Conceptualization and Validation of a De Novo Cost-Effectiveness Analysis for Pulmonary Hypertension Associated With Interstitial Lung Disease

OBJECTIVES: Pulmonary hypertension (PH) associated with interstitial lung disease (PH-ILD) is a rare and severe disease classified as World Health Organisation PH Group 3. The 16-week INCREASE randomised control trial (NCT02630316) and open-label extension (NCT02633293) evaluated the safety and efficacy of inhaled treprostinil in PH-ILD. No PH-ILD treatments have been appraised by health technology assessment agencies and no published cost-effectiveness analyses are available. We aimed to conceptualise and validate a conceptual model that appropriately captured the progressive nature of PH-ILD and efficacy of inhaled treprostinil.

METHODS: The de-novo model conceptualisation was based on a targeted literature review, consideration of the INCREASE endpoints and external natural history data. The chosen model structure was validated through individual interviews and advisory boards with pulmonologists and health-economic experts.

RESULTS: The use of health states based on disease progression events or severity was considered. Severity was disregarded because no disease staging systems are established in PH-ILD and the WHO Functional Class was not recorded in the INCREASE trial, nor could be inferred from other endpoints. Instead, a cohort partitioned survival model (PSM), incorporating time-to-event data from INCREASE was considered most appropriate. Health states were based on disease progression events representing clinical worsening including death, decreased six-minute walking distance of ≥15% from baseline (to capture PH decline), decreased predicted forced vital capacity of ≥10% from baseline (to capture lung function decline), cardiopulmonary hospitalisation, lung-disease exacerbation and lung transplants. The structure allowed for long-term extrapolation of survival and disease progression using clinically meaningful, PH-ILD-specific events, whilst directly utilising INCREASE endpoints. The PSM approach directly informs health state occupation and avoids the need to specify individual transitions – this will further facilitate validation of the results.

CONCLUSIONS: A cohort PSM using patient-level data from INCREASE to inform survival curves predicting the time to disease progression events and death was considered most appropriate.