4. Discussion

We assessed the quality of MBW measurements collected in clinical routine before and after implementing mandatory real-time quality control in two centers. We provided the MBW operators with a simplified quality control matrix according to current guidelines and assessed whether real-time quality control of MBW data during the measurement improves test acceptability in routine clinical testing. Following implementation of real-time quality control, acceptability of MBW measurements improved from 68 to 84% in Bern and from 50% to 90% in the validation center Zurich, and resulted in excellent agreement between the operator and reviewer.
Implementing mandatory real-time quality control improved overall test acceptability and the ability of MBW operators to recognize and perform good quality MBW measurements in routine clinical testing. Before implementation of real-time quality control criteria, operators reported outcomes to the clinicians with trials that should have been rejected according to current quality control guidelines11,13-15. After implementing mandatory real-time quality control and providing the simplified guidelines in Bern, operators were able to correctly identify all technically not acceptable trials. These findings were validated in a center with less experience in MBW measurements (Zurich). These results indicate that performing standardized real-time quality control is feasible in the clinical setting, improves overall test acceptability, and results in good agreement between reviewer and operator.
Our quality control guidelines are a simplified version of the current MBW consensus guidelines, preschool technical standards, and quality control guidelines by Jensen et al. and Saunders et al 11 14 13,15. Due to the complexity and time consuming nature of the MBW test, there has been a focus on the need for detailed retrospective quality control and central over-reading by highly experienced MBW researchers. While central over-reading is a suitable approach for large, multi-centre research studies15,17, clinical MBW testing requires immediate reporting of outcomes. Therefore, real-time quality control by the operator is the only way to ensure that good quality MBW outcomes are used for clinical interpretation. Our quality control criteria (Table 1) provide simplified guidelines for MBW trial grading, trial acceptability, and test acceptability. They can be applied at the time of the measurement to allow immediate reporting of MBW data in clinics. To maintain high quality MBW data in the clinical setting we suggest that centres implement regular training sessions with operators, provide updates of recent literature in the field, and perform regular over-reading of random subsets of MBW measurements in order to provide feedback to the operators.
We found that systematic quality control of MBW measurements by an experienced reviewer did not lead to differences in mean LCI or FRC values compared with values reported without quality control. These findings are similar to those reported by Jensen et al, who found no differences in mean LCI values following qualitative and quantitative review13. However, while both studies reported no significant differences in outcomes on a population level, reporting outcomes from technically not acceptable MBW trials can impact the outcomes of a test occasion on an individual level18. In the clinical setting, longitudinal changes in MBW outcomes from one visit to the next are likely to influence treatment decisions. Therefore, it is essential to ensure that only data from good quality MBW trials are reported.
In the clinical setting, trials with highly irregular breathing pattern (D grade) but meeting technical acceptability criteria are a frequent challenge. Especially in younger children, where individual attention may be limited and time restricted in busy outpatient clinics, collecting three immaculate trials can be challenging. We, therefore, decided to follow a pragmatic approach and include them for outcome reporting if at least one good quality trial (A, B or C trial) within the same test occasion was available and LCI and FRC are within 10% of this acceptable trial. The consensus statement suggests repeatability of FRC to be within 25% of the median FRC of technically acceptable trials11 and states that trials exceeding those limits should prompt further investigation. We did not find an influence on MBW outcomes and test variability when including these D grade trials and variability was substantially lower than the limits proposed by the consensus statement11. However, these criteria need to be validated in a larger dataset.
There are some limitations to our quality control criteria. Trial grading is based on qualitative criteria for breathing pattern that are inherently subjective. Users need to have familiarity with the MBW test to differentiate between minimally and moderately variable breathing pattern. However, many of these subjective criteria will only influence whether a trial is graded as either A, B or C, all of which are technically acceptable trials. Our criteria could have been simplified further to only include acceptable, questionable, or rejected trial grading, however, MBW operators stated that it was beneficial to be able to recognize what constitutes a technically perfect trial and understand which deviations can still be accepted. Further, our quality control criteria are only applicable to N2MBW measurements. This study was performed using Ecomedics equipment and Spiroware 3.2.1 software whereby flow-volume loops, nitrogen, oxygen, and carbon dioxide signals are visible during and after the measurement. It is unclear how easily these criteria can be applied to data collected in other devices, software, and alternative tracer gases.
Conclusion
Real-time quality control of MBW data at the time of the measurement is feasible in the clinical setting and results in improved test acceptability with excellent agreement between MBW operators and experienced reviewers. Our quality control criteria provide simplified guidelines for MBW trial grading, trial acceptability, test acceptability, and outcome reporting that can be applied even in centers with less experience in MBW methodology. Applying these criteria by MBW operators at the time of the measurement ensures that only good quality MBW outcomes are reported to the clinician.