Psychometric Properties of the Living With Pulmonary Fibrosis (L-PF) Questionnaire in Patients With Idiopathic Pulmonary Fibrosis

OBJECTIVES: Idiopathic pulmonary fibrosis (IPF) is characterized by progressive fibrosis, abnormal lung function, and burdensome symptoms that impair patients’ quality of life. Living with Pulmonary Fibrosis (L-PF) is a 44-item patient-reported outcome measure (PROM) assessing symptoms and impacts of pulmonary fibrosis. We evaluated the psychometric properties of L-PF-44 in an IPF clinical trial population (FIBRONEER™-IPF; NCT05321069).

METHODS: Our analyses were based on clinical trial data from FIBRONEER™-IPF (N=1,177). Confirmatory factor analysis (CFA) was performed to verify the pre-specified factor structure of L-PF. Symptoms domain was represented as a second-order factor including cough, dyspnea, and fatigue sub-domains. Impacts domain was represented as a single factor correlated with the symptoms domain. The appropriateness of this factor solution was evaluated with model fit indices (comparative fit index [CFI], root mean square error of approximation [RMSEA], and standardized root mean squared residual [SRMR]). To allow for cross-validation of the model, CFA was performed sequentially: first using the calibration dataset with ∼50% of randomly selected baseline data, second using the validation dataset with the remaining 50% of baseline data, and third using the full available sample at Week 12.

RESULTS:

The factor structure of L-PF showed adequate model fit for baseline calibration, baseline validation, and Week 12 analyses (CFI 0.890, 0.897, 0.905; RMSEA 0.101, 0.102, 0.104; SRMR 0.102, 0.101, 0.101, respectively). Standardized factor loadings were acceptable for all items, ranging from 0.573–0.979 for symptoms and 0.478–0.923 for impacts. High second-order factor loadings (range across analyses for cough: 0.753–0.783; dyspnea: 0.802–0.829; fatigue: 0.897–0.915) confirmed that symptoms sub-domains were consistent with each other.

CONCLUSIONS:

This study provided initial evidence for the structural validity of L-PF in an IPF clinical trial population. Further psychometric properties of L-PF domains and sub-domains (construct validity, reliability, responsiveness) are ongoing and will be presented at the conference.