In 1997, the United States lifted a ban on prescription-drug ads that are pitched directly at the public rather than at doctors or other health professionals.
The change prompted pharmaceutical companies to invest heavily in retail advertising, and annual spending in this sector now averages more than five billion dollars (four billion euros).
Advocates of direct-to-public drug advertising say it raises awareness and encourages the treatment of neglected conditions.
Critics counter that it artificially drives up demand for expensive novel drugs that have not been tested over the long haul.
Both sides, though, have assumed that the marketing technique works -- but the new study suggests they have got it flat wrong.
"People tend to think that if direct-to-consumer advertising wasn't effective, Pharma wouldn't be doing it," said Harvard Medical School professor Stephen Soumerai, who led the investigation.
"But as it turns out, decisions to market directly to consumers are based on scant data."
Even when sales increase, it has been nearly impossible to isolate the impact of targeting consumers from, say, marketing campaigns directed at doctors and health professionals.
Ideally, one would want to compare two otherwise identical populations before and after one of them is exposed to direct-to-consumer ads for certain drugs in order to measure the difference.
In the United States that is not possible, so Soumerai came up with a clever idea: Canada.
Direct-to-consumer ads for prescription drugs are illegal north of the US border, but English-speaking Canadians are still massively exposed to pharma advertising on American television and radio, and in US magazines.
Their French-speaking neighbours in the Canadian province of Quebec, however, are not, which provided the investigators with a useful group for making a comparison.
The study focused on three drugs: Enbrel, for arthritis; Nasonex, for allergies; and Zelnorm, for irritable bowel syndrome.
All three had been on the market in both countries for at least one year before the direct-to-consumer campaigns began in the United States.
The question was simple: once the ads kicked in, did sales increase more quickly in English-speaking Canada than in Quebec?
Based on data from a registry of 2,700 pharmacies, the answer for the first two drugs was a clear "no", while for the third drug, sales initially leapt in the Canadian anglophone market but then fell back.
The study was published online on Tuesday by the British Medical Journal (BMJ).
Sales of Zelnorm initially jumped 40 percent, but within a couple of years prescription rates in both groups were the same again.
Given that Zelnorm was the only drug available for that syndrome, it might have sold as well without the direct-to-consumer ads.
Soumerai thinks the ads failed because the process of buying prescription drugs is not as easy or spontaneous as buying over-the-counter medication such as aspirin.
A person has to see an ad, get motivated, contact their doctor, show up for an appointment, communicate both the condition and the drug, convince the doctor to prescribe it, and then actually fill the prescription, which is also likely to carry an out-of-pocket cost, he explained.
The United States is one of only two countries in the world -- along with New Zealand -- where direct-to-consumer ads for prescription medication is legal.
In March 2007, Zelnorm was pulled from the market due to concerns about increased risk of heart attack and stroke.