Network Meta-analysis of the Efficacy of Different Music Therapy Interventions for Delirium in Adult Intensive Care Unit Patients

Background and aims: Delirium is common in adult intensive care unit (ICU) patients and is associated with worse outcomes.Music-based interventions are increasingly used, but the relative effectiveness of different programs remains unclear.We conducted a network meta-analysis (NMA) to compare multiple music therapy interventions for ICU delirium.Data sources: We searched major Chinese and English databases (e.g., CNKI, PubMed, and the Cochrane Library) from inception to 21 August 2025.Study selection: Randomized controlled trials comparing music-based interventions with usual care (UC) in adults (18 years) in the ICU were eligible.Interventions were categorized as simple music (SM) therapy, traditional Chinese medicine (TCM) five-element music, family voice stimulation combined with music (FS), and combined music (CM) therapy (music combined with another nonpharmacological component).Data synthesis: A frequentist random-effects NMA (R, netmeta) was performed.Compared with UC, all music-based interventions were associated with a reduced delirium incidence CM: Risk ratio (RR) = 0.43; TCM: RR = 0.45; FS: RR = 0.48; SM: RR = 0.51.For delirium duration, CM mean difference (MD) = -3.21 days and TCM (MD = -2.89days) showed statistically significant reductions, whereas SM and FS did not.The network ranking suggested CM had the highest probability of benefit; however, between-intervention differences were mainly supported by indirect comparisons and small trials.Heterogeneity and potential small-study effects were observed.Conclusion: Music-based interventions may reduce ICU delirium incidence, and some programs may shorten delirium duration.Because evidence is limited by heterogeneity, possible publication bias, and a paucity of head-to-head trials, rankings should be interpreted cautiously.Well-designed, multicenter RCTs with standardized protocols are needed before firm recommendations can be made.