Assessment of inter- and intra-rater reliability of objective cough frequency in patients with chronic cough

Abstract

Background: Clinical trials of antitussive therapies increasingly use sound recording devices to assess cough frequency as an objective measure of efficacy. The VitaloJAK™ (Vitalograph, Buckingham, UK) comprises a digital sound recorder and a computer algorithm to filter out silence and most non-cough sounds from 24-h ambulatory recordings. Trained human cough analysts (CAs) listen to and inspect audio waveforms in the filtered recordings, tagging individual coughs which are tallied automatically. Recordings unsuitable for filtering are analyzed for the full 24 h. The reliability of Vitalograph CAs has not been previously documented.
Objective: To describe the intra- and inter-rater reliability of cough frequency measurements assessed by Vitalograph CAs from filtered 24-h cough recordings.
Methods: The recordings came from Phase 2 trials of gefipixant in patients with chronic cough. For intra-rater reliability, 13 CAs reviewed 1 randomly-selected recording two times, at least 1 month apart. For inter-rater reliability, 11 recordings were randomly selected and each was reviewed by 5 randomly-selected CAs. Intra-class correlations (ICC) to assess for absolute agreement were used to gauge reliability.
Results: Across the 81 recording assessments, the median 24-h cough count was 356 (IQR: 167, 656). ICCs for both intra- and inter-rater reliability were found to be outstanding [0.999 (95% CI: 0.996, 1.000 and 0.999 (95% CI:0.997, 1.000), respectively].
Conclusions: The cough frequencies assessed by Vitalograph CAs were highly reliable. These results support the use of the VitaloJAK cough monitoring system to assess cough frequency in clinical trials.