Continuous versus categorical data for ROC analysis: Some quantitative considerations

Rationale and Objectives. Several authors have encouraged the use of a quas i-continuous rating scale for data collection in receiver operating charact eristic (ROC) curve analysis of diagnostic modalities, rather than rating s cales based on five to seven ordinal categories or levels of suspicion. Alt hough many investigators have gone over to this method, a discussion of the issues continues. The present work provides a quantitative analysis from t he viewpoint of measurement science. Materials and Methods. A simple model of the effect of data discretization or quantization on the measurement of the variance of noisy data was develo ped. Then Monte Carlo simulations of multiple-reader. multiple-case ROC exp eriments were performed and analyzed in terms of components-of-variance mod els to investigate the effect of data quantization in that more complex set ting. Results. For single-reader studies, discretization into five categories can reduce the precision of ROC measurements by a large amount, The effect may be attenuated in multireader studies. Conclusion, More precise measurements of diagnostic detection performance a nd thus more efficient use of resources are served by good measurement meth ods. These are promoted by the use of a quasi-continuous raring scale in RO C studies.