Maybe Pharmaceutical Reps Actually Aren't Bribing Doctors

Megan McArdle.

"Is your doctor willing to sell you out for the price of a sandwich? That’s the implication a lot of people seem to have taken from a new study published by JAMA Internal Medicine. It looked at the prescribing behavior of physicians for whom pharmaceutical reps bought meals, and found that those who got even a meal worth less than $20 -- think Olive Garden or Potbelly, not Le Bernardin -- were significantly more likely to prescribe brand-name medications that had a cheaper alternative.

I’m a little less sure what to make of the study. For one thing, its design doesn’t really control for reverse causation -- the possibility that pharma reps were buying sandwiches for the folks who already prescribed their drugs more, because those folks were more interested in attending industry-sponsored events on that drug. We should also question the assumptions that the doctors who weren’t bought sandwiches are prescribing the “right” amount of these drugs, and that any deviation from this prescribing behavior represents an unjustified bias.

Let’s lay out an alternative story: The prescription drugs are superior to the older or cheaper alternatives in some significant way -- perhaps they’re more effective in certain subgroups of patients, perhaps they have a better dosing schedule, perhaps their side effects are more easily tolerated, or they have fewer nasty potential interactions with other drugs the doctors might be prescribing. Doctors, however, are busy. And like all of us, they can get stuck in a rut, tending to prescribe the same old things they’ve always written for various conditions. If a pharma rep could sit down with these doctors and inform them about the benefits of their newer brand-name drug, these physicians would write a lot more prescriptions for it.

But as we just noted, doctors are busy, so you’re unlikely to get 30 minutes of their time -- unless you buy them lunch. Or as the study itself says: “If events where industry-sponsored meals are provided affect prescribing by informing physicians about new evidence and clinical guidelines, then the receipt of sponsored meals may benefit patient care. If physicians, however, choose to attend industry events where information is provided about drugs they already prefer, then meals may have no affect on prescribing patterns. If, alternatively, meals change physicians’ prescribing practices as a result of promotional influence, either by encouraging future use or rewarding an ongoing preference for the promoted drug, this would be cause for concern.”

In other words, while the study does seem to suggest that buying a doctor a meal correlates with that doctor writing more prescriptions of your fancy drug (a result consistent with other studies), it doesn’t tell us what the mechanism of action is. “Doctors who are bought sandwiches by pharma reps then prescribe their drugs” sounds nefarious; “doctors who attend presentations on drugs are more likely to prescribe them” somewhat less so. While the easy, intuitive response is that they’re bribing doctors to write for their drugs, the frequent wrongness of easy, intuitive responses is exactly why we have a whole elaborate social science system, building theories through series of studies, rather than just checking our gut every time we want an answer to some question.

And indeed, a straight bribe actually seems the least likely explanation for this behavior. Doctors' incomes have been under a lot of pressure in recent years. But they are not under so much pressure that these folks are in desperate need of someone to buy them an inexpensive sandwich a couple of times a month. (The average cost of a meal in the study was $12-18 -- a respectable weekday lunch, no doubt, but one that most physicians can afford to pay for themselves without much trouble.) It seems implausible that a lot of cardiologists (average income almost $400,000) were thinking to themselves: “Sure, this brand-name prescription is probably completely unnecessary, but if I stop prescribing it, who will buy my next ham on rye?”

So there's the most innocuous theory: that doctors who receive meals are also receiving information that helps them do their jobs. There's the most cynical theory: bribery. Somewhere in between is the possibility that gifts create the drive for reciprocity.

By giving physicians something -- anything -- pharmaceutical sales reps trigger a deep human instinct for reciprocating. This impulse is what small-group economies run on; instead of the measured exchange of trade, foragers embed themselves in open-ended networks of favor trading. I do something for you now, and later, you do something for me (the "what" and "when" to be determined). Though this stuff is often derided as corruption or old boys’ networks when it happens in modern economies, some amount of it is essential to keep any human group running -- and it is often what restrains the more cutthroat temptations of the marketplace.

Interestingly, the "gift" that triggers this impulse doesn’t even have to be something you particularly want. In one study, for example, gifts to a charity went up substantially when the solicitation letters included a gift of postcards and envelopes, even though, judging from my own household’s behavior, I’d guess that few of these gifts were actually used. On the other hand, it does seem that the open-ended nature of the gift is important. Another study found that nonprofits' thank-you gifts to donors actually made subsequent donations go down, perhaps because it made the relationship feel like a quid pro quo. People like doing things for others. They don’t like paying $250 for a charity-themed mug.

In some sense, then, physicians might be acting out of a laudable instinct -- responding to generosity with generosity of their own. But the action itself is not socially desirable, because few of us want our medical treatment to be the currency in someone else’s favor-trading network.

The question is how you figure out whether this is what’s going on -- and then, perhaps, what to do about it. As the postcard example suggests, the gifts can be of entirely negligible value or even use, and they still trigger this deep human instinct. That’s exactly what this study seems to show: A past crackdown on expensive gifts to physicians still left us with a sizeable problem. If doctors had been primarily motivated by expensive meals as bribes, then the switch to $15 sandwiches should have driven the effectiveness of this technique to near zero. Instead they remained effective.

Physicians' ethical guidelines could outright forbid contact with pharmaceutical or medical-device representatives, but that would create more problems than it solved. Who wants to deprive physicians of information from folks who know a lot about the product?  Or perhaps we could forbid them to accept gifts of even negligible value, not so much as a pen or a sandwich eaten on the lap while watching a presentation … while accepting the risk that doctors who buy their own sandwiches and Bics will also stop making the time to learn about new drugs."