Abstract
Background Patient-Reported Outcomes (PROs), including quality of life or symptoms, are now frequently used as endpoints in RCTs testing novel drugs. To provide robust evidence to inform treatment decisions, it is critical that PRO findings are accurately and timely published.
Aims The primary objective of this study was to identify opportunities for improvement in the quality of PRO reporting of RCTs of multiple myeloma (MM) and lymphoma (Lym) by reviewing papers published over the last 10 years. Secondary objectives were to evaluate the timing of the publication of PRO results, relative to the primary RCT publication, and to investigate whether the quality of PRO reporting has improved over time.
Methods We analyzed all the RCTs with a PRO endpoint conducted in MM and Lym, published between Jan 2015 and Feb 2025. Data were obtained from the IMPROvE project, whose main aim is to investigate the magnitude of a possible open-label bias of PRO endpoints in RCTs investigating anti-cancer treatments. Two reviewers independently extracted data using a pre-piloted form in REDCap. For this analysis, data extracted from each RCT included trial characteristics and the CONSORT-PRO checklist. The latter comprises five PRO-specific extension items and nine PRO-specific elaborations to CONSORT-2010 items. These items are related, for example, to the methodological quality of the PRO assessment. To evaluate the quality of PRO reporting over time we calculated the CONSORT-PRO score following published scoring recommendations. The final score ranged from 0 to 14 (PROs as secondary endpoint) or 15 (PROs as primary endpoint), with a higher score indicating a higher quality of PRO reporting. PROs are generally used as secondary or exploratory endpoint, often not included in the primary RCT publication but in a separate PRO publication. We therefore assessed the timing of the publication of a secondary paper on PRO results using median and interquartile range (IQR). Univariable and multivariable linear regression analyses were performed using the CONSORT-PRO score as dependent variable and time as independent variable. Potential confounders in the multivariable model included disease type, study setting and sponsorship, type of experimental treatment and mode of administration, secondary paper on PROs, and PRO results favoring the experimental group.
Results We assessed 83 RCTs conducted in MM (n=55) and Lym (n=28), the majority of which (78%) had an open-label study design. Most were supported by industry (88%) and were conducted in an international setting (77%). Immunotherapy (either used alone or in conjunction with other therapies) was the most frequently studied experimental therapy (70%). Progression-free survival was the primary endpoint in 76% of RCTs, while only one study (1.2%) had PROs as primary endpoint. A statistically significant difference between treatment arms in any PRO scale was found in 55% of RCTs. Overall, the median CONSORT-PRO score for the analyzed studies was 9.5 (IQR 4.5-12) and was similar between MM (9.5; IQR 4.5-12) and Lym studies (9.8; IQR 4.8-11.8). Although generally the RCTs reported PRO results for each domain (78%), cited evidence of PRO instrument validity (77%), and interpreted PROs in relation to clinical outcomes (73%), the reporting of some key PRO information was suboptimal. For example, few RCTs reported details on PRO hypotheses (16%), mode of PRO administration (33%), or the statistical approach for dealing with missing data (47%). More than half (59%) of RCTs published a secondary paper on PRO results. In this subset of RCTs, the median time between the publication of the primary RCT paper and a secondary publication on PROs was 1.8 (IQR 1-3.3) years. We observed no significant improvement over time in the quality of PRO reporting, in either the univariable (β= -0.04; p=0.79) or multivariable analyses (β= -0.07; p=0.4). The only factor remaining significantly associated with a higher quality of PRO reporting in the multivariable analysis was having a secondary paper on PRO study results (β= 7.22; p<0.001).
Conclusion The publication of a separate PRO manuscript is often performed when including PROs in MM and Lym RCTs, and this is also associated with higher quality reporting. However, efforts are needed to encourage timelier publication of secondary PRO study reports, in order to maximize the chances that such data are fully considered in treatment decision making.
