To the Editors:
We read with interest the recent publication by Miller and Pedersen 1 on determining the best way of expressing lung function impairment for prediction of mortality. The authors ranked the methods of expressing forced expiratory volume in 1 s (FEV1). It was concluded that the FEV1 quotient (Q) was the best predictor. We believe that this outcome is in need for some nuances. This ranking is, in our view, only suitable when the parameters under investigation are all valid and when based on proper testing. We feel that this is not the case. We will illustrate this via the FEV1 standardised residuals (SR) and focus on the receiver operating curve (ROC) comparison chosen for the ranking.
The authors point at the phenomenon that the number of residual standard deviation (RSD) values an elderly subject can decline is less than that of a younger subject. Based …