A new individual patient data meta-analysis in Nature Communications reports that intravenous phage therapy for difficult-to-treat bacterial infections has an estimated probability of clinical improvement, alongside a relatively low reported adverse-reaction rate in the aggregated evidence base. The analysis synthesized data from 60 studies across nine high-income countries. Across 122 patients, the authors estimated a 65.8% probability of reported clinical improvement, while microbiological eradication was estimated at 55.3% among 106 evaluable patients. Safety reporting suggested an adverse-reaction probability of 9.2% among 168 patients, with reactions predominantly mild and self-limited. The review also found that combining intravenous delivery with local phage administration was associated with more frequent clinical improvement than intravenous delivery alone, while the authors emphasized the evidence remains fragmented and heterogeneous, with many underlying studies limited to case series or uncontrolled designs.