Skip to content

I’ve warned you! But I shouldn’t have

Among close friends, or even within the family, the use of SSRI’s (selective serotonin reuptake inhibitors) can be a delicate topic, it may come with connotations of depression, suicidal behaviour, and can be emotionally marginalizing. A new scientific review may further entangle this already vexing situation, in the study (Isacsson, G. & Alhner, J. 2013) it was found that the presence of suicide warnings in SSRI’s may have indirectly increased suicide rates due to an intricate form of cultural Information Hazard.

Despite being a review, the study only faintly points toward a certain interpretational direction, but it doesn’t clearly guarantee it. With that prior caveat in mind, Isacsson and Alhner gesture that the warnings mentioning suicide risk have caused a plateau in SSRI consumption in Sweden, and that during this period, the rate of suicides among those who didn’t take SSRI’s increased, and more so than among those who did. Once the label was in place, it made the decision not to take SSRI’s worse than it was before. This is a complicated process, and it is worth analysing it further.

In both the no-warning, and the yes-warning situations, you had three populations, A, who wouldn’t take the prescription regardless of anything else, B, who would take it as long as no suicide warnings were inscribed, and C, who would take them anyway. A comparison was made between C plus B and, separately, A, in the no-warning case, factoring out of course as many irrelevant variables as possible, and it was concluded that warning the C plus B population would be a good move to avoid suicides. Yet, in the yes-warning situation, it turns out that group B (those who would actually shift from taking to not taking) ended up more likely to commit suicide, contrary to prior expectation.

In this case, the warning itself can be considered an Information Hazard, and can be qualified as belonging to the following categories developed in the linked paper above by Nick Bostrom:

Temptation hazard: The warning tempts group B, who should not make the shift, to make the shift from accepting to denying prescription.

Knowing-too-much hazard: Knowing that SSRI’s can downstream into suicidal behaviour, despite true prima facie, is an undesirable piece of knowledge to have, making one more vulnerable to the very problem the knowledge should help with.

Although unusual, this is not an unheard of situation, its reverse even has a popular name: self-fulfilling prophecy. This is a case where a prophecy is self defeating. Not very frequently Medical Ethics intersects the field of logic and philosophy of language in relevant ways, but in this case, a comparison with the liar’s paradox is worth a note. In the same vein as “this sentence is false” is false when it is true, and true when it is false, so a suicide warning appears to be ethically desirable when absent, and ethically undesirable when present. Unfortunately, in this case, what results is not a beautiful paradox like Newcomb’s, but the worst of possible worlds.

As the quantifiable data on human cultural transmission increases, we are likely to find more and more “ethical conundrums” of this kind, and it would be well advised to educate ourselves with a vocabulary and intuitions to deal with these preventively, because sometimes, it is better to find ways to deal with what’s inside the box before finding out it belongs to Pandora.


Bostrom, N. (2011). Information hazards: A typology of potential harms from knowledge. Review of Contemporary Philosophy, (10), 44-79.

Isacsson, G. & Alhner, J. (2013). Antidepressants and the risk of suicide in young persons – prescription trends and toxicological analyses. Acta Psychiatric Scandinavia, 1-7.

Share on