Reporting errors in studies of the diagnostic performance of self-administered questionnaires: Extent of the problem, recommendations for standardized presentation of results, and implications for the peer review process.

Abstract

Diagnostic efficiency statistics include sensitivity, specificity, and positive and negative predictive power. In reviewing the literature on the performance of self-report questionnaires to screen for depression, we found errors in several published articles in which these statistics were computed. To determine the extent of this problem we examined all studies of the diagnostic performance of self-report scales published between 1980 and 1991 in the Journal of Consulting and Clinical Psychology and Psychological Assessment: A Journal of Consulting and Clinical Psychology. We found 26 relevant studies: 9 had an error in the calculation of diagnostic efficiency statistics and 3 made calculations based on unconventional definitions of the terms. Moreover, no study reported all 4 diagnostic statistics together with the total and chance-corrected level of agreement between the scale and the diagnostic gold standard