Caren van Roekel

94 Chapter 3 Quality of Life (GHS), 5 functioning scales and 22 symptom scores were derived. All but two items are scored on 4-point Likert scales (1: not at all, 2: a little, 3: quite a bit, 4: very much). The two other items are scored on a 7-point linear analogue scale. The raw subscale scores are transformed to a 0-100 scale, where a high score in a functioning scale represents unimpaired functioning and high score in a symptom scale represents a high level of symptomatology. The functioning scales are: physical functioning (PF), role functioning (RF), emotional functioning (EF), cognitive functioning (CF) and social functioning (SF). The symptom scales are: fatigue (FA), nausea and vomiting (NV), pain (PA), dyspnea (DY), insomnia (SL), appetite loss (AP), constipation (CO), diarrhea (DI), financial difficulties (FI)(QLQ-C30); and eating (LMNutri), fatigue (LMCFati), pain (LMCPA), emotional problems (LMCEp), weight loss (LMCWL), taste (LMCTA), dry mouth (LMCDM), sore mouth/tongue (LMCSM), peripheral neuropathy (LMCPN), jaundice (LMCJ), contact with friends (LMCFr), talking about feelings (LMCFeelings), and sex life (LMCSx) (QLQ-LMC21). Patients received the questionnaires at baseline, 6 weeks and 3 months after treatment. Follow-up in the HEPAR II study was longer, so those patients also received the questionnaires at 6, 9 and 12 months after treatment. The last included 26 patients of the HEPAR II study received an extra questionnaire 1 week after treatment to better reflect patients’ transient symptoms shortly after treatment (30, 31). Response assessment Response assessment was based on contrast-enhanced CT at 3 months post- treatment, according to the Response Evaluation Criteria in Solid Tumours (RECIST) version 1.1 (34). Scoring and statistical analysis Scoring of the questionnaires was performed according to the scoring manual provided by the EORTC (EORTC scoring manual). Missing values were imputed using multiple imputation. Internal consistency of the multi-item scales was determined using Cronbach’s alpha. Kolmogrov-Smirnov and Shapiro-Wilk tests were carried out for all categories at the different time points and showed that the data were not normally distributed (p≤0.001).