Practicable Frequent Hemodialysis: A Proposal to Meet the Needs of Patients and the Requirements of Medicare

Frequent Dialysis and the National Institutes of Health Trials 

Hundreds of articles have appeared for more than 30 years reporting the benefits of frequent hemodialysis (HD) compared with 3-treatment-weekly in-center hemodialysis (conventional HD) without a single opposite conclusion or report of significant risks. However, several articles have appeared recently1, 2 supporting the National Institutes of Health (NIH) Frequent Dialysis Network: Daily Trial and Frequent Dialysis Network: Nocturnal Trial as necessary precedents to a Medicare frequent dialysis funding decision for the conventional HD population. These articles acknowledge many of the benefits, but they characterize the reported studies as small or observational and call for higher quality evidence from randomized clinical trials (a randomized trial of frequent nocturnal dialysis with favorable results has since been published3). Rocco1 and Kliger2 suggested by analogy that just as the HEMO Study4 did not seem to confirm the consensus of nephrologists about longer treatments, a randomized clinical trial of frequent dialysis might be similarly surprising. They conceded that the NIH trials, each with only 250 subjects, lacked sufficient power to establish a survival advantage. Kliger2 acknowledged that more than 3,500 subjects in the daily study and 5,000 in the nocturnal study would be required to have even 90% confidence in measuring a 30% decrease in mortality, likely placing that goal out of reach. Agar5 noted that conventional HD itself has never been subjected to a randomized trial.

The NIH reports starting and ending dates of January 2006 to July 2009 and January 2006 to January 2010 for the daily6 and nocturnal trials.7 However, both trials were still recruiting patients as of April 2008, despite reduction from 250 to 150 subjects in the nocturnal trial, which may further jeopardize these schedules. Even viewed optimistically, this will apparently delay a policy decision to extend Medicare coverage for frequent dialysis for the conventional HD population until at least 2011. If a proven survival advantage then is seen as required, the delay will be many years longer. Some might argue that many medical advances have substantial life-saving potential, but cannot be responsibly used without convincing evidence of both their efficacy and acceptable level of risk. However, frequent dialysis differs from “many new medical advances” in 2 important respects. It has long been used with favorable results, albeit on a tiny scale (<1% of dialysis patients), and through decades of experience, no significant adverse effects have been reported.

Resolving the Dilemma 

Should Medicare then abandon waiting for higher quality evidence and initiate immediate funding for frequent dialysis for conventional HD patients? I suggest this is not the right question. Doubling weekly treatments from 3 to 6 would halve US in-center dialysis capacity (4 or 5 treatments would have nearly the same effect). Given the resultant staffing, facility, and equipment needs, such a decision would require years to implement and create major short-term conflicts about allocation of resources. Dialysis patients already consume a disproportionate share of Medicare, representing 1.7% of Medicare patients but consuming 6.7% of Medicare's resources.8 Doubling treatments and hence costs seems politically naive, even if it were practical. “When they say it isn't about money, it's about money!” (source unknown).

There is a much more appropriate question; how can we make frequent dialysis and its benefits available to conventional HD patients and at the same time control and prospectively reduce Medicare costs? This goal would be politically realistic, the proverbial “win-win.” The answer is for Medicare to take steps to accelerate the rate of conversion from conventional HD to frequent dialysis at home. The long decline in home dialysis has modestly reversed in the last several years. The leading manufacturer of home machines reports use by 2,481 patients as of March 31, 2008, still much less than 1% of the total dialysis population.9 Only a minority of dialysis centers have initiated home programs. Most recognize the benefits, but see the front-end costs as daunting. Medicare now allows an extra $20/d added to the composite rate for up to 25 days, a fraction of the initial investment required to organize and recruit, educate, train, and “hand-hold” new home patients. This investment represents a barrier to home dialysis for many centers despite the promise of a favorable long-term return.

I propose that Medicare accelerate conversion from conventional HD to frequent dialysis by initiating 1-time payments to centers of, for example, $15,000 for each patient upon successfully completing 5- or 6-times-weekly dialysis at home or in a self-care unit for, for example, 45 days. Payment delay will encourage appropriate patient selection, effective training, and timely support. Other rules must be considered for returning dropouts, retraining after failed transplantations, and so on while keeping in mind that any rule increasing center-perceived risk will reduce incentives and hence conversions.

Medicare would also require reporting of designated preconversion and postconversion clinical, cost, and social data to a central research database at a qualified independent institution. Among the most crucial data would be hospitalization and drug use, providing visibility into the return on Medicare's investment in 1-time subsidies to centers. If positive, Medicare could increase its subsidy to accelerate conversions and hence cost savings to Medicare and benefits to more patients. Conversely, subsidies could be reduced or eliminated if a persistent negative return is noted. Given the modest numbers associated with even an accelerated conversion rate, this subsidy likely would be completely offset or negligibly small in the context of the total Medicare end-stage renal disease–related budget and thus provide a practicable solution to the frequent dialysis decision dilemma. The dictionary defines “practicable” as “capable of being effected, done, or put into practice; feasible.”10

Patient preconversion and postconversion employment, school, disability, and public assistance status would also be monitored in the database. The enhanced well-being widely reported by frequent dialysis patients may be expected to return many as productive, rather than dependent, members of society.

A fundamental flaw with most government subsidies for any purpose is that it is politically difficult to later end dependency of their beneficiaries. This proposal avoids the dependency problem because it involves only a 1-time payment for each converted patient. Reducing or eliminating future subsidies would not alter or reduce the care any patient is receiving. We have witnessed some centers self-financing the front-end costs based on their own analysis of financial feasibility; it seems likely that other centers responding to the 1-time subsidy would reach a similar conclusion after their startup expenses were behind them. I have argued elsewhere11 that the very survival of dialysis centers may ultimately depend on conversion of conventional HD patients to home HD because the differential inflation rates between center labor and revenue can be countered best by substitution of unpaid home patient/caregiver labor for in-center labor.

The immediate labor savings to the center will be substantially offset by increased dialysis machine costs (1/patient instead of 1/multiple patients) and increased consumable costs (5 to 6 times weekly instead of 3 times weekly). However, this substitution of technology and materials costs for labor costs is important. Although unit labor costs inexorably inflate over time, unit technology costs typically decrease. Although labor costs rarely reflect economies of scale, home dialysis machines and consumables are susceptible to such economies. Thus, Medicare may also expect improvement in dialysis cost trends to accompany a shift from in-center to frequent dialysis at home.

The proposed 1-time subsidy to centers will also stimulate home dialysis equipment and consumables markets not by subsidizing their costs but by increasing market growth rates. This, in turn, will stimulate private investment. Given the miniscule size of the market, even limited efforts by the major dialysis equipment manufacturers to offer innovative designs for the home have been evident in only the past few months. For the same reason, potential breakthrough renal replacement ideas from the universities (eg, nanotechnology based) have attracted minimal attention to date from the venture capital community.

Finally, from the perspective of dialysis centers, the risk of inadequate future Medicare reimbursement rates is partially shifted to vendors, for which pricing for equipment, maintenance, and consumables must remain compatible with such rates as a necessary condition of remaining in business. This, combined with shifting an ever-increasing proportion of labor costs to home patients and caregivers, should strengthen dialysis organizations.

Because only a subset of the conventional HD population will benefit from the proposed Medicare 1-time subsidy, is such a policy discriminatory and hence an inappropriate use of government funds? Conversion begins with self-selection by patients, just as with organ transplantation. Not every patient is qualified, just as with transplant applicants. However, unlike transplants, Medicare will be able to manage supply and demand through adjustment of the conversion subsidy. If centers follow a simple first-come, first-serve rule in selecting qualified applicants, discrimination will be avoided and fairness will be preserved.

Undue coercion, reflected by a high dropout rate, should largely be prevented by the penalty implicit in delaying 1-time subsidy payments until 45 days after conversion.

Research Consequences 

This proposal will not answer questions concerning conversion of conventional HD patients as a whole to frequent dialysis because it does not call for randomization, questions which I believe are moot. Instead, the research target becomes self-selected qualified members of the conventional HD population. The most relevant questions will be answered by comparison of preconversion and postconversion data; clinical, economic, and social consequences for each patient; economic impact on the center; effect on Medicare costs; and, finally, impact on society. The growing national database will become an increasingly rich source for investigators. Technology, frequency, and treatment time differences will be illuminated. The efficacy and cost-effectiveness of the Medicare 1-time subsidy policy would become increasingly clear.

Some may argue that simultaneously changing 2 variables, 3 to 5 or 6 times weekly and in-center to home, is poor research design, a difficulty avoided in 1 of the NIH trials. A further confounding issue is that the presently dominant home system, the NxStage System One (NxStage Medical, Inc, Lawrence, MA), uses a saturated dialysate system; its blood flow rate is much greater than its dialysate rate, in contrast to conventional HD systems where the opposite relationship prevails. Only limited data are available about the effect of this difference.12 It may be difficult to ever sort out the independent effects of these and perhaps other variables. However, such distinctions should not deter research addressing the central policy issue for patients and Medicare alike, validating the oft-reported favorable clinical and economic consequences of converting self-selected conventional HD patients to frequent dialysis at home.

Summing Up 

I urge shifting Medicare policy for frequent dialysis from waiting for a future “yes or no” answer to emerge from the NIH trials with its problematic operational and political implications to adopting a 1-time subsidy for each patient converted from conventional HD to frequent dialysis at home or self-care, together with mandatory monitoring of clinical, cost, and social consequences. This will accelerate conversion of patients to frequent dialysis and brighten their presently grim outlook and poor quality of life, lessen the revenue-cost strain for dialysis organizations, and stimulate research and private investment in new renal replacement technologies. Costs to Medicare are modest, controllable, and easily discontinued without affecting already converted patients. Comparison of preconversion and postconversion clinical status, Medicare costs, and social impact will yield data of increasing quality as conversions grow.

With a practicable alternative for home dialysis–qualified patients, we can no longer justify delaying their chance for longer and better lives.