This study is not secondarily reported [information], where you're asking somebody their opinion or asking somebody [about their experience] post visit. This is linguistically analyzed conversation.

Describe a typical conversation.

COLUMBIA-WALSH: In one conversation, the physician talks to a female patient almost in the form of a medical questionnaire—which is how they're trained. They ask very close-ended questions, such as, "When you get pain, does it do this?" "Yes." "Does it interrupt this?" "Yes." At the end of it, the patient walked away with an over-the-counter recommendation.

How do we train physicians to ask, in the same amount of time, these types of questions—and [in ways that get the] patient to feel they have the empowerment to share that [information] without being rushed? It's not just a medical conversation, it's how the illness is affecting their life.

GATTUSO: A lot of times, rather than saying, "Ask your doctor about this brand," an ad can simply say, "Tell your doctor about the impact of this condition on your life." That will lead to the right kind of discussion and the right kind of medication. [Companies] need to repair some of the things in the dialogue that people say out of habit.

So, is DTC advertising not working?

GATTUSO: Look at it from a marketing point of view. If you say, "If I invest X amount of dollars, do I get a return?" the answer in many successful campaigns is yes, it does move the needle in terms of brand prescribing.

We're simply saying that it doesn't appear that a high percentage of patients are going to the doctor and directly saying, "I saw X brand on TV and that's what I want." And a lot of the critics of DTC have suggested that that's the case.

Does DTC affect the conversations between patients and physicians about a drug's risk?

DTC does not appear to negatively impact the level of risk–benefit discussion. There's no appreciable difference when a DTC brand is prescribed in terms of the risks and benefits that the physician would typically talk about.

We looked at every single mention of every medication in those three categories, which covered about 98 percent of all prescriptions written in those disease areas. But the risk–benefit discussions wasn't the classic, "These are your risks of treatment, these are the benefits of treatment. These are the risks of non-treatment, these are the benefits of non-treatment." We were very generous in what we coded as a risk or a benefit— even with that, most mentions of medications simply didn't have a risk or a benefit associated with them.

In the disease categories where companies spent money on DTC [allergy and cholesterol], more of those conversations had some form of risk–benefit conversation compared with in hypertension, which had significantly less DTC spend. So you can say, at least directionally, that DTC is not removing the risk–benefit conversation. But in all those categories, the risk–benefit discussion could be made much more relevant, stronger, and richer through a more evolved form of DTC, where it's based on the understanding of the dynamics of the conversation.