15250-m-cuypers

97 PCPCC study protocol 5 Comparable to health skills, the relation of personality to beneficial DA-usage will be investigated. Some studies suggest a link between personality and treatment choice 51,52 . Following these studies, some relevant aspects of personality will be taken into account: hospital anxiety and depression (HADS-scale) 53 , prostate specific anxiety (MAX-PC) 54,55 , optimism (Life Orientation Test – Revised) 56 , the big five personality dimensions (Big Five Inventory-10) 57 , information seeking preferences (subscale from the Autonomy preference index) 58 and maximization tendencies (Maximization scale) 59 . Sociodemographic variables and additional healthcare utilization Standard sociodemographics will be asked on age, marital status, occupation, and education. Also, patients will be asked to report any additional healthcare utilization (general practitioner or other medical specialist) in the past 12months to assess whether this affects the decision-making process. Healthcare providers’ evaluation Healthcare providers in the intervention arm will be asked their opinion about implementation of the DA in qualitative interviews as well as questionnaires at the end of patient inclusion (approximately after 12 months). This questionnaire will be based on the MIDI-instrument 60 , which is developed for the evaluation of implementing an innovation in a healthcare setting. The questionnaire will focus on usage of the DA (for example ‘Did you offer the DA to all eligible patients?’ ) and evaluate the pros and cons of the DA with help of statements (for example ‘ The DA is practical in use’ ). Healthcare providers in the control condition will be asked to evaluate the current information provision and decision-making processes, their expectation of DA-usage andmotivation for implementation. Sample size calculation The sample size for this study is determined by power analysis with decisional conflict as the primary measure. To be able to detect a clinically relevant minimum effect size 61 of .50, power is set at .80 and alpha at .05. With 19 hospitals (clusters) that agreed to participate, it is needed to estimate the intra-class coefficient (ICC). The ICC assesses the proportion of variance explained by clusters. Higher ICC values decrease effective sample size and statistical power. ICC ranges from 0 to .1 are considered common in medical literature 62 . A more specific review of ICC values in (cluster) RCTs with psychosocial measures is provided by Bell and McKenzie 63 , which also included a cluster RCT evaluating a group support tool for prostate cancer patients 64 . The median estimated value for 82 longitudinal ICCs from 15 included studies was 0.0007, and the