Engineering

Statistics

Civil engineering

Mathematics

Lung function

Internal medicine

Grading (engineering)

Medicine

Lung

European Respiratory Journal

Background: In population studies, automated portable devices are used to measure lung function on location. A build-in automated quality algorithm selects technically correct lung function data for analysis. null Aim: To evaluate the performance of intergraded quality control in lung function testing equipment. null Methods: Lung function measurements in 6137 children (age 6 -13) were collected from the HITEA, FRESH and the PIAMA study. All measurements were performed with a portable device, which uses an algorithm based on between-manoeuvre acceptability criteria from the ATS/ERS in an adapted grading system (NDD EasyOne). Data with A, B and C grades are acceptable for data analysis while data with grades D or F are rejected. All tests were manually checked by a certified respiratory function technologist using the same grading system to evaluate decisions made by the algorithm. Furthermore, data from the PIAMA study (n=1530) was used to determine the absolute effect of manual revision on mean FEV 1 null and FVC. null Results: Manual revision resulted in 2411 adjustments (39%): 374 (16%) to a lower and 1346 (56%) to a higher quality grade. Of all original measurements 21% was rejected by automated quality control. From these, 685 measurements (53%) could be saved for analysis after manual revision. In addition, 110 accepted manoeuvres were rejected (2%) Manuel revision resulted in an absolute change in FEV 1 null and FVC of respectively 189 and 240 ml. null Conclusion: Manual revision saves a large number of measurements and changes lung function values. Adaptation of the algorithm may improve the performance of automated grading of lung function measurements.

Manual revision of automated lung function grading results in an increase of accepted data and a relevant change in FEV1 and FVC