6,828 and $7,377, respectively; costs for patients with peritoneal catheters are $833.
figure 11.24 In 2006, the highest per person per month vascular access procedure costs by physician specialty were incurred for radiologic and surgery services, at $35 and $32, respectively. Monthly nephrology costs totaled $18, while anesthesiology costs were lowest, at $8.
figure 11.25 Per person per year adjusted vascular access costs for access infections are highest in patients with a catheter, at nearly $2,500 in 2005. These costs fell in 2003 and 2004, but in 2005 rose 21 percent. Costs for patients with an arteriovenous fistula or graft have been more stable over time, though each increased 26–27 percent in 2005, reaching $240 and $775, respectively.
table 11.b & figure 11.26 Predicting costs with the CMS-HCC model reduces the difference between actual costs and those based on the AAPCC method. Actual Medicare costs for ESRD patients with diabetes and CHF are nearly double that of ESRD patients without diabetes and CHF. Compared to patients younger than 65, older patients cost 10–35 percent more.
figure 11.1 period prevalent ESRD patients; Medicare all ages, Medstat age <65. ESA costs identified from revenue & HCPCS codes in Medicare & Medstat claims.
figure 11.2 total ESRD expenditures are from paid claims (Table K.2) as well as estimated costs for HMO & organ acquisition. ESRD costs in 2006 are inflated by 2 percent to account for costs incurred but not reported. Total Medicare expenditures obtained from the CMS Office of Financial Management, Division of Budget. figure 11.3 December 31 point prevalent ESRD patients; non-Medicare status determined from payor sequence. figure 11.4 Medicare costs obtained from claims files, & include all Medicare as primary payor claims as well as amounts paid by Medicare as secondary payor (MSP). Medicare patient obligations estimated as the difference between the Medicare payment & the Medicare allowable cost, HMO costs estimated as the number of HMO months times the Medicare AAPCC rate, & organ acquisition costs estimated as $25,000 per transplant. Non-Medicare estimate includes all non-Medicare patients (using AAPCC), the primary payor estimate for MSP patients, & estimated patient obligations. figure 11.5 total Medicare ESRD costs obtained from claims data, & include all Medicare as primary payor claims as well as amounts paid by Medicare as secondary payor. PPPY expenditures include only patients with Medicare as primary payor. figure 11.6 ESRD spending obtained from Medicare ESRD claims, & includes all Medicare as primary payor claims as well as amounts paid by Medicare as secondary payor. figure 11.7 period prevalent ESRD patients. Modalities determined using Model 2 methodology, as described in Appendix A. Includes Medicare paid claims for ESRD patients, starting at first ESRD service date & continuing until death or the end of the study period. Patients with Medicare as secondary payor are included. figure 11.8 period prevalent ESRD patients. Modalities determined using Model 2 methodology, as described in Appendix A; patients with Medicare as secondary payor excluded. figure 11.9–11 Medicare: patients age 67 & older, initiating in 2005, with Medicare as primary pay0r. Medstat: ESRD patients age <65, initiating in 2005. For Figure 11.11, vascular access defined as “pure” (see Appendix A) & cardiovascular hospital cause determined using principal diagnosis. figure 11.12 organ acquisition costs computed from CMS hospital reports (CMS form 2552-96). Transplant costs estimated from the 5 percent Medicare sample; include inpatient costs only.
figures 11.13 Medicare: period prevalent ESRD patients with Medicare as primary payor. Medstat: period prevalent ESRD patients age <65 with enrollment for the entire study period. Modalities determined using Model 2 methodology as described in Appendix A. Costs are per person per month inpatient plus outpatient for each calendar year. figure 11.14 Medicare: period prevalent ESRD patients with Medicare as primary payor; costs are per person per month physician/supplier for each calendar year. Medstat: period prevalent ESRD patients age <65 with enrollment for the entire study period; costs are per person per month physician net for each calendar year. Modalities determined using Model 2 methodology as described in Appendix A. table 11.a period prevalent ESRD patients, 2006. Costs are per person per month for calendar year 2006. figure 11.15 period prevalent dialysis patients. Same methods as those used in Reference Table K.2; see Appendix A for details. figure 11.16 period prevalent dialysis patients, with Medicare as primary payor, & who survive on dialysis for the entire calendar year, 2006. PPPY calculated for patients who receive one type of vitamin D or iron preparation exclusively during calendar year 2006. figure 11.17 period prevalent dialysis patients with Medicare as primary payor, 2006, by HSA, unadjusted. Excludes patients residing in Puerto Rico & the Territories. ESA: erythropoiesis stimulating agent. For details on how to read the map legends, please see page xvi in the Introduction.
figure 11.18, 11.19, & 11.20 period prevalent dialysis patients, 2006; excludes non-Medicare, Medicare HMO, & Medicare as secondary payor patients. As-treated model. Comprehensive diabetic monitoring includes at least four HbA1c tests per year, at least two lipid tests per year, & at least one diabetic eye examination per year.
figure 11.21 period prevalent hemodialysis patients. Costs determined as the payment amount for each line-level placement event billed through the physician/ supplier claims. The sharp drop in 2004 is due to changes in the Medicare reimbursement policy. figures 11.22–23 dialysis patients from the 1999–2006 ESRD CPM data with Medicare as primary payor & vascular access data. Intent-to-treat model. Vascular access type in use in December prior to cost years 1999–2006. For Figure 11.23, costs include “pure” inpatient & outpatient claims & physician/supplier access costs; see Appendix A for definition of “pure” inpatient & outpatient claims. figure 11.24 period prevalent hemodialysis patients. Costs determined as the payment amount for each line-level access procedure billed through physician/supplier claims. figure 11.25 prevalent hemodialysis patients, 1998–2005, in both the ESRD CPM & USRDS databases. Access determined from ESRD CPM data; vascular access infections identified from inpatient Medicare claims; adjusted for inflation to dollars in 2000. table 11.b & figure 11.26 December 31 point prevalent dialysis patients, 2006. Medicare payment includes inpatient, skilled nursing, home health, outpatient, physician/supplier, & durable medical equipment. M+C risk adjustment model; actual PPPM cost in table not weighted by follow-up time. Definitions for diabetes & CHF based on CMS-HCC model.