Page 66 - The efficacy and effectiveness of psychological treatments for eating disorders - Elske van den Berg

66 Chapter 3
Comparing both samples used in the final analyses, ‘eligible for treatment’, BMI > 17.5 and being treated within the timeframe of this study (N = 239 2012-2014, N = 320 2015-2017), the same significant difference with regard to the diagnosis distri- bution was found in the 2015-2017 sample (χ2 = 22.51, p <.001), except for the diagno- sis anorexia nervosa, which was equally distributed over both samples. In addition, the percentage of laxatives misuse is significantly higher in the 2012-2014 sample (Fisher’s exact test, p = .013)
In 2012-2014, mean time frame between start and EOT measurements was 268 days (SD = 138); in 2015-2017 mean timeframe was 195 days (SD = 83), a significant difference (MW-U=5106; p < .001). Mean treatment episode in 2012-2014 was 300 days (SD = 162), in 2015-2017 240 days (SD = 90), a significant difference (MW-U = 30559, p < .001).
Treatment outcomes for eating disorder and general psychopathology
Table 3 presents outcomes of the 2012-2014 and 2015-2017 samples from baseline to EOT. Eating disorder pathology and general psychopathology declined significantly in both samples (2012-2014, t = 9.24, p <.001; 2015-2017, t = 21.70, p <.001). Eating disorder behaviours also declined significantly (all p <.001). Linear mixed model anal- yses, adjusted for differences in diagnosis distribution and baseline EDE-Q scores, showed no significant differences between both samples with regard to the decrease of eating disorder pathology (EMD = -.05, p = .736) or general psychopathology (EMD = -1.86, p = .539). In 2012-2014 however, significantly fewer patients reported remaining binges after treatment (36.4% vs 56.2%; χ2 = 8.27, p = .002). Effect sizes for both samples were large (2012-2014, d = 1.19; 2015-2017, d = 1.41). Linear mixed model analyses on an imputed ITT dataset did not show a significant difference between both samples with regard to the decrease of eating disorder pathology either (EMD = -.12, p = .440).
Remission rates
With regard to remission rates, UK and Dutch norms were used for the purpose of comparison. Remission rate for responders was defined as having an EDE-Q glob- al score under one SD above community mean, i.e. below 2.77 (UK norms) or 1.79 (Dutch norms). In 2012-2014, 60.2% (UK norms) and 40.9% (Dutch norms) achieved remission. In 2015-2017, remission rate was 62.1% (UK norms) respectively 40.6% (Dutch norms); not a significant difference (UK norms, p = .760; Dutch norms, p = .965). Imputed ITT analyses did not show a significant difference either; 59.0% 2012- 2014 versus 60.9% in 2015-2017 (F(1, 1063.07) = .45, p =.504).