Skip to main content
. 2020 Sep 8;77(12):1–6. doi: 10.1001/jamaneurol.2020.3241

Figure 1. Predicted Cumulative Probabilities of Disease Progression for Miglustat Use vs No Use From Longitudinal Models .

Figure 1.

A, Shaded regions indicate 95% CI for American Speech-Language-Hearing Association National Outcome Measurement System swallowing domain (ASHA-NOMS; range, 0 to 6 where 0 indicates ability to eat independently and 6 indicates swallowing is not safe by mouth) and Rosenbek aspiration-penetration scale (PAS; range, 0 to 4 where 0 indicates no obvious aspiration risk or penetration and 4 indicates profound aspiration). B, Analyses were performed using generalized linear models for repeated measures with independent working correlations for ordinal multinomial models. Each model also adjusted for baseline levels of the outcome variable. The full model was based on a priori hypothesis, and the final model retained variables following iterative model fitting, which are reflected in panel A. ECO indicates early childhood onset (age <6 years); LCO, late childhood onset (age ≥6 and <15 years); NA, not applicable; OR, odds ratio.