A fundamental problem with contemporary prostate cancer treatment in the United States is that our system for reimbursement rewards technology rather than outcomes, with the result that economic incentives drive treatment decisions that should instead be based on clinical risk and patient-centered outcomes.
Proton beam therapy is perhaps the most extreme example of perverse incentives. There has never been a single published study that has shown any clinical benefit in terms of oncologic outcomes or quality of life for proton therapy over contemporary photon-based radiation – or over surgery, for that matter – for prostate cancer. In fact, not only is proton therapy no better than intensity-modulated radiation therapy (IMRT), a recent large-scale study suggested it has greater rectal toxicity (ASCO GU, Abstract 3).
By Dr. Matthew R. Cooperberg
Yet Medicare and other payers reimburse for proton beam therapy extraordinarily generously – far more than for any other treatment. Researchers have argued that even if proton therapy allowed dose-escalation up to 90 Gy, it would still not be cost-effective (J. Clin. Oncol. 2007;25:3603-8).
Fortunately, proton beam therapy for prostate cancer is still relatively rare, though this may change as new facilities are constructed across the county. IMRT, on the other hand, the next most expensive treatment, is used very commonly for low-risk disease, again despite absence of any evidence of advantage over – or even equivalence to – brachytherapy, a much more economical option. IMRT generally requires that patients attend treatment sessions daily for up to 8 weeks. Another abstract at the ASCO GU symposium suggested that this required number of treatment sessions could safely be reduced to as few as five total. But as long as payers continue to reimburse radiation on a per-fraction basis, this strategy seems unlikely to gain traction among providers, no matter what the convenience benefit may be to the patient.
Though not discussed directly at the ASCO GU session, robot-assisted prostatectomy has been another example of rapid dissemination of a novel technology in advance of clinical data demonstrating superiority. As with the competing radiation modalities, large-scale studies so far have not demonstrated improved outcomes for robot-assisted over open surgery in broad community practice. However, in the case of surgery there is substantial literature showing improved outcomes in centers with very high surgical volumes. A fundamental difference, too, between the spread of IMRT and that of robot-assisted surgery is that the latter has been driven in part by hospitals competing for market share – but not by reimbursement incentives. In most cases, robot-assisted prostatectomy is paid no differently than open surgery, and the additional costs of the surgery are assumed by the hospitals rather than payers. Neither surgical modality approaches IMRT, to say nothing of proton-beam therapy – in terms of cost.
The evolution of practice patterns for prostate cancer is in a sense emblematic of the failings of the U.S. health care system as a whole. Treatment is frequently supply sensitive: driven by availability of technology and system capacity rather than clinical need; this situation is reinforced by economic incentives that reward maximal rather than efficient care, and that do not emphasize achievement – or even assessment – of high-quality outcomes. Responsibility for this situation is shared among payers, providers, hospitals, manufacturers of equipment, and patients. The public tends to misplace faith in new technology, assuming that if it’s newer and more expensive, it must be better. Marketing has of course played a major role as well, and there are unquestionably instances of misleading and even false advertising of these new technologies.
We need more high-quality comparative effectiveness research (CER) of prostate cancer treatments. Indeed, prostate cancer has been identified as one of the highest-priority areas for comparative effectiveness research by the Agency for Healthcare Research and Quality, the National Institutes of Health, and the Institute of Medicine. Despite saber-rattling by some members of Congress, CER will someday by necessity inform at least some insurance coverage decisions.
But the research has to be done well. There is no denying that prostate cancer studies are difficult; they are long and they are expensive if done properly. They require collecting patient-reported data with validated quality of life instruments prospectively, regularly, and consistently across all centers and treatment modalities. Relying on administrative coding data or on physician-reported toxicities to assess complications is not sufficient – patients must report their own outcomes. And metastasis and mortality outcomes – the only end points useful for comparing surgery and radiation – take years to reach and must be associated with rich clinical detail for adequate risk adjustment.