Issue link: https://read.uberflip.com/i/1503820
screening purposes. 28 Participants' confidence in using the Internet and their English proficiency are rated on a 5‐point scale ranging from 'not confident' to 'very confident'. The primary outcome measure was the Hospital Anxiety and Depression Scale. 29 Total score (HADS‐T) and depression (HADS‐D) and anxiety (HADS‐A) subscales were used. Subscales have seven questions each. All items are rated between 0 and 3. Scores 0–7 are normal, 8–10 borderline case, and 11–21 abnormal case. A single item on the European Organization for Research and Treatment of Cancer Quality of Life Core Questionnaire (EORTC‐ QLQ) assessed the overall health‐related life quality. 30 The item is rated on a 7‐point scale ranging from 1 'very poor' to 7 'excellent'. The six‐item version of the Breast Cancer Worry Scale (CWS) measured concerns about cancer recurrence and its impact on daily functioning. 31 Responses are rated on a 4‐point scale ranging from 1 'never' to 4 'almost always'. Higher scores indicate more frequent worries about breast cancer coming back. Two coping dimensions, active and avoidant coping, in the 28‐ item Brief Coping Orientation to Problems Encountered (Brief COPE) scale assessed breast cancer coping strategies. 32 Responses are rated on a 4‐point scale ranging from 1 'I haven't been doing this at all' to 4 'I have been doing this a lot'. Perceived social support was measured using the eight‐item emotional/informational support subscale of the Medical Outcomes Study Social Support Survey (MOS‐SSS). 33 Responses are rated on a 5‐point scale ranging from 1 'none of the time' to 5 'all of the time'; higher scores indicate better perceived social support. 2.5 | Sample size The sample size requirement reported in the protocol 34 was recal- culated for a two‐arm pilot RCT. Assuming a medium effect size (f = 0.30) with a power of 0.80 and a 0.05 alpha, the required sample size was 20 per group, and 60 in total (2:1 allocation ratio). Assuming an attrition rate of 25%, 75 participants were required. 2.6 | Statistical analyses IBM SPSS Statistics Version 21 was used for all statistical analyses. Group differences in baseline characteristics were analysed using Chi‐squared and t‐tests. Using Linear Mixed Modelling (LMM) with Intention‐to‐Treat (ITT), the groups were compared on the outcome measures over time. Restricted Maximum Likelihood (REML) esti- mation was used to account for missing data. Analyses were con- ducted separately for each outcome variable with time, group, and interaction as fixed effects. Significant effects were followed up using pairwise contrasts. The LMM assumes that the data are missing at random, which was investigated using Binary Logistic Regression. Non‐completers' and completers' data were compared using ANOVA and Chi‐square tests. Within and between‐group effect sizes were calculated. For Hedge's g, an effect size of 0.20 is considered small, 0.50 moderate, and 0.80 large. A reliable change index (RCI) 35 assessed clinically meaningful changes at post‐intervention and follow‐up on the HADS‐T. A person was considered 'recovered' if the pre‐and post‐test difference was greater than the RCI and if the post‐test score is lower than the cut‐ off point of 15 on HADS‐T. A person was considered 'improved' if there was a statistical change but the values did not pass the cut‐off, 'unimproved' if there was no statistical change regardless of whether the cut‐off was crossed, and 'deteriorated' if there was a statistically significant worsening in symptoms. 36 Four mediation models using the PROCESS macro for SPSS 37 tested whether active coping mediates the intervention effect on HADS‐T and Quality of Life scale (QOL) and whether avoidant coping mediates the intervention effect on HADS‐T and QOL. As suggested for clinical research, 38 difference scores for each mediator and dependent variable were calculated. Time 1 scores of each depen- dent variable were added as a covariate. We also tested whether the time since finishing treatment (0–24 months, 24–48 months, >48 months) moderated the relationship between the treatment (iCBT vs. TAU) and HADS‐T using PROCESS V4.2. A bootstrapping method (n = 5000 re‐samples) with a 95% confidence interval 39 and significance testing at p < 0.05 was used for all analyses. 3 | RESULTS 3.1 | Participants One hundred and eighteen survivors clicked on the questionnaire (see Figure 1 for participant flow). Of these, 42(36%) did not com- plete the questionnaires or were ineligible to proceed. Seventy six participants were randomised to iCBT (n = 53) or TAU (n = 23). The research response was high. Out of 49 iCBT participants, 38(77.6%) completed the post‐intervention, and 30(61.2%) completed the follow‐up questionnaires. Out of 23 TAU participants, 19(82.6%) completed the post‐intervention and 17(73.9%) completed the follow‐up questionnaires. 3.2 | Descriptive statistics and baseline comparisons Table 1 presents descriptive statistics. No significant differences between iCBT and TAU were found on any characteristics at base- line, except for hormonal therapy. A greater percentage of the TAU group (95.2%) received hormonal therapy than iCBT group (68.2%). Regarding the clinical anxiety at baseline, approximately half (37, 51.4%) met the criteria for abnormal case, 21(29.2%) borderline normal, and 14(19.4%) normal. For clinical depression, the majority (46, 63.9%) were normal, 17(23.6%) borderline normal, and only 9 (12.5%) were abnormal case. Internal reliabilities were acceptable for 448 - AKKOL‐SOLAKOGLU AND HEVEY 10991611, 2023, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pon.6097, Wiley Online Library on [28/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License