The style of early clinical research reporting: what are we saying and how do we say it?

Abstract

Phase II studies involving novel chemotherapy treatment are often first reported to the scientific community as a published abstract. This study was designed to determine how authors report their phase II abstract data and what types of conclusions are made. All 1992 phase II colorectal cancer abstracts from the 1992 Proceedings of the American Society of Clinical Oncology were reviewed and analyzed for descriptive and quantitative data and conclusions. The response rate was the most commonly reported response statistic (28/29), with few studies reporting confidence intervals (4/29), median duration of response (5/29), or median survival (9/29). Toxicity data was quantitative in 21 trials and qualitative or absent in 8 trials. Conclusions about the toxicity data were ambiguous or absent in five trials. Conclusions about the response data included inappropriate use of terms, indirect or ambiguous language, or no conclusion. These results indicate that the reporting of phase II trials in abstract form is highly variable and plagued by problems that make data interpretation difficult. Recommendations for change are suggested.