Lessons learned in conducting youth suicide prevention research in emergency departments

To address youth suicide, we recruited seven emergency departments (EDs) for what we believe is the first controlled trial of an intervention to promote safer firearm and medication storage after a child was seen in an ED by a behavioural health clinician. We provided training and a common protocol that required behavioural health clinicians to counsel about reducing access to household firearms and medication and provide locking devices. This paper examines how we addressed implementation challenges and considers how lessons learnt might inform future studies and interventions.

We tracked enrolment numbers and documented, through project records, the challenges in recruiting and developing partnerships with hospitals, as well as implementing the intervention and carrying out data collection.

We encountered challenges identifying hospitals with sufficient patient volumes to meet our analytic requirements, obtaining contact information from families in the ED and providing supplies to the sites. These challenges were compounded by lack of uniformity in data systems, making it difficult to estimate total ED encounters meeting our eligibility criteria, and by differences in patterns of behavioural healthcare delivery across sites. The strategies we devised to address these challenges included creating visual materials that appealed to parents' altruistic desire to help other families, laminated 'cheat sheets' and hang tags for clinician badges reminding them of the key points of the intervention and contracting with a distribution centre to coordinate shipping.

Despite the challenges noted, we found that the behavioural health clinicians in the EDs followed the protocol and found it useful in engaging families in discussions about both firearm and medication storage. Several hospitals intend to continue the intervention on their own as the new usual care, suggesting that the challenges encountered can be and are worth tackling.