Breast cancer bone cancer

On April 15th, 1999, the American Society of Clinical Oncology (ASCO) took the unusual step of publishing abstracts of four critical clinical trials of autologous bone marrow transplant (ABMT) for breast cancer prior to their formal presentation at the ASCO annual meeting in May. The debate regarding the efficacy of and insurance coverage for ABMT for breast cancer has been raging since the widespread dissemination of ABMT in the early 1990s. This debate has involved virtually all parties in the health care equation: patients, physicians, payers, hospitals, the media, lawyers, and legislators.

Underlying the debate was the expectation that the results of randomized clinical trials reported at the ASCO meeting would finally provide scientific data to resolve the frequently emotional controversy. Unfortunately. in contrast to the high expectations for ABMT for breast cancer, the results of these eagerly awaited trials failed to show a significant impact in patients with either metastatic breast cancer or those at high risk for metastatic disease. Unfortunately, in contrast to the high expectations that the results of clinical trials would resolve the insurance coverage controversy regarding ABMT for breast cancer, the preliminary nature of some of the results precludes its resolution. So what have we learned?

One issue that there appears to be universal agreement on is the importance of participation in clinical trials, specifically randomized phase III trials. In fact, it was the promising results of phase II trials, which frequently reported results in comparison to historical controls, which prompted randomized trials. The discrepant results between phase II and III trials emphasize the importance of randomized trials using contemporary controls. The second issue with universal agreement was that slow accrual to these randomized trials delayed the results. The role of the payer in patient accrual is a complicated one, which could be described as "damned if you don't, damned if you do."

Damned if you don't, damned if you do

In its early days, ABMT for breast cancer was considered an investigative therapy by many payers, and thus considered ineligible for coverage. One of the first salvos in the debate were challenges to this interpretation, with providers and patients arguing that ABMT represented state-of-the-art" care or was "widely accepted by the practicing physician community." These challenges, widely reported in the media, resulted in lawsuits and state mandates for coverage in ten states.

Participation in clinical trials essentially requires three Ingredients: (1) physicians willing to refer patients to clinical trials: (2) patients willing to participate and accept the uncertainty of randomization: and (3) a mechanism to pay for the therapy. By the mid 1990s many payers began to routinely provide coverage for ABMT for breast cancer, removing what was perceived as the principle hurdle to conducting clinical trials. Thus the stage was seemingly set for rapidly forging ahead with the necessary research,

However, in its "damned if you do" or "be careful what you ask for" role, the widespread coverage for ABMT may have perversely inhibited research. Based on their contract language, many payers cannot make a distinction regarding the setting of an ABMT. For example, health plan coverage for ABMT implicitly suggests that it is no longer considered investigational and would be medically necessary for an individual patient.

Participation in a clinical trial contradicts that interpretation. While many health plans may have been interested in limiting coverage of ABMT for breast cancer to clinical trials, for most plans there was no contractual basis to do so, and widespread insurance coverage tends to dampen enthusiasm for participation in clinical trials.

For example, in the early recruitment period for the clinical trials, it was estimated that less than 1 percent of patients with metastatic breast cancer who received an ABMT did so within the context of a randomized clinical trial. The comparable figure for patients with high risk breast cancer was 11 percent. (1) The General Office of Accounting reported that the wide availability of ABMT is possibly the major reason for the poor accrual to the clinical trials. (2) Zujewski and Abrams, from the National Cancer Institute, further state, "This problem [accrual] has been further aggravated by the willingness of insurers to pay for this therapy, whether or not a patient enters a randomized trial..." (1)

Fueling the bias

With coverage not restricted to clinical trials, many patients opted not to be part of a randomized study for fear of being randomized to the "standard" treatment arm, which was widely perceived to be ineffective, The initial debate regarding insurance coverage may also have influenced the perception of the superiority of the ABMT arm. Stating that ABMT was "state-of-the-art" or "widely accepted" as a means of lobbying for coverage may have also fueled the growing positive bias towards ABMT among both physicians and patients and the negative bias towards standard therapy. Resultant state mandates further reinforced the perception that ABMT should be considered as standard therapy.

Although certainly not totally responsible. the more relaxed coverage criteria by payers in the mid-1990s also coincided with an explosion of facilities offering the treatment and the rapid increase in patients undergoing ABMT. It is estimated that there are some 350 health care organizations offering some form of ABMT, which includes academic, community, and proprietary institutions. Some 30,000 autologous transplants were performed in 1998, with 30 percent of them for breast cancer. This growth of ABMT capabilities, in part stimulated by widespread insurance coverage, represents a considerable infrastructure supporting ABMT for breast cancer; it is estimated that there is now a 25 to 50 percent excess service capacity for ABMT. (3)

At the very least, such capacity carries an implicit bias towards ABMT. In response to the growing demand for ABMT, some insurers have set up transplant networks for ABMTs, focusing on high quality institutions and providers. Criteria for participation in the network typically require a certain volume of transplants performed per year, potentially creating a bias toward ABMT.

In 1995, focus groups held at the ASCO meeting explored the reasons behind the slow accrual to the breast cancer trials. One of the reasons cited was competition within medical centers between participation in phase II versus phase III trials. In most phase II trials, all patients receive an ABMT. In randomized phase III trials, institutions may lose half of their prospective ABMT candidates to the non-transplant conventional therapy arm. (4) In addition, single institution phase II trials may present less administrative burdens compared to the multi-institution phase III trials. Patient-driven reasons for slow accrual may be related to perceived physician bias towards ABMT, travel costs, travel time, and insurance coverage issues.

No resolution

Another point of agreement regarding the results of the clinical trials is that they are still preliminary in nature, and like many clinical trials, these probably raise more issues than they answer. Therefore, anybody expecting that these trials would resolve the debate regarding the efficacy of ABMT for breast cancer is likely to be sorely disappointed. As the researchers pointed out, long-term results of ABMT for breast cancer may still show a positive outcome, particularly in patients with high-risk breast cancer. Additionally, further subsets of patients who respond optimally to ABMT may still be identified.

There will be continued research interest in refining the combination of drugs used and the timing of therapy. For example, in the one clinical trial from South Africa that did report a treatment advantage of ABMT in patients at high risk for metastatic breast cancer, the ABMT arm included drugs not typically used in this country. Additionally, unlike the trials in this country, the South African patients did not achieve initial rounds of induction chemotherapy designed to reduce tumor burden to a minimum prior to the high dose regimen. These observations led to the suggestion that initial rounds of induction chemotherapy may induce drug resistance to the subsequent high dose therapy and thus should be eliminated.