Thursday, June 25, 2009

Will Comparative Effectiveness Research lead to Rationing?

Senators Ron Kyl (R-AZ) and Mitch McConnell (R-KY) recently introduced the ''Preserving Access to Targeted, Individualized, and Effective New Treatments and Services Act of 2009" (S1259), a legislative proposal that Sen. Kyl's office said was designed to prevent comparative effectiveness research (CER) from being used to ration healthcare. The economic stimulus package passed earlier this year contained $1.1 Billion for such research; the House Democrats' draft health reform plan now being debated provides an additional $100 Million per year for the next three years for CER; the research is a fundamental building block of the President's approach to health reform. Are we inviting rationing of health care when we support the reform effort?

A few Republican Senators aren't the only ones nervous about CER. AdvaMed, the leading medical device industry trade association, is cautiously supportive, so long as CER is not "used by Medicare, insurance companies, or other public or private payers to deny coverage". PhRMA is worried that CER could take treatment decisionmaking out of the hands of physicians and patients. BIO "is concerned that comparative effectiveness information may be used strictly as a means to contain costs, rather than deliver health care value by improving patient health outcomes". None of these associations uses the dreaded R word, but their concern is clear - in the wrong hands (i.e. the government's hands), CER could be used to justify insurers' decisions not to pay for services and technologies doctors want to prescribe and companies want to sell. And isn't that rationing?

Well ... No, not by any rational use of the term. Formally, rationing is the limitation of the amount of some scarce commodity that is available to an individual, family or community - think about food or gas rationing during WW II. Rationing can be overt and explicit policy in response to unavoidable supply limitations, as in those examples, or it can be indirect, through the purposeful limitation of supply - think about planned limitations on kidney dialysis treatment capacity in Britain until just a few years ago, or waiting lists for artificial hip procedures in Canada on account of conscious budget allocation decisions. Any purposeful management of a supply deficit for a good or service that people want and need can fairly be characterized as rationing. We've typically not done anything like that in healthcare in the United States, and CER doesn't open the door to it. CER won't artificially limit the supply of technologies or services; CER won't limit the amount of healthcare, either segmentally or in toto, available to individuals; CER won't arbitrarily impose fixed budget ceilings that translate into service shortages.

What CER does promise to do is to gather reliable information about the relative merits of different treatment options for particular classes of patients under defined circumstances. The information would be used in different ways by different health system stakeholders. It is pretty universally agreed that:

doctors would and should use CER findings to make better therapy choices or recommendations to patients; and
patients would and should use CER findings to more effectively evaluate physicians' advice and to be better informed participants in decisionmaking about their healthcare.

What seems to be in dispute is whether health insurers should be allowed to use CER findings to define the circumstances under which they will or will not pay for specified technologies or services. Let's look at the language of S1259:

Purpose:Notwithstanding any other provision of law, the Secretary of Health and Human Services--
(1) shall not use data obtained from the conduct of comparative effectiveness research, including such research that is conducted or supported using funds appropriated under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5), to deny coverage of an item or service under a Federal health care program (as defined in section 1128B(f) of the Social Security Act (42 U.S.C. 1320a-7b(f)))

This is pretty remarkable. Medicare and Medicaid, faced with data from a CER study indicating that Treatment A is clearly superior to Treatment B for a particular clinical problem for all patient groups under all conditions studied, would be prohibited from denying coverage to the clearly inferior treatment. I'll say it again differently: Medicare would be required by law to cover and pay for a treatment option known to be inferior because the data establishing inferiority were collected in a CER study.

This is nonsense. We want insurers, including Medicare and Medicaid, to use the best available data to cover what works and not cover what doesn't. And that is exactly what Medicare and every private insurer has been doing for years. The coverage analyses and decisions, often with detailed documentation of the determinants of the decision, are available online for anyone to review: when and where Medicare will pay for an implantable left ventricular assist device; under what circumstances Aetna will or will not pay for continuous mobile cardiac monitoring to help diagnose cardiac rhythm disorders; etc. Making these decisions is what insurers do, and they would be irresponsible - with our well-being as well as with our tax monies or premium payments - if they didn't use the best available information to make them.

Stakeholders potentially adversely (financially) affected by insurance coverage decisions have historically had recourse to two arguments: that the insurers are "interfering with the practice of medicine" (sometimes transposed into "interfering with the sanctity of the physician-patient relationship"); or that they are rationing care. As we've come to understand that docors are as human, fallible and self-interested as the rest of us, the first argument has lost its bite. Now, those who see wide dissemination of better information as a threat are left to play the rationing card. Get below the surface, and it is absurd - but we'll see whether it retains any power.

Wednesday, June 10, 2009

The reimbursement challenge for advanced prosthetics

We are aging, and inevitably that means that we - or at least some of our parts - are wearing out. Increasingly, there are demands for wider availability of "replacement parts" to restore physical functionality and quality of life. And while tissue engineering and stem cell research may hold the promise of biological replacement parts, grown to order from the patient's own cells, clinical realization of that promise remains in the fairly distant future. For years to come, advances in prosthetic devices will be the dominant mode for replacing the function of non-functional body parts or organs. Our imagination tends to be grabbed by the technology projects that aim at vital organ replacement - Abiomed's AbioCor Total Replacement Heart (apparently, excuse the expression, moribund), the various artificial pancreas projects that receive publicity from time to time (such as this one), liver replacement technologies, wearable artificial kidneys, etc. - but the real current action seems to be in the orthopedic arena, where basic technology is in place, but there are exciting innovations in the pipeline.

It seems reasonable to ask: "Who will pay for these prosthetic devices, and for which patients?"
The "who" is fairly clear. The population in need is primarily elderly and the dominant payer will be Medicare. The Medicare program has a long-standing and well-understood prosthetic device benefit (see Section 1834(h) of the Social Security Act), which provides payment for devices that replace the functionality of permanently non-functional body parts or organs. The benefit covers the obvious orthopedic prostheses for amputees, but also extends to implanted replacement knees, hips and other joints, external and implanted mechanical circulatory support devices, total parenteral nutrition (i.e. replacement of non-functional gut), ostomy and colostomy procedures and supplies, and a host of other technologies. For covered prosthetic devices, Medicare pays for any medical procedures required to initiate device use, the device itself, and any supplies and equipment required for ongoing functionality. To meet the "permanence" standard, the program requires clinical evidence that the impairment in function be "of long and indefinite duration" - the potential for recovery of function at some indeterminable future time is not disqualifying. For clinical procedures, implanted devices and professional services provided under the prosthetic device benefit, Medicare pays under the various payment schemes for hospitals and physicians. External prostheses and prosthetic device benefit supplies and equipment are paid under the rules and procedures for durable medical equipment, with a 20% patient copay obligation.

Coverage for prosthetic devices - determination of whether Medicare will pay for a particular device class, and if so under what circumstances - is subject to the conventional coverage standard of "reasonable and necessary". As usual, it is in the application of this standard that the potential for controversy resides. Even for vital organs, where complete failure means death, Medicare has struggled with three difficult questions:
  1. What is the threshold - the degree of organ failure - for coverage under the prosthetic device benefit;
  2. When is a device good enough - i.e. a sufficiently effective replacement - to warrant coverage; and
  3. Is there a performance level for a prosthetic device beyond which incremental improvement is defined as unnecessary and therefore non-covered? How good a replacement are we willing to provide?
The most common implanted prosthetic devices are replacement hips and knees. We all know someone with one or more of these (I myself have two replacement knees). Qualification for coverage is straightforward - without the replacement the patient cannot walk; with it he/she can; every insurer finds it reasonable and necessary to cover replacement of these joints to restore mobility. And while the devices have been incrementally improved over the years to wear better, last longer, get placed easier and more precisely, etc. - all unequivically good things for patients, surgeons and insurers, there have been no substantial changes in the degree of functional capability they provide, and therefore no significant coverage issues raised. Insurers might resist a hypothetical new model knee that provides 10% increased functional longevity at a 50% increase in cost, but that value calculation is itself pretty straightforward.

A recent newsletter report about replacement ankles raised a significant coverage policy question. There are several FDA approved ankle arthroplasty (replacement) devices, a fair number of orthopedic surgeons want to do favor them and want to use them, but insurance coverage is rare. Dr. Michael Pinzur, writing in Foot & Ankle International, the official journal of the American Orthopaedic Foot and Ankle Society, wants to know why:
  • "It seems curious that the FDA agrees with the [foot and ankle society] that total ankle replacement is a reasonable treatment option . . . while several insurance providers do not find ankle replacement as a reasonable treatment option for ankle arthritis," and
  • "Should insurance companies make decisions on what treatments are appropriate and what treatments are deemed experimental?"
It turns out, of course, that FDA has never opined that total ankle replacement is a reasonable treatment option. Ankle replacement systems are class II devices. FDA has cleared at least three through the 510k process, but that signifies nothing about "reasonable" (a word and a consideration wholly absent from FDA marketing approval processes) and very little about anything else beyond the fact that the systems were deemed "substantially equivalent" to systms already on the market.

Medicare has no formal coverage statements on ankle arthroplasty, indicating that coverage may be granted on a case by case basis by the medical directors of the various fiscal intermediaries. But several major private insurers ( e.g. Anthem, Cigna,) provide detailed reviews of ankle replacement and why they don't cover the procedure: lack of reliable supportive clinical data (in part because FDA did not require clinical results for market clearance), uncertainty of superiority over surgical options, high percentage of re-ops required, lack of data based guidance concerning circumstances/sandards for use, etc. In short, advocates for ankle replacement haven't yet done what insurers reasonably expect them to do - make a strong clinical case for why and when the technology provides benefits beyond the available treatment alternatives. Contrary to Dr. Pinzur's view, this kind of decision is precisely what insurers always do, and what we need them to do.

The third question - when is a prosthesis too advanced to be necessary - is raised by recent developments in artificial lower limb prostheses, with the advent of computer- assisted joints which allow prostheses to perform in increasingly natural ways. New knee and ankle prostheses (like this one) under development will allow recipients almost perfectly normal looking gaits, much improved balance on uneven surfaces, and greatly decreased workload for given amounts of movement (translating into dramatic improvement in stamina). These prostheses will raise the ceiling on amputees' athletic performance and will provide them with capabilities increasingly approaching full normality, but functional and aesthetic. Wow! - but will insurers pay?

The answer, I think, is "Maybe someday, but not very soon." Insurers pay for prostheses that are reasonable and necessary to restore function, and the restoration target is predicated on the functional level deemed "normal" for the individual. We can see this in current coverage policies for non-computerized limbs, where Medicare and private insurers all provide prostheses capable of supporting the functional level deemed obtainable by the individual patient absent the amputation. A patient bedbound or wheelchair-bound for reasons unrelated to an amputation would not be covered for a prosthetic limb; one capable assisted ambulation becvause of cardiac conditions will be covered for the most basic prosthesis; one capable of ambulation on an uneven surface for extended periods of time will qualify for a more advanced prosthesis. Aesthetics do not enter the equation; neither does enhanced mobility for athletic or avocational purposes. Insurers will pay what is necessary for technology to help an amputee maximize his/her independance, maximize mobility in normal day to day activities, and get back to work ... but they will not pay more for appearances, or to facilitate running a marathon or climbing a mountain.

Normal function is socially defined. As these advanced technologies become available and their potential is more widely understood, it is likely that it will become increasingly difficult for insurers to maintain the "necessity" line where it is today. Technology invariably carries rising expectations with it. What is readily available and desirable becomes required. I think that will happen with advanced prostheses, aided by (relatively) decreasing unit costs as volumes grow. But not for a while, perhaps for a long while.