Can Money Really Be No Object When Cancer Care Is the Subject?

Drug prices are the most rapidly increasing component of American health care costs, and cancer drugs in particular have been leading this meteoric rise. For medical care in general and cancer treatment in particular however, we have, at least in the United States, embraced the notion that such care should be, unlike virtually all other goods and services, above discussion of, and indeed even consideration of, cost. In fact, the US Food and Drug Administration does not consider cost in the evaluation and approval process of new drugs, while Medicare acts without ability to directly negotiate prices of drugs or consider cost-effectiveness information, thereby essentially compelling Medicare to purchase the drugs the US Food and Drug Administration approves at any price that a pharmaceutical company chooses to set. In parallel, we have created a health care system heavily dependent on third party payment, in which few individuals are paying with their own money, and even fewer could afford to do so. The result of this well-intentioned but ultimately unrealistic structure is our current health care morass, in which costs are spiraling out of control, a disconnect has been created between cost and value, and the descriptive word unsustainable is used with such frequency that we have become numb to its meaning; we decline to consider what follows when our unsustainable system, once no longer sustained, fails, and what havoc will be wreaked in terms of even further widening of health care disparities and lack of access to necessary care when that happens. As we have begun to slowly acknowledge a need to at least consider costs and value in cancer care, the research and reporting of cost effectiveness has assumed a more prominent role, both in the academic literature and in the national discussion. Indeed this has advanced to the point that the stated theme of the 2014 American Society of Clinical Oncology Annual Meeting was a consideration of cost and value. This represents an encouraging trend, since the topic of cost and consideration of the financial toxicity that a treatment could inflict on a patient has previously been taboo to the degree that we find ourselves acknowledging that many doctors feel uncomfortable discussing costs of medical treatments with patients. Considering the degree to which we are comfortable routinely asking our patients about their bowel, bladder, and sexual function, the concept that we should be uncomfortable with any topic because it is perceived to be of a too personal nature would seem incongruous. In the article that accompanies this editorial, Goldstein et al have presented a cost-effectiveness analysis of the addition of bevacizumab to the most commonly used standard chemotherapy for the treatment of metastatic colorectal cancer). They conclude that “bevacizumab provides minimal incremental benefit at high incremental cost per quality-adjusted life-year in both the first and second-line settings of metastatic colorectal cancer.” Relevant questions for us as readers of their analysis would be: do their data support these conclusions, are there flaws in the data and/or methodology, and if so, do they weaken the conclusions, and finally, how should an analysis such as this be used to help make care better for our patients? Let’s start by reviewing the methodology. While a prospective analysis based on actual individual patient care and cost data would have been preferable, no such analyses were done on the initial trials. Goldstein et al have therefore chosen to do a retrospective analysis based on available published literature, and have imputed the costs of management of major toxicities based on their reported incidences in the pivotal trials and the average expected costs of managing such events, and imputed costs of drugs based on average patient size and median duration of reported drug use. Given that the trials used for their analyses are large, randomized, phase III trials, these assumptions seem reasonable and are not likely to meaningfully detract from the interpretability of their data. Furthermore, as their analysis found that the major driver in cost differential was the cost of the bevacizumab, and there were relatively minor differences in the incidences of severe toxicities between study arms, any discrepancies between actual and imputed costs associated with adverse events on the bevacizumabcontaining versus the non–bevacizumab-containing arms are not likely to differ between arms and so are unlikely to impact the reliability of the analysis. One possible concern with the data presented would stem not from the analyses performed by Goldstein et al, but rather from the limitations of the pivotal trials on which their analyses are based. As lead investigator of the N016966 first-line study used by Goldstein et al in their analysis, I will note that the N016966 trial was a 2 2 randomization, first between capecitabine and oxaliplatin versus infusional fluorouracil, leucovorin, and oxaliplatin (FOLFOX), and then between bevacizumab or placebo. Thus, technically the costeffectiveness analysis should have considered the cost of capecitabine as well as fluorouracil and leucovorin. Again, however, given the randomized nature of the study and even use of capecitabine in both arms, this issue is unlikely to meaningfully influence the costeffectiveness analysis. More importantly, in the N016966 trial, based on investigator and/or patient preference, a substantial number of patients came off therapy, including bevacizumab therapy, before progression of disease. It is not clear to what degree continuation of JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 33 NUMBER 10 APRIL 1 2015