Evaluation of the quality of clinical research studies of magnetic resonance angiography: 1991-1994.

To evaluate the quality of clinical research studies of magnetic resonance angiography (MRA), the authors reviewed studies of carotid, renal, and lower extremity MRA published in English during 1991-1994. The purpose, design, and implementation of the study, and the type of statistical information presented, were evaluated. The results were compared with those of previous studies that assessed the quality of MRI research. Fifty-nine studies of renal (n = 10), carotid (n = 29), and lower extremity (n = 20) MRA were identified; 30 met the inclusion criteria for the analysis. In these papers, presentation of a reference standard was routine, observer blinding was usual, information often was given about patient selection and exclusion, sample sizes were always given, and parameter estimates and statistical analyses were common. However, evidence of research planning was not routine, measurement of interobserver variability was occasional, variability of parameter estimates was usually not estimated, and validation of statistical assumptions was uncommon. On balance the authors conclude that the scientific quality of recent investigations of MRA is better than that of older studies of MRI.