Mentioning side effects in drug advertisements

[Boyo_Jim]

I read somewhere that advertisements for a particular drug must also note potential side effects. Please let me know if this is incorrect.

My specific question is how frequently a side effect must occur in the test subject population to make such notice necessary. Presumably a single occurrence my be coincidental and not necessarily linked to the drug being tested. I would guess that it must occur at a rate that would be statistically significant. However, that doesn’t seem plausible either because me limited understanding of stats tells me that’s 5%, and I just can’t imagine anyone marketing a drug that kills 5% of the people who take it and simply noting that death is possible.

[Duckster]

From the National Conference of State Legislatures - Marketing and Direct-to-Consumer Advertising (DTCA) of Pharmaceuticals (Updated November 2010)

Direct-to-consumer advertising (DTCA) by pharmaceutical companies has always been somewhat controversial, with the US and New Zealand being the only two developed countries that allow such promotion. Direct-to-consumer advertisements in the US are broadcast on television and radio, are displayed on billboards, appear in the pages of magazines and newspapers, and are on the internet. The average American television viewer, for instance, can expect to spend as much as 16 hours per year watching such ads (Frosch et al. 2007).

Expansion of DTCA into broadcast media is a relatively recent phenomenon in the US For a number of years, the Food and Drug Administration (FDA) had guidelines requiring the advertiser to provide detailed information on usage and risks that is contained in the drug’s FDA-approved product label insert, thereby confining ads to print form. Expansion into broadcast DTCA was precipitated by the FDA’s regulatory changes in 1997 and in 1999. The new regulations require broadcast ads to include only “major statements” of the risks and benefits of the drug along with directions to alternate information sources for full disclosure. This clarification of what constitutes adequate disclosure removed a major barrier that had initially made TV and radio advertisements infeasible. DTCA spending subsequently increased from $150 million in 1993 to $4.24 billion in 2005. Specifically there was no broadcast advertising in 1993, but it now comprises the primary form of DTCA – amounting to $2.55 billion in 2005.

http://www.voxeu.org/index.php?q=node/5135

FDA - Guidance for Industry Presenting Risk Information in Prescription Drug and Medical Device Promotion - DRAFT GUIDANCE (PDF Warning)

[Nametag]

There is no specific number. As described in the Draft Guidance above, the probability and the severity of risks are weighed, as well as the approved use of the drug, the probable patient population, and the target of the ads. A serious side effect (like death) will usually be mentioned regardless of the actual risk, common side effects will be mentioned regardless of their severity, and moderate side effects will be mentioned or not, depending on their likelihood – given who is likely to be taking the drug, and for what. Also be aware that advertising is not subject to pre-approval.

[Marley23]

@Boyo_Jim

I read somewhere that advertisements for a particular drug must also note potential side effects. Please let me know if this is incorrect.

I might be a little behind the times here, but I think the rule is a bit more nuanced than that: the ad has to include the side effects only if it makes specific claims about what the drug does. Other than that I agree with Nametag.

[Boyo_Jim]

@Nametag

There is no specific number. As described in the Draft Guidance above, the probability and the severity of risks are weighed, as well as the approved use of the drug, the probable patient population, and the target of the ads. A serious side effect (like death) will usually be mentioned regardless of the actual risk, common side effects will be mentioned regardless of their severity, and moderate side effects will be mentioned or not, depending on their likelihood – given who is likely to be taking the drug, and for what. Also be aware that advertising is not subject to pre-approval.

I should have figured. It sounds like they’ve drawn up a set of rules that are more or less unenforceable.

[Boyo_Jim]

@Marley23

I might be a little behind the times here, but I think the rule is a bit more nuanced than that: the ad has to include the side effects only if it makes specific claims about what the drug does. Other than that I agree with Nametag.

As I recall, requirement kicked in if a specific drug was named. I have seen ads by drug companies saying things like, “To learn more about options for treating your (whatever), go to this website…” But if they name the product, they have to specify some details including something about risk.

[Boyo_Jim]

Expansion of DTCA into broadcast media is a relatively recent phenomenon in the US For a number of years, the Food and Drug Administration (FDA) had guidelines requiring the advertiser to provide detailed information on usage and risks that is contained in the drug’s FDA-approved product label insert, thereby confining ads to print form. Expansion into broadcast DTCA was precipitated by the FDA’s regulatory changes in 1997 and in 1999. The new regulations require broadcast ads to include only “major statements” of the risks and benefits of the drug along with directions to alternate information sources for full disclosure. This clarification of what constitutes adequate disclosure removed a major barrier that had initially made TV and radio advertisements infeasible…

This doesn’t seem to me a “clarification” at all. It seems like a deliberate blurring, protecteing manufacturers from claims by consumers that they do not provide adequate information on their ads.

I thank you folks for your answers, though. This isn’t the right forum to discuss whether or not this is a clarification.

I consider my question answered.

[Crazyhorse]

@Boyo_Jim

As I recall, requirement kicked in if a specific drug was named. I have seen ads by drug companies saying things like, “To learn more about options for treating your (whatever), go to this website…” But if they name the product, they have to specify some details including something about risk.

According to the FDA Website ‘reminder ads’ where the drug name is used but no product claims are made don’t need to list any side effects.

The answer to the OP’s question is also at that link, in a very long-winded and complex breakdown based on print ads, television ads, radio ads, “reminder ads” “product claim ads”, etc. the answer is 'It varies".

[Canadjun]

@Boyo_Jim

My specific question is how frequently a side effect must occur in the test subject population to make such notice necessary. Presumably a single occurrence my be coincidental and not necessarily linked to the drug being tested. I would guess that it must occur at a rate that would be statistically significant. However, that doesn’t seem plausible either because me limited understanding of stats tells me that’s 5%, and I just can’t imagine anyone marketing a drug that kills 5% of the people who take it and simply noting that death is possible.

You are misunderstanding statistical significance. Assuming you are using 5% as your determination of statistical significance (that’s the most common percentage, but isn’t the only one used), that doesn’t mean that the side effect occurs 5% of the time, it means that there is a 5% likelihood that the side effects you observed occurred purely by chance (i.e. not related to the drug), and determining that requires knowing how often it occurs in the general population and comparing that to how often it occurs when the drug is administered.

[Boyo_Jim]

@Canadjun

You are misunderstanding statistical significance. Assuming you are using 5% as your determination of statistical significance (that’s the most common percentage, but isn’t the only one used), that doesn’t mean that the side effect occurs 5% of the time, it means that there is a 5% likelihood that the side effects you observed occurred purely by chance (i.e. not related to the drug), and determining that requires knowing how often it occurs in the general population and comparing that to how often it occurs when the drug is administered.

Does this mean:
1 If a condition would normally occur in 20% of the general population over the same time frame of the study, and it occurs in 21% of the test subjects, it is significant?

2.If a condition would normally occur in 20% of the general population over the same time frame of the study, and it occurs in 25% of the test subjects, it is significant?

3. Something else?

[Canadjun]

Depends how many subjects you are looking at. The more subjects you are looking at, the smaller the difference can be while still being statistically significant.

[Boyo_Jim]

Let’s just say a large number – 10,000?

What I’m trying to get at is what exactly the “significance” is bound to. 20% up to 21% represents a 5% increase of occurrence (in my mind, that is), 20% up to 25% means that 5% more of the total population experiences the effect, and represents an increase of occurrence of 20%.

At least that’s my intention by picking these numbers. Please correct me if I’m wrong.

[Canadjun]

When you are talking about statistical significance, a 10.01% occurrence of an adverse event for those taking the drug vs a 10% occurrence of an adverse event for those taking the placebo (or nothing at all) may in fact be significant at a 5% level of confidence if you have enough subjects (you’ll presumably need many thousands of subjects to distinguish such a small difference).

An entirely different question is whether an adverse event occurs x% more frequently when the drug is taken vs when it isn’t taken. That is a different issue entirely. If the adverse event is of little consequence (e.g. mild nausea) compared to what the drug is intended to do (e.g. kill drug resistant germs) then the frequency can be very high. If it is of major consequence (e.g. death) then obviously the frequency must be very low. That is a different question (although both questions are important) than whether what you are seeing is due to chance (i.e. statistical significance).