New ProPublica analysis correlates physician payments with higher brand prescribing

The saying “correlation does not imply causation” comes to mind when looking at the data presented by ProPublica, the nonprofit investigative-news organization that has been running the “Dollars for Docs” effort for the past six years. Its studies, now propelled by the Open Payments System of CMS, which releases data on industry (phama and med device) spending on physicians annually, have shown a variety of patterns in physician reimbursements (for meals, speaker engagements, consulting work and other “transfers of value”) and their apparent effect on prescribing practices.

In the latest report, “Now There’s Proof,” Propublica matched the prescribing patterns under Medicare Part D with physicians across five medical specialties: psychiatry, cardiovascular disease, family medicine, internal medicine and ophthalmology (this ordering roughly corresponds to higher percentages of physicians in that specialty receiving payments with their branded drug prescribing, with highly reimbursed ophthamologists prescribing branded drugs almost 40% more frequently than unreimbursed ophthamologists). ProPublica’s methodology required it to match—sometimes by hand—doctor identifiers in Medicare Part D with the data on payments collected in the Open Payments system; out of a universe of over 1 million prescribing HCPs, 150,323 physicians in the five specialties, and who made more than 1,000 Part D prescriptions, were analyzed.

The data show a fairly linear relationship: the more money spent on a physician, the more prescribing of brand-name drugs. Overall, the data indicate that something like three-quarters of these specialists receive some transfer of value from industry in 2014 (ProPublica did not publish an exact figure), but that needs to be correlated with the universe of all prescribing HCPs including those under 1,000 Part D prescriptions. And the study authors, Charles Orenstein and Ryan Growchowski, note that “This analysis was not designed to resolve the question of causality,” with Orenstein and other writers going on to say that “ProPublica’s analysis doesn’t prove industry payments sway doctors to prescribe particular drugs, or even a particular company’s drugs. Rather, it shows that payments are associated with an approach to prescribing that, writ large, benefits drug companies’ bottom line.”

One surprising result of the analysis: there is considerable geographic variation in frequency of payments: higher the farther south and east a state is (see map). Minnesota and Vermont are the lowest two states, with Vermont being the lowest of all: only 23.5% of prescribers receiving an industry payment in that state. Conversely, 90.3% of prescribers in Nevada received a payment.

The ProPublica news article of the report cites a number of physicians, as well as PhRMA, commenting on the study, with a PhRMA spokesperson noting that when they are “working together,” physicians and industry “can improve patient care.” For its part, ProPublica seems fixated on prescribing branded drugs over generics, even though nearly 90% of all prescriptions in the US are generic, and payers and pharmacy benefit managers focus on getting generic prescribing rates as high as possible. Since the Open Payments system was put in place, industry has been revamping how they manage payments to prescribers and at least some in the prescriber ranks are trying to keep themselves off Open Payments lists. For sure, with everyone watching, everyone is proceeding more carefully.

Finally, it’s worth noting that the ProPublica database itself could be a poor (industry) man’s market research tool: the organization is selling its analysis for commercial purposes for an unspecified price.