Chitty and colleagues' recent investigation ^{Reference Chitty, Dobbins, Dawson, Isbister and Buckley1} into the association between psychotropic medication use and alcohol consumption during emergency department presentations for self-poisoning raises an interesting perspective on the putative role of psycho-pharmacology in reducing risky alcohol use among those at risk of self-harm and suicide.

Episodes of attempted suicide resulting in hospital presentation may underestimate the true extent of psychotropic medication and alcohol co-ingestion across the community, given recent findings suggesting that just over half of patients treated by ambulance paramedics following an episode of self-harm and/or attempted suicide are transported to hospital. ^{Reference Matsuyama, Kitamura, Kiyohara, Hayashida, Kawamura and Iwami2}

Using data from our ongoing study of psychiatric presentations to ambulance services, ^{Reference Lloyd, Gao, Heilbronn and Lubman3} we extracted information on all episodes of non-fatal self-poisoning in the state of Victoria, Australia, from January 2012 to December 2016 (N = 24 726). In contrast to Chitty and colleagues, we found that, overall, use of psychotropic medications was associated with an increased, not decreased, risk of alcohol co-consumption in the self-poisoning episode (odds ratio (OR) = 1.35, 95% CI 1.28–1.42).

While anticonvulsants (OR = 0.74, 95% CI 0.65–0.84), antipsychotics (OR = 0.81, 95% CI 0.75–0.86) and psychostimulants (OR = 0.52, 95% CI 0.32–0.85) were associated with a decreased risk of alcohol co-consumption, in contrast to Chitty and colleagues, we found that benzodiazepines (OR = 1.60, 95% CI 1.52–1.69) were associated with an increased risk of alcohol co-consumption. Additionally, we found no significant association between antidepressant use and risk of alcohol co-consumption for these presentations (OR = 1.04, 95% CI 0.97–1.11).

Importantly, however, we found that medical severity may modify these associations. Specifically, most associations were reduced to non-significance when considering those not requiring hospital treatment following the self-poisoning episode: all psychotropic medication classes (OR = 1.12, 95% CI 0.76–1.65), anticonvulsants (OR = 0.39, 95% CI 0.09–1.80), antidepressants (OR = 1.05, 95% CI 0.63–1.77), antipsychotics (OR = 0.81, 95% CI 0.48–1.36), benzodiazepines (OR = 1.40, 95% CI 0.94–2.07) and psychostimulants (OR = 0.44, 95% CI 0.02–9.21).

This highlights the importance of considering the breadth of services that people who engage in self-harm come into contact with, so as to provide a fuller picture of the treatment needs of this population and how these may vary as a consequence of medical severity.