Concordance between patient reported and clinical outcomes in randomized controlled trials (RCTs) of cancer treatment

Aims: Understanding the added value of patient-reported outcomes (PROs) to traditional clinical endpoints (e.g. survival or response to therapy) in randomized controlled trials (RCTs) is critical to better inform clinical decision-making. We assessed concordance between the results of clinical endpoints and PROs in RCTs, and evaluated how authors summarized treatment recommendations based on the combined datasets. Methods: A systematic review in PubMed/Medline and the Cochrane Library identified cancer RCTs with a PRO endpoint published from January 2004 to December 2016 in the top four cancer sites: breast, colorectal, lung and prostate cancer. Two independent reviewers recorded information on trial characteristics, PROs and clinical endpoints and level of concordance. For clinical endpoints, we defined the study outcome as “better” or “worse” if more than half of the endpoints of the RCT reported in the publication favored the experimental or the control arm, respectively. For PROs, we defined “better” or “worse” if more than half of the specific scales/domains of PRO instruments used favored the experimental or the control arm, respectively. Final treatment recommendations, reported by authors themselves, were also extracted. Results: Overall, 480 RCTs were identified and, based on predefined criteria, 417 RCTs were included in the concordance analysis. Overall survival (19.4%) and progression-free survival (14.1%) were the most frequent clinical primary endpoints. In one third of the studies, PROs supported traditional clinical endpoints favouring the experimental treatment. In 23.7% of cases both clinical and PROs yielded similar results for each treatment group. However, in the large majority of studies (76.3%) PROs provided information that contrasted with the clinical findings and thus yielded additional information on overall treatment effectiveness. Details are provided in Table 1. When there were no differences in clinical endpoints between treatment arms, and PRO favoured the experimental arm, final trial recommendations basically always included discussion of PRO findings. Conclusions: In the large majority of RCTs that we evaluated, PROs contributed information enabling a more comprehensive evaluation of overall treatment effectiveness of the treatments being tested