Oncologists who receive payments from industry may be more likely to prescribe non-recommended and low-value treatments. That is the assertion of a paper published in The BMJ that analyzed data on claims and physician payments to assess whether drugmakers have a negative influence on prescribing habits.
Earlier studies have found U.S. physicians, who collectively pocket $2 billion a year from drug and device companies, are more likely to prescribe medicines from companies that write them checks. However, the studies lacked insights into whether the seeming effects on prescribing practices are positive or negative for patients. In theory, docs who receive pharma dollars may be more likely to prescribe the right drugs.
To assess the impact of payments, physicians at Memorial Sloan Kettering Cancer Center and other sites looked at the drugs prescribed to Medicare beneficiaries from 2014 to 2019. The study focused on four drugs that guidelines either discourage or state they offer no benefits over cheaper products.
Comparing the Medicare data to a payment database enabled the researchers to show whether physicians who received cash from pharma companies were more likely to prescribe the drugs. Prescription rates were higher in the payment cohort for three of the four medicines.
The difference between the prescribing patterns of doctors was biggest for denosumab, a drug sold by Amgen as Xgeva, in castration-sensitive prostate cancer. U.S. guidelines recommend against the use of denosumab in the setting. Yet, 49.5% of patients whose doctors took pharma payments received the drug, compared to 31.4% of their counterparts whose oncologist hadn’t received payment.
Controlling for patients’ characteristics and the calendar year reduced the gap between the two cohorts slightly. Accounting for other characteristics of the physicians, such as their specialization, had a bigger effect, closing the gap between the prescription rates in the two cohorts to 7.4 percentage points.
The study linked physician payments to higher rates of prescriptions for two other non-recommended or low-value drugs—granulocyte colony-stimulating factors for patients at low risk for neutropenic fever and nab-paclitaxel for cancers with no evidence of superiority over paclitaxel—but the differences were smaller than in the denosumab analysis.
The potential negative effect on prescribing habits evaporated completely in an analysis of generics and biosimilars. The use of branded drugs when an off-patent alternative was lower among patients whose oncologists received industry payments, 83.5%, than their peers whose doctors didn’t take payments, 88.3%.
The authors acknowledge multiple limitations that mean the study cannot infer a causal link between the payments and prescribing practices. They nonetheless argue that the findings raise concerns about the quality of patient care that suggest “it may be appropriate to reexamine the current status of personal payments from the drug industry to physicians.”