Prediction of normal values in pulmonary function testing.

Appropriate use of reference equations for predicting normal values is important for maximizing the clinical and research usefulness of pulmonary function testing. Because of the diversity of clinical and research uses of pulmonary function testing, predictive equations derived from different populations may be needed for specific applications. Similarly, different limits of normalcy other than the traditional 95 per cent confidence limits may be more appropriate for some applications. Unanswered questions remain about the effects of race and altitude on predicted values for many pulmonary function parameters. In order to establish confidence that the predictive equations, lower limits of normalcy, and testing methodology are appropriate for specific patient populations, it is suggested that each clinical laboratory obtain PFT measurements on 10 to 20 individuals considered free of disease and representative of their patient population and compare these results with the prediction equations selected for use. Although the testing of such a small sample of normal subjects will not assist in the selection of optimal equations if the differences between equations are minor, it will detect grossly inappropriate predictive values and limits of normalcy as well as identify problems with testing methodology.