Effects of media stories of hope and recovery on suicidal ideation and help-seeking attitudes and intentions: systematic review and meta-analysis

BACKGROUND: There is strong evidence that suicides increase after media stories about suicides by celebrities, particularly those that highlight the suicide method (the Werther effect). Much less is known about the Papageno effect-the protective effects of media stories of hope and recovery from suicidal crises. A synthesis of the retrievable evidence is lacking. We aim to summarise findings from randomised controlled trials about the effects of stories of hope and recovery on individuals with some degree of vulnerability to suicide. METHODS: For this systematic review and individual participant data meta-analysis, we searched PubMed (including MEDLINE), Scopus, Embase, PsycInfo, Web of Science, and Google Scholar published from inception to Sept 6, 2021, without language restrictions. We included trials that reported suicidal ideation (the primary outcome) or help-seeking attitudes or intentions (the secondary outcome) and tested a media narrative of hope and recovery. Studies were excluded if they did not feature a clearly positive story of hope and recovery, or had a control group exposed to suicide-related stimulus material. We contacted the lead or senior authors of all original studies to obtain participant-level data for this study. The primary analysis was restricted to individuals with some vulnerability to suicide. Risk of bias was assessed using the Cochrane risk-of-bias tool for randomised trials. The study is registered with PROSPERO, number CRD42020221341. FINDINGS: Our search yielded 7347 records. 3920 records were screened by title and abstract, and 25 full-text records assessed for eligibility. There were eight eligible studies with 2350 participants for which individual participant data were sought. For suicidal ideation, six studies met the inclusion criteria for the primary analysis. Follow-up responses were available for 569 (90%) of 633 participants who were randomised with high vulnerability (345 [55%] allocated to the intervention group and 2