3 A meta-analysis on the association between EA and BPP 99 Table 2. Results of Meta-Analyses for Study Design, Type of Alexithymia Instrument, and Components of Alexithymia Subgroup n r 95% C.I. Z p Study design1 [Q(3) = 18.862; p < 0.001] BPD vs. HC 15 .518 .411 - .611 8.227 <.001 BPD vs. PC 4 .409 .123 -.633 2.736 .006 Healthy subjects 14 .236 .103 - .361 3.438 .001 Patient subjects 8 .231 .050 - .393 2.492 .013 Instrument [Q(1) = 0.682; p = 0.409] Alexithymia 18 .393 .280 - .496 6.375 <.001 Deficit in Emotional Awareness 21 .330 .222 - .430 5.735 <.001 Components of alexithymia Difficulty Identifying Emotions 11 .461 .266 - .620 4.327 <.001 Difficulty Describing Emotions 11 .327 .140 - .491 3.359 <.001 Externally Oriented Thinking 10 .193 .049 - .329 2.621 0.009 1 Study by Lizeretti et al. (2012) was split for this analysis (BPD vs HC / BPD vs PC) Various analyses were conducted to assess publication bias. The distribution on the funnel plot was somewhat skewed to the right, suggesting a small bias towards publishing studies with positive results. However, Egger’s test of intercept was found to be nonsignificant, t(37) = 1.2; p = 0.221. Focusing on studies to the left of the mean, Duval and Tweedie’s [70] trim and fill analysis demonstrated that no studies needed to be trimmed or filled. Therefore, we concluded there was no evidence for publication bias.
RkJQdWJsaXNoZXIy MjY0ODMw