Mariken Stegmann

Figure 2.  Study timeline. In most cases the time between both consultations was 1 week.  Questionnaires were mostly completed directly after the second consultation. OPT = Outcome  Prioritisation Tool.  Outcomes  Patient‐reported age, sex, education level, and social network data were recorded. Data on  the tumour type, performance score (Eastern Cooperative Oncology Group classification 19 ),  and comorbidities (Charlson score 20 ) were extracted from hospital records.  The primary outcome was patient empowerment, as measured by the decision self‐efficacy  (DSE) scale. The DSE scale comprises 11 items that are scored 0–4. 21  The scores were  transformed to a scale ranging from 0 (no self‐efficacy) to 100 (high self‐efficacy). 22  No cut‐ off value is described in the literature, but a score of <75 was defined as low empowerment  and <50 defined as very low empowerment, based on consensus in the research group.  Secondary outcomes focused on symptoms of fatigue, anxiety, and depression. These  outcomes were chosen based on their high prevalence in patients with cancer and their  major impact on their lives. It has also been suggested that these outcomes are associated  with patient empowerment. 23,24  Fatigue was measured using the multidimensional fatigue  inventory (MFI‐20), which ranges from 0 to 80. 25,26  Based on earlier research in older  populations, the authors defined cut off scores of 58 and 73 to represent mild and severe  fatigue, respectively. 27–29  Symptoms of anxiety and depression were measured using the  Hospital Anxiety and Depression Scale (HADS‐A and HADS‐D). These subscales each include  seven questions, resulting in a total score that ranges from 0 to 21. 30  Subscale cut‐off scores  of ≥8 and ≥11 represent mild and severe symptoms, respectively. 31  Finally, the experiences  of patients and GPs with the OPT were explored using questions based on former  evaluations of this instrument. 32 Sample size  The authors aimed to include 80 patients in each group based on a difference of at least four points  on the DSE (effect size or Cohen’s d = 0.44),22 two‐sided testing, an α of 0.05, and a β of 0.20 (P =  0.80). A low rate of loss to followupwas expected because comparison between groups was  performed immediately after consultations. Therefore, the authors aimed to include 84 patients per  group to allow for a 5% loss.  Statistical methods  All data were entered in a secured digital data management system, pseudonymised, and extracted  to IBMSPSS Statistics (version 25) and STATA/SE (version 15). Descriptive statistics were used to  describe the participants and to compare groups at baseline. The effects of the intervention on the  DSE, MFI‐20, and HADS scores were tested by linear regression for continuous measures, and by  logistic regression for dichotomous measures. Odds ratios (ORs) and 95% confidence intervals (CIs)  4 43 The OPTion randomised controlled trial