Table 3. Efficacy outcomes for individual studies.

Study publication (Study number/Trial registration)	Primary endpoint: correct use of inhaler	Secondary endpoint: ease-of-use of Inhaler	Overall (critical and non-critical) errors using inhaler
van der Palen et al. 2016 (200301/ NCT02184624) [15]	After reading the PIL only, fewer patients had ≥1 critical error using the placebo ELLIPTA DPI compared with DISKUS (5% vs. 44%), MDI (13% vs. 60%), Turbuhaler (8% vs. 44%), HandiHaler (14% vs. 48%), or Breezhaler (13% vs. 46%) (all p<0.001). After reading the PIL, the majority of patients made no errors using ELLIPTA DPI (57–70% across the 5 substudies), thus not requiring instruction from the nurse	A larger proportion of patients in each substudy rated ELLIPTA DPI “very easy” or “easy” to use compared with DISKUS (97% vs. 60%), MDI (92% vs. 44%), Turbuhaler (96% vs. 55%), HandiHaler (98% vs. 38%), or Breezhaler (94% vs. 55%) (all p<0.001)	After reading the PIL only, fewer patients with COPD had ≥1 overall error using ELLIPTA DPI compared with DISKUS (30% vs. 65%), MDI (31% vs. 85%), Turbuhaler (31% vs. 71%), HandiHaler (43% vs. 62%), or Breezhaler (31% vs. 56%) (all p<0.001)
Feldman et al. 2019 (201071/NCT02586493) [16]	In the COPD study, 258 (97%) patients demonstrated correct use of the placebo ELLIPTA DPI at the end of the study	Among subjects demonstrating the correct use of ELLIPTA DPI at study end, 93% of COPD subjects rated the inhaler as “easy” or “very easy” to use	In the COPD study, 8 (3%) patients made an error using ELLIPTA DPI at the end of the study
van der Palen et al. 2018 (206215/ NCT02982187) [17]	In both substudies, significantly fewer patients using the placebo ELLIPTA DPI made ≥1 critical error after reading the PIL compared with patients using DISKUS + HandiHaler or Turbuhaler + HandiHaler. In each substudy, 9% (n = 7/80) of patients made ≥1 critical error using the placebo ELLIPTA DPI vs. 75% (n = 60/80) using DISKUS + HandiHaler in substudy 1 (p<0.001) and 73% (n = 58/79) of patients using Turbuhaler + HandiHaler in substudy 2 (p<0.001)	Not analyzed	Significantly fewer patients using the placebo ELLIPTA DPI had ≥1 overall error in the first attempt after reading the PIL, compared with patients using other inhalers (24% vs. 80% with DISKUS + HandiHaler and 22% vs. 80% with Turbuhaler + HandiHaler) (both p<0.001)
Kerwin et al. 2020 (206901/NCT03227445) [18]	The number of patients demonstrating errors (critical and non-critical errors) was 9 (4%) for the placebo ELLIPTA DPI, 14 (6%) for DISKUS, and 27 (12%) for HandiHaler in the ITT population (n = 239) at Day 28	Not analyzed	Per primary endpoint column data

COPD, chronic obstructive pulmonary disease; DPI, dry-powder inhaler; ITT, intent-to-treat; MDI, metered dose inhaler; PIL, patient information leaflet.