Aaron P. Mitchell
Oncologists who received general payments from a pharmaceutical company appeared more likely to prescribe a drug marketed by that company, according to a research letter published in JAMA Internal Medicine.
“This is the first study to show an association between pharmaceutical industry payments and increased prescribing for oncology drugs, as well as drugs in other areas of practice,” Aaron P. Mitchell, MD, fellow in the division of hematology and oncology at University of North Carolina at Chapel Hill School of Medicine, told HemOnc Today.
The pharmaceutical industry pays approximately $7 billion annually to physicians and teaching hospitals. Previous research has shown an association between industry payments and prescribing patterns of higher-cost brand-name drugs in other medical fields. However, whether this trend persisted for selection of oncology treatments had been unclear.
Mitchell and colleagues evaluated data from the CMS Open Payments and the Medicare Part D Prescriber Public Use File databases to evaluate any association between payments to oncologists in 2013 and the drug selections they made for patient treatment in 2014.
For the analysis, researchers chose therapeutics that were FDA approved, recommended by the National Comprehensive Cancer Network, and had been prescribed by at least 10 physicians in 2014. Investigators focused their analysis on sorafenib (Nexavar, Bayer), sunitinib (Sutent, Pfizer) and pazopanib (Votrient, Novartis) for the treatment of metastatic renal cell carcinoma, as well as dasatinib (Sprycel; Bristol-Myers Squibb, Otsuka), imatinib (Gleevec, Novartis) and nilotinib (Tasigna, Novartis) for the treatment of chronic myeloid leukemia.
“These two cancer types had several different orally available drugs to choose from that had equivalent FDA and NCCN recommendations,” Mitchell said. “Because we were studying physician choice between drugs, we wanted to study situations where oncologists would have multiple different — yet similar — treatments to choose from, and these cancers had enough approved drugs in order for us to make a comparison.”
The pharmaceutical industry pays approximately $7 billion annually to physicians and teaching hospitals.
The study included oncologists who filled a minimum 20 prescriptions among the three drugs within each cancer type in 2014. Researchers linked the prescriptions to general and research payments determined by the Open Payments system, which records any payment a physician receives from a pharmaceutical company.
General payments include sponsored meals, consultant fees, honoraria for speaking engagements, gifts, travel/lodging and educational grants. Research payments comprised those made to physicians identified as principal investigators on clinical trials. These payments include anything related to basic science and clinical research, Mitchell said.
“The Open Payments system divides all payments from companies into three categories: general payments, research payments and ownership interests,” he said. “Ownership interests are so few and far between that we just focused on the two payments for the analysis.”
During the study period, 354 physicians prescribed drugs for metastatic renal cell carcinoma and 2,225 prescribed drugs for CML.
For metastatic renal cell carcinoma, results showed physicians appeared more likely to prescribe a drug if they received general payments only (OR = 2.05; 95% CI, 1.34-3.14), research payments only (OR = 1.84; 95% 1.25-2.7) or either payment type (OR = 1.84; 95% CI, 1.25-2.7).
For CML, results showed physicians appeared more likely to prescribe a drug if they received general payments only (OR = 1.29; 95% CI, 1.13-1.47) or either payment type (OR = 1.31; 95% CI, 1.14-1.48).
Among oncologists who prescribed the evaluated drugs for metastatic renal cell carcinoma, 9% received research payments from the manufacturer and 25.1% received general payments in both 2013 and 2014.
Among oncologists who prescribed the evaluated drugs for CML, 3.8% received research payments from the manufacturer and 39.5% received general payments.
“We can’t say for sure there is an association between research payments and prescribing. There looks to be a less consistent association [for CML], and it is not as strong as for general payments,” Mitchell said
“There are many academic oncologists who depend on pharmaceutical industry funding for their research so, to some degree, it is reassuring that we didn’t see a clear demonstration of funding for research as a potentially concerning effect on prescribing,” he added.
An analysis of individual drugs showed physicians who received general payments appeared more likely than those who did not receive general payments to prescribe sunitinib (50.5% vs. 34.4%; P = .01), dasatinib (13.8% vs. 11.4%; P = .02) or nilotinib (15.4% vs. 12.5%; P = .01).
Researchers observed the opposite trend for imatinib (72.4% vs. 75.5%; P = .02).
Because imatinib and nilotinib are made by the same manufacturer, researchers suggested this may reflect a strategy by Novartis to promote switching to nilotinib before imatinib’s patent expired in 2015.
“We firmly disagree with the characterization ... that Novartis’ promotional activities for Tasigna (nilotinib) were connected to the anticipated patent expiration of Gleevec (imatinib),” Julie Masow, spokeswoman for Novartis, told HemOnc Today. “Novartis did — and continues to — promote Tasigna for Philadelphia chromosome-positive chronic myeloid leukemia patients appropriately based on the results of a head-to-head clinical trial, which demonstrated Tasigna’s superiority over Gleevec.”
At ASH Annual Meeting and Exposition in 2012, a study reporting the differences in molecular response among patients with Philadelphia chromosome-positive CML treated with nilotinib or imatinib was included in the peer-reviewed scientific program.
“The long-term phase 3 data showed that Philadelphia chromosome-positive chronic myeloid leukemia patients on Tasigna achieved a significantly deeper molecular response versus Gleevec for both newly diagnosed patients and those who switched to Tasigna,” Masow said. “These new findings were significant for the community and may have played a role in changing physician CML prescribing practices during the JAMA study’s period compared to the time before this phase 3 data [were] available.”
Mitchell and colleagues acknowledged several limitations of their study. These included the observational design, potential inaccuracies in Open Payments data and lack of generalizability to other cancers. However, the greatest limitation was the limited time period, Mitchell said.
“With it being such a short snapshot of time, we can’t really work out anything in terms of the order of payments to doctors and prescribing that would allow us to get us to get closer to the question of causality,” Mitchell said. “What we see is something we certainly want to describe as an association and not an effect of industry payments on prescribing, because we can’t say that yet. If more years of data become available, then we can study changes in doctor’s prescriptions over time, which we can’t yet do based on this limitation.”
The overall association between industry payments and prescribing may be driven by one of two hypotheses: the bribe hypothesis — in which an industry partner gives money to a physician and then the physician starts using the drug — or a reward hypothesis, in which a physician decides by their own accord that they prefer one drug over another. Then, the company rewards physicians with increased payments.
“The first scenario is a big concern that would very much challenge us to reconsider holding these types of relationships with the pharmaceutical industries,” Mitchell said. “Patients want treatment in an unbiased setting, where their preferences for treatment and medical needs are the only things influencing treatment decision and no other influences.”
Masow said Novartis “does not make payments to health care professionals for the purpose of influencing prescribing decisions,” and that there are no valid data that establish payments from pharmaceutical companies — such as those described in the study by Mitchell and colleagues — influence prescribing decisions.
“Advances in patient care often result from discussions among pharmaceutical companies, physicians and teaching hospitals about the details of data developed in clinical trials,” Masow said. “These interactions help ensure that medicines are used safely and that new research findings are applied to clinical practice. Novartis appropriately works with expert physicians to set up meetings where they can discuss clinical trial data with their peers to help educate them about the appropriate, data-driven use of FDA-approved medicines.” – by Melinda Stevens
For more information:
Aaron P. Mitchell, MD, can be reached at firstname.lastname@example.org.
Julie Masow can be reached at email@example.com.
Disclosure: The authors report no relevant financial disclosures.