Abstract
Quality-of-life and psychosocial oncology studies that have low participation might have less precision, less statistical power, and can have non-response bias. In this systematic Review, we searched MEDLINE, Embase, and PsycInfo, for paediatric studies published in 2010-15 and adults studies published 2014-15. Studies were eligible if they were original studies published in a peer-reviewed journal; recruited children (aged 0-18 years at diagnosis) with cancer or their parents, or adult patients with cancer; and assessed psychosocial outcomes, including quality of life, depression, anxiety, wellbeing, distress, coping, or adjustment as a primary or secondary outcome. We assessed participation reporting quality, calculated percentages of participation achieved, and measured the influence of study design and participant characteristics. We reviewed 311 studies including a total of 87 240 adults, children, and parents. Mean participation across studies was more than 70% (paediatric participation was 72% and adult participation was 74%). Many studies did not report data essential for the assessment of participation, especially for non-respondents. Studies using a longitudinal cohort design had higher participation than randomised trials. In paediatric studies, recruitment of participants at diagnosis, face to face, and with the use of short questionnaires yielded higher participation. Other study design characteristics (method of data collection, who enrolled the participants, and incentives) and patient characteristics (cancer type, patient or parent age, and sex) did not affect participation in either paediatric or adult studies. Researchers can use these data to improve reporting quality and make evidence-based choices to maximise participation in future studies.