Reader Comments

Post a new comment on this article

Relative Risk

Posted by JasonGIlesMD on 13 Jan 2011 at 14:09 GMT

It is hard for me to tell if the findings of the study are correlative or causal. Are people with a greater propensity to violence more likely to be on the drugs implicated in the study? What is the relationship to alcohol?

Further, reporting rare adverse events such as these in terms of relative risk gives the impression of outsized risk. I believe there are changes in behavior observed in patients on the medications cited but saying something rare is now seven times more likely is not that clinically useful.

I would like to see how the expected number of violent adverse events was estimated as well.

No competing interests declared.

Re: Relative Risk

DrugSafetyResearch replied to JasonGIlesMD on 13 Jan 2011 at 21:18 GMT

Are people with greater propensity to violence more likely to be taking the drugs implicated?

While there was no evidence for 85% of all drugs, those implicated involved a wide variety of patient populations. It included an asthma drug (montelukast), a malaria drug (mefloquine) where a propensity seems remote. One might expect some propensity for smoking cessation drugs, but in Table 2 we observed large differences between the 3 drugs used in the same population. On the other hand, we observed low PRR values for some antipsychotic drugs even though some members of this patient population might have a propensity for violence.

We hope that the clinical use of this information would be to alert physicians that patient or caregiver complaints of violent thoughts or rage could indeed be related to the implicated drugs and to take appropriate action.

The expected value for any given drug was 0.25% of the total number of serious adverse event reports of all types for that drug.

Competing interests declared: Delcared in the original article.