Should we have confidence in radiology papers?

This study aims to assess whether Confidence Intervals are used appropriately in papers published in Clinical Radiology, and compares 50 consecutive papers in this journal with similar samples in Radiology and the British Medical Journal (BMJ). The utilization or omission of Confidence Intervals and statistical hypothesis testing is analysed. The results are discussed using χ 2 tests and 95% Confidence Intervals for the difference between proportions. This illustrates the important additional quantitative information that Confidence Intervals can convey, particularly when evaluating the degree of uncertainty related to a result from a small sample or subgroup in a study.