Lung function measurement in general practice. General practice measurements compared with laboratory measurements during the DIMCA trial

Background. Lung function measurement in the general practice setting (GPS) is of growing importance. Objective. The aim of this study was to compare results from a lung function laboratory (LFL) with those in a GPS. Methods. Comparisons were made for decline calculated from GPS and LFL measurements and intra-individual paired measurements. Test characteristics of the spirometer used in the GPS were also assessed. Results. The mean decline in lung function was: -0.037 I/year [95% confidence interval (CI) -0.202 to 0.129] from LFL data and -0.027 I/year (95% CI -0.242 to 0.188) from GPS data. The mean intra-individual difference was -0.0025 I (95% CI-0.493 to 0.488 I). The test characteristics of the spirometer used in GPS did not meet all of the American Thoracic Society guidelines; in particular, the random error was too large. The difference in assessments between measurement in LFL and GPS were such that misclassification might occur if slopes were calculated. Conclusions. Repeated measurement in this study showed that GPS measurements are not interchangeable with LFL measurements. Therefore, one has to be cautious when interchanging lung function data from an LFL and a GPS.