Skip to main content
. 2020 Dec 10;49(3):352–369. doi: 10.1017/S135246582000079X

Table 2.

Post-CBT diagnostic outcomes reported in the included texts

Author/year Sample size on whom CBT recovery outcomes are based1 % recovered from primary SoAD post-CBT % recovered from other primary anxiety disorders post-CBT Authors’ conclusions regarding association between primary diagnosis and likelihood of recovery
Barrett et al. (1996) 53 62% 73% No significant association between primary diagnosis and recovery rates3
Shortt et al. (2001) 48 56% 72% No significant association between primary diagnosis and recovery rates3
Wergeland et al. (2014, 2016)* 179 27% 42% A primary SoAD or SAD diagnosis, compared with primary GAD, significantly reduced the odds of recovery post-CBT
Arendt et al. (2016)* 101 29% 65% Participants with primary SoAD were significantly less likely to be recovered at post-treatment than those with any other primary anxiety diagnosis
Creswell et al. (2017)* 622 50% 45% Significant association between primary GAD and post-treatment severity of primary diagnosis (reflecting greater improvement in severity of primary GAD compared with other primary anxiety diagnoses)
Villabo et al. (2018)* 165 55% 63% Not reported
Suveg et al. (2018) 92 Not reported Not reported No significant difference in recovery rates by primary diagnoses, although co-morbid primary diagnoses were used, and co-morbidity was high (>90%)3
Silk et al. (2018) 90 Not reported Not reported Pre-treatment diagnosis did not significantly predict or moderate recovery post-CBT3
Thirlwall et al. (2013, 2017)* 96 38% 49% Participants with primary GAD were significantly more likely to have recovered from their primary diagnosis at post-treatment than those with all other primary anxiety disorders
Stjerneklar et al. (2019)* 32 21% 50% Not reported

*Included in meta-analysis. 1Sample sizes listed here refer only to young people who met diagnostic criteria for primary anxiety diagnoses included in this review. 2Participants with sub-clinical primary diagnoses (n = 4) excluded. 3Statistical power to detect a between-group effect not reported.