Osborne RH, Nelson L, Elsworth GR, Fehnel S, Williams N, Bender RH, Ziemiecki R, Gymnopoulou E, De Paepe E, Norcross L, Heijnen E, Ispas G, Cheret A, Comeaux C, Chan E, Scott JA. Psychometric evaluation of the Respiratory Infection Intensity and Impact Questionnaire Version 2 (RiiQ™v2) symptom scores for respiratory syncytial virus. Poster presented at the Virtual ISPOR 2021 Conference; May 2021.

OBJECTIVES: The Respiratory Infection Intensity and Impact Questionnaire Version 2 (RiiQ™v2) is a patient-reported outcome tool to measure symptoms (respiratory and systemic) and impacts of respiratory syncytial virus (RSV) infection. This study evaluated the reliability, validity, and responsiveness of two key RiiQ™v2 symptom scores: lower respiratory tract symptom score (LRT-SS; cough, wheeze, phlegm, short of breath) and upper respiratory tract symptom score (URT-SS; sore throat, nasal congestion) to support their use as clinical trial endpoints.

METHODS: Data were collected in patients with acute respiratory infection (ARI) from a Phase 2b study in 60 confirmed RSV-positive and 1,615 RSV-negative non-hospitalized individuals with no coinfections, and two observational studies in hospitalized RSV patients (n=20; n=100). Descriptive statistics, confirmatory factor analysis (CFA), test-retest intraclass correlation coefficients (ICCs), construct validity correlations (with clinician-reported signs/symptoms scores [CSS] and patient global impression of severity [PGIS]), known-groups validity (based on comorbidities and PGIS), and responsiveness (correlations of change scores) were evaluated.

RESULTS: At ARI onset, higher proportions of RSV-positive patients had moderate-to-severe respiratory symptoms (≥3 on a 1 to 4 scale) than RSV-negative patients (cough: 60%/35%; phlegm: 28%/12%; nasal congestion: 56%/49%; sore throat: 28%/21%). CFA loadings (>0.40) along with several adequate goodness-of-fit indices supported planned summary scores. LRT and URT ICCs supported test-retest reliability. Moderate-to-strong correlations between LRT-SS, the CSS and PGIS support construct validity and consistency between clinician and patient ratings; correlations with URT-SS were moderate among RSV-negative patients but small among RSV-positive patients. Correlations with change in PGIS >0.30 supported responsiveness.

CONCLUSIONS: The LRT domain showed stronger psychometric properties than the URT domain, possibly resulting from higher variability in the distribution of URT scores. Results show initial support for the reliability, validity, and responsiveness of the RiiQ™v2 for use in clinical studies.

Share on: