Impact of PASI response on work productivity and the effect of risankizumab on indirect costs using machine learning in patients with moderate-to-severe psoriasis

Objectives To compare the impact of Psoriasis Area and Severity Index (PASI) response on total work productivity impairment (TWPI) in patients with moderate-to-severe psoriasis; to compare TWPI and associated indirect costs among patients treated with risankizumab, adalimumab, ustekinumab, and placebo. Methods Data from REVEAL (adalimumab phase III trial) were used to assess differences in trial-observed TWPI across PASI response cohorts. A machine learning model used REVEAL data to predict TWPI for patients in the risankizumab trials. These values were used to estimate work loss hours and work impairment-related indirect costs for each treatment cohort. Results Among REVEAL patients (N = 741), TWPI in the PASI 100, 90-99, 75-89 cohorts was lower than the PASI <75 cohort (p < .05); mean TWPI was lowest with PASI 100 (1.7%) vs. 90-99 (2.5%) vs. 75-89 (4.8%) vs. <75 (14.3%). There was a significant (p < .0001) monotonic relationship between higher PASI response and lower TWPI. In the risankizumab trials (N = 2046), incremental TWPI relative to risankizumab was 3.4%/week for ustekinumab/adalimumab, and 17.1%/week for placebo; incremental indirect cost savings for risankizumab were $2179/year vs. adalimumab, $2321/year vs. ustekinumab, and $11,284/year vs. placebo. Conclusions Higher PASI responses were associated with reduced TWPI. Risankizumab was associated with less work impairment/indirect costs vs. ustekinumab/adalimumab/placebo.