Chapter 3 98 3.3 Sensitivity and publication bias We assessed whether correlations differed for study design, measurement instruments, and components of alexithymia. Results can be found in Table 2. Correlations between BPP and emotional awareness differed significantly among different study designs (Q(3) = 18.862; p < 0.001). Fifteen studies comparing patients with an official diagnosis of BPD to healthy controls yielded an average strong correlation of 0.518 (95% CI [0.411, 0.611]; Z = 8.227; p < 0.001). Four studies comparing BPD patients to people with other psychiatric diagnoses had an average moderate correlation of 0.409 (95% CI [0.123, 0.633]; Z = 2.736; p = 0.006). For the 14 studies in healthy subjects, the average small correlation was .236 (95% CI [0.103, 0.361]; Z = 3.438; p = .001). Among eight studies with patient groups, the average small correlation was .231 (95% CI [0.050, 0.398]; Z = 2.492; p = .013). No significant differences were found for studies measuring emotional awareness versus studies measuring alexithymia (Q(1) = 0.682; p = .409 (see the second row under “Instrument” in Table 2). Study design and the use of emotional awareness versus alexithymia measures are unrelated (chi2(3) = 5.35; p = .148). Since measures of NA have been found to correlate substantially to measures of alexithymia (such as the TAS-20) [105, 106], we also assessed the effect of NA as a possible confounder. Seven studies included a direct or related measure of NA (i.e., anxiety, depression, PTSD) as a covariate. Two studies found a small to moderate positive association between BPP and emotional awareness when NA was added to the analysis as a covariate, four found no significant association, and one study found a small negative association. The overall association in this selection was nonsignificant (r = .068; 95% CI [–0.128, 0.259]; Z = 0.678, p = 0.498). Ten studies presented findings on associations between BPP and three components of alexithymia: difficulty in identifying feelings, difficulty in describing feelings, and externally oriented thinking; and one study presented findings on the first two components (see the last row of Table 2). The strongest association was found between BPP and difficulty in identifying feelings (r = .461; 95% [CI 0.266, 0.620]; Z = 4.327; p < .001), followed by difficulty in describing feelings (r = .327; 95% CI [0.140, 0.491]; Z = 3.359; p = .001) and externally oriented thinking (r = .106; 95% CI [0.059, 0.153]; Z = 4.406; p < 0.0001).
RkJQdWJsaXNoZXIy MjY0ODMw