ERS/ATS WORKSHOP REPORT SERIES Reproducibility of lung volume measurements

For a number of reasons, it is important to know and understand test variability or reproducibility. All pulmonary function tests are subject to: 1) technical variation related to instrument procedure, calibration, observer, subject, and their interaction; 2) biological variation; and 3) variation caused by dysfunction or disease [1]. A knowledge and understanding of test variability can enhance the interpretation of the results through the separation of variability of interest (signal) from other sources of variation (noise) both within a subject and within a population [2]. By quantifying the variability or reproducibility of a measurement for both normal subjects and patients with disease, the interpretation of the results is enhanced. Test reproducibility can be a useful indicator of the quality of a laboratory's measurements and procedures, and changes in procedures have been shown to result in significant reductions in test variability [3]. Similarly, a trend towards an increasing variability may indicate an equipment malfunction or procedural problems. Because the sensitivity of a test can be improved with reductions in test variability, attempts should be made to monitor and control test variability. Test reproducibility can also be useful in assessing the adequacy of a test in a particular subject. If multiple determinations of a test are made, test reproducibility provides an indication of the level of confidence in an individual's results. However, test reproducibility should not be used to exclude subjects from a study or to avoid interpreting the results. Rather, test reproducibility should be considered when the test results are interpreted. Method of expressing variability