Treatment patterns, healthcare resource utilization, and costs among patients with idiopathic pulmonary fibrosis treated with antifibrotic medications in US-based commercial and Medicare Supplemental claims databases: a retrospective cohort study

Background Pirfenidone and nintedanib are antifibrotic therapies which slow disease progression in idiopathic pulmonary fibrosis (IPF), an irreversible, progressive lung disease with poor prognosis. We compared adherence, persistence, and healthcare costs between patients initiating one of the two therapies. Methods We used the IBM Watson Health Commercial and Medicare Supplemental claims databases to select patients with IPF with >= 1 pharmacy claim for pirfenidone or nintedanib between 10/1/2014 and 6/30/2018. Adherence (proportion of days covered >= 0.80) and persistence (time to a gap of >= 60 days without medication or switch to the other antifibrotic medication) based on the days' supply and service date fields on claims were measured over a variable-length follow-up period. Healthcare costs, all-cause and respiratory-related, were measured over the persistent period and a fixed 12-month follow-up period. Inverse probability of treatment weights were applied to models comparing adherence, persistence, and costs between the two cohorts. Results Overall, 799 pirfenidone patients and 656 nintedanib patients were identified. Similar proportions of patients were adherent in both cohorts (pirfenidone = 49% vs. nintedanib = 51%) and there was no significant difference in the odds of being adherent after weighting (odds ratio = 1.1,p = 0.513). The proportions of patients who discontinued/switched were also similar (pirfenidone = 41% vs. nintedanib 43%); however, in a weighted model, the hazards of discontinuation/switching was lower for the pirfenidone cohort (hazard ratio = 0.8,p = 0.032). While patients were persistent on therapy, weighted all-cause healthcare costs were comparable (pirfenidone = $11,272 vs. nintedanib = $11,987 per-patient per-month;p = 0.115), but weighted respiratory-related costs were significantly lower for the pirfenidone cohort ($9015 vs. $10,167 per-patient per-month,p < 0.001). Weighted annual total all-cause and respiratory-related healthcare costs were comparable between cohorts over the fixed 12-month follow-up period, but the pirfenidone cohort had significantly lower weighted annual mean antifibrotic drug costs than the nintedanib cohort ($68,850 vs. $77,033,p = 0.007). Conclusions Pirfenidone use was associated with longer time to discontinuation/switch, lower antifibrotic drug costs, and lower respiratory-related total costs compared to nintedanib use.