VAS was the least favoured scale and should be used cautiously in this population. Most participants had a scale preference with high intrapatient consistency between scales. CRS was preferred for appetite loss and tiredness and NRS for pain. Consideration should be given to individualised cancer symptom assessment according to patient scale preference.

69 Background: Systematic assessment in cancer is conducted by a variety of instruments. Such assessment is important as comprehensive instruments detect more symptoms than casual clinical evaluation. In choosing assessment scales for polysymptomatic cancer patients one must consider the burden of assessment to ensure satisfactory completion rates. This study investigated patient preference and clinical utility of symptom assessment scales. Methods: A prospective survey was conducted in an Irish palliative medicine inpatient unit.Consecutive cancer admissions were recruited within 7 days. Patients’ preferences were elicited with regards to 3 symptom assessment scales; categorical response (CRS), numerical rating (NRS), and visual analogue (VAS), across 3 common symptoms; appetite loss, pain, and tiredness. Participants selected their preferred scale per symptom. We determined the clinical utility of each scale, defined by ease of completion as judged by an observer. Results: 100 participants wererecruited,aged 38-93 years (x̅ = 71 years; SD=11.6). Median European Cooperative Oncology Group (ECOG) score was 2 (range 0-4). Participants preferred CRS for appetite loss (48%) and tiredness (40%), and NRS for pain (44%). VAS was consistently the least preferred measure. Scale preference was fully consistent across symptoms for 52% of patients, with just 4% choosing a different scale per symptom. There was moderate agreement between participant scale preference and ease of completion as determined by observer (Pain: K=0.486; Fatigue: K=0.452; Appetite loss K=0.364). Conclusions: (1) Most participants had a specific scale preference which was consistent across symptoms, (2) CRS was preferred overall, (3) Participants did not need to experience a symptom to have a preference, (4) VAS should be used with caution in hospice clinical care or research, (4) Symptom assessment scales should be carefully selected for clinical and research purposes.