Letter from Kidney Care Partners (of which ANNA is a member) to the Centers for Medicare and Medicaid Services regarding Physicians' Fee Schedule Final Rule for Calendar Year 2005, Patient Characteristic Adjustments for the ESRD Composite Rate Payment
KIDNEY CARE PARTNERS
December 9, 2004
The Honorable Mark McClellan, M.D.
Centers for Medicare and Medicaid Services
U.S. Department of Health and Human Services
Hubert H. Humphrey Building
200 Independence Avenue, SW
Washington, DC 20201
Re: Comments on the Centers for Medicare and Medicaid Services’ Physicians’ Fee Schedule Final Rule for Calendar Year 2005, Patient Characteristic Adjustments for the ESRD Composite Rate Payment
Dear Dr. McClellan:
Thank you for taking the time to meet with us and for your thoughtful consideration of our issues. As we mentioned to you, Kidney Care Partners (KCP) remains concerned about the case-mix adjustment for the Medicare dialysis composite rate payment system as described in the Centers for Medicare and Medicaid Services’ (CMS) Final Rule for the Calendar Year 2005 Physicians’ Fee Schedule (Final Rule). As you know, KCP is a coalition of members of the kidney care community including renal patient advocates, nephrologists, dialysis care professionals, providers, and suppliers that works to improve the quality of care of individuals with irreversible kidney failure, known as End Stage Renal Disease (ESRD).
When we last met, we highlighted some of our issues with the case-mix system and promised to share what we learned with CMS. You acknowledged your concerns about the low R values associated with the case-mix adjustors and their low predictive power of the costs for estimating dialysis treatments. KCP has reviewed these issues carefully. Our analysis has benefited from the work conducted by Dr. J. Michael Lazarus, the Chief Medical Officer of one of our member organizations, Fresenius Medical Care North America (FMCNA). This letter summarizes Dr. Lazarus’ findings, as well as other policy and operational concerns of KCP members, regarding the proposed case-mix adjustment system for ESRD.
In our comments on the Proposed Rule, we voiced specific concerns about CMS prematurely implementing a case-mix adjustment system that has not been clinically tested and verified to predict resource use by dialysis patients during a dialysis treatment. We also expressed concern about the lack of transparency regarding the underlying data and methodology (including statistical measures of significance) for the selected case-mix adjustors. We sought additional information to permit a full evaluation of the analysis upon which CMS relied when selecting these adjusters for the case-mix system. Most importantly, we strongly urged CMS to exercise its discretion as it had with other prospective payment systems to delay implementation until such testing and verification occurred. The Final Rule neither addressed these concerns nor provided sufficient information for stakeholders to evaluate CMS’s analysis.
We acknowledge that CMS took the advice of KCP in its selection of weight rather than gender as a case-mix measure. However, there is no clinical basis for the use of both body mass index (BMI) and body surface area (BSA) as indicators. We also are concerned about CMS’s implementation approach, which relies on inappropriate data sources for this information. KCP strongly urges you to delay implementation of the case-mix system described in the Final Rule until CMS has appropriate patient level data, which is not currently available. If delay is not possible, CMS should use patient age and BSA as adjusters and correct for the effects of aging on body size. In addition, CMS should develop specific case-mix multipliers to predict resource use at the low and high ends of the weight spectrum.
We also would like to reiterate a concern voiced in our original comment letters on the Proposed and Final Rules regarding CMS’s budget neutrality calculation within the case-mix system. We continue to believe that the poor predictive power of the case-mix adjustors seriously calls into question the budget neutrality adjustment in the Final Rule. CMS is relying on the case-mix adjustment system to redistribute the dollars appropriately to facilities in order to meet the resource needs of their patients. Because of the low predictive power of the case-mix adjustors, the dialysis payment system will remain seriously misalignedresources will not be distributed appropriately to meet individual patients’ needs and CMS risks under-funding the system.
To assure the appropriate alignment of resources with dialysis patient needs, KCP urges CMS to:
- Delay implementation of the case-mix adjustment system in ESRD;
- Collect appropriate patient-level data for the case-mix system and then refine the age and weight risk adjustment measures;
- Refine its case-mix adjusters at the lower and higher ranges of weight; and
- Adjust the case-mix measures and case-mix conversion factor to assure the case-mix system is developed in a budget neutral manner.
I. CMS Has Authority To Delay Implementation
CMS has delayed implementing prospective payment systems (PPS) in the past, even when Congress has established specific deadlines. For example, the agency delayed implementation of the Hospital Outpatient PPS, the Home Health PPS, and the Skilled Nursing Facility PPS. The same reasons for delaying these systems exist for delaying the case-mix adjustment to the ESRD composite rate payment: CMS needs additional time beyond that allocated by Congress to ensure the development of an appropriate case-mix methodology that accurately predicts resource needs and that is less likely to cause unintended negative consequences. Consistent with this historical approach, CMS should take the time it needs to also get the ESRD payment system right. KCP members firmly believe getting the policy right is a greater priority than meeting an arbitrary and unrealistic legislative deadline.
II. Cost Report Data Do Not Accurately Reflect Patient-Level Clinical Characteristics and Resource Needs - CMS Should Identify Other Sources of Data for the Case-Mix Adjustment System in ESRD
We understand that the analyses that led to the case-mix adjusters in the Final Rule were initially derived from data collected by CMS’s Dialysis Facility Cost Reports (HCFA Form 265-94), which were used to identify facilities with higher and lower costs. Using these data, CMS then attempted to identify predictors of those facility level costs by analyzing claims for patients in the various categories of facilities. We have carefully reviewed the Cost Report form (See Appendix 1) and believe the use of the data from this form to determine a relationship between dialysis costs and clinical outcomes or conditions is inappropriate. Cost Reports reflect facility, not patient characteristics because treatment in facilities is based on groups of patients and all facility costs are allocated at aggregate, not at patient specific, levels.
In general, the Cost Report has very limited information about treatment and patient characteristics. Data found on the Cost Reports include: 1) the average number of treatments per week a patient receives dialysis; 2) the average time of patient dialysis treatment; 3) the number of patient shifts at the facility in a typical week; 4) the number of hours per shift in a typical week; 5) the total number of treatments provided; and 6) the type of dialyzer (though this data element is severely outdated). These data are presented as “averages” and “typical” weeks and do not reflect patient specific data. The Cost Report captures information on the number of employees and salaries for direct patient care. The personnel and the nursing time devoted to dialysis is presented in the aggregate and patient-level nursing labor is not further detailed. Finally, the Cost Report includes data on the number of units of Epogen® (EPO) furnished, drug expense, supply expense, laboratory expense, whole blood and packed red blood cells, and hepatitis vaccine.
Cost Reports were designed to reflect characteristics of facility treatment practices in the aggregate and not patient-specific resource needs. The Cost Reports contain the average age, average BSA, and average BMI of patients treated at the facility. Data from Cost Reports cannot be broken down to provide patient specific characteristics--a critical element for the development of a case-mix adjustment system.
The Cost Reports can only crudely be related to clinical outcomes. During the past 30 years, reimbursement for dialysis has never been related to the severity of the patient condition or the consumption of nursing or physician effort (the major cost factors). Patients have never been characterized by their clinical conditions or the severity of their conditions. CMS’s 2003 report, “The Design, Development and Implementation of an Improved Medicare Outpatient ESRD Prospective Payment System” (Report) further reinforces the shortcomings of these data:
The analyses are based on the principle that each facility efficiently adjusts the level of treatment resources to reflect appropriate care for the type of patients being treated at that facility. In fact, the ability to tailor treatment levels to the case-mix at each facility is constrained by the current composite rate payment system, which reimburses the same amount regardless of the characteristics of the patients. Thus, the estimates from the analyses reported here are likely to under estimate the actual variation in costs associated with delivery of ideal care to various case-mix patients.
The Report’s analysis indicates that the case-mix adjustments developed from the data derived from Cost Reports accounted for less than one percent of the cost difference among dialysis facilities.
Additionally, KCP’s analysis of these issues has benefited from the work of Dr. Lazarus, as mentioned previously. Dr. Lazarus conducted an analysis of the case-mix adjustors identified in the Final Rule using Cost Report data from 62,000 ESRD patients treated in FMCNA facilities during July-September 2004. This analysis confirms the weak association between cost and the case-mix measures identified in CMS’s Final Rule (Appendix 2). These measures account generally for less than one-half of one percent of the cost differences among patients when the same costs are presumed associated with each patient treated in a dialysis unit. When each dialysis facility, instead of each patient, is used as the unit of analysis, however, the average age, average BSA, and other adjustors in the analysis account for a much higher fraction. The fact that facility-based statistical models do so much better than patient-based models further illustrates the weakness of the analysis on which the Final Rule is based.
The lack of patient-specific data in the Cost Reports has led CMS to establish a system that contains case-mix measures that account for only a tiny fraction of the total cost per treatment (less than one percent). While the Cost Reports can be used to determine the amount of dialysis, they do not provide a breakdown or characterization of these costs relative to specific patient characteristics. The costs of a dialysis facility may be related to many other factors, such as physician compensation, bad debt, equipment overhead, fixed overhead, etc., which likely have much greater impact on overall facility costs than the patient’s medical condition, at least as determined by the data that is captured on the Cost Report.
KCP urges CMS to comply with its Congressional mandate to construct a patient specific case-mix adjusted payment mechanism that accurately predicts resource use in dialysis centers. At this point, the agency simply does not have the patient specific cost data readily at hand, nor did it undertake steps to collect these data. Until such data are available, CMS should not implement a case-mix adjustment system for the ESRD program.
III. Once Appropriate Data Are Available, CMS Should Refine the Case-Mix Adjusters of Weight and Age
Because CMS used the Cost Reports as the basis for its analysis, the agency drew incorrect conclusions that will lead to serious reimbursement problems that may negatively impact patient care. The most dramatic indication of these unintended consequences is illustrated in the Final Rule itself. The Final Rule indicates that CMS will reimburse providers for the care of a healthy 18-year-old male in Baltimore at $170.80, while an 82-year-old, malnourished, female patient in Baltimore will be reimbursed at $139.24 (Appendix 3). The system rewards caring for young, healthier patients and fails to acknowledge that sicker, older patients may need more care and consume more resources.
This example is not an isolated one. Consider a 22-year-old male student with a BSA of 1.94 and a BMI of 27.0 who lives in Baltimore with his parents. He drives to the dialysis unit for his evening treatments and arrives on time, weighs himself prior to treatment, cleans his own access site, goes to his chair, and leaves promptly at the end of each treatment after weighing himself again. Consider also a 69-year-old diabetic female with a below-the-knee amputation who comes to the unit from an extended care facility in a community van, waits in a wheel chair to be taken for treatment, is assisted with the weighing process, undergoes her treatment, and is picked up by the van after being assisted by staff to the facility lobby. Both patients are treated for 3.5 hours. He requires this length of time to achieve an acceptable URR. She is treated for the same length of time because she has gained weight between treatments and does not tolerate rapid ultrafiltration. He is treated with a dialyzer that is somewhat larger than the one used to treat her.
The true operating costs assigned to those two patients are not the same even though they are presumed to be in the “by-patient” analysis of Cost Report data. She consumes far more direct patient care (which is 36 percent of total cost/Rx) and other costs such as space, benefits, etc. (which is 49 percent of total cost/Rx) than he does. The younger larger patient does use a larger dialyzer, so the supply costs (which are 15 percent of total cost/Rx) might be slightly higher, but other disposable supply costs will be the same. As this example demonstrates, cost data from the facility Cost Reports do not reflect costs allocated to individual patients. A case-mix system based on the use of Cost Report data will be fraught with these types of inequities.
To resolve these problems, KCP suggests that CMS modify certain reimbursement multipliers. We suggest that CMS use one and two standard deviations at the low and high end of the distribution curves of either BSA (Figure 9) or BMI (Figure 10) to indicate patients for whom a multiplier of reimbursement would be applied. This approach is summarized in Table 1, which shows the upper and lower one and two standard deviations of either BSA or BMI of the distribution curves. Different levels of multipliers or adjustments should be developed for each of these standard deviation levels within each specific weight outcome. A detailed analysis of these issues is presented below.
A. Body Size
By using both BMI and BSA in combination (see Appendices 4 & 5), CMS skews the analysis. Both of these measures utilize body weight and height. There is no clinical or scientific rationale for using both BSA and BMI. In Figure 1, we provide information on BMI from a population of 98,478 ESRD patients from FMCNA facilities obtained in July-September 2004. The distribution of the percent of patients by BMI is severely skewed to the right because the height is “squared” in the denominator. The median of this population of BMIs is 26 kg/m2 and the mean 27.3 kg/m2. The upper and lower one standard deviation and two standard deviations of the population are indicated in the figure. The skewed distribution makes BMI a less preferable measure than BSA, the distribution of which is shown in Figure 2. Here, there is a normal distribution with a bell shaped curve, the median is 1.85 m2 and the mean is 1.83 m2. The upper and lower one and two standard deviations of the distribution curve are noted in the figure as well. We are not aware that either one of these measures has been shown to be superior to the other in judging outcomes in ESRD dialysis patients. Certainly, there is no support for using both.
A recent paper published in Kidney International (Appendix 7, Ref. 1), with data taken from FMCNA facilities, suggests that either measure can be used as a predictor of the risk of death. In this publication, it is shown that increasing weight, increasing BMI, and increasing BSA all have very similar curves when related to log-hazard of the probability of death (Figure 3). Of the three, BSA has the highest linear R2 value, which suggests that it might be slightly preferable. All three measures demonstrate that increasing body weight, however measured, is associated with a reduction in the risk of death. This has been supported by several other clinical investigations (Appendix 7, Refs. 2-5).
CMS has recommended the use of a BMI of 18.5 kg/m2 as a level below which a multiplier for reimbursement will be applied. Examination of these death risk curves suggests that a level of BMI of 18.5 kg/m2 has a log-hazard risk of death of approximately 0.6 (i.e., 82 percent increase in death risk). A much more tolerable risk of death for application of this modifier would be at approximately 0.2 (22 percent increase in risk of death) which would occur at a BMI of approximately 25 kg/m2.
B. Dialysis Intensity
Although it appears that CMS relied on a recent study (Appendix 7, Ref. 1) that indicated that both increased body size and dialysis intensity are positively correlated with increased survival (Appendix 7, Refs. 3-5), CMS misapplied these associations in the Final Rule. According to Lowrie, et al., increasing dialysis intensity, whether measured by urea clearance (K), time on dialysis (t) or the clearance times time (Kt), is associated with improved survival (Figure 4). The finding that increased dialysis intensity is associated with improved survival has also been demonstrated in several other investigations (Appendix 7, Refs. 6-13). In further analyses of the same data using a surface plot, increased Kt and increased BSA both are shown to be associated with decreased risk of death (Figure 5). When BMI is exchanged for BSA, or dialysis time or clearance alone is substituted for Kt, similar surface plot outcomes are found.
To a certain degree, CMS is correct that large patients have increased costs and better outcomes because increased body size is associated with increased dialysis intensity as indicated by three treatment parameters. Under this theory, sick, frail patients who have poorer outcomes are much more likely to be smaller (by BSA and BMI) and, based on dialysis time and dialyzer size alone, would cost less. Yet, such sick, frail patients require much more intensive nursing services, greater non-EPO, non-Iron, non-Vitamin D drug therapy (primarily antibiotics) and suffer an increased incidence of missed dialysis treatments due to hospitalizations. Hospitals, extended care facilities, rehabilitation centers, nursing homes and other medical settings use acuity or sickness levels to adjust efforts and reimbursement for sicker, frailer, time-consuming patients. Although there are no studies that illustrate this effect, physicians and nurses’ experience in dialysis facilities, as well as the settings above, would strongly support such a finding. Thus, costs (labor, supplies, and medications) other than dialysis time and dialyzer size are significantly increased in small, sick patients who also should have a multiplier applied for reimbursement. The Final Rule simply does not take these other factors into account.
With regard to age, we are concerned about the development of the age category of multipliers in Table 13 (Appendix 6). Research has shown that as normal humans become older, they have a slight reduction in BSA or BMI or Total Body Weight (TBW, also a measure of size) (Appendix 7, Ref. 14). Data on ESRD patients from the FMCNA database demonstrate a slight decrease in BSA in the age brackets above 45-59 (Figure 11) and a decrease in BMI beyond the 60-69 age brackets (Figure 12). Following the previous reasoning, if larger patients require larger surface dialyzers and longer dialysis time, one might conclude from the s that younger patients, who are larger, cost more. However, older patients, certainly those 70 and greater, are generally smaller in body size, but are more likely to have complicating illnesses and multiple co-morbidities. Also, they will require more intensive nursing services and greater drug therapy (other than Iron, EPO and Vitamin D). They will have more missed treatments due to increased hospitalizations. These patients, in fact, probably make up the major portion of the small, frail patients discussed previously. Therefore, we believe it is necessary to adjust the age categories for body size when using age as a case-mix adjuster.
IV. As Described in the Final Rule, the Case-Mix Adjustment System Will Misalign Resources with Patient Needs and Seriously Risk Under Funding Dialysis Payments
The lack of appropriate data and the analysis problems described above create a case-mix adjustment system that will result in CMS providing lower payments for treatment of the sickest patients. (See Figure 13). This serious flaw creates significant concerns about the potential for adverse selection of patients based on reimbursement.
Further, because the case-mix system will not completely replace the dollars removed from the ESRD program in CMS’s budget neutrality calculation, Medicare risks seriously under-funding ESRD services. We recently learned that the patient weights used in the analysis of body size and the budget neutrality factor that is derived from that are based on data from the Chronic Disease Medical Evidence Report Form (HCFA-2728). We believe this approach is analytically flawed because the HCFA 2728 Form is filed only when a patient first begins therapy for ESRD and is not routinely updated thereafter. As a result, the weight could have been taken between one month and several years before the time period for which payment is now being determined. Further, clinical data have shown that patient weight upon entering dialysis treatment could be nearly 4-5 kg higher than weight in subsequent time periods. This means the budget neutrality factor is underestimated and the case-mix system will never be able to redistribute dollars appropriately and achieve budget neutralitythe system will always be under-funded.
The case-mix system should reimburse the medical community appropriately based on the level of care each individual patient needs. For example, CMS should reimburse providers more for providing more care to patients in need of it. The elderly have a greater incidence of ESRD and the mean age of ESRD patients is 63 years of age, with the fastest growing decade being that above age 70. Yet, the system has potentially created financial incentives that would lead to lower acceptance rates for such patients and over time lead to a smaller ESRD population. Thus, KCP is extremely concerned that the case-mix adjustment approach described in the Final Rule will lead to reduced access to dialysis care for the frail elderly. We do not believe this was the intent of the Congressional mandate or of CMS.
If implemented, the case-mix adjustment system in the Final Rule would result in large differences in Medicare payment rates as much as $60.00 per treatment as noted earlier. Analyses using multivariable models suggest that the maximum difference of cost among patients within a practical range of age (20 90) and body size is only about $9.00 per treatment and the coefficient of variation of case-mix adjusted cost is only about 0.8 percent (Appendix 2 Figure 9). The associated text, which compares the case-mix adjustment factor for age to the range of age-adjusted cost from our analysis, illustrates the excessive payment rate differences required by this Final Rule.
V. KCP Recommendations:
- CMS should delay implementation of the case-mix adjustment system in ESRD. KCP recommends that the case-mix adjustment to the dialysis composite rate be delayed until CMS identifies a source of data other than the Dialysis Facility Cost Report or obtains specific information about how various services and patient conditions contribute to costs. Review of the case-mix literature as provided in the report, “An Expanded Medicare Outpatient End-stage Renal Disease Prospective Payment System-Phase 1” indicates in numerous studies that there are few strong case-mix relationships using the Dialysis Facility Cost Reports.
- Once appropriate data are used, CMS should refine the case-mix adjusters of weight and age. We recommend that CMS use age as a case-mix adjuster, but age should be corrected for body size changes that occur with aging.
- CMS should refine its case-mix adjusters at the lower and higher ranges of weight. We recommend that BSA be used as the measure of body size and general health and that there be a low-weight one standard deviation multiplier at 1.5 m2 and a 2 standard deviation multiplier at 1.3 m2 with a high-weight one standard deviation multiplier at 2.0 m2 and a 2 standard deviation multiplier at 2.4 m2.
- CMS should adjust the case-mix conversion factor to assure the case-
mix system is developed in a budget neutral manner. Our analysis shows that CMS potentially underestimated both the case-mix adjustment conversion factor and the ability of the case-mix system to redistribute dollars in a budget neutral manner.
It is critical that CMS reconsider the analysis, which led to the promulgation of the Final Rule on case-mix adjustment for the dialysis composite rate payment. Because the case-mix adjustment system is due to go into effect on April 1, we look forward to discussing our concerns with you at your earliest possible convenience.
Kathleen E. Means
cc: Hubert Kuhn, CMS; Laurence Wilson, CMS; Liz Richter, CMS
American Kidney Fund
American Nephrology Nurses Association
American Regent, Inc.
Baxter Healthcare Corporation
Bone Care International
California Dialysis Council
Centers for Dialysis Care
Fresenius Medical Care North America
Medical Education Institute
National Kidney Foundation
National Renal Administrators Association
Northwest Kidney Centers
Physicians Dialysis, Inc.
Renal Care Group
Renal Physicians Association
Renal Support Network
Satellite Health Care
Sigma-Tau Pharmaceuticals, Inc.
Watson Pharma, Inc.