Clinical epidemiological research on suicide-related behaviors (SRBs) is characterized by weak associations, weak effect sizes of clinical interventions, and strong effect sizes of some means-restriction interventions. These results have led several recent JAMA Psychiatry articles to conclude that nonclinical interventions must be central components in any successful multimodal SRB prevention strategy. I agree. But I disagree with the pessimism about components of this strategy in 2 of these recent articles: an editorial by Hoge about the low value of existing SRB-focused clinical trials and an article by Belsher et al about the low value of machine-learning (ML) methods ...
Language: en
", language="en", issn="2168-622X", doi="10.1001/jamapsychiatry.2019.1238", url="http://dx.doi.org/10.1001/jamapsychiatry.2019.1238" }