3.4 Agreement between MBW operator and reviewer after implementing
real-time quality control
After implementing real-time quality control, the agreement between
operator and reviewer was high in both centers. For test acceptability
agreement was 97% (ĸ = 0.9, p < 0.001) in Bern and 100% (ĸ =
1.0, p < 0.001) in Zurich. In Bern, only one test occasion was
rejected by the reviewer but not by the operator. The remaining four
test occasions rejected by the reviewer were also rejected by the
operator. In Zurich, all test occasions rejected by the reviewer were
also rejected by the operator.
For trial grading, agreement was 68% (ĸ = 0.6, p < 0.001) in
Bern and 73% (ĸ = 0.6, p < 0.001) in Zurich. In Bern, the
operators were able to recognize all technically invalid trials (F
grade). All F grade trials consisting of leaks were identified by the
operator. The only trials in which the end of test criteria was not met
were trials that were prematurely terminated by the operator. All the
trials evaluated satisfied the start of test criteria, which meant that
operators were consistently waiting enough time between trials. Three
trials were given a C grade by the operator and a D grade by the
reviewer due to irregular breathing pattern. However, these trials were
then excluded by both as not being accompanied by a second good quality
trial. Only one trial with a D grade was rejected by the reviewer but
not by the operator. In Zurich, only one F grade trial that did not meet
the end of test criteria was not recognized by the operator. Leaks were
all correctly identified by the operator. As in Bern all trials
satisfied the start of test criteria. One trial was given a D grade by
the reviewer and a C grade by the operator, however, both were rejected
for final reporting as not being accompanied by a second good quality
trial.