“…Analysis of the comparative efficacy of SSRIs and tricyclic antidepressants in in-patients (judged likely to be a more severely affected group) provided a slightly larger estimate of effect favouring tricyclic antidepressants, though again non-significant. The overall estimate of effect in this grouping of studies 8,16,27,35,44,47,67,69,73,74,77,80,81,82,86,88,91,94,98,99,105,108,111,118,120,127 (using a random effects model) was 0.10 (95% CI -0.072 to 0.272, Q = 49.1, d.f. = 22, P = 0.008).…”