Purpose: The quality of the supportive and palliative oncology literature is unclear. We examined five indicators of study quality in the supportive and palliative oncology literature.
Methods: We systematically searched MEDLINE, PsychInfo, EMBASE, ISI Web of Science, and CINAHL for original studies related to "palliative care" and "cancer" during the first 6 months of 2004 and 2009. For each study, we extracted the study size, the sample size calculation, and other study characteristics. We also determined the Consolidated Standards of Reporting Trials (CONSORT) overall quality score, the key methodologic index, and the Jadad score for randomized controlled trials (RCTs).
Results: A total of 840 studies were included for sample size analysis, and 44 RCTs were identified for quality of reporting analysis. The median sample size was 70 for RCTs, 112 for cohort studies, and 200 for cross-sectional studies. Sample size calculations were most frequently reported in RCTs; however, 29/44 (66%) RCTs had no reporting, 5/44 (11%) had partial reporting, and 10/44 (23%) had full reporting. The median CONSORT overall quality score was 9 (interquartile range 7-11), key methodologic index was 0.50 (interquartile range 0-1.75), and Jadad score was 2 (interquartile range 1-3), suggesting low quality of reporting. We found no differences in the quality scores between 2004 and 2009. Key methodologic index score ≥ 2 was associated with the presence of funding from any source (10/25 vs. 1/19, P = 0.013) and funding from pharmaceutical industry (5/9 vs. 6/35, P = 0.03).
Conclusions: We identified deficiencies in the quality of supportive and palliative oncology RCTs.