Feasibility of using an automated quality control algorithm for spirometry

Spirometry tests that meet quality control criteria are essential for accurate interpretation; however, reviewing individual maneuvers for quality in large population-based studies introduces barriers to achieving this goal. The objective of this study was to explore the use of a novel automated artificial intelligence software (ArtiQ.QC) to apply the 2019 American Thoracic Society/European Respiratory Society spirometry quality control criteria to data collected as part of the Canadian Longitudinal Study on Aging (CLSA). Individual spirometry maneuvers (ie, flow-volume and time data) from the CLSA were imported into ArtiQ.QC. Each maneuver was evaluated for technical acceptability according to the ATS/ERS 2019 standard. Quality grades were compared between those provided by the spirometer software and the ArtiQ.QC grade. Of the 21,795 spirometry test sessions, 15,079 (69.2%) were technically acceptable and repeatable (Grade A, B or C) for both forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC). Of 67,908 maneuvers, 25,598 were deemed technically unacceptable, with 87.8% of these failing to meet end of forced expiration criteria. The proportion of individuals below the lower limit of normal for FEV1 and FVC was lower when ArtiQ.QC evaluation was applied compared with those provided by the software. AI-based quality control algorithms are a feasible and efficient way to ensure high quality spirometry data in research studies.