15250-m-cuypers

51 HRQoL changes prior to Pca treatment onset 3 2.2.3 Psychological factors As possible moderating variables, three measures for individual differences measures were included. First, the Big Five Inventory-10 (BFI-10) was included to measure extraversion , agreeableness , conscientiousness , neuroticism , and openness; also known as the ‘big five’ personality traits 28 . The BFI-10 was included in t0. With only two items per trait, low reliability scores were found (α<.50), which is common for this scale 29 . A subsequent confirmatory factor analysis confirmed five underlying factors, with each set of two items per trait yielding highest factor loadings. Secondly, dispositional optimism, a generalized expectation that good things will happen, was assessed with the Life Orientation Test-Revised (LOT-R) 30 . Some minor textual adjustments were made to an existing and previously validated Dutch version of the LOT-R 31 . Scale reliability was sufficient (α=.67). Thirdly, the Decision Self-Efficacy Scale was used as a subjective measure of the perceived ability to make a healthcare decision 32 . Rather than focusing on one specific decision, the goal of this scale was to measure feelings of self-confidence in a healthcare setting. The scale was included at t0 to measure a person’s baseline decisional self- efficacy before the distress fromdiagnosis. In absence of an existing and validated Dutch version of this scale, a forward-backward translation was made by two researchers and the result was evaluated and consented on by two other researchers who were not involved to the translation. Scale reliability was good (α=.85). 2.3 Statistical analysis Descriptive statistics arepresentedasmeans and standarddeviations (SD) for continuous variables and as frequencies and percentages for categorical variables. Mean HRQoL scores at t0 were compared to the scores obtained at t1 using paired-samples t-tests. The association between personality traits and HRQoL scores were assessed using bivariate correlation analyses (Pearson’s). Linear regression modelling was carried out with global health as dependent variable and personality characteristics as independent variables, controlling for age, education, PSA levels and diagnosis (dummy variable; for t0 only). All analyses were performed using SPSS version 22.0 (Statistical Package for Social Sciences, Chicago, IL, USA). P -values <.05 were considered statistically significant.