Major Finding: The use of GnRH agonists as androgen deprivation therapy for prostate cancer rose dramatically from the 1990s until 2004, an interval in which Medicare reimbursement for the drugs was profitable for physicians. It then declined markedly in 2004 and 2005, when the government's reimbursement policy was changed and use of the drugs was no longer as profitable.
Data Source: A cohort study using data on 54,925 patients with prostate cancer who were enrolled in the SEER database 1994-2005.
Disclosures: This study was funded by the American Cancer Society. Dr. Shahinian reported working as a consultant to Amgen.
For men with prostate cancer, use of gonadotropin-releasing hormone agonists rose dramatically in the 1990s, when Medicare reimbursement for the drugs was highly profitable for physicians, and dropped just as dramatically after 2004, when reimbursement was drastically lowered, according to an analysis of Medicare data.
The recent reductions in use were most profound among patients for whom the drugs were probably not beneficial and therefore inappropriate. In contrast, among the types of patients for whom the GnRH agonists' benefit has been established, use did not change with reimbursement level, said Dr. Vahakn B. Shahinian of the University of Michigan, Ann Arbor, and his associates.
“Our findings suggest that reductions in reimbursement may influence the delivery of care in a potentially beneficial way, with even the modest changes in 2004 [in reimbursement policy] associated with a substantial decrease in the use of inappropriate therapy,” they noted.
The investigators used data from the Surveillance, Epidemiology, and End Results (SEER) Medicare database on older cancer patients to test their hypothesis that, given the 50% cut over 2 years in reimbursement to physicians administering the GnRH agonists, the use of the androgen deprivation therapy “would decline markedly for indications for which there was limited evidence of efficacy” but would continue to be used for truly evidence-based indications.
The researchers categorized the use of androgen-deprivation therapy in 54,925 prostate cancer patients seen from 1994 through 2005 as inappropriate, appropriate, or discretionary – the last category being for therapy of uncertain benefit because of insufficient evidence or because reasonable alternatives were available.
“Reimbursement for GnRH agonists per monthly dose fell from $356 in 2003 to $311 in 2004 and to $176 in 2005,” Dr. Shahinian and colleagues noted.
The rate of inappropriate use of the drugs increased steadily from 30% in 1994 to a peak of 45% in 2002, then dropped precipitously, according to the analysis. “In the inappropriate-use group, there was a dramatic drop in rates … from 39% in the fourth quarter of 2003 to 30% in the first quarter of 2004, with a continued decline to 22% by the end of 2005,” the researchers said (N. Engl. J. Med. 2010;363:1822-32).
In the discretionary-use group, the rate of use also was highest in 2003, “gradually declined in 2004, and dropped more markedly in 2005,” they added. Rates of use did not decline in the appropriate-use group.
“These findings are consistent with previous research on the influence of financial incentives on the delivery of health care,” the investigators said, adding that “Financial incentives are most likely to have an effect on physicians' behavior in cases in which medical uncertainty exists, as opposed to cases in which care is clearly lifesaving.”
The authors acknowledged that a guideline change by American Society of Clinical Oncology in 2004 may have influenced physicians to reach for GnRH for localized high-risk disease.
Additionally, it is possible that increasing recognition of the adverse effects of androgen-deprivation therapy may have contributed to some of the reductions in use of the drugs.
One major study published in 2005 demonstrated a link between the treatment and fracture risk. Since the changes in reimbursement policy roughly coincided with this publication, “it is difficult to separate out the contributions of these influences,” the researchers noted.
The corollary to their findings is that reimbursement policies should be carefully crafted to avoid inadvertently providing incentives for care for which no clear benefit has been established, they added.