When Industry Pays for Meals, Physicians Are More Likely to Prescribe Certain Brand-name Drugs


Doctors who receive at least one industry-sponsored meal costing less than $20 on average are more likely to prescribe the most popular brand-name drug in its class compared with those who don’t get a free lunch, according to a new analysis of data on nearly 300,000 individual physicians culled from the Open Payments Program.

Implications. When Industry Pays for Meals, Physicians Are More Likely to Prescribe Certain Brand-name Drugs

Though the paper specifies that the “findings represent an association, not a cause-and-effect relationship,” this caveat does not erase the real-world implications, senior author R. Adams Dudley, MD (University of California, San Francisco), told TCTMD.

“Whether or not this is cause and effect, it is easy for doctors to underestimate how important this is for patients,” he said. “We spend in this country an extra $75 billion on brand-name drugs for which there is an equivalent generic, and patients pay about a third of that. And so, whenever possible, doctors should be trying to prescribe the generic.”

To TCTMD, lead author Colette DeJong (University of California, San Francisco) explained that for Medicare Part D patients, the median copay for a preferred generic drug is $1 per month, while average monthly prices for preferred and nonpreferred brand-name drugs are $38 and $80, respectively.

The costs can be “much, much higher for some patients” and can “really mount up” for elderly patients taking a lot of specialty drugs, she commented.

For their study, published online June 20, 2016, in JAMA Internal Medicine, DeJong and colleagues looked at 279,669 physicians who wrote more than 20 Medicare prescriptions in one of four classes: statins; cardioselective beta-blockers; ACE inhibitors and angiotensin-receptor blockers (ARBs); and selective serotonin and serotonin-norepinephrine reuptake inhibitors (SSRIs and SNRIs).

Using data from the Open Payments Program for August through December 2013, the researchers identified which physicians had received industry-sponsored meals promoting the most-prescribed brand-name drug in each class: rosuvastatin (Crestor; AstraZeneca), nebivolol (Bystolic; Forest Laboratories), olmesartan (Benicar; Daiichi Sankyo), and desvenlafaxine (Pristiq; Pfizer). In all, the physicians received 63,524 payments associated with these drugs. Nearly all (95%) of the payments came in the form of meals, which had a mean value of less than $20.

Doctors who received just a single meal promoting each brand-name drug were then more likely to prescribe it. Differences remained significant when adjusted for physicians’ specialty, sex, region, practice size, number of years since graduating medical school, rural or urban practice setting, median household income for their zip code, prescription volume for the drug class of interest, overall rate of brand-name drug prescription, and percentage of prescriptions written for low-income subsidy and Medicare Advantage beneficiaries.

Table. When Industry Pays for Meals, Physicians Are More Likely to Prescribe Certain Brand-name Drugs


Additional meals and costlier meals strengthened the observed relationships.

What Can One Doctor Do?

A big question is how physicians should alter their behavior, if at all.

Asked whether they should refuse lunches offered to them, or simply avoid the educational events where they are served, DeJong said, “it’s a lot to ask of individual physicians to end the system as it is now.

“I think that when you’re at an educational conference, it’s hard for physicians not to go to the lunch, not to hear what the drug reps have to say,” she noted. “When you’re in private practice, it’s hard to turn away a drug rep you’ve known for 15 years who’s bringing lunch for your entire staff who might be expecting it that day.”

Rather, the emphasis should be on developing better approaches on a large scale “for disseminating educational information to physicians,” DeJong suggested.

Something that both individual physicians and professional societies such as the American College of Cardiology could do “is advocate for an alternative educational system,” Dudley added. “Right now, it really is true that as new drugs come out, there isn’t a formalized mechanism for teaching people about [them].”

What with billions of dollars in potential savings at stake, he proposed, the Centers for Medicare & Medicaid Services might volunteer to pay for educational programs about new drugs and then see how this affects prescribing patterns, both in terms of over- and underuse. The idea would be to increase transparency, Dudley said, with fewer commercial interests involved.

In an editor’s note accompanying the paper, Editor at Large Robert Steinbrook, MD, encourages less industry involvement in the practice of medicine.

“Other than research support, product development, and bona fide consulting related to specific research programs and projects, it is already evident that there are few reasons for physicians to have financial associations with industry,” he comments. “Outright gifts, such as meals, may be legal, but why should physicians either expect or accept them?”

Steinbrook suggests there are “inherent tensions” among the motivations of healthcare companies, the independence and integrity of physicians, and the need for affordable care. “If drug and device manufacturers were to stop sending money to physicians for promotional speaking, meals, and other activities without clear medical justifications and invest more in independent bona fide research on safety, effectiveness, and affordability, our patients and the healthcare system would be better off,” he asserts.


Related Stories:

Sources
  • DeJong C, Aguilar T, Tseng C-W, et al. Pharmaceutical industry-sponsored meals and physician prescribing patterns for Medicare beneficiaries. JAMA Intern Med. 2016;Epub ahead of print.

  • Steinbrook R. Industry payments to physicians and prescribing of brand-name drugs. JAMA Intern Med. 2016;Epub ahead of print.

Disclosures
  • DeJong, Dudley, and Steinbrook report no relevant conflicts of interest.

Comments