Variability of quality-of-life measurements and reporting in randomised controlled trials of pancreatic cancer: a systematic review.

Objectives: This systematic review aims to evaluate the methodology used in pancreatic cancer (PC) randomised controlled trials (RCTs) measuring quality of life (QOL) and focuses on the type, frequency, survey compliance and duration of these assessments.

Design: Systematic review of PC RCTs measuring QOL.

Data sources: A search of PubMed.gov and ClinicalTrials.gov was conducted for PC RCTs measuring QOL from inception to 21 March 2023. Only phase III RCTs were included. Studies were excluded if QOL was not measured, the study was phase I/II, in the second-line setting or unavailable in English. Data were independently extracted by two reviewers in a standardised fashion.

Primary and secondary outcome measures: Primary outcomes included the type of QOL instrument used, the timing and frequency of assessments, methods of analysis and survey completion rates (SCRs) over time. Secondary outcomes included patient demographics, significant QOL improvements and the frequency of trials measuring QOL.

Results: Out of 269 studies screened, 54 RCTs were identified, and 24 measured QOL (involving 11 229 patients). Instruments used included the EORTC QLQ-C30 (n=15), FACT-HEP (n=3), Spitzer-QOL-Index (n=2), EQ-5D (n=2), LASA (n=1) and FACT-PA (n=1). Most trials assessed QOL until disease progression or death (10/24), with 4-week intervals being the most common (7/24). SCRs were reported in 15/24 trials, with disease stage influencing SCRs over time. In trials with metastatic, locally advanced/metastatic, and resectable disease, the median times to reach a 50% response rate-defined as the point where the number of surveys completed was half of the enrolled participants-were 12.41 weeks (n=2), 14.14 weeks (n=10), and 54.2 weeks (n=3), respectively." Only 2/24 trials reported significant QOL improvements between treatment arms. Patient age was reported in all trials, while race/ethnicity was only reported in 4/24 trials.

Conclusions: Significant variability exists in the timing, methods and reporting of QOL assessments in PC trials. There is a need for further research to assess the implications of missing data and consider the temporality of QOL assessment in patients with advanced cancers and poor prognosis.