LAUSR

s. PLoS One 2019;14:e0224697. Evaluation of patient-reported outcome (PRO) data to assess understanding and interpretability Hannah Birkett, Aliyah Mansur, Matthew Reaney, Emily Ruzich, Jayne Spink, David James and France Ginchereau IQVIA, Reading, UK; IQVIA, London, UK; IQVIA, Cambridge, MA, USA; Prostate Cancer Research, London, UK; IQVIA, Rockville, MD, USA Objective: Increased focus on PRO data in clinical trials is intended to share insights that help people make more informed care decisions. Research into patient-centric data visualisation is needed to support this decision-making. We used varied graphical formats to summarise clinical trial PRO data among people with prostate cancer (PC) to determine how format affects understanding and data interpretation. Research design and methods: Thirty people with PC recruited through Prostate Cancer Research (UK advocacy group) participated in 7-day online asynchronous research. Through days 1–6, participants saw 6/36 possible PRO/format combinations (1/day, covering all 6 PRO concepts/formats). Participants used free text to describe their interpretation of PRO data and investigators rated their understanding (scale of 1 [clear understanding] to 4 [no understanding]). Participants also rated ease of understanding and design clarity on 5-point scales. On day 7, one PRO concept was shown in all 6 formats and preferences ranked. Results: Participants most accurately interpreted results shown in Sankey diagrams (mean [SD] 1.37 [0.84]), followed by bar charts (1.46 [0.86]), line graphs (1.48 [1.03]) and pie charts (1.58 [0.90]). Pie charts, bar charts and line graphs were rated highest for ease of understanding and clarity. 86.7% of participants ranked pie ABSTRACTS S49