Skip to main content
. 2022 Oct 19;79(12):1162–1172. doi: 10.1001/jamapsychiatry.2022.3352

Table 2. Moderator Analysis by Age, Sex, Sample Size, Region, Randomization, Type of ECT Treatment, and Presence of Psychotic Features Using Meta-Regression Analysesa.

Moderator k Coefficient (95% CI) SMD (95% CI) P value
No. of total patients 6 0.002 (−0.003 to 0.006) .51b
Age 6 0.018 (−0.030 to 0.066) .47b
Male sex (%) 6 −4.490 (−4.687 to 1.708) .36b
Region .25c
Asia/Middle East 4 NA −0.839 (−1.171 to −0.507)
Europe 2 NA −0.589 (−0.850 to −0.327)
Randomization .86c
Yes 5 NA −0.692 (−0.912 to −0.472)
No 1 NA −0.636 (−1.211 to −0.061)
Type of ECT .12c
Right unilateral 2 NA −0.589 (−0.850 to −0.327)
Bilateral 3 NA −0.655 (−1.048 to −0.262)
Mixed 2 NA −1.297 (−1.916 to −0.677)
Psychotic features .78c
Yes 2 NA −0.709 (−0.974 to −0.443)
No 4 NA −0.649 (−0.974 to −0.324)

Abbreviations: ECT, electroconvulsive therapy; k, number of effect sizes; NA, not applicable; SMD, standardized mean difference (Hedges g).

a

Coefficient refers to the meta-regression coefficient. A restricted maximum likelihood was used.

b

Continuous moderators.

c

Categorical moderators.