“…We identified nine meta‐analyses where the authors reported that they conducted formal subgroup analyses (Table 1, Table S3; Agabio et al., 2023; Agabio, Baldwin, et al., 2021; Agosti et al., 2012; Donoghue et al., 2015; Feinn & Kranzler, 2005; Hartwell et al., 2020; Lejoyeux & Lehert, 2011; Mason & Lehert, 2012; Pierce et al., 2018), of which three compared acamprosate to placebo for AUD (Lejoyeux & Lehert, 2011) or AD (Donoghue et al., 2015; Mason & Lehert, 2012), and four compared naltrexone to placebo for AD (Agosti et al., 2012; Donoghue et al., 2015; Feinn & Kranzler, 2005; Hartwell et al., 2020), and three compared baclofen to placebo for AD (Pierce et al., 2018) or AUD (Agabio et al., 2023; Agabio, Pisanu, et al., 2021). Across 61 individual subgroup analyses reported in these nine meta‐analyses (median [IQR] 6 [2, 8] per meta‐analysis), 33 (54%) were based on baseline participant‐level characteristics and 28 (46%) were based on trial‐level characteristics.…”