Need for Increased Biostatistical Consultation

To the Editor.— Perusing the scientific literature reveals, far too often, that even our most renowned journals unfortunately contain gross mathematical and language usage errors of statistical import.1-5 Examples of the former error include using a statistical procedure incorrectly to analyze data and even choosing the wrong statistical test with which to perform the analysis. Two illustrations of frequently appearing statistical language usage mistakes are the use of the word significant in medical writing without statistical support4 and the improper use of the terms sensitivity and specificity.5